Commit
·
dc6b582
1
Parent(s):
15421f1
Upload folder using huggingface_hub
Browse filesThis view is limited to 50 files because it contains too many changes.
See raw diff
- checkpoint-10000/config.json +34 -0
- checkpoint-10000/pytorch_model.bin +3 -0
- checkpoint-10000/trainer_state.json +616 -0
- checkpoint-100000/config.json +34 -0
- checkpoint-100000/pytorch_model.bin +3 -0
- checkpoint-100000/trainer_state.json +0 -0
- checkpoint-105000/config.json +34 -0
- checkpoint-105000/pytorch_model.bin +3 -0
- checkpoint-105000/trainer_state.json +0 -0
- checkpoint-110000/config.json +34 -0
- checkpoint-110000/pytorch_model.bin +3 -0
- checkpoint-110000/trainer_state.json +0 -0
- checkpoint-115000/config.json +34 -0
- checkpoint-115000/pytorch_model.bin +3 -0
- checkpoint-115000/trainer_state.json +0 -0
- checkpoint-120000/config.json +34 -0
- checkpoint-120000/pytorch_model.bin +3 -0
- checkpoint-120000/trainer_state.json +0 -0
- checkpoint-125000/config.json +34 -0
- checkpoint-125000/pytorch_model.bin +3 -0
- checkpoint-125000/trainer_state.json +0 -0
- checkpoint-15000/config.json +34 -0
- checkpoint-15000/pytorch_model.bin +3 -0
- checkpoint-15000/trainer_state.json +916 -0
- checkpoint-20000/config.json +34 -0
- checkpoint-20000/pytorch_model.bin +3 -0
- checkpoint-20000/trainer_state.json +1216 -0
- checkpoint-25000/config.json +34 -0
- checkpoint-25000/pytorch_model.bin +3 -0
- checkpoint-25000/trainer_state.json +1516 -0
- checkpoint-30000/config.json +34 -0
- checkpoint-30000/pytorch_model.bin +3 -0
- checkpoint-30000/trainer_state.json +1816 -0
- checkpoint-35000/config.json +34 -0
- checkpoint-35000/pytorch_model.bin +3 -0
- checkpoint-35000/trainer_state.json +2116 -0
- checkpoint-40000/config.json +34 -0
- checkpoint-40000/pytorch_model.bin +3 -0
- checkpoint-40000/trainer_state.json +2416 -0
- checkpoint-45000/config.json +34 -0
- checkpoint-45000/pytorch_model.bin +3 -0
- checkpoint-45000/trainer_state.json +2716 -0
- checkpoint-5000/config.json +34 -0
- checkpoint-5000/pytorch_model.bin +3 -0
- checkpoint-5000/trainer_state.json +316 -0
- checkpoint-50000/config.json +34 -0
- checkpoint-50000/pytorch_model.bin +3 -0
- checkpoint-50000/trainer_state.json +3016 -0
- checkpoint-55000/config.json +34 -0
- checkpoint-55000/pytorch_model.bin +3 -0
checkpoint-10000/config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "Davlan/afro-xlmr-base",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"XLMRobertaModel"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 768,
|
| 13 |
+
"id2label": {
|
| 14 |
+
"0": "LABEL_0"
|
| 15 |
+
},
|
| 16 |
+
"initializer_range": 0.02,
|
| 17 |
+
"intermediate_size": 3072,
|
| 18 |
+
"label2id": {
|
| 19 |
+
"LABEL_0": 0
|
| 20 |
+
},
|
| 21 |
+
"layer_norm_eps": 1e-05,
|
| 22 |
+
"max_position_embeddings": 514,
|
| 23 |
+
"model_type": "xlm-roberta",
|
| 24 |
+
"num_attention_heads": 12,
|
| 25 |
+
"num_hidden_layers": 12,
|
| 26 |
+
"output_past": true,
|
| 27 |
+
"pad_token_id": 1,
|
| 28 |
+
"position_embedding_type": "absolute",
|
| 29 |
+
"torch_dtype": "float32",
|
| 30 |
+
"transformers_version": "4.29.0",
|
| 31 |
+
"type_vocab_size": 1,
|
| 32 |
+
"use_cache": true,
|
| 33 |
+
"vocab_size": 250002
|
| 34 |
+
}
|
checkpoint-10000/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:294c8f011978da665b20fa2692cae1fa538f826e5e5cb60fb246fe8523cfe1a9
|
| 3 |
+
size 1112246250
|
checkpoint-10000/trainer_state.json
ADDED
|
@@ -0,0 +1,616 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 3.1938677738741617,
|
| 5 |
+
"global_step": 10000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.03,
|
| 12 |
+
"learning_rate": 2.8744809964867454e-07,
|
| 13 |
+
"loss": 91.176,
|
| 14 |
+
"step": 100
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.06,
|
| 18 |
+
"learning_rate": 6.068348770360907e-07,
|
| 19 |
+
"loss": 64.0684,
|
| 20 |
+
"step": 200
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.1,
|
| 24 |
+
"learning_rate": 9.26221654423507e-07,
|
| 25 |
+
"loss": 27.1061,
|
| 26 |
+
"step": 300
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.13,
|
| 30 |
+
"learning_rate": 1.242414564037049e-06,
|
| 31 |
+
"loss": 8.0198,
|
| 32 |
+
"step": 400
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.16,
|
| 36 |
+
"learning_rate": 1.561801341424465e-06,
|
| 37 |
+
"loss": 4.7691,
|
| 38 |
+
"step": 500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.19,
|
| 42 |
+
"learning_rate": 1.8811881188118813e-06,
|
| 43 |
+
"loss": 3.4387,
|
| 44 |
+
"step": 600
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.22,
|
| 48 |
+
"learning_rate": 2.2005748961992977e-06,
|
| 49 |
+
"loss": 2.7027,
|
| 50 |
+
"step": 700
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.26,
|
| 54 |
+
"learning_rate": 2.5199616735867137e-06,
|
| 55 |
+
"loss": 2.4431,
|
| 56 |
+
"step": 800
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.29,
|
| 60 |
+
"learning_rate": 2.8393484509741297e-06,
|
| 61 |
+
"loss": 2.0534,
|
| 62 |
+
"step": 900
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.32,
|
| 66 |
+
"learning_rate": 3.1587352283615465e-06,
|
| 67 |
+
"loss": 1.8537,
|
| 68 |
+
"step": 1000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.35,
|
| 72 |
+
"learning_rate": 3.4781220057489625e-06,
|
| 73 |
+
"loss": 1.642,
|
| 74 |
+
"step": 1100
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.38,
|
| 78 |
+
"learning_rate": 3.797508783136378e-06,
|
| 79 |
+
"loss": 1.5092,
|
| 80 |
+
"step": 1200
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.42,
|
| 84 |
+
"learning_rate": 4.1168955605237945e-06,
|
| 85 |
+
"loss": 1.4302,
|
| 86 |
+
"step": 1300
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.45,
|
| 90 |
+
"learning_rate": 4.436282337911211e-06,
|
| 91 |
+
"loss": 1.3527,
|
| 92 |
+
"step": 1400
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.48,
|
| 96 |
+
"learning_rate": 4.755669115298627e-06,
|
| 97 |
+
"loss": 1.2495,
|
| 98 |
+
"step": 1500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.51,
|
| 102 |
+
"learning_rate": 5.075055892686043e-06,
|
| 103 |
+
"loss": 1.2039,
|
| 104 |
+
"step": 1600
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.54,
|
| 108 |
+
"learning_rate": 5.39444267007346e-06,
|
| 109 |
+
"loss": 1.1438,
|
| 110 |
+
"step": 1700
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.57,
|
| 114 |
+
"learning_rate": 5.713829447460876e-06,
|
| 115 |
+
"loss": 1.1056,
|
| 116 |
+
"step": 1800
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.61,
|
| 120 |
+
"learning_rate": 6.033216224848291e-06,
|
| 121 |
+
"loss": 1.0436,
|
| 122 |
+
"step": 1900
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.64,
|
| 126 |
+
"learning_rate": 6.352603002235708e-06,
|
| 127 |
+
"loss": 1.0225,
|
| 128 |
+
"step": 2000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.67,
|
| 132 |
+
"learning_rate": 6.671989779623123e-06,
|
| 133 |
+
"loss": 0.9856,
|
| 134 |
+
"step": 2100
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.7,
|
| 138 |
+
"learning_rate": 6.9913765570105406e-06,
|
| 139 |
+
"loss": 0.9412,
|
| 140 |
+
"step": 2200
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.73,
|
| 144 |
+
"learning_rate": 7.310763334397957e-06,
|
| 145 |
+
"loss": 0.9075,
|
| 146 |
+
"step": 2300
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.77,
|
| 150 |
+
"learning_rate": 7.630150111785373e-06,
|
| 151 |
+
"loss": 0.8752,
|
| 152 |
+
"step": 2400
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.8,
|
| 156 |
+
"learning_rate": 7.949536889172788e-06,
|
| 157 |
+
"loss": 0.854,
|
| 158 |
+
"step": 2500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.83,
|
| 162 |
+
"learning_rate": 8.268923666560205e-06,
|
| 163 |
+
"loss": 0.8282,
|
| 164 |
+
"step": 2600
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.86,
|
| 168 |
+
"learning_rate": 8.588310443947621e-06,
|
| 169 |
+
"loss": 0.8074,
|
| 170 |
+
"step": 2700
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.89,
|
| 174 |
+
"learning_rate": 8.907697221335038e-06,
|
| 175 |
+
"loss": 0.7976,
|
| 176 |
+
"step": 2800
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.93,
|
| 180 |
+
"learning_rate": 9.227083998722454e-06,
|
| 181 |
+
"loss": 0.7955,
|
| 182 |
+
"step": 2900
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.96,
|
| 186 |
+
"learning_rate": 9.54647077610987e-06,
|
| 187 |
+
"loss": 0.7502,
|
| 188 |
+
"step": 3000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.99,
|
| 192 |
+
"learning_rate": 9.865857553497285e-06,
|
| 193 |
+
"loss": 0.7455,
|
| 194 |
+
"step": 3100
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 1.02,
|
| 198 |
+
"learning_rate": 1.0185244330884702e-05,
|
| 199 |
+
"loss": 0.7177,
|
| 200 |
+
"step": 3200
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 1.05,
|
| 204 |
+
"learning_rate": 1.050463110827212e-05,
|
| 205 |
+
"loss": 0.724,
|
| 206 |
+
"step": 3300
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 1.09,
|
| 210 |
+
"learning_rate": 1.0824017885659535e-05,
|
| 211 |
+
"loss": 0.6975,
|
| 212 |
+
"step": 3400
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 1.12,
|
| 216 |
+
"learning_rate": 1.114340466304695e-05,
|
| 217 |
+
"loss": 0.7054,
|
| 218 |
+
"step": 3500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 1.15,
|
| 222 |
+
"learning_rate": 1.1462791440434366e-05,
|
| 223 |
+
"loss": 0.71,
|
| 224 |
+
"step": 3600
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 1.18,
|
| 228 |
+
"learning_rate": 1.1782178217821782e-05,
|
| 229 |
+
"loss": 0.6825,
|
| 230 |
+
"step": 3700
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 1.21,
|
| 234 |
+
"learning_rate": 1.2101564995209199e-05,
|
| 235 |
+
"loss": 0.6792,
|
| 236 |
+
"step": 3800
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 1.25,
|
| 240 |
+
"learning_rate": 1.2420951772596615e-05,
|
| 241 |
+
"loss": 0.6794,
|
| 242 |
+
"step": 3900
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 1.28,
|
| 246 |
+
"learning_rate": 1.2740338549984032e-05,
|
| 247 |
+
"loss": 0.6672,
|
| 248 |
+
"step": 4000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 1.31,
|
| 252 |
+
"learning_rate": 1.3059725327371449e-05,
|
| 253 |
+
"loss": 0.6356,
|
| 254 |
+
"step": 4100
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 1.34,
|
| 258 |
+
"learning_rate": 1.3379112104758865e-05,
|
| 259 |
+
"loss": 0.6503,
|
| 260 |
+
"step": 4200
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 1.37,
|
| 264 |
+
"learning_rate": 1.369849888214628e-05,
|
| 265 |
+
"loss": 0.6495,
|
| 266 |
+
"step": 4300
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 1.41,
|
| 270 |
+
"learning_rate": 1.4017885659533696e-05,
|
| 271 |
+
"loss": 0.6428,
|
| 272 |
+
"step": 4400
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 1.44,
|
| 276 |
+
"learning_rate": 1.4337272436921113e-05,
|
| 277 |
+
"loss": 0.6171,
|
| 278 |
+
"step": 4500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 1.47,
|
| 282 |
+
"learning_rate": 1.4656659214308529e-05,
|
| 283 |
+
"loss": 0.6084,
|
| 284 |
+
"step": 4600
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 1.5,
|
| 288 |
+
"learning_rate": 1.4976045991695944e-05,
|
| 289 |
+
"loss": 0.6222,
|
| 290 |
+
"step": 4700
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 1.53,
|
| 294 |
+
"learning_rate": 1.529543276908336e-05,
|
| 295 |
+
"loss": 0.5971,
|
| 296 |
+
"step": 4800
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 1.56,
|
| 300 |
+
"learning_rate": 1.5614819546470777e-05,
|
| 301 |
+
"loss": 0.5745,
|
| 302 |
+
"step": 4900
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 1.6,
|
| 306 |
+
"learning_rate": 1.5934206323858196e-05,
|
| 307 |
+
"loss": 0.5753,
|
| 308 |
+
"step": 5000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 1.63,
|
| 312 |
+
"learning_rate": 1.625359310124561e-05,
|
| 313 |
+
"loss": 0.6045,
|
| 314 |
+
"step": 5100
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 1.66,
|
| 318 |
+
"learning_rate": 1.6572979878633027e-05,
|
| 319 |
+
"loss": 0.6006,
|
| 320 |
+
"step": 5200
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 1.69,
|
| 324 |
+
"learning_rate": 1.6892366656020443e-05,
|
| 325 |
+
"loss": 0.6003,
|
| 326 |
+
"step": 5300
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 1.72,
|
| 330 |
+
"learning_rate": 1.7211753433407858e-05,
|
| 331 |
+
"loss": 0.5688,
|
| 332 |
+
"step": 5400
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 1.76,
|
| 336 |
+
"learning_rate": 1.7531140210795274e-05,
|
| 337 |
+
"loss": 0.5662,
|
| 338 |
+
"step": 5500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 1.79,
|
| 342 |
+
"learning_rate": 1.785052698818269e-05,
|
| 343 |
+
"loss": 0.5788,
|
| 344 |
+
"step": 5600
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 1.82,
|
| 348 |
+
"learning_rate": 1.8169913765570105e-05,
|
| 349 |
+
"loss": 0.5541,
|
| 350 |
+
"step": 5700
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 1.85,
|
| 354 |
+
"learning_rate": 1.8489300542957524e-05,
|
| 355 |
+
"loss": 0.5656,
|
| 356 |
+
"step": 5800
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 1.88,
|
| 360 |
+
"learning_rate": 1.880868732034494e-05,
|
| 361 |
+
"loss": 0.5466,
|
| 362 |
+
"step": 5900
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 1.92,
|
| 366 |
+
"learning_rate": 1.9128074097732355e-05,
|
| 367 |
+
"loss": 0.563,
|
| 368 |
+
"step": 6000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 1.95,
|
| 372 |
+
"learning_rate": 1.9447460875119774e-05,
|
| 373 |
+
"loss": 0.5603,
|
| 374 |
+
"step": 6100
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 1.98,
|
| 378 |
+
"learning_rate": 1.976684765250719e-05,
|
| 379 |
+
"loss": 0.5579,
|
| 380 |
+
"step": 6200
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 2.01,
|
| 384 |
+
"learning_rate": 2.0086234429894605e-05,
|
| 385 |
+
"loss": 0.5346,
|
| 386 |
+
"step": 6300
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 2.04,
|
| 390 |
+
"learning_rate": 2.0405621207282017e-05,
|
| 391 |
+
"loss": 0.5038,
|
| 392 |
+
"step": 6400
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 2.08,
|
| 396 |
+
"learning_rate": 2.0725007984669436e-05,
|
| 397 |
+
"loss": 0.5121,
|
| 398 |
+
"step": 6500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 2.11,
|
| 402 |
+
"learning_rate": 2.1044394762056855e-05,
|
| 403 |
+
"loss": 0.5135,
|
| 404 |
+
"step": 6600
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 2.14,
|
| 408 |
+
"learning_rate": 2.1363781539444267e-05,
|
| 409 |
+
"loss": 0.504,
|
| 410 |
+
"step": 6700
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 2.17,
|
| 414 |
+
"learning_rate": 2.1683168316831686e-05,
|
| 415 |
+
"loss": 0.5119,
|
| 416 |
+
"step": 6800
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 2.2,
|
| 420 |
+
"learning_rate": 2.20025550942191e-05,
|
| 421 |
+
"loss": 0.4986,
|
| 422 |
+
"step": 6900
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 2.24,
|
| 426 |
+
"learning_rate": 2.2321941871606518e-05,
|
| 427 |
+
"loss": 0.504,
|
| 428 |
+
"step": 7000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 2.27,
|
| 432 |
+
"learning_rate": 2.2641328648993933e-05,
|
| 433 |
+
"loss": 0.5024,
|
| 434 |
+
"step": 7100
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 2.3,
|
| 438 |
+
"learning_rate": 2.296071542638135e-05,
|
| 439 |
+
"loss": 0.479,
|
| 440 |
+
"step": 7200
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 2.33,
|
| 444 |
+
"learning_rate": 2.3280102203768764e-05,
|
| 445 |
+
"loss": 0.492,
|
| 446 |
+
"step": 7300
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 2.36,
|
| 450 |
+
"learning_rate": 2.3599488981156183e-05,
|
| 451 |
+
"loss": 0.4928,
|
| 452 |
+
"step": 7400
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 2.4,
|
| 456 |
+
"learning_rate": 2.39188757585436e-05,
|
| 457 |
+
"loss": 0.4862,
|
| 458 |
+
"step": 7500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 2.43,
|
| 462 |
+
"learning_rate": 2.4238262535931014e-05,
|
| 463 |
+
"loss": 0.5148,
|
| 464 |
+
"step": 7600
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 2.46,
|
| 468 |
+
"learning_rate": 2.4557649313318433e-05,
|
| 469 |
+
"loss": 0.4922,
|
| 470 |
+
"step": 7700
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 2.49,
|
| 474 |
+
"learning_rate": 2.4877036090705846e-05,
|
| 475 |
+
"loss": 0.4882,
|
| 476 |
+
"step": 7800
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 2.52,
|
| 480 |
+
"learning_rate": 2.5196422868093264e-05,
|
| 481 |
+
"loss": 0.4871,
|
| 482 |
+
"step": 7900
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 2.56,
|
| 486 |
+
"learning_rate": 2.5515809645480677e-05,
|
| 487 |
+
"loss": 0.4967,
|
| 488 |
+
"step": 8000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 2.59,
|
| 492 |
+
"learning_rate": 2.5835196422868096e-05,
|
| 493 |
+
"loss": 0.4834,
|
| 494 |
+
"step": 8100
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 2.62,
|
| 498 |
+
"learning_rate": 2.615458320025551e-05,
|
| 499 |
+
"loss": 0.4769,
|
| 500 |
+
"step": 8200
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 2.65,
|
| 504 |
+
"learning_rate": 2.6473969977642927e-05,
|
| 505 |
+
"loss": 0.4776,
|
| 506 |
+
"step": 8300
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 2.68,
|
| 510 |
+
"learning_rate": 2.6793356755030342e-05,
|
| 511 |
+
"loss": 0.4861,
|
| 512 |
+
"step": 8400
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 2.71,
|
| 516 |
+
"learning_rate": 2.711274353241776e-05,
|
| 517 |
+
"loss": 0.486,
|
| 518 |
+
"step": 8500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 2.75,
|
| 522 |
+
"learning_rate": 2.7432130309805177e-05,
|
| 523 |
+
"loss": 0.4745,
|
| 524 |
+
"step": 8600
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 2.78,
|
| 528 |
+
"learning_rate": 2.7751517087192592e-05,
|
| 529 |
+
"loss": 0.4805,
|
| 530 |
+
"step": 8700
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 2.81,
|
| 534 |
+
"learning_rate": 2.807090386458001e-05,
|
| 535 |
+
"loss": 0.5003,
|
| 536 |
+
"step": 8800
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 2.84,
|
| 540 |
+
"learning_rate": 2.8390290641967424e-05,
|
| 541 |
+
"loss": 0.4639,
|
| 542 |
+
"step": 8900
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 2.87,
|
| 546 |
+
"learning_rate": 2.8709677419354843e-05,
|
| 547 |
+
"loss": 0.4763,
|
| 548 |
+
"step": 9000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 2.91,
|
| 552 |
+
"learning_rate": 2.9029064196742255e-05,
|
| 553 |
+
"loss": 0.4727,
|
| 554 |
+
"step": 9100
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 2.94,
|
| 558 |
+
"learning_rate": 2.9348450974129674e-05,
|
| 559 |
+
"loss": 0.4839,
|
| 560 |
+
"step": 9200
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 2.97,
|
| 564 |
+
"learning_rate": 2.966783775151709e-05,
|
| 565 |
+
"loss": 0.4769,
|
| 566 |
+
"step": 9300
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 3.0,
|
| 570 |
+
"learning_rate": 2.9987224528904505e-05,
|
| 571 |
+
"loss": 0.4641,
|
| 572 |
+
"step": 9400
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 3.03,
|
| 576 |
+
"learning_rate": 3.030661130629192e-05,
|
| 577 |
+
"loss": 0.4282,
|
| 578 |
+
"step": 9500
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 3.07,
|
| 582 |
+
"learning_rate": 3.0625998083679336e-05,
|
| 583 |
+
"loss": 0.4057,
|
| 584 |
+
"step": 9600
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 3.1,
|
| 588 |
+
"learning_rate": 3.094538486106676e-05,
|
| 589 |
+
"loss": 0.4192,
|
| 590 |
+
"step": 9700
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 3.13,
|
| 594 |
+
"learning_rate": 3.126477163845417e-05,
|
| 595 |
+
"loss": 0.4186,
|
| 596 |
+
"step": 9800
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 3.16,
|
| 600 |
+
"learning_rate": 3.158415841584159e-05,
|
| 601 |
+
"loss": 0.4086,
|
| 602 |
+
"step": 9900
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 3.19,
|
| 606 |
+
"learning_rate": 3.1903545193229005e-05,
|
| 607 |
+
"loss": 0.4194,
|
| 608 |
+
"step": 10000
|
| 609 |
+
}
|
| 610 |
+
],
|
| 611 |
+
"max_steps": 125240,
|
| 612 |
+
"num_train_epochs": 40,
|
| 613 |
+
"total_flos": 0.0,
|
| 614 |
+
"trial_name": null,
|
| 615 |
+
"trial_params": null
|
| 616 |
+
}
|
checkpoint-100000/config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "Davlan/afro-xlmr-base",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"XLMRobertaModel"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 768,
|
| 13 |
+
"id2label": {
|
| 14 |
+
"0": "LABEL_0"
|
| 15 |
+
},
|
| 16 |
+
"initializer_range": 0.02,
|
| 17 |
+
"intermediate_size": 3072,
|
| 18 |
+
"label2id": {
|
| 19 |
+
"LABEL_0": 0
|
| 20 |
+
},
|
| 21 |
+
"layer_norm_eps": 1e-05,
|
| 22 |
+
"max_position_embeddings": 514,
|
| 23 |
+
"model_type": "xlm-roberta",
|
| 24 |
+
"num_attention_heads": 12,
|
| 25 |
+
"num_hidden_layers": 12,
|
| 26 |
+
"output_past": true,
|
| 27 |
+
"pad_token_id": 1,
|
| 28 |
+
"position_embedding_type": "absolute",
|
| 29 |
+
"torch_dtype": "float32",
|
| 30 |
+
"transformers_version": "4.29.0",
|
| 31 |
+
"type_vocab_size": 1,
|
| 32 |
+
"use_cache": true,
|
| 33 |
+
"vocab_size": 250002
|
| 34 |
+
}
|
checkpoint-100000/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0ac23e739f01083499bb41488fd3971c50df1c724a9a8ea243d03d25f7be2103
|
| 3 |
+
size 1112246250
|
checkpoint-100000/trainer_state.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-105000/config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "Davlan/afro-xlmr-base",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"XLMRobertaModel"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 768,
|
| 13 |
+
"id2label": {
|
| 14 |
+
"0": "LABEL_0"
|
| 15 |
+
},
|
| 16 |
+
"initializer_range": 0.02,
|
| 17 |
+
"intermediate_size": 3072,
|
| 18 |
+
"label2id": {
|
| 19 |
+
"LABEL_0": 0
|
| 20 |
+
},
|
| 21 |
+
"layer_norm_eps": 1e-05,
|
| 22 |
+
"max_position_embeddings": 514,
|
| 23 |
+
"model_type": "xlm-roberta",
|
| 24 |
+
"num_attention_heads": 12,
|
| 25 |
+
"num_hidden_layers": 12,
|
| 26 |
+
"output_past": true,
|
| 27 |
+
"pad_token_id": 1,
|
| 28 |
+
"position_embedding_type": "absolute",
|
| 29 |
+
"torch_dtype": "float32",
|
| 30 |
+
"transformers_version": "4.29.0",
|
| 31 |
+
"type_vocab_size": 1,
|
| 32 |
+
"use_cache": true,
|
| 33 |
+
"vocab_size": 250002
|
| 34 |
+
}
|
checkpoint-105000/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c2ab5df5740a5d60afcb315107feae275fe5798e24d6474cdf84e1c63a6bf623
|
| 3 |
+
size 1112246250
|
checkpoint-105000/trainer_state.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-110000/config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "Davlan/afro-xlmr-base",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"XLMRobertaModel"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 768,
|
| 13 |
+
"id2label": {
|
| 14 |
+
"0": "LABEL_0"
|
| 15 |
+
},
|
| 16 |
+
"initializer_range": 0.02,
|
| 17 |
+
"intermediate_size": 3072,
|
| 18 |
+
"label2id": {
|
| 19 |
+
"LABEL_0": 0
|
| 20 |
+
},
|
| 21 |
+
"layer_norm_eps": 1e-05,
|
| 22 |
+
"max_position_embeddings": 514,
|
| 23 |
+
"model_type": "xlm-roberta",
|
| 24 |
+
"num_attention_heads": 12,
|
| 25 |
+
"num_hidden_layers": 12,
|
| 26 |
+
"output_past": true,
|
| 27 |
+
"pad_token_id": 1,
|
| 28 |
+
"position_embedding_type": "absolute",
|
| 29 |
+
"torch_dtype": "float32",
|
| 30 |
+
"transformers_version": "4.29.0",
|
| 31 |
+
"type_vocab_size": 1,
|
| 32 |
+
"use_cache": true,
|
| 33 |
+
"vocab_size": 250002
|
| 34 |
+
}
|
checkpoint-110000/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:23b3e5de38dddb5b95c9dff2473e8a4eb364518075b45b9a8459ea1b08188dd3
|
| 3 |
+
size 1112246250
|
checkpoint-110000/trainer_state.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-115000/config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "Davlan/afro-xlmr-base",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"XLMRobertaModel"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 768,
|
| 13 |
+
"id2label": {
|
| 14 |
+
"0": "LABEL_0"
|
| 15 |
+
},
|
| 16 |
+
"initializer_range": 0.02,
|
| 17 |
+
"intermediate_size": 3072,
|
| 18 |
+
"label2id": {
|
| 19 |
+
"LABEL_0": 0
|
| 20 |
+
},
|
| 21 |
+
"layer_norm_eps": 1e-05,
|
| 22 |
+
"max_position_embeddings": 514,
|
| 23 |
+
"model_type": "xlm-roberta",
|
| 24 |
+
"num_attention_heads": 12,
|
| 25 |
+
"num_hidden_layers": 12,
|
| 26 |
+
"output_past": true,
|
| 27 |
+
"pad_token_id": 1,
|
| 28 |
+
"position_embedding_type": "absolute",
|
| 29 |
+
"torch_dtype": "float32",
|
| 30 |
+
"transformers_version": "4.29.0",
|
| 31 |
+
"type_vocab_size": 1,
|
| 32 |
+
"use_cache": true,
|
| 33 |
+
"vocab_size": 250002
|
| 34 |
+
}
|
checkpoint-115000/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4f78199abdb1e7e7ef07adc5060981f13fc373a35f8968383da22ae6f547651b
|
| 3 |
+
size 1112246250
|
checkpoint-115000/trainer_state.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-120000/config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "Davlan/afro-xlmr-base",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"XLMRobertaModel"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 768,
|
| 13 |
+
"id2label": {
|
| 14 |
+
"0": "LABEL_0"
|
| 15 |
+
},
|
| 16 |
+
"initializer_range": 0.02,
|
| 17 |
+
"intermediate_size": 3072,
|
| 18 |
+
"label2id": {
|
| 19 |
+
"LABEL_0": 0
|
| 20 |
+
},
|
| 21 |
+
"layer_norm_eps": 1e-05,
|
| 22 |
+
"max_position_embeddings": 514,
|
| 23 |
+
"model_type": "xlm-roberta",
|
| 24 |
+
"num_attention_heads": 12,
|
| 25 |
+
"num_hidden_layers": 12,
|
| 26 |
+
"output_past": true,
|
| 27 |
+
"pad_token_id": 1,
|
| 28 |
+
"position_embedding_type": "absolute",
|
| 29 |
+
"torch_dtype": "float32",
|
| 30 |
+
"transformers_version": "4.29.0",
|
| 31 |
+
"type_vocab_size": 1,
|
| 32 |
+
"use_cache": true,
|
| 33 |
+
"vocab_size": 250002
|
| 34 |
+
}
|
checkpoint-120000/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c22aa854643305ed7f6b163c95afca7977dee986c5d87d314303ec18c922f21f
|
| 3 |
+
size 1112246250
|
checkpoint-120000/trainer_state.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-125000/config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "Davlan/afro-xlmr-base",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"XLMRobertaModel"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 768,
|
| 13 |
+
"id2label": {
|
| 14 |
+
"0": "LABEL_0"
|
| 15 |
+
},
|
| 16 |
+
"initializer_range": 0.02,
|
| 17 |
+
"intermediate_size": 3072,
|
| 18 |
+
"label2id": {
|
| 19 |
+
"LABEL_0": 0
|
| 20 |
+
},
|
| 21 |
+
"layer_norm_eps": 1e-05,
|
| 22 |
+
"max_position_embeddings": 514,
|
| 23 |
+
"model_type": "xlm-roberta",
|
| 24 |
+
"num_attention_heads": 12,
|
| 25 |
+
"num_hidden_layers": 12,
|
| 26 |
+
"output_past": true,
|
| 27 |
+
"pad_token_id": 1,
|
| 28 |
+
"position_embedding_type": "absolute",
|
| 29 |
+
"torch_dtype": "float32",
|
| 30 |
+
"transformers_version": "4.29.0",
|
| 31 |
+
"type_vocab_size": 1,
|
| 32 |
+
"use_cache": true,
|
| 33 |
+
"vocab_size": 250002
|
| 34 |
+
}
|
checkpoint-125000/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8b92502575ab1a45ada94bf14f2e51029ec598081fab56cd7660d21d30aff386
|
| 3 |
+
size 1112246250
|
checkpoint-125000/trainer_state.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-15000/config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "Davlan/afro-xlmr-base",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"XLMRobertaModel"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 768,
|
| 13 |
+
"id2label": {
|
| 14 |
+
"0": "LABEL_0"
|
| 15 |
+
},
|
| 16 |
+
"initializer_range": 0.02,
|
| 17 |
+
"intermediate_size": 3072,
|
| 18 |
+
"label2id": {
|
| 19 |
+
"LABEL_0": 0
|
| 20 |
+
},
|
| 21 |
+
"layer_norm_eps": 1e-05,
|
| 22 |
+
"max_position_embeddings": 514,
|
| 23 |
+
"model_type": "xlm-roberta",
|
| 24 |
+
"num_attention_heads": 12,
|
| 25 |
+
"num_hidden_layers": 12,
|
| 26 |
+
"output_past": true,
|
| 27 |
+
"pad_token_id": 1,
|
| 28 |
+
"position_embedding_type": "absolute",
|
| 29 |
+
"torch_dtype": "float32",
|
| 30 |
+
"transformers_version": "4.29.0",
|
| 31 |
+
"type_vocab_size": 1,
|
| 32 |
+
"use_cache": true,
|
| 33 |
+
"vocab_size": 250002
|
| 34 |
+
}
|
checkpoint-15000/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3cb359e3eb6774638f32324a9d457669c6c0969aca5d121ad0ed41ef9f8b5982
|
| 3 |
+
size 1112246250
|
checkpoint-15000/trainer_state.json
ADDED
|
@@ -0,0 +1,916 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 4.790801660811242,
|
| 5 |
+
"global_step": 15000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.03,
|
| 12 |
+
"learning_rate": 2.8744809964867454e-07,
|
| 13 |
+
"loss": 91.176,
|
| 14 |
+
"step": 100
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.06,
|
| 18 |
+
"learning_rate": 6.068348770360907e-07,
|
| 19 |
+
"loss": 64.0684,
|
| 20 |
+
"step": 200
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.1,
|
| 24 |
+
"learning_rate": 9.26221654423507e-07,
|
| 25 |
+
"loss": 27.1061,
|
| 26 |
+
"step": 300
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.13,
|
| 30 |
+
"learning_rate": 1.242414564037049e-06,
|
| 31 |
+
"loss": 8.0198,
|
| 32 |
+
"step": 400
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.16,
|
| 36 |
+
"learning_rate": 1.561801341424465e-06,
|
| 37 |
+
"loss": 4.7691,
|
| 38 |
+
"step": 500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.19,
|
| 42 |
+
"learning_rate": 1.8811881188118813e-06,
|
| 43 |
+
"loss": 3.4387,
|
| 44 |
+
"step": 600
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.22,
|
| 48 |
+
"learning_rate": 2.2005748961992977e-06,
|
| 49 |
+
"loss": 2.7027,
|
| 50 |
+
"step": 700
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.26,
|
| 54 |
+
"learning_rate": 2.5199616735867137e-06,
|
| 55 |
+
"loss": 2.4431,
|
| 56 |
+
"step": 800
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.29,
|
| 60 |
+
"learning_rate": 2.8393484509741297e-06,
|
| 61 |
+
"loss": 2.0534,
|
| 62 |
+
"step": 900
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.32,
|
| 66 |
+
"learning_rate": 3.1587352283615465e-06,
|
| 67 |
+
"loss": 1.8537,
|
| 68 |
+
"step": 1000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.35,
|
| 72 |
+
"learning_rate": 3.4781220057489625e-06,
|
| 73 |
+
"loss": 1.642,
|
| 74 |
+
"step": 1100
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.38,
|
| 78 |
+
"learning_rate": 3.797508783136378e-06,
|
| 79 |
+
"loss": 1.5092,
|
| 80 |
+
"step": 1200
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.42,
|
| 84 |
+
"learning_rate": 4.1168955605237945e-06,
|
| 85 |
+
"loss": 1.4302,
|
| 86 |
+
"step": 1300
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.45,
|
| 90 |
+
"learning_rate": 4.436282337911211e-06,
|
| 91 |
+
"loss": 1.3527,
|
| 92 |
+
"step": 1400
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.48,
|
| 96 |
+
"learning_rate": 4.755669115298627e-06,
|
| 97 |
+
"loss": 1.2495,
|
| 98 |
+
"step": 1500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.51,
|
| 102 |
+
"learning_rate": 5.075055892686043e-06,
|
| 103 |
+
"loss": 1.2039,
|
| 104 |
+
"step": 1600
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.54,
|
| 108 |
+
"learning_rate": 5.39444267007346e-06,
|
| 109 |
+
"loss": 1.1438,
|
| 110 |
+
"step": 1700
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.57,
|
| 114 |
+
"learning_rate": 5.713829447460876e-06,
|
| 115 |
+
"loss": 1.1056,
|
| 116 |
+
"step": 1800
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.61,
|
| 120 |
+
"learning_rate": 6.033216224848291e-06,
|
| 121 |
+
"loss": 1.0436,
|
| 122 |
+
"step": 1900
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.64,
|
| 126 |
+
"learning_rate": 6.352603002235708e-06,
|
| 127 |
+
"loss": 1.0225,
|
| 128 |
+
"step": 2000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.67,
|
| 132 |
+
"learning_rate": 6.671989779623123e-06,
|
| 133 |
+
"loss": 0.9856,
|
| 134 |
+
"step": 2100
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.7,
|
| 138 |
+
"learning_rate": 6.9913765570105406e-06,
|
| 139 |
+
"loss": 0.9412,
|
| 140 |
+
"step": 2200
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.73,
|
| 144 |
+
"learning_rate": 7.310763334397957e-06,
|
| 145 |
+
"loss": 0.9075,
|
| 146 |
+
"step": 2300
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.77,
|
| 150 |
+
"learning_rate": 7.630150111785373e-06,
|
| 151 |
+
"loss": 0.8752,
|
| 152 |
+
"step": 2400
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.8,
|
| 156 |
+
"learning_rate": 7.949536889172788e-06,
|
| 157 |
+
"loss": 0.854,
|
| 158 |
+
"step": 2500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.83,
|
| 162 |
+
"learning_rate": 8.268923666560205e-06,
|
| 163 |
+
"loss": 0.8282,
|
| 164 |
+
"step": 2600
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.86,
|
| 168 |
+
"learning_rate": 8.588310443947621e-06,
|
| 169 |
+
"loss": 0.8074,
|
| 170 |
+
"step": 2700
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.89,
|
| 174 |
+
"learning_rate": 8.907697221335038e-06,
|
| 175 |
+
"loss": 0.7976,
|
| 176 |
+
"step": 2800
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.93,
|
| 180 |
+
"learning_rate": 9.227083998722454e-06,
|
| 181 |
+
"loss": 0.7955,
|
| 182 |
+
"step": 2900
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.96,
|
| 186 |
+
"learning_rate": 9.54647077610987e-06,
|
| 187 |
+
"loss": 0.7502,
|
| 188 |
+
"step": 3000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.99,
|
| 192 |
+
"learning_rate": 9.865857553497285e-06,
|
| 193 |
+
"loss": 0.7455,
|
| 194 |
+
"step": 3100
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 1.02,
|
| 198 |
+
"learning_rate": 1.0185244330884702e-05,
|
| 199 |
+
"loss": 0.7177,
|
| 200 |
+
"step": 3200
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 1.05,
|
| 204 |
+
"learning_rate": 1.050463110827212e-05,
|
| 205 |
+
"loss": 0.724,
|
| 206 |
+
"step": 3300
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 1.09,
|
| 210 |
+
"learning_rate": 1.0824017885659535e-05,
|
| 211 |
+
"loss": 0.6975,
|
| 212 |
+
"step": 3400
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 1.12,
|
| 216 |
+
"learning_rate": 1.114340466304695e-05,
|
| 217 |
+
"loss": 0.7054,
|
| 218 |
+
"step": 3500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 1.15,
|
| 222 |
+
"learning_rate": 1.1462791440434366e-05,
|
| 223 |
+
"loss": 0.71,
|
| 224 |
+
"step": 3600
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 1.18,
|
| 228 |
+
"learning_rate": 1.1782178217821782e-05,
|
| 229 |
+
"loss": 0.6825,
|
| 230 |
+
"step": 3700
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 1.21,
|
| 234 |
+
"learning_rate": 1.2101564995209199e-05,
|
| 235 |
+
"loss": 0.6792,
|
| 236 |
+
"step": 3800
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 1.25,
|
| 240 |
+
"learning_rate": 1.2420951772596615e-05,
|
| 241 |
+
"loss": 0.6794,
|
| 242 |
+
"step": 3900
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 1.28,
|
| 246 |
+
"learning_rate": 1.2740338549984032e-05,
|
| 247 |
+
"loss": 0.6672,
|
| 248 |
+
"step": 4000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 1.31,
|
| 252 |
+
"learning_rate": 1.3059725327371449e-05,
|
| 253 |
+
"loss": 0.6356,
|
| 254 |
+
"step": 4100
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 1.34,
|
| 258 |
+
"learning_rate": 1.3379112104758865e-05,
|
| 259 |
+
"loss": 0.6503,
|
| 260 |
+
"step": 4200
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 1.37,
|
| 264 |
+
"learning_rate": 1.369849888214628e-05,
|
| 265 |
+
"loss": 0.6495,
|
| 266 |
+
"step": 4300
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 1.41,
|
| 270 |
+
"learning_rate": 1.4017885659533696e-05,
|
| 271 |
+
"loss": 0.6428,
|
| 272 |
+
"step": 4400
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 1.44,
|
| 276 |
+
"learning_rate": 1.4337272436921113e-05,
|
| 277 |
+
"loss": 0.6171,
|
| 278 |
+
"step": 4500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 1.47,
|
| 282 |
+
"learning_rate": 1.4656659214308529e-05,
|
| 283 |
+
"loss": 0.6084,
|
| 284 |
+
"step": 4600
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 1.5,
|
| 288 |
+
"learning_rate": 1.4976045991695944e-05,
|
| 289 |
+
"loss": 0.6222,
|
| 290 |
+
"step": 4700
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 1.53,
|
| 294 |
+
"learning_rate": 1.529543276908336e-05,
|
| 295 |
+
"loss": 0.5971,
|
| 296 |
+
"step": 4800
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 1.56,
|
| 300 |
+
"learning_rate": 1.5614819546470777e-05,
|
| 301 |
+
"loss": 0.5745,
|
| 302 |
+
"step": 4900
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 1.6,
|
| 306 |
+
"learning_rate": 1.5934206323858196e-05,
|
| 307 |
+
"loss": 0.5753,
|
| 308 |
+
"step": 5000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 1.63,
|
| 312 |
+
"learning_rate": 1.625359310124561e-05,
|
| 313 |
+
"loss": 0.6045,
|
| 314 |
+
"step": 5100
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 1.66,
|
| 318 |
+
"learning_rate": 1.6572979878633027e-05,
|
| 319 |
+
"loss": 0.6006,
|
| 320 |
+
"step": 5200
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 1.69,
|
| 324 |
+
"learning_rate": 1.6892366656020443e-05,
|
| 325 |
+
"loss": 0.6003,
|
| 326 |
+
"step": 5300
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 1.72,
|
| 330 |
+
"learning_rate": 1.7211753433407858e-05,
|
| 331 |
+
"loss": 0.5688,
|
| 332 |
+
"step": 5400
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 1.76,
|
| 336 |
+
"learning_rate": 1.7531140210795274e-05,
|
| 337 |
+
"loss": 0.5662,
|
| 338 |
+
"step": 5500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 1.79,
|
| 342 |
+
"learning_rate": 1.785052698818269e-05,
|
| 343 |
+
"loss": 0.5788,
|
| 344 |
+
"step": 5600
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 1.82,
|
| 348 |
+
"learning_rate": 1.8169913765570105e-05,
|
| 349 |
+
"loss": 0.5541,
|
| 350 |
+
"step": 5700
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 1.85,
|
| 354 |
+
"learning_rate": 1.8489300542957524e-05,
|
| 355 |
+
"loss": 0.5656,
|
| 356 |
+
"step": 5800
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 1.88,
|
| 360 |
+
"learning_rate": 1.880868732034494e-05,
|
| 361 |
+
"loss": 0.5466,
|
| 362 |
+
"step": 5900
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 1.92,
|
| 366 |
+
"learning_rate": 1.9128074097732355e-05,
|
| 367 |
+
"loss": 0.563,
|
| 368 |
+
"step": 6000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 1.95,
|
| 372 |
+
"learning_rate": 1.9447460875119774e-05,
|
| 373 |
+
"loss": 0.5603,
|
| 374 |
+
"step": 6100
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 1.98,
|
| 378 |
+
"learning_rate": 1.976684765250719e-05,
|
| 379 |
+
"loss": 0.5579,
|
| 380 |
+
"step": 6200
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 2.01,
|
| 384 |
+
"learning_rate": 2.0086234429894605e-05,
|
| 385 |
+
"loss": 0.5346,
|
| 386 |
+
"step": 6300
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 2.04,
|
| 390 |
+
"learning_rate": 2.0405621207282017e-05,
|
| 391 |
+
"loss": 0.5038,
|
| 392 |
+
"step": 6400
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 2.08,
|
| 396 |
+
"learning_rate": 2.0725007984669436e-05,
|
| 397 |
+
"loss": 0.5121,
|
| 398 |
+
"step": 6500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 2.11,
|
| 402 |
+
"learning_rate": 2.1044394762056855e-05,
|
| 403 |
+
"loss": 0.5135,
|
| 404 |
+
"step": 6600
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 2.14,
|
| 408 |
+
"learning_rate": 2.1363781539444267e-05,
|
| 409 |
+
"loss": 0.504,
|
| 410 |
+
"step": 6700
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 2.17,
|
| 414 |
+
"learning_rate": 2.1683168316831686e-05,
|
| 415 |
+
"loss": 0.5119,
|
| 416 |
+
"step": 6800
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 2.2,
|
| 420 |
+
"learning_rate": 2.20025550942191e-05,
|
| 421 |
+
"loss": 0.4986,
|
| 422 |
+
"step": 6900
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 2.24,
|
| 426 |
+
"learning_rate": 2.2321941871606518e-05,
|
| 427 |
+
"loss": 0.504,
|
| 428 |
+
"step": 7000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 2.27,
|
| 432 |
+
"learning_rate": 2.2641328648993933e-05,
|
| 433 |
+
"loss": 0.5024,
|
| 434 |
+
"step": 7100
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 2.3,
|
| 438 |
+
"learning_rate": 2.296071542638135e-05,
|
| 439 |
+
"loss": 0.479,
|
| 440 |
+
"step": 7200
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 2.33,
|
| 444 |
+
"learning_rate": 2.3280102203768764e-05,
|
| 445 |
+
"loss": 0.492,
|
| 446 |
+
"step": 7300
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 2.36,
|
| 450 |
+
"learning_rate": 2.3599488981156183e-05,
|
| 451 |
+
"loss": 0.4928,
|
| 452 |
+
"step": 7400
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 2.4,
|
| 456 |
+
"learning_rate": 2.39188757585436e-05,
|
| 457 |
+
"loss": 0.4862,
|
| 458 |
+
"step": 7500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 2.43,
|
| 462 |
+
"learning_rate": 2.4238262535931014e-05,
|
| 463 |
+
"loss": 0.5148,
|
| 464 |
+
"step": 7600
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 2.46,
|
| 468 |
+
"learning_rate": 2.4557649313318433e-05,
|
| 469 |
+
"loss": 0.4922,
|
| 470 |
+
"step": 7700
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 2.49,
|
| 474 |
+
"learning_rate": 2.4877036090705846e-05,
|
| 475 |
+
"loss": 0.4882,
|
| 476 |
+
"step": 7800
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 2.52,
|
| 480 |
+
"learning_rate": 2.5196422868093264e-05,
|
| 481 |
+
"loss": 0.4871,
|
| 482 |
+
"step": 7900
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 2.56,
|
| 486 |
+
"learning_rate": 2.5515809645480677e-05,
|
| 487 |
+
"loss": 0.4967,
|
| 488 |
+
"step": 8000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 2.59,
|
| 492 |
+
"learning_rate": 2.5835196422868096e-05,
|
| 493 |
+
"loss": 0.4834,
|
| 494 |
+
"step": 8100
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 2.62,
|
| 498 |
+
"learning_rate": 2.615458320025551e-05,
|
| 499 |
+
"loss": 0.4769,
|
| 500 |
+
"step": 8200
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 2.65,
|
| 504 |
+
"learning_rate": 2.6473969977642927e-05,
|
| 505 |
+
"loss": 0.4776,
|
| 506 |
+
"step": 8300
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 2.68,
|
| 510 |
+
"learning_rate": 2.6793356755030342e-05,
|
| 511 |
+
"loss": 0.4861,
|
| 512 |
+
"step": 8400
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 2.71,
|
| 516 |
+
"learning_rate": 2.711274353241776e-05,
|
| 517 |
+
"loss": 0.486,
|
| 518 |
+
"step": 8500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 2.75,
|
| 522 |
+
"learning_rate": 2.7432130309805177e-05,
|
| 523 |
+
"loss": 0.4745,
|
| 524 |
+
"step": 8600
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 2.78,
|
| 528 |
+
"learning_rate": 2.7751517087192592e-05,
|
| 529 |
+
"loss": 0.4805,
|
| 530 |
+
"step": 8700
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 2.81,
|
| 534 |
+
"learning_rate": 2.807090386458001e-05,
|
| 535 |
+
"loss": 0.5003,
|
| 536 |
+
"step": 8800
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 2.84,
|
| 540 |
+
"learning_rate": 2.8390290641967424e-05,
|
| 541 |
+
"loss": 0.4639,
|
| 542 |
+
"step": 8900
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 2.87,
|
| 546 |
+
"learning_rate": 2.8709677419354843e-05,
|
| 547 |
+
"loss": 0.4763,
|
| 548 |
+
"step": 9000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 2.91,
|
| 552 |
+
"learning_rate": 2.9029064196742255e-05,
|
| 553 |
+
"loss": 0.4727,
|
| 554 |
+
"step": 9100
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 2.94,
|
| 558 |
+
"learning_rate": 2.9348450974129674e-05,
|
| 559 |
+
"loss": 0.4839,
|
| 560 |
+
"step": 9200
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 2.97,
|
| 564 |
+
"learning_rate": 2.966783775151709e-05,
|
| 565 |
+
"loss": 0.4769,
|
| 566 |
+
"step": 9300
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 3.0,
|
| 570 |
+
"learning_rate": 2.9987224528904505e-05,
|
| 571 |
+
"loss": 0.4641,
|
| 572 |
+
"step": 9400
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 3.03,
|
| 576 |
+
"learning_rate": 3.030661130629192e-05,
|
| 577 |
+
"loss": 0.4282,
|
| 578 |
+
"step": 9500
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 3.07,
|
| 582 |
+
"learning_rate": 3.0625998083679336e-05,
|
| 583 |
+
"loss": 0.4057,
|
| 584 |
+
"step": 9600
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 3.1,
|
| 588 |
+
"learning_rate": 3.094538486106676e-05,
|
| 589 |
+
"loss": 0.4192,
|
| 590 |
+
"step": 9700
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 3.13,
|
| 594 |
+
"learning_rate": 3.126477163845417e-05,
|
| 595 |
+
"loss": 0.4186,
|
| 596 |
+
"step": 9800
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 3.16,
|
| 600 |
+
"learning_rate": 3.158415841584159e-05,
|
| 601 |
+
"loss": 0.4086,
|
| 602 |
+
"step": 9900
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 3.19,
|
| 606 |
+
"learning_rate": 3.1903545193229005e-05,
|
| 607 |
+
"loss": 0.4194,
|
| 608 |
+
"step": 10000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 3.23,
|
| 612 |
+
"learning_rate": 3.221973810284254e-05,
|
| 613 |
+
"loss": 0.4443,
|
| 614 |
+
"step": 10100
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 3.26,
|
| 618 |
+
"learning_rate": 3.2539124880229965e-05,
|
| 619 |
+
"loss": 0.4217,
|
| 620 |
+
"step": 10200
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 3.29,
|
| 624 |
+
"learning_rate": 3.2858511657617374e-05,
|
| 625 |
+
"loss": 0.4275,
|
| 626 |
+
"step": 10300
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 3.32,
|
| 630 |
+
"learning_rate": 3.3177898435004796e-05,
|
| 631 |
+
"loss": 0.4422,
|
| 632 |
+
"step": 10400
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 3.35,
|
| 636 |
+
"learning_rate": 3.3497285212392205e-05,
|
| 637 |
+
"loss": 0.4227,
|
| 638 |
+
"step": 10500
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 3.39,
|
| 642 |
+
"learning_rate": 3.381667198977963e-05,
|
| 643 |
+
"loss": 0.4191,
|
| 644 |
+
"step": 10600
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 3.42,
|
| 648 |
+
"learning_rate": 3.413605876716704e-05,
|
| 649 |
+
"loss": 0.4309,
|
| 650 |
+
"step": 10700
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 3.45,
|
| 654 |
+
"learning_rate": 3.445544554455446e-05,
|
| 655 |
+
"loss": 0.4079,
|
| 656 |
+
"step": 10800
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 3.48,
|
| 660 |
+
"learning_rate": 3.4774832321941874e-05,
|
| 661 |
+
"loss": 0.4281,
|
| 662 |
+
"step": 10900
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 3.51,
|
| 666 |
+
"learning_rate": 3.509421909932929e-05,
|
| 667 |
+
"loss": 0.4137,
|
| 668 |
+
"step": 11000
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 3.55,
|
| 672 |
+
"learning_rate": 3.5413605876716705e-05,
|
| 673 |
+
"loss": 0.4168,
|
| 674 |
+
"step": 11100
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 3.58,
|
| 678 |
+
"learning_rate": 3.573299265410412e-05,
|
| 679 |
+
"loss": 0.4186,
|
| 680 |
+
"step": 11200
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 3.61,
|
| 684 |
+
"learning_rate": 3.605237943149154e-05,
|
| 685 |
+
"loss": 0.4325,
|
| 686 |
+
"step": 11300
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 3.64,
|
| 690 |
+
"learning_rate": 3.637176620887895e-05,
|
| 691 |
+
"loss": 0.4422,
|
| 692 |
+
"step": 11400
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 3.67,
|
| 696 |
+
"learning_rate": 3.6691152986266374e-05,
|
| 697 |
+
"loss": 0.4269,
|
| 698 |
+
"step": 11500
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 3.7,
|
| 702 |
+
"learning_rate": 3.701053976365379e-05,
|
| 703 |
+
"loss": 0.4368,
|
| 704 |
+
"step": 11600
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 3.74,
|
| 708 |
+
"learning_rate": 3.7329926541041205e-05,
|
| 709 |
+
"loss": 0.4207,
|
| 710 |
+
"step": 11700
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 3.77,
|
| 714 |
+
"learning_rate": 3.764931331842862e-05,
|
| 715 |
+
"loss": 0.4365,
|
| 716 |
+
"step": 11800
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 3.8,
|
| 720 |
+
"learning_rate": 3.7968700095816036e-05,
|
| 721 |
+
"loss": 0.4394,
|
| 722 |
+
"step": 11900
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 3.83,
|
| 726 |
+
"learning_rate": 3.828808687320345e-05,
|
| 727 |
+
"loss": 0.4351,
|
| 728 |
+
"step": 12000
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 3.86,
|
| 732 |
+
"learning_rate": 3.860747365059087e-05,
|
| 733 |
+
"loss": 0.4113,
|
| 734 |
+
"step": 12100
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 3.9,
|
| 738 |
+
"learning_rate": 3.892686042797828e-05,
|
| 739 |
+
"loss": 0.4261,
|
| 740 |
+
"step": 12200
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 3.93,
|
| 744 |
+
"learning_rate": 3.92462472053657e-05,
|
| 745 |
+
"loss": 0.4311,
|
| 746 |
+
"step": 12300
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 3.96,
|
| 750 |
+
"learning_rate": 3.956563398275312e-05,
|
| 751 |
+
"loss": 0.4246,
|
| 752 |
+
"step": 12400
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 3.99,
|
| 756 |
+
"learning_rate": 3.988502076014053e-05,
|
| 757 |
+
"loss": 0.4294,
|
| 758 |
+
"step": 12500
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 4.02,
|
| 762 |
+
"learning_rate": 3.997728805138579e-05,
|
| 763 |
+
"loss": 0.3854,
|
| 764 |
+
"step": 12600
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 4.06,
|
| 768 |
+
"learning_rate": 3.994180063167608e-05,
|
| 769 |
+
"loss": 0.3656,
|
| 770 |
+
"step": 12700
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 4.09,
|
| 774 |
+
"learning_rate": 3.9906313211966366e-05,
|
| 775 |
+
"loss": 0.3836,
|
| 776 |
+
"step": 12800
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 4.12,
|
| 780 |
+
"learning_rate": 3.987082579225665e-05,
|
| 781 |
+
"loss": 0.3664,
|
| 782 |
+
"step": 12900
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 4.15,
|
| 786 |
+
"learning_rate": 3.9835338372546935e-05,
|
| 787 |
+
"loss": 0.3884,
|
| 788 |
+
"step": 13000
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 4.18,
|
| 792 |
+
"learning_rate": 3.979985095283722e-05,
|
| 793 |
+
"loss": 0.3764,
|
| 794 |
+
"step": 13100
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 4.22,
|
| 798 |
+
"learning_rate": 3.976436353312751e-05,
|
| 799 |
+
"loss": 0.3733,
|
| 800 |
+
"step": 13200
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 4.25,
|
| 804 |
+
"learning_rate": 3.97288761134178e-05,
|
| 805 |
+
"loss": 0.3682,
|
| 806 |
+
"step": 13300
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 4.28,
|
| 810 |
+
"learning_rate": 3.969338869370809e-05,
|
| 811 |
+
"loss": 0.371,
|
| 812 |
+
"step": 13400
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 4.31,
|
| 816 |
+
"learning_rate": 3.965790127399837e-05,
|
| 817 |
+
"loss": 0.3741,
|
| 818 |
+
"step": 13500
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 4.34,
|
| 822 |
+
"learning_rate": 3.9622413854288656e-05,
|
| 823 |
+
"loss": 0.3757,
|
| 824 |
+
"step": 13600
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 4.38,
|
| 828 |
+
"learning_rate": 3.9586926434578944e-05,
|
| 829 |
+
"loss": 0.3652,
|
| 830 |
+
"step": 13700
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 4.41,
|
| 834 |
+
"learning_rate": 3.955143901486923e-05,
|
| 835 |
+
"loss": 0.3892,
|
| 836 |
+
"step": 13800
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 4.44,
|
| 840 |
+
"learning_rate": 3.951630646935662e-05,
|
| 841 |
+
"loss": 0.3893,
|
| 842 |
+
"step": 13900
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 4.47,
|
| 846 |
+
"learning_rate": 3.948081904964691e-05,
|
| 847 |
+
"loss": 0.3823,
|
| 848 |
+
"step": 14000
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 4.5,
|
| 852 |
+
"learning_rate": 3.944533162993719e-05,
|
| 853 |
+
"loss": 0.3654,
|
| 854 |
+
"step": 14100
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 4.54,
|
| 858 |
+
"learning_rate": 3.9409844210227476e-05,
|
| 859 |
+
"loss": 0.3791,
|
| 860 |
+
"step": 14200
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 4.57,
|
| 864 |
+
"learning_rate": 3.9374356790517764e-05,
|
| 865 |
+
"loss": 0.3918,
|
| 866 |
+
"step": 14300
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 4.6,
|
| 870 |
+
"learning_rate": 3.933886937080805e-05,
|
| 871 |
+
"loss": 0.3713,
|
| 872 |
+
"step": 14400
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 4.63,
|
| 876 |
+
"learning_rate": 3.930338195109834e-05,
|
| 877 |
+
"loss": 0.376,
|
| 878 |
+
"step": 14500
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 4.66,
|
| 882 |
+
"learning_rate": 3.926789453138863e-05,
|
| 883 |
+
"loss": 0.3682,
|
| 884 |
+
"step": 14600
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 4.69,
|
| 888 |
+
"learning_rate": 3.9232407111678915e-05,
|
| 889 |
+
"loss": 0.374,
|
| 890 |
+
"step": 14700
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 4.73,
|
| 894 |
+
"learning_rate": 3.91969196919692e-05,
|
| 895 |
+
"loss": 0.3774,
|
| 896 |
+
"step": 14800
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 4.76,
|
| 900 |
+
"learning_rate": 3.9161432272259484e-05,
|
| 901 |
+
"loss": 0.3806,
|
| 902 |
+
"step": 14900
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 4.79,
|
| 906 |
+
"learning_rate": 3.912594485254977e-05,
|
| 907 |
+
"loss": 0.3886,
|
| 908 |
+
"step": 15000
|
| 909 |
+
}
|
| 910 |
+
],
|
| 911 |
+
"max_steps": 125240,
|
| 912 |
+
"num_train_epochs": 40,
|
| 913 |
+
"total_flos": 0.0,
|
| 914 |
+
"trial_name": null,
|
| 915 |
+
"trial_params": null
|
| 916 |
+
}
|
checkpoint-20000/config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "Davlan/afro-xlmr-base",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"XLMRobertaModel"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 768,
|
| 13 |
+
"id2label": {
|
| 14 |
+
"0": "LABEL_0"
|
| 15 |
+
},
|
| 16 |
+
"initializer_range": 0.02,
|
| 17 |
+
"intermediate_size": 3072,
|
| 18 |
+
"label2id": {
|
| 19 |
+
"LABEL_0": 0
|
| 20 |
+
},
|
| 21 |
+
"layer_norm_eps": 1e-05,
|
| 22 |
+
"max_position_embeddings": 514,
|
| 23 |
+
"model_type": "xlm-roberta",
|
| 24 |
+
"num_attention_heads": 12,
|
| 25 |
+
"num_hidden_layers": 12,
|
| 26 |
+
"output_past": true,
|
| 27 |
+
"pad_token_id": 1,
|
| 28 |
+
"position_embedding_type": "absolute",
|
| 29 |
+
"torch_dtype": "float32",
|
| 30 |
+
"transformers_version": "4.29.0",
|
| 31 |
+
"type_vocab_size": 1,
|
| 32 |
+
"use_cache": true,
|
| 33 |
+
"vocab_size": 250002
|
| 34 |
+
}
|
checkpoint-20000/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b203af42a7c9b516535edd3dc0e9e4bc1dfdd52ad7e9b263a3de2beebb7ebddb
|
| 3 |
+
size 1112246250
|
checkpoint-20000/trainer_state.json
ADDED
|
@@ -0,0 +1,1216 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 6.387735547748323,
|
| 5 |
+
"global_step": 20000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.03,
|
| 12 |
+
"learning_rate": 2.8744809964867454e-07,
|
| 13 |
+
"loss": 91.176,
|
| 14 |
+
"step": 100
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.06,
|
| 18 |
+
"learning_rate": 6.068348770360907e-07,
|
| 19 |
+
"loss": 64.0684,
|
| 20 |
+
"step": 200
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.1,
|
| 24 |
+
"learning_rate": 9.26221654423507e-07,
|
| 25 |
+
"loss": 27.1061,
|
| 26 |
+
"step": 300
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.13,
|
| 30 |
+
"learning_rate": 1.242414564037049e-06,
|
| 31 |
+
"loss": 8.0198,
|
| 32 |
+
"step": 400
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.16,
|
| 36 |
+
"learning_rate": 1.561801341424465e-06,
|
| 37 |
+
"loss": 4.7691,
|
| 38 |
+
"step": 500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.19,
|
| 42 |
+
"learning_rate": 1.8811881188118813e-06,
|
| 43 |
+
"loss": 3.4387,
|
| 44 |
+
"step": 600
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.22,
|
| 48 |
+
"learning_rate": 2.2005748961992977e-06,
|
| 49 |
+
"loss": 2.7027,
|
| 50 |
+
"step": 700
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.26,
|
| 54 |
+
"learning_rate": 2.5199616735867137e-06,
|
| 55 |
+
"loss": 2.4431,
|
| 56 |
+
"step": 800
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.29,
|
| 60 |
+
"learning_rate": 2.8393484509741297e-06,
|
| 61 |
+
"loss": 2.0534,
|
| 62 |
+
"step": 900
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.32,
|
| 66 |
+
"learning_rate": 3.1587352283615465e-06,
|
| 67 |
+
"loss": 1.8537,
|
| 68 |
+
"step": 1000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.35,
|
| 72 |
+
"learning_rate": 3.4781220057489625e-06,
|
| 73 |
+
"loss": 1.642,
|
| 74 |
+
"step": 1100
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.38,
|
| 78 |
+
"learning_rate": 3.797508783136378e-06,
|
| 79 |
+
"loss": 1.5092,
|
| 80 |
+
"step": 1200
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.42,
|
| 84 |
+
"learning_rate": 4.1168955605237945e-06,
|
| 85 |
+
"loss": 1.4302,
|
| 86 |
+
"step": 1300
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.45,
|
| 90 |
+
"learning_rate": 4.436282337911211e-06,
|
| 91 |
+
"loss": 1.3527,
|
| 92 |
+
"step": 1400
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.48,
|
| 96 |
+
"learning_rate": 4.755669115298627e-06,
|
| 97 |
+
"loss": 1.2495,
|
| 98 |
+
"step": 1500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.51,
|
| 102 |
+
"learning_rate": 5.075055892686043e-06,
|
| 103 |
+
"loss": 1.2039,
|
| 104 |
+
"step": 1600
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.54,
|
| 108 |
+
"learning_rate": 5.39444267007346e-06,
|
| 109 |
+
"loss": 1.1438,
|
| 110 |
+
"step": 1700
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.57,
|
| 114 |
+
"learning_rate": 5.713829447460876e-06,
|
| 115 |
+
"loss": 1.1056,
|
| 116 |
+
"step": 1800
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.61,
|
| 120 |
+
"learning_rate": 6.033216224848291e-06,
|
| 121 |
+
"loss": 1.0436,
|
| 122 |
+
"step": 1900
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.64,
|
| 126 |
+
"learning_rate": 6.352603002235708e-06,
|
| 127 |
+
"loss": 1.0225,
|
| 128 |
+
"step": 2000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.67,
|
| 132 |
+
"learning_rate": 6.671989779623123e-06,
|
| 133 |
+
"loss": 0.9856,
|
| 134 |
+
"step": 2100
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.7,
|
| 138 |
+
"learning_rate": 6.9913765570105406e-06,
|
| 139 |
+
"loss": 0.9412,
|
| 140 |
+
"step": 2200
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.73,
|
| 144 |
+
"learning_rate": 7.310763334397957e-06,
|
| 145 |
+
"loss": 0.9075,
|
| 146 |
+
"step": 2300
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.77,
|
| 150 |
+
"learning_rate": 7.630150111785373e-06,
|
| 151 |
+
"loss": 0.8752,
|
| 152 |
+
"step": 2400
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.8,
|
| 156 |
+
"learning_rate": 7.949536889172788e-06,
|
| 157 |
+
"loss": 0.854,
|
| 158 |
+
"step": 2500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.83,
|
| 162 |
+
"learning_rate": 8.268923666560205e-06,
|
| 163 |
+
"loss": 0.8282,
|
| 164 |
+
"step": 2600
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.86,
|
| 168 |
+
"learning_rate": 8.588310443947621e-06,
|
| 169 |
+
"loss": 0.8074,
|
| 170 |
+
"step": 2700
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.89,
|
| 174 |
+
"learning_rate": 8.907697221335038e-06,
|
| 175 |
+
"loss": 0.7976,
|
| 176 |
+
"step": 2800
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.93,
|
| 180 |
+
"learning_rate": 9.227083998722454e-06,
|
| 181 |
+
"loss": 0.7955,
|
| 182 |
+
"step": 2900
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.96,
|
| 186 |
+
"learning_rate": 9.54647077610987e-06,
|
| 187 |
+
"loss": 0.7502,
|
| 188 |
+
"step": 3000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.99,
|
| 192 |
+
"learning_rate": 9.865857553497285e-06,
|
| 193 |
+
"loss": 0.7455,
|
| 194 |
+
"step": 3100
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 1.02,
|
| 198 |
+
"learning_rate": 1.0185244330884702e-05,
|
| 199 |
+
"loss": 0.7177,
|
| 200 |
+
"step": 3200
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 1.05,
|
| 204 |
+
"learning_rate": 1.050463110827212e-05,
|
| 205 |
+
"loss": 0.724,
|
| 206 |
+
"step": 3300
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 1.09,
|
| 210 |
+
"learning_rate": 1.0824017885659535e-05,
|
| 211 |
+
"loss": 0.6975,
|
| 212 |
+
"step": 3400
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 1.12,
|
| 216 |
+
"learning_rate": 1.114340466304695e-05,
|
| 217 |
+
"loss": 0.7054,
|
| 218 |
+
"step": 3500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 1.15,
|
| 222 |
+
"learning_rate": 1.1462791440434366e-05,
|
| 223 |
+
"loss": 0.71,
|
| 224 |
+
"step": 3600
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 1.18,
|
| 228 |
+
"learning_rate": 1.1782178217821782e-05,
|
| 229 |
+
"loss": 0.6825,
|
| 230 |
+
"step": 3700
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 1.21,
|
| 234 |
+
"learning_rate": 1.2101564995209199e-05,
|
| 235 |
+
"loss": 0.6792,
|
| 236 |
+
"step": 3800
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 1.25,
|
| 240 |
+
"learning_rate": 1.2420951772596615e-05,
|
| 241 |
+
"loss": 0.6794,
|
| 242 |
+
"step": 3900
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 1.28,
|
| 246 |
+
"learning_rate": 1.2740338549984032e-05,
|
| 247 |
+
"loss": 0.6672,
|
| 248 |
+
"step": 4000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 1.31,
|
| 252 |
+
"learning_rate": 1.3059725327371449e-05,
|
| 253 |
+
"loss": 0.6356,
|
| 254 |
+
"step": 4100
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 1.34,
|
| 258 |
+
"learning_rate": 1.3379112104758865e-05,
|
| 259 |
+
"loss": 0.6503,
|
| 260 |
+
"step": 4200
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 1.37,
|
| 264 |
+
"learning_rate": 1.369849888214628e-05,
|
| 265 |
+
"loss": 0.6495,
|
| 266 |
+
"step": 4300
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 1.41,
|
| 270 |
+
"learning_rate": 1.4017885659533696e-05,
|
| 271 |
+
"loss": 0.6428,
|
| 272 |
+
"step": 4400
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 1.44,
|
| 276 |
+
"learning_rate": 1.4337272436921113e-05,
|
| 277 |
+
"loss": 0.6171,
|
| 278 |
+
"step": 4500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 1.47,
|
| 282 |
+
"learning_rate": 1.4656659214308529e-05,
|
| 283 |
+
"loss": 0.6084,
|
| 284 |
+
"step": 4600
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 1.5,
|
| 288 |
+
"learning_rate": 1.4976045991695944e-05,
|
| 289 |
+
"loss": 0.6222,
|
| 290 |
+
"step": 4700
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 1.53,
|
| 294 |
+
"learning_rate": 1.529543276908336e-05,
|
| 295 |
+
"loss": 0.5971,
|
| 296 |
+
"step": 4800
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 1.56,
|
| 300 |
+
"learning_rate": 1.5614819546470777e-05,
|
| 301 |
+
"loss": 0.5745,
|
| 302 |
+
"step": 4900
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 1.6,
|
| 306 |
+
"learning_rate": 1.5934206323858196e-05,
|
| 307 |
+
"loss": 0.5753,
|
| 308 |
+
"step": 5000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 1.63,
|
| 312 |
+
"learning_rate": 1.625359310124561e-05,
|
| 313 |
+
"loss": 0.6045,
|
| 314 |
+
"step": 5100
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 1.66,
|
| 318 |
+
"learning_rate": 1.6572979878633027e-05,
|
| 319 |
+
"loss": 0.6006,
|
| 320 |
+
"step": 5200
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 1.69,
|
| 324 |
+
"learning_rate": 1.6892366656020443e-05,
|
| 325 |
+
"loss": 0.6003,
|
| 326 |
+
"step": 5300
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 1.72,
|
| 330 |
+
"learning_rate": 1.7211753433407858e-05,
|
| 331 |
+
"loss": 0.5688,
|
| 332 |
+
"step": 5400
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 1.76,
|
| 336 |
+
"learning_rate": 1.7531140210795274e-05,
|
| 337 |
+
"loss": 0.5662,
|
| 338 |
+
"step": 5500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 1.79,
|
| 342 |
+
"learning_rate": 1.785052698818269e-05,
|
| 343 |
+
"loss": 0.5788,
|
| 344 |
+
"step": 5600
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 1.82,
|
| 348 |
+
"learning_rate": 1.8169913765570105e-05,
|
| 349 |
+
"loss": 0.5541,
|
| 350 |
+
"step": 5700
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 1.85,
|
| 354 |
+
"learning_rate": 1.8489300542957524e-05,
|
| 355 |
+
"loss": 0.5656,
|
| 356 |
+
"step": 5800
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 1.88,
|
| 360 |
+
"learning_rate": 1.880868732034494e-05,
|
| 361 |
+
"loss": 0.5466,
|
| 362 |
+
"step": 5900
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 1.92,
|
| 366 |
+
"learning_rate": 1.9128074097732355e-05,
|
| 367 |
+
"loss": 0.563,
|
| 368 |
+
"step": 6000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 1.95,
|
| 372 |
+
"learning_rate": 1.9447460875119774e-05,
|
| 373 |
+
"loss": 0.5603,
|
| 374 |
+
"step": 6100
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 1.98,
|
| 378 |
+
"learning_rate": 1.976684765250719e-05,
|
| 379 |
+
"loss": 0.5579,
|
| 380 |
+
"step": 6200
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 2.01,
|
| 384 |
+
"learning_rate": 2.0086234429894605e-05,
|
| 385 |
+
"loss": 0.5346,
|
| 386 |
+
"step": 6300
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 2.04,
|
| 390 |
+
"learning_rate": 2.0405621207282017e-05,
|
| 391 |
+
"loss": 0.5038,
|
| 392 |
+
"step": 6400
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 2.08,
|
| 396 |
+
"learning_rate": 2.0725007984669436e-05,
|
| 397 |
+
"loss": 0.5121,
|
| 398 |
+
"step": 6500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 2.11,
|
| 402 |
+
"learning_rate": 2.1044394762056855e-05,
|
| 403 |
+
"loss": 0.5135,
|
| 404 |
+
"step": 6600
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 2.14,
|
| 408 |
+
"learning_rate": 2.1363781539444267e-05,
|
| 409 |
+
"loss": 0.504,
|
| 410 |
+
"step": 6700
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 2.17,
|
| 414 |
+
"learning_rate": 2.1683168316831686e-05,
|
| 415 |
+
"loss": 0.5119,
|
| 416 |
+
"step": 6800
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 2.2,
|
| 420 |
+
"learning_rate": 2.20025550942191e-05,
|
| 421 |
+
"loss": 0.4986,
|
| 422 |
+
"step": 6900
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 2.24,
|
| 426 |
+
"learning_rate": 2.2321941871606518e-05,
|
| 427 |
+
"loss": 0.504,
|
| 428 |
+
"step": 7000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 2.27,
|
| 432 |
+
"learning_rate": 2.2641328648993933e-05,
|
| 433 |
+
"loss": 0.5024,
|
| 434 |
+
"step": 7100
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 2.3,
|
| 438 |
+
"learning_rate": 2.296071542638135e-05,
|
| 439 |
+
"loss": 0.479,
|
| 440 |
+
"step": 7200
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 2.33,
|
| 444 |
+
"learning_rate": 2.3280102203768764e-05,
|
| 445 |
+
"loss": 0.492,
|
| 446 |
+
"step": 7300
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 2.36,
|
| 450 |
+
"learning_rate": 2.3599488981156183e-05,
|
| 451 |
+
"loss": 0.4928,
|
| 452 |
+
"step": 7400
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 2.4,
|
| 456 |
+
"learning_rate": 2.39188757585436e-05,
|
| 457 |
+
"loss": 0.4862,
|
| 458 |
+
"step": 7500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 2.43,
|
| 462 |
+
"learning_rate": 2.4238262535931014e-05,
|
| 463 |
+
"loss": 0.5148,
|
| 464 |
+
"step": 7600
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 2.46,
|
| 468 |
+
"learning_rate": 2.4557649313318433e-05,
|
| 469 |
+
"loss": 0.4922,
|
| 470 |
+
"step": 7700
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 2.49,
|
| 474 |
+
"learning_rate": 2.4877036090705846e-05,
|
| 475 |
+
"loss": 0.4882,
|
| 476 |
+
"step": 7800
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 2.52,
|
| 480 |
+
"learning_rate": 2.5196422868093264e-05,
|
| 481 |
+
"loss": 0.4871,
|
| 482 |
+
"step": 7900
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 2.56,
|
| 486 |
+
"learning_rate": 2.5515809645480677e-05,
|
| 487 |
+
"loss": 0.4967,
|
| 488 |
+
"step": 8000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 2.59,
|
| 492 |
+
"learning_rate": 2.5835196422868096e-05,
|
| 493 |
+
"loss": 0.4834,
|
| 494 |
+
"step": 8100
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 2.62,
|
| 498 |
+
"learning_rate": 2.615458320025551e-05,
|
| 499 |
+
"loss": 0.4769,
|
| 500 |
+
"step": 8200
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 2.65,
|
| 504 |
+
"learning_rate": 2.6473969977642927e-05,
|
| 505 |
+
"loss": 0.4776,
|
| 506 |
+
"step": 8300
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 2.68,
|
| 510 |
+
"learning_rate": 2.6793356755030342e-05,
|
| 511 |
+
"loss": 0.4861,
|
| 512 |
+
"step": 8400
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 2.71,
|
| 516 |
+
"learning_rate": 2.711274353241776e-05,
|
| 517 |
+
"loss": 0.486,
|
| 518 |
+
"step": 8500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 2.75,
|
| 522 |
+
"learning_rate": 2.7432130309805177e-05,
|
| 523 |
+
"loss": 0.4745,
|
| 524 |
+
"step": 8600
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 2.78,
|
| 528 |
+
"learning_rate": 2.7751517087192592e-05,
|
| 529 |
+
"loss": 0.4805,
|
| 530 |
+
"step": 8700
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 2.81,
|
| 534 |
+
"learning_rate": 2.807090386458001e-05,
|
| 535 |
+
"loss": 0.5003,
|
| 536 |
+
"step": 8800
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 2.84,
|
| 540 |
+
"learning_rate": 2.8390290641967424e-05,
|
| 541 |
+
"loss": 0.4639,
|
| 542 |
+
"step": 8900
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 2.87,
|
| 546 |
+
"learning_rate": 2.8709677419354843e-05,
|
| 547 |
+
"loss": 0.4763,
|
| 548 |
+
"step": 9000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 2.91,
|
| 552 |
+
"learning_rate": 2.9029064196742255e-05,
|
| 553 |
+
"loss": 0.4727,
|
| 554 |
+
"step": 9100
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 2.94,
|
| 558 |
+
"learning_rate": 2.9348450974129674e-05,
|
| 559 |
+
"loss": 0.4839,
|
| 560 |
+
"step": 9200
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 2.97,
|
| 564 |
+
"learning_rate": 2.966783775151709e-05,
|
| 565 |
+
"loss": 0.4769,
|
| 566 |
+
"step": 9300
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 3.0,
|
| 570 |
+
"learning_rate": 2.9987224528904505e-05,
|
| 571 |
+
"loss": 0.4641,
|
| 572 |
+
"step": 9400
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 3.03,
|
| 576 |
+
"learning_rate": 3.030661130629192e-05,
|
| 577 |
+
"loss": 0.4282,
|
| 578 |
+
"step": 9500
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 3.07,
|
| 582 |
+
"learning_rate": 3.0625998083679336e-05,
|
| 583 |
+
"loss": 0.4057,
|
| 584 |
+
"step": 9600
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 3.1,
|
| 588 |
+
"learning_rate": 3.094538486106676e-05,
|
| 589 |
+
"loss": 0.4192,
|
| 590 |
+
"step": 9700
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 3.13,
|
| 594 |
+
"learning_rate": 3.126477163845417e-05,
|
| 595 |
+
"loss": 0.4186,
|
| 596 |
+
"step": 9800
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 3.16,
|
| 600 |
+
"learning_rate": 3.158415841584159e-05,
|
| 601 |
+
"loss": 0.4086,
|
| 602 |
+
"step": 9900
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 3.19,
|
| 606 |
+
"learning_rate": 3.1903545193229005e-05,
|
| 607 |
+
"loss": 0.4194,
|
| 608 |
+
"step": 10000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 3.23,
|
| 612 |
+
"learning_rate": 3.221973810284254e-05,
|
| 613 |
+
"loss": 0.4443,
|
| 614 |
+
"step": 10100
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 3.26,
|
| 618 |
+
"learning_rate": 3.2539124880229965e-05,
|
| 619 |
+
"loss": 0.4217,
|
| 620 |
+
"step": 10200
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 3.29,
|
| 624 |
+
"learning_rate": 3.2858511657617374e-05,
|
| 625 |
+
"loss": 0.4275,
|
| 626 |
+
"step": 10300
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 3.32,
|
| 630 |
+
"learning_rate": 3.3177898435004796e-05,
|
| 631 |
+
"loss": 0.4422,
|
| 632 |
+
"step": 10400
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 3.35,
|
| 636 |
+
"learning_rate": 3.3497285212392205e-05,
|
| 637 |
+
"loss": 0.4227,
|
| 638 |
+
"step": 10500
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 3.39,
|
| 642 |
+
"learning_rate": 3.381667198977963e-05,
|
| 643 |
+
"loss": 0.4191,
|
| 644 |
+
"step": 10600
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 3.42,
|
| 648 |
+
"learning_rate": 3.413605876716704e-05,
|
| 649 |
+
"loss": 0.4309,
|
| 650 |
+
"step": 10700
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 3.45,
|
| 654 |
+
"learning_rate": 3.445544554455446e-05,
|
| 655 |
+
"loss": 0.4079,
|
| 656 |
+
"step": 10800
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 3.48,
|
| 660 |
+
"learning_rate": 3.4774832321941874e-05,
|
| 661 |
+
"loss": 0.4281,
|
| 662 |
+
"step": 10900
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 3.51,
|
| 666 |
+
"learning_rate": 3.509421909932929e-05,
|
| 667 |
+
"loss": 0.4137,
|
| 668 |
+
"step": 11000
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 3.55,
|
| 672 |
+
"learning_rate": 3.5413605876716705e-05,
|
| 673 |
+
"loss": 0.4168,
|
| 674 |
+
"step": 11100
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 3.58,
|
| 678 |
+
"learning_rate": 3.573299265410412e-05,
|
| 679 |
+
"loss": 0.4186,
|
| 680 |
+
"step": 11200
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 3.61,
|
| 684 |
+
"learning_rate": 3.605237943149154e-05,
|
| 685 |
+
"loss": 0.4325,
|
| 686 |
+
"step": 11300
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 3.64,
|
| 690 |
+
"learning_rate": 3.637176620887895e-05,
|
| 691 |
+
"loss": 0.4422,
|
| 692 |
+
"step": 11400
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 3.67,
|
| 696 |
+
"learning_rate": 3.6691152986266374e-05,
|
| 697 |
+
"loss": 0.4269,
|
| 698 |
+
"step": 11500
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 3.7,
|
| 702 |
+
"learning_rate": 3.701053976365379e-05,
|
| 703 |
+
"loss": 0.4368,
|
| 704 |
+
"step": 11600
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 3.74,
|
| 708 |
+
"learning_rate": 3.7329926541041205e-05,
|
| 709 |
+
"loss": 0.4207,
|
| 710 |
+
"step": 11700
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 3.77,
|
| 714 |
+
"learning_rate": 3.764931331842862e-05,
|
| 715 |
+
"loss": 0.4365,
|
| 716 |
+
"step": 11800
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 3.8,
|
| 720 |
+
"learning_rate": 3.7968700095816036e-05,
|
| 721 |
+
"loss": 0.4394,
|
| 722 |
+
"step": 11900
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 3.83,
|
| 726 |
+
"learning_rate": 3.828808687320345e-05,
|
| 727 |
+
"loss": 0.4351,
|
| 728 |
+
"step": 12000
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 3.86,
|
| 732 |
+
"learning_rate": 3.860747365059087e-05,
|
| 733 |
+
"loss": 0.4113,
|
| 734 |
+
"step": 12100
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 3.9,
|
| 738 |
+
"learning_rate": 3.892686042797828e-05,
|
| 739 |
+
"loss": 0.4261,
|
| 740 |
+
"step": 12200
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 3.93,
|
| 744 |
+
"learning_rate": 3.92462472053657e-05,
|
| 745 |
+
"loss": 0.4311,
|
| 746 |
+
"step": 12300
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 3.96,
|
| 750 |
+
"learning_rate": 3.956563398275312e-05,
|
| 751 |
+
"loss": 0.4246,
|
| 752 |
+
"step": 12400
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 3.99,
|
| 756 |
+
"learning_rate": 3.988502076014053e-05,
|
| 757 |
+
"loss": 0.4294,
|
| 758 |
+
"step": 12500
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 4.02,
|
| 762 |
+
"learning_rate": 3.997728805138579e-05,
|
| 763 |
+
"loss": 0.3854,
|
| 764 |
+
"step": 12600
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 4.06,
|
| 768 |
+
"learning_rate": 3.994180063167608e-05,
|
| 769 |
+
"loss": 0.3656,
|
| 770 |
+
"step": 12700
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 4.09,
|
| 774 |
+
"learning_rate": 3.9906313211966366e-05,
|
| 775 |
+
"loss": 0.3836,
|
| 776 |
+
"step": 12800
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 4.12,
|
| 780 |
+
"learning_rate": 3.987082579225665e-05,
|
| 781 |
+
"loss": 0.3664,
|
| 782 |
+
"step": 12900
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 4.15,
|
| 786 |
+
"learning_rate": 3.9835338372546935e-05,
|
| 787 |
+
"loss": 0.3884,
|
| 788 |
+
"step": 13000
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 4.18,
|
| 792 |
+
"learning_rate": 3.979985095283722e-05,
|
| 793 |
+
"loss": 0.3764,
|
| 794 |
+
"step": 13100
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 4.22,
|
| 798 |
+
"learning_rate": 3.976436353312751e-05,
|
| 799 |
+
"loss": 0.3733,
|
| 800 |
+
"step": 13200
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 4.25,
|
| 804 |
+
"learning_rate": 3.97288761134178e-05,
|
| 805 |
+
"loss": 0.3682,
|
| 806 |
+
"step": 13300
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 4.28,
|
| 810 |
+
"learning_rate": 3.969338869370809e-05,
|
| 811 |
+
"loss": 0.371,
|
| 812 |
+
"step": 13400
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 4.31,
|
| 816 |
+
"learning_rate": 3.965790127399837e-05,
|
| 817 |
+
"loss": 0.3741,
|
| 818 |
+
"step": 13500
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 4.34,
|
| 822 |
+
"learning_rate": 3.9622413854288656e-05,
|
| 823 |
+
"loss": 0.3757,
|
| 824 |
+
"step": 13600
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 4.38,
|
| 828 |
+
"learning_rate": 3.9586926434578944e-05,
|
| 829 |
+
"loss": 0.3652,
|
| 830 |
+
"step": 13700
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 4.41,
|
| 834 |
+
"learning_rate": 3.955143901486923e-05,
|
| 835 |
+
"loss": 0.3892,
|
| 836 |
+
"step": 13800
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 4.44,
|
| 840 |
+
"learning_rate": 3.951630646935662e-05,
|
| 841 |
+
"loss": 0.3893,
|
| 842 |
+
"step": 13900
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 4.47,
|
| 846 |
+
"learning_rate": 3.948081904964691e-05,
|
| 847 |
+
"loss": 0.3823,
|
| 848 |
+
"step": 14000
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 4.5,
|
| 852 |
+
"learning_rate": 3.944533162993719e-05,
|
| 853 |
+
"loss": 0.3654,
|
| 854 |
+
"step": 14100
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 4.54,
|
| 858 |
+
"learning_rate": 3.9409844210227476e-05,
|
| 859 |
+
"loss": 0.3791,
|
| 860 |
+
"step": 14200
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 4.57,
|
| 864 |
+
"learning_rate": 3.9374356790517764e-05,
|
| 865 |
+
"loss": 0.3918,
|
| 866 |
+
"step": 14300
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 4.6,
|
| 870 |
+
"learning_rate": 3.933886937080805e-05,
|
| 871 |
+
"loss": 0.3713,
|
| 872 |
+
"step": 14400
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 4.63,
|
| 876 |
+
"learning_rate": 3.930338195109834e-05,
|
| 877 |
+
"loss": 0.376,
|
| 878 |
+
"step": 14500
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 4.66,
|
| 882 |
+
"learning_rate": 3.926789453138863e-05,
|
| 883 |
+
"loss": 0.3682,
|
| 884 |
+
"step": 14600
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 4.69,
|
| 888 |
+
"learning_rate": 3.9232407111678915e-05,
|
| 889 |
+
"loss": 0.374,
|
| 890 |
+
"step": 14700
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 4.73,
|
| 894 |
+
"learning_rate": 3.91969196919692e-05,
|
| 895 |
+
"loss": 0.3774,
|
| 896 |
+
"step": 14800
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 4.76,
|
| 900 |
+
"learning_rate": 3.9161432272259484e-05,
|
| 901 |
+
"loss": 0.3806,
|
| 902 |
+
"step": 14900
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 4.79,
|
| 906 |
+
"learning_rate": 3.912594485254977e-05,
|
| 907 |
+
"loss": 0.3886,
|
| 908 |
+
"step": 15000
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 4.82,
|
| 912 |
+
"learning_rate": 3.909045743284006e-05,
|
| 913 |
+
"loss": 0.3742,
|
| 914 |
+
"step": 15100
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 4.85,
|
| 918 |
+
"learning_rate": 3.905497001313035e-05,
|
| 919 |
+
"loss": 0.3714,
|
| 920 |
+
"step": 15200
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 4.89,
|
| 924 |
+
"learning_rate": 3.9019482593420636e-05,
|
| 925 |
+
"loss": 0.3721,
|
| 926 |
+
"step": 15300
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 4.92,
|
| 930 |
+
"learning_rate": 3.8983995173710924e-05,
|
| 931 |
+
"loss": 0.3716,
|
| 932 |
+
"step": 15400
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 4.95,
|
| 936 |
+
"learning_rate": 3.8948507754001205e-05,
|
| 937 |
+
"loss": 0.3728,
|
| 938 |
+
"step": 15500
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 4.98,
|
| 942 |
+
"learning_rate": 3.891302033429149e-05,
|
| 943 |
+
"loss": 0.384,
|
| 944 |
+
"step": 15600
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 5.01,
|
| 948 |
+
"learning_rate": 3.887753291458178e-05,
|
| 949 |
+
"loss": 0.3415,
|
| 950 |
+
"step": 15700
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 5.05,
|
| 954 |
+
"learning_rate": 3.884204549487207e-05,
|
| 955 |
+
"loss": 0.3064,
|
| 956 |
+
"step": 15800
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 5.08,
|
| 960 |
+
"learning_rate": 3.880655807516236e-05,
|
| 961 |
+
"loss": 0.3164,
|
| 962 |
+
"step": 15900
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 5.11,
|
| 966 |
+
"learning_rate": 3.8771070655452645e-05,
|
| 967 |
+
"loss": 0.3215,
|
| 968 |
+
"step": 16000
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 5.14,
|
| 972 |
+
"learning_rate": 3.873558323574293e-05,
|
| 973 |
+
"loss": 0.3234,
|
| 974 |
+
"step": 16100
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 5.17,
|
| 978 |
+
"learning_rate": 3.870009581603322e-05,
|
| 979 |
+
"loss": 0.3119,
|
| 980 |
+
"step": 16200
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 5.21,
|
| 984 |
+
"learning_rate": 3.866460839632351e-05,
|
| 985 |
+
"loss": 0.3145,
|
| 986 |
+
"step": 16300
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 5.24,
|
| 990 |
+
"learning_rate": 3.8629120976613796e-05,
|
| 991 |
+
"loss": 0.3333,
|
| 992 |
+
"step": 16400
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 5.27,
|
| 996 |
+
"learning_rate": 3.8593633556904084e-05,
|
| 997 |
+
"loss": 0.3195,
|
| 998 |
+
"step": 16500
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 5.3,
|
| 1002 |
+
"learning_rate": 3.855814613719437e-05,
|
| 1003 |
+
"loss": 0.3159,
|
| 1004 |
+
"step": 16600
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 5.33,
|
| 1008 |
+
"learning_rate": 3.852265871748465e-05,
|
| 1009 |
+
"loss": 0.3189,
|
| 1010 |
+
"step": 16700
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 5.37,
|
| 1014 |
+
"learning_rate": 3.848717129777494e-05,
|
| 1015 |
+
"loss": 0.317,
|
| 1016 |
+
"step": 16800
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 5.4,
|
| 1020 |
+
"learning_rate": 3.845168387806523e-05,
|
| 1021 |
+
"loss": 0.3231,
|
| 1022 |
+
"step": 16900
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 5.43,
|
| 1026 |
+
"learning_rate": 3.841619645835552e-05,
|
| 1027 |
+
"loss": 0.3328,
|
| 1028 |
+
"step": 17000
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 5.46,
|
| 1032 |
+
"learning_rate": 3.8380709038645805e-05,
|
| 1033 |
+
"loss": 0.3282,
|
| 1034 |
+
"step": 17100
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 5.49,
|
| 1038 |
+
"learning_rate": 3.8345221618936086e-05,
|
| 1039 |
+
"loss": 0.3169,
|
| 1040 |
+
"step": 17200
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 5.53,
|
| 1044 |
+
"learning_rate": 3.8309734199226374e-05,
|
| 1045 |
+
"loss": 0.3156,
|
| 1046 |
+
"step": 17300
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 5.56,
|
| 1050 |
+
"learning_rate": 3.827424677951666e-05,
|
| 1051 |
+
"loss": 0.3313,
|
| 1052 |
+
"step": 17400
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 5.59,
|
| 1056 |
+
"learning_rate": 3.823875935980695e-05,
|
| 1057 |
+
"loss": 0.3181,
|
| 1058 |
+
"step": 17500
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 5.62,
|
| 1062 |
+
"learning_rate": 3.820327194009724e-05,
|
| 1063 |
+
"loss": 0.334,
|
| 1064 |
+
"step": 17600
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 5.65,
|
| 1068 |
+
"learning_rate": 3.8167784520387526e-05,
|
| 1069 |
+
"loss": 0.3313,
|
| 1070 |
+
"step": 17700
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 5.69,
|
| 1074 |
+
"learning_rate": 3.8132297100677814e-05,
|
| 1075 |
+
"loss": 0.3331,
|
| 1076 |
+
"step": 17800
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 5.72,
|
| 1080 |
+
"learning_rate": 3.80968096809681e-05,
|
| 1081 |
+
"loss": 0.3405,
|
| 1082 |
+
"step": 17900
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 5.75,
|
| 1086 |
+
"learning_rate": 3.806132226125839e-05,
|
| 1087 |
+
"loss": 0.3345,
|
| 1088 |
+
"step": 18000
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 5.78,
|
| 1092 |
+
"learning_rate": 3.802583484154868e-05,
|
| 1093 |
+
"loss": 0.3295,
|
| 1094 |
+
"step": 18100
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 5.81,
|
| 1098 |
+
"learning_rate": 3.7990347421838965e-05,
|
| 1099 |
+
"loss": 0.3244,
|
| 1100 |
+
"step": 18200
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 5.84,
|
| 1104 |
+
"learning_rate": 3.795486000212925e-05,
|
| 1105 |
+
"loss": 0.3346,
|
| 1106 |
+
"step": 18300
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 5.88,
|
| 1110 |
+
"learning_rate": 3.791972745661663e-05,
|
| 1111 |
+
"loss": 0.3317,
|
| 1112 |
+
"step": 18400
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 5.91,
|
| 1116 |
+
"learning_rate": 3.788424003690692e-05,
|
| 1117 |
+
"loss": 0.3219,
|
| 1118 |
+
"step": 18500
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 5.94,
|
| 1122 |
+
"learning_rate": 3.784875261719721e-05,
|
| 1123 |
+
"loss": 0.3167,
|
| 1124 |
+
"step": 18600
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 5.97,
|
| 1128 |
+
"learning_rate": 3.781326519748749e-05,
|
| 1129 |
+
"loss": 0.33,
|
| 1130 |
+
"step": 18700
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 6.0,
|
| 1134 |
+
"learning_rate": 3.777777777777778e-05,
|
| 1135 |
+
"loss": 0.3211,
|
| 1136 |
+
"step": 18800
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 6.04,
|
| 1140 |
+
"learning_rate": 3.7742290358068066e-05,
|
| 1141 |
+
"loss": 0.2794,
|
| 1142 |
+
"step": 18900
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 6.07,
|
| 1146 |
+
"learning_rate": 3.7706802938358354e-05,
|
| 1147 |
+
"loss": 0.2675,
|
| 1148 |
+
"step": 19000
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 6.1,
|
| 1152 |
+
"learning_rate": 3.767131551864864e-05,
|
| 1153 |
+
"loss": 0.2676,
|
| 1154 |
+
"step": 19100
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 6.13,
|
| 1158 |
+
"learning_rate": 3.763582809893893e-05,
|
| 1159 |
+
"loss": 0.2693,
|
| 1160 |
+
"step": 19200
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 6.16,
|
| 1164 |
+
"learning_rate": 3.760034067922922e-05,
|
| 1165 |
+
"loss": 0.2669,
|
| 1166 |
+
"step": 19300
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 6.2,
|
| 1170 |
+
"learning_rate": 3.7564853259519506e-05,
|
| 1171 |
+
"loss": 0.2744,
|
| 1172 |
+
"step": 19400
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 6.23,
|
| 1176 |
+
"learning_rate": 3.7529365839809794e-05,
|
| 1177 |
+
"loss": 0.282,
|
| 1178 |
+
"step": 19500
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 6.26,
|
| 1182 |
+
"learning_rate": 3.749387842010008e-05,
|
| 1183 |
+
"loss": 0.2848,
|
| 1184 |
+
"step": 19600
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 6.29,
|
| 1188 |
+
"learning_rate": 3.745839100039037e-05,
|
| 1189 |
+
"loss": 0.2783,
|
| 1190 |
+
"step": 19700
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 6.32,
|
| 1194 |
+
"learning_rate": 3.742290358068066e-05,
|
| 1195 |
+
"loss": 0.2843,
|
| 1196 |
+
"step": 19800
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 6.36,
|
| 1200 |
+
"learning_rate": 3.738741616097094e-05,
|
| 1201 |
+
"loss": 0.2696,
|
| 1202 |
+
"step": 19900
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 6.39,
|
| 1206 |
+
"learning_rate": 3.7351928741261226e-05,
|
| 1207 |
+
"loss": 0.2932,
|
| 1208 |
+
"step": 20000
|
| 1209 |
+
}
|
| 1210 |
+
],
|
| 1211 |
+
"max_steps": 125240,
|
| 1212 |
+
"num_train_epochs": 40,
|
| 1213 |
+
"total_flos": 0.0,
|
| 1214 |
+
"trial_name": null,
|
| 1215 |
+
"trial_params": null
|
| 1216 |
+
}
|
checkpoint-25000/config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "Davlan/afro-xlmr-base",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"XLMRobertaModel"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 768,
|
| 13 |
+
"id2label": {
|
| 14 |
+
"0": "LABEL_0"
|
| 15 |
+
},
|
| 16 |
+
"initializer_range": 0.02,
|
| 17 |
+
"intermediate_size": 3072,
|
| 18 |
+
"label2id": {
|
| 19 |
+
"LABEL_0": 0
|
| 20 |
+
},
|
| 21 |
+
"layer_norm_eps": 1e-05,
|
| 22 |
+
"max_position_embeddings": 514,
|
| 23 |
+
"model_type": "xlm-roberta",
|
| 24 |
+
"num_attention_heads": 12,
|
| 25 |
+
"num_hidden_layers": 12,
|
| 26 |
+
"output_past": true,
|
| 27 |
+
"pad_token_id": 1,
|
| 28 |
+
"position_embedding_type": "absolute",
|
| 29 |
+
"torch_dtype": "float32",
|
| 30 |
+
"transformers_version": "4.29.0",
|
| 31 |
+
"type_vocab_size": 1,
|
| 32 |
+
"use_cache": true,
|
| 33 |
+
"vocab_size": 250002
|
| 34 |
+
}
|
checkpoint-25000/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9bfd42151b5d4a1f83bc6961a6a82e14e9b306ee29f29f55addd2f3149897c88
|
| 3 |
+
size 1112246250
|
checkpoint-25000/trainer_state.json
ADDED
|
@@ -0,0 +1,1516 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 7.984669434685404,
|
| 5 |
+
"global_step": 25000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.03,
|
| 12 |
+
"learning_rate": 2.8744809964867454e-07,
|
| 13 |
+
"loss": 91.176,
|
| 14 |
+
"step": 100
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.06,
|
| 18 |
+
"learning_rate": 6.068348770360907e-07,
|
| 19 |
+
"loss": 64.0684,
|
| 20 |
+
"step": 200
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.1,
|
| 24 |
+
"learning_rate": 9.26221654423507e-07,
|
| 25 |
+
"loss": 27.1061,
|
| 26 |
+
"step": 300
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.13,
|
| 30 |
+
"learning_rate": 1.242414564037049e-06,
|
| 31 |
+
"loss": 8.0198,
|
| 32 |
+
"step": 400
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.16,
|
| 36 |
+
"learning_rate": 1.561801341424465e-06,
|
| 37 |
+
"loss": 4.7691,
|
| 38 |
+
"step": 500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.19,
|
| 42 |
+
"learning_rate": 1.8811881188118813e-06,
|
| 43 |
+
"loss": 3.4387,
|
| 44 |
+
"step": 600
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.22,
|
| 48 |
+
"learning_rate": 2.2005748961992977e-06,
|
| 49 |
+
"loss": 2.7027,
|
| 50 |
+
"step": 700
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.26,
|
| 54 |
+
"learning_rate": 2.5199616735867137e-06,
|
| 55 |
+
"loss": 2.4431,
|
| 56 |
+
"step": 800
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.29,
|
| 60 |
+
"learning_rate": 2.8393484509741297e-06,
|
| 61 |
+
"loss": 2.0534,
|
| 62 |
+
"step": 900
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.32,
|
| 66 |
+
"learning_rate": 3.1587352283615465e-06,
|
| 67 |
+
"loss": 1.8537,
|
| 68 |
+
"step": 1000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.35,
|
| 72 |
+
"learning_rate": 3.4781220057489625e-06,
|
| 73 |
+
"loss": 1.642,
|
| 74 |
+
"step": 1100
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.38,
|
| 78 |
+
"learning_rate": 3.797508783136378e-06,
|
| 79 |
+
"loss": 1.5092,
|
| 80 |
+
"step": 1200
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.42,
|
| 84 |
+
"learning_rate": 4.1168955605237945e-06,
|
| 85 |
+
"loss": 1.4302,
|
| 86 |
+
"step": 1300
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.45,
|
| 90 |
+
"learning_rate": 4.436282337911211e-06,
|
| 91 |
+
"loss": 1.3527,
|
| 92 |
+
"step": 1400
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.48,
|
| 96 |
+
"learning_rate": 4.755669115298627e-06,
|
| 97 |
+
"loss": 1.2495,
|
| 98 |
+
"step": 1500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.51,
|
| 102 |
+
"learning_rate": 5.075055892686043e-06,
|
| 103 |
+
"loss": 1.2039,
|
| 104 |
+
"step": 1600
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.54,
|
| 108 |
+
"learning_rate": 5.39444267007346e-06,
|
| 109 |
+
"loss": 1.1438,
|
| 110 |
+
"step": 1700
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.57,
|
| 114 |
+
"learning_rate": 5.713829447460876e-06,
|
| 115 |
+
"loss": 1.1056,
|
| 116 |
+
"step": 1800
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.61,
|
| 120 |
+
"learning_rate": 6.033216224848291e-06,
|
| 121 |
+
"loss": 1.0436,
|
| 122 |
+
"step": 1900
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.64,
|
| 126 |
+
"learning_rate": 6.352603002235708e-06,
|
| 127 |
+
"loss": 1.0225,
|
| 128 |
+
"step": 2000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.67,
|
| 132 |
+
"learning_rate": 6.671989779623123e-06,
|
| 133 |
+
"loss": 0.9856,
|
| 134 |
+
"step": 2100
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.7,
|
| 138 |
+
"learning_rate": 6.9913765570105406e-06,
|
| 139 |
+
"loss": 0.9412,
|
| 140 |
+
"step": 2200
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.73,
|
| 144 |
+
"learning_rate": 7.310763334397957e-06,
|
| 145 |
+
"loss": 0.9075,
|
| 146 |
+
"step": 2300
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.77,
|
| 150 |
+
"learning_rate": 7.630150111785373e-06,
|
| 151 |
+
"loss": 0.8752,
|
| 152 |
+
"step": 2400
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.8,
|
| 156 |
+
"learning_rate": 7.949536889172788e-06,
|
| 157 |
+
"loss": 0.854,
|
| 158 |
+
"step": 2500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.83,
|
| 162 |
+
"learning_rate": 8.268923666560205e-06,
|
| 163 |
+
"loss": 0.8282,
|
| 164 |
+
"step": 2600
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.86,
|
| 168 |
+
"learning_rate": 8.588310443947621e-06,
|
| 169 |
+
"loss": 0.8074,
|
| 170 |
+
"step": 2700
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.89,
|
| 174 |
+
"learning_rate": 8.907697221335038e-06,
|
| 175 |
+
"loss": 0.7976,
|
| 176 |
+
"step": 2800
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.93,
|
| 180 |
+
"learning_rate": 9.227083998722454e-06,
|
| 181 |
+
"loss": 0.7955,
|
| 182 |
+
"step": 2900
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.96,
|
| 186 |
+
"learning_rate": 9.54647077610987e-06,
|
| 187 |
+
"loss": 0.7502,
|
| 188 |
+
"step": 3000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.99,
|
| 192 |
+
"learning_rate": 9.865857553497285e-06,
|
| 193 |
+
"loss": 0.7455,
|
| 194 |
+
"step": 3100
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 1.02,
|
| 198 |
+
"learning_rate": 1.0185244330884702e-05,
|
| 199 |
+
"loss": 0.7177,
|
| 200 |
+
"step": 3200
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 1.05,
|
| 204 |
+
"learning_rate": 1.050463110827212e-05,
|
| 205 |
+
"loss": 0.724,
|
| 206 |
+
"step": 3300
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 1.09,
|
| 210 |
+
"learning_rate": 1.0824017885659535e-05,
|
| 211 |
+
"loss": 0.6975,
|
| 212 |
+
"step": 3400
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 1.12,
|
| 216 |
+
"learning_rate": 1.114340466304695e-05,
|
| 217 |
+
"loss": 0.7054,
|
| 218 |
+
"step": 3500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 1.15,
|
| 222 |
+
"learning_rate": 1.1462791440434366e-05,
|
| 223 |
+
"loss": 0.71,
|
| 224 |
+
"step": 3600
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 1.18,
|
| 228 |
+
"learning_rate": 1.1782178217821782e-05,
|
| 229 |
+
"loss": 0.6825,
|
| 230 |
+
"step": 3700
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 1.21,
|
| 234 |
+
"learning_rate": 1.2101564995209199e-05,
|
| 235 |
+
"loss": 0.6792,
|
| 236 |
+
"step": 3800
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 1.25,
|
| 240 |
+
"learning_rate": 1.2420951772596615e-05,
|
| 241 |
+
"loss": 0.6794,
|
| 242 |
+
"step": 3900
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 1.28,
|
| 246 |
+
"learning_rate": 1.2740338549984032e-05,
|
| 247 |
+
"loss": 0.6672,
|
| 248 |
+
"step": 4000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 1.31,
|
| 252 |
+
"learning_rate": 1.3059725327371449e-05,
|
| 253 |
+
"loss": 0.6356,
|
| 254 |
+
"step": 4100
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 1.34,
|
| 258 |
+
"learning_rate": 1.3379112104758865e-05,
|
| 259 |
+
"loss": 0.6503,
|
| 260 |
+
"step": 4200
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 1.37,
|
| 264 |
+
"learning_rate": 1.369849888214628e-05,
|
| 265 |
+
"loss": 0.6495,
|
| 266 |
+
"step": 4300
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 1.41,
|
| 270 |
+
"learning_rate": 1.4017885659533696e-05,
|
| 271 |
+
"loss": 0.6428,
|
| 272 |
+
"step": 4400
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 1.44,
|
| 276 |
+
"learning_rate": 1.4337272436921113e-05,
|
| 277 |
+
"loss": 0.6171,
|
| 278 |
+
"step": 4500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 1.47,
|
| 282 |
+
"learning_rate": 1.4656659214308529e-05,
|
| 283 |
+
"loss": 0.6084,
|
| 284 |
+
"step": 4600
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 1.5,
|
| 288 |
+
"learning_rate": 1.4976045991695944e-05,
|
| 289 |
+
"loss": 0.6222,
|
| 290 |
+
"step": 4700
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 1.53,
|
| 294 |
+
"learning_rate": 1.529543276908336e-05,
|
| 295 |
+
"loss": 0.5971,
|
| 296 |
+
"step": 4800
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 1.56,
|
| 300 |
+
"learning_rate": 1.5614819546470777e-05,
|
| 301 |
+
"loss": 0.5745,
|
| 302 |
+
"step": 4900
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 1.6,
|
| 306 |
+
"learning_rate": 1.5934206323858196e-05,
|
| 307 |
+
"loss": 0.5753,
|
| 308 |
+
"step": 5000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 1.63,
|
| 312 |
+
"learning_rate": 1.625359310124561e-05,
|
| 313 |
+
"loss": 0.6045,
|
| 314 |
+
"step": 5100
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 1.66,
|
| 318 |
+
"learning_rate": 1.6572979878633027e-05,
|
| 319 |
+
"loss": 0.6006,
|
| 320 |
+
"step": 5200
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 1.69,
|
| 324 |
+
"learning_rate": 1.6892366656020443e-05,
|
| 325 |
+
"loss": 0.6003,
|
| 326 |
+
"step": 5300
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 1.72,
|
| 330 |
+
"learning_rate": 1.7211753433407858e-05,
|
| 331 |
+
"loss": 0.5688,
|
| 332 |
+
"step": 5400
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 1.76,
|
| 336 |
+
"learning_rate": 1.7531140210795274e-05,
|
| 337 |
+
"loss": 0.5662,
|
| 338 |
+
"step": 5500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 1.79,
|
| 342 |
+
"learning_rate": 1.785052698818269e-05,
|
| 343 |
+
"loss": 0.5788,
|
| 344 |
+
"step": 5600
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 1.82,
|
| 348 |
+
"learning_rate": 1.8169913765570105e-05,
|
| 349 |
+
"loss": 0.5541,
|
| 350 |
+
"step": 5700
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 1.85,
|
| 354 |
+
"learning_rate": 1.8489300542957524e-05,
|
| 355 |
+
"loss": 0.5656,
|
| 356 |
+
"step": 5800
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 1.88,
|
| 360 |
+
"learning_rate": 1.880868732034494e-05,
|
| 361 |
+
"loss": 0.5466,
|
| 362 |
+
"step": 5900
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 1.92,
|
| 366 |
+
"learning_rate": 1.9128074097732355e-05,
|
| 367 |
+
"loss": 0.563,
|
| 368 |
+
"step": 6000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 1.95,
|
| 372 |
+
"learning_rate": 1.9447460875119774e-05,
|
| 373 |
+
"loss": 0.5603,
|
| 374 |
+
"step": 6100
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 1.98,
|
| 378 |
+
"learning_rate": 1.976684765250719e-05,
|
| 379 |
+
"loss": 0.5579,
|
| 380 |
+
"step": 6200
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 2.01,
|
| 384 |
+
"learning_rate": 2.0086234429894605e-05,
|
| 385 |
+
"loss": 0.5346,
|
| 386 |
+
"step": 6300
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 2.04,
|
| 390 |
+
"learning_rate": 2.0405621207282017e-05,
|
| 391 |
+
"loss": 0.5038,
|
| 392 |
+
"step": 6400
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 2.08,
|
| 396 |
+
"learning_rate": 2.0725007984669436e-05,
|
| 397 |
+
"loss": 0.5121,
|
| 398 |
+
"step": 6500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 2.11,
|
| 402 |
+
"learning_rate": 2.1044394762056855e-05,
|
| 403 |
+
"loss": 0.5135,
|
| 404 |
+
"step": 6600
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 2.14,
|
| 408 |
+
"learning_rate": 2.1363781539444267e-05,
|
| 409 |
+
"loss": 0.504,
|
| 410 |
+
"step": 6700
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 2.17,
|
| 414 |
+
"learning_rate": 2.1683168316831686e-05,
|
| 415 |
+
"loss": 0.5119,
|
| 416 |
+
"step": 6800
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 2.2,
|
| 420 |
+
"learning_rate": 2.20025550942191e-05,
|
| 421 |
+
"loss": 0.4986,
|
| 422 |
+
"step": 6900
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 2.24,
|
| 426 |
+
"learning_rate": 2.2321941871606518e-05,
|
| 427 |
+
"loss": 0.504,
|
| 428 |
+
"step": 7000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 2.27,
|
| 432 |
+
"learning_rate": 2.2641328648993933e-05,
|
| 433 |
+
"loss": 0.5024,
|
| 434 |
+
"step": 7100
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 2.3,
|
| 438 |
+
"learning_rate": 2.296071542638135e-05,
|
| 439 |
+
"loss": 0.479,
|
| 440 |
+
"step": 7200
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 2.33,
|
| 444 |
+
"learning_rate": 2.3280102203768764e-05,
|
| 445 |
+
"loss": 0.492,
|
| 446 |
+
"step": 7300
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 2.36,
|
| 450 |
+
"learning_rate": 2.3599488981156183e-05,
|
| 451 |
+
"loss": 0.4928,
|
| 452 |
+
"step": 7400
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 2.4,
|
| 456 |
+
"learning_rate": 2.39188757585436e-05,
|
| 457 |
+
"loss": 0.4862,
|
| 458 |
+
"step": 7500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 2.43,
|
| 462 |
+
"learning_rate": 2.4238262535931014e-05,
|
| 463 |
+
"loss": 0.5148,
|
| 464 |
+
"step": 7600
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 2.46,
|
| 468 |
+
"learning_rate": 2.4557649313318433e-05,
|
| 469 |
+
"loss": 0.4922,
|
| 470 |
+
"step": 7700
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 2.49,
|
| 474 |
+
"learning_rate": 2.4877036090705846e-05,
|
| 475 |
+
"loss": 0.4882,
|
| 476 |
+
"step": 7800
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 2.52,
|
| 480 |
+
"learning_rate": 2.5196422868093264e-05,
|
| 481 |
+
"loss": 0.4871,
|
| 482 |
+
"step": 7900
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 2.56,
|
| 486 |
+
"learning_rate": 2.5515809645480677e-05,
|
| 487 |
+
"loss": 0.4967,
|
| 488 |
+
"step": 8000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 2.59,
|
| 492 |
+
"learning_rate": 2.5835196422868096e-05,
|
| 493 |
+
"loss": 0.4834,
|
| 494 |
+
"step": 8100
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 2.62,
|
| 498 |
+
"learning_rate": 2.615458320025551e-05,
|
| 499 |
+
"loss": 0.4769,
|
| 500 |
+
"step": 8200
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 2.65,
|
| 504 |
+
"learning_rate": 2.6473969977642927e-05,
|
| 505 |
+
"loss": 0.4776,
|
| 506 |
+
"step": 8300
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 2.68,
|
| 510 |
+
"learning_rate": 2.6793356755030342e-05,
|
| 511 |
+
"loss": 0.4861,
|
| 512 |
+
"step": 8400
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 2.71,
|
| 516 |
+
"learning_rate": 2.711274353241776e-05,
|
| 517 |
+
"loss": 0.486,
|
| 518 |
+
"step": 8500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 2.75,
|
| 522 |
+
"learning_rate": 2.7432130309805177e-05,
|
| 523 |
+
"loss": 0.4745,
|
| 524 |
+
"step": 8600
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 2.78,
|
| 528 |
+
"learning_rate": 2.7751517087192592e-05,
|
| 529 |
+
"loss": 0.4805,
|
| 530 |
+
"step": 8700
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 2.81,
|
| 534 |
+
"learning_rate": 2.807090386458001e-05,
|
| 535 |
+
"loss": 0.5003,
|
| 536 |
+
"step": 8800
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 2.84,
|
| 540 |
+
"learning_rate": 2.8390290641967424e-05,
|
| 541 |
+
"loss": 0.4639,
|
| 542 |
+
"step": 8900
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 2.87,
|
| 546 |
+
"learning_rate": 2.8709677419354843e-05,
|
| 547 |
+
"loss": 0.4763,
|
| 548 |
+
"step": 9000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 2.91,
|
| 552 |
+
"learning_rate": 2.9029064196742255e-05,
|
| 553 |
+
"loss": 0.4727,
|
| 554 |
+
"step": 9100
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 2.94,
|
| 558 |
+
"learning_rate": 2.9348450974129674e-05,
|
| 559 |
+
"loss": 0.4839,
|
| 560 |
+
"step": 9200
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 2.97,
|
| 564 |
+
"learning_rate": 2.966783775151709e-05,
|
| 565 |
+
"loss": 0.4769,
|
| 566 |
+
"step": 9300
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 3.0,
|
| 570 |
+
"learning_rate": 2.9987224528904505e-05,
|
| 571 |
+
"loss": 0.4641,
|
| 572 |
+
"step": 9400
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 3.03,
|
| 576 |
+
"learning_rate": 3.030661130629192e-05,
|
| 577 |
+
"loss": 0.4282,
|
| 578 |
+
"step": 9500
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 3.07,
|
| 582 |
+
"learning_rate": 3.0625998083679336e-05,
|
| 583 |
+
"loss": 0.4057,
|
| 584 |
+
"step": 9600
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 3.1,
|
| 588 |
+
"learning_rate": 3.094538486106676e-05,
|
| 589 |
+
"loss": 0.4192,
|
| 590 |
+
"step": 9700
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 3.13,
|
| 594 |
+
"learning_rate": 3.126477163845417e-05,
|
| 595 |
+
"loss": 0.4186,
|
| 596 |
+
"step": 9800
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 3.16,
|
| 600 |
+
"learning_rate": 3.158415841584159e-05,
|
| 601 |
+
"loss": 0.4086,
|
| 602 |
+
"step": 9900
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 3.19,
|
| 606 |
+
"learning_rate": 3.1903545193229005e-05,
|
| 607 |
+
"loss": 0.4194,
|
| 608 |
+
"step": 10000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 3.23,
|
| 612 |
+
"learning_rate": 3.221973810284254e-05,
|
| 613 |
+
"loss": 0.4443,
|
| 614 |
+
"step": 10100
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 3.26,
|
| 618 |
+
"learning_rate": 3.2539124880229965e-05,
|
| 619 |
+
"loss": 0.4217,
|
| 620 |
+
"step": 10200
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 3.29,
|
| 624 |
+
"learning_rate": 3.2858511657617374e-05,
|
| 625 |
+
"loss": 0.4275,
|
| 626 |
+
"step": 10300
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 3.32,
|
| 630 |
+
"learning_rate": 3.3177898435004796e-05,
|
| 631 |
+
"loss": 0.4422,
|
| 632 |
+
"step": 10400
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 3.35,
|
| 636 |
+
"learning_rate": 3.3497285212392205e-05,
|
| 637 |
+
"loss": 0.4227,
|
| 638 |
+
"step": 10500
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 3.39,
|
| 642 |
+
"learning_rate": 3.381667198977963e-05,
|
| 643 |
+
"loss": 0.4191,
|
| 644 |
+
"step": 10600
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 3.42,
|
| 648 |
+
"learning_rate": 3.413605876716704e-05,
|
| 649 |
+
"loss": 0.4309,
|
| 650 |
+
"step": 10700
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 3.45,
|
| 654 |
+
"learning_rate": 3.445544554455446e-05,
|
| 655 |
+
"loss": 0.4079,
|
| 656 |
+
"step": 10800
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 3.48,
|
| 660 |
+
"learning_rate": 3.4774832321941874e-05,
|
| 661 |
+
"loss": 0.4281,
|
| 662 |
+
"step": 10900
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 3.51,
|
| 666 |
+
"learning_rate": 3.509421909932929e-05,
|
| 667 |
+
"loss": 0.4137,
|
| 668 |
+
"step": 11000
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 3.55,
|
| 672 |
+
"learning_rate": 3.5413605876716705e-05,
|
| 673 |
+
"loss": 0.4168,
|
| 674 |
+
"step": 11100
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 3.58,
|
| 678 |
+
"learning_rate": 3.573299265410412e-05,
|
| 679 |
+
"loss": 0.4186,
|
| 680 |
+
"step": 11200
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 3.61,
|
| 684 |
+
"learning_rate": 3.605237943149154e-05,
|
| 685 |
+
"loss": 0.4325,
|
| 686 |
+
"step": 11300
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 3.64,
|
| 690 |
+
"learning_rate": 3.637176620887895e-05,
|
| 691 |
+
"loss": 0.4422,
|
| 692 |
+
"step": 11400
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 3.67,
|
| 696 |
+
"learning_rate": 3.6691152986266374e-05,
|
| 697 |
+
"loss": 0.4269,
|
| 698 |
+
"step": 11500
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 3.7,
|
| 702 |
+
"learning_rate": 3.701053976365379e-05,
|
| 703 |
+
"loss": 0.4368,
|
| 704 |
+
"step": 11600
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 3.74,
|
| 708 |
+
"learning_rate": 3.7329926541041205e-05,
|
| 709 |
+
"loss": 0.4207,
|
| 710 |
+
"step": 11700
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 3.77,
|
| 714 |
+
"learning_rate": 3.764931331842862e-05,
|
| 715 |
+
"loss": 0.4365,
|
| 716 |
+
"step": 11800
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 3.8,
|
| 720 |
+
"learning_rate": 3.7968700095816036e-05,
|
| 721 |
+
"loss": 0.4394,
|
| 722 |
+
"step": 11900
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 3.83,
|
| 726 |
+
"learning_rate": 3.828808687320345e-05,
|
| 727 |
+
"loss": 0.4351,
|
| 728 |
+
"step": 12000
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 3.86,
|
| 732 |
+
"learning_rate": 3.860747365059087e-05,
|
| 733 |
+
"loss": 0.4113,
|
| 734 |
+
"step": 12100
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 3.9,
|
| 738 |
+
"learning_rate": 3.892686042797828e-05,
|
| 739 |
+
"loss": 0.4261,
|
| 740 |
+
"step": 12200
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 3.93,
|
| 744 |
+
"learning_rate": 3.92462472053657e-05,
|
| 745 |
+
"loss": 0.4311,
|
| 746 |
+
"step": 12300
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 3.96,
|
| 750 |
+
"learning_rate": 3.956563398275312e-05,
|
| 751 |
+
"loss": 0.4246,
|
| 752 |
+
"step": 12400
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 3.99,
|
| 756 |
+
"learning_rate": 3.988502076014053e-05,
|
| 757 |
+
"loss": 0.4294,
|
| 758 |
+
"step": 12500
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 4.02,
|
| 762 |
+
"learning_rate": 3.997728805138579e-05,
|
| 763 |
+
"loss": 0.3854,
|
| 764 |
+
"step": 12600
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 4.06,
|
| 768 |
+
"learning_rate": 3.994180063167608e-05,
|
| 769 |
+
"loss": 0.3656,
|
| 770 |
+
"step": 12700
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 4.09,
|
| 774 |
+
"learning_rate": 3.9906313211966366e-05,
|
| 775 |
+
"loss": 0.3836,
|
| 776 |
+
"step": 12800
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 4.12,
|
| 780 |
+
"learning_rate": 3.987082579225665e-05,
|
| 781 |
+
"loss": 0.3664,
|
| 782 |
+
"step": 12900
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 4.15,
|
| 786 |
+
"learning_rate": 3.9835338372546935e-05,
|
| 787 |
+
"loss": 0.3884,
|
| 788 |
+
"step": 13000
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 4.18,
|
| 792 |
+
"learning_rate": 3.979985095283722e-05,
|
| 793 |
+
"loss": 0.3764,
|
| 794 |
+
"step": 13100
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 4.22,
|
| 798 |
+
"learning_rate": 3.976436353312751e-05,
|
| 799 |
+
"loss": 0.3733,
|
| 800 |
+
"step": 13200
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 4.25,
|
| 804 |
+
"learning_rate": 3.97288761134178e-05,
|
| 805 |
+
"loss": 0.3682,
|
| 806 |
+
"step": 13300
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 4.28,
|
| 810 |
+
"learning_rate": 3.969338869370809e-05,
|
| 811 |
+
"loss": 0.371,
|
| 812 |
+
"step": 13400
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 4.31,
|
| 816 |
+
"learning_rate": 3.965790127399837e-05,
|
| 817 |
+
"loss": 0.3741,
|
| 818 |
+
"step": 13500
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 4.34,
|
| 822 |
+
"learning_rate": 3.9622413854288656e-05,
|
| 823 |
+
"loss": 0.3757,
|
| 824 |
+
"step": 13600
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 4.38,
|
| 828 |
+
"learning_rate": 3.9586926434578944e-05,
|
| 829 |
+
"loss": 0.3652,
|
| 830 |
+
"step": 13700
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 4.41,
|
| 834 |
+
"learning_rate": 3.955143901486923e-05,
|
| 835 |
+
"loss": 0.3892,
|
| 836 |
+
"step": 13800
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 4.44,
|
| 840 |
+
"learning_rate": 3.951630646935662e-05,
|
| 841 |
+
"loss": 0.3893,
|
| 842 |
+
"step": 13900
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 4.47,
|
| 846 |
+
"learning_rate": 3.948081904964691e-05,
|
| 847 |
+
"loss": 0.3823,
|
| 848 |
+
"step": 14000
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 4.5,
|
| 852 |
+
"learning_rate": 3.944533162993719e-05,
|
| 853 |
+
"loss": 0.3654,
|
| 854 |
+
"step": 14100
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 4.54,
|
| 858 |
+
"learning_rate": 3.9409844210227476e-05,
|
| 859 |
+
"loss": 0.3791,
|
| 860 |
+
"step": 14200
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 4.57,
|
| 864 |
+
"learning_rate": 3.9374356790517764e-05,
|
| 865 |
+
"loss": 0.3918,
|
| 866 |
+
"step": 14300
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 4.6,
|
| 870 |
+
"learning_rate": 3.933886937080805e-05,
|
| 871 |
+
"loss": 0.3713,
|
| 872 |
+
"step": 14400
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 4.63,
|
| 876 |
+
"learning_rate": 3.930338195109834e-05,
|
| 877 |
+
"loss": 0.376,
|
| 878 |
+
"step": 14500
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 4.66,
|
| 882 |
+
"learning_rate": 3.926789453138863e-05,
|
| 883 |
+
"loss": 0.3682,
|
| 884 |
+
"step": 14600
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 4.69,
|
| 888 |
+
"learning_rate": 3.9232407111678915e-05,
|
| 889 |
+
"loss": 0.374,
|
| 890 |
+
"step": 14700
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 4.73,
|
| 894 |
+
"learning_rate": 3.91969196919692e-05,
|
| 895 |
+
"loss": 0.3774,
|
| 896 |
+
"step": 14800
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 4.76,
|
| 900 |
+
"learning_rate": 3.9161432272259484e-05,
|
| 901 |
+
"loss": 0.3806,
|
| 902 |
+
"step": 14900
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 4.79,
|
| 906 |
+
"learning_rate": 3.912594485254977e-05,
|
| 907 |
+
"loss": 0.3886,
|
| 908 |
+
"step": 15000
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 4.82,
|
| 912 |
+
"learning_rate": 3.909045743284006e-05,
|
| 913 |
+
"loss": 0.3742,
|
| 914 |
+
"step": 15100
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 4.85,
|
| 918 |
+
"learning_rate": 3.905497001313035e-05,
|
| 919 |
+
"loss": 0.3714,
|
| 920 |
+
"step": 15200
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 4.89,
|
| 924 |
+
"learning_rate": 3.9019482593420636e-05,
|
| 925 |
+
"loss": 0.3721,
|
| 926 |
+
"step": 15300
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 4.92,
|
| 930 |
+
"learning_rate": 3.8983995173710924e-05,
|
| 931 |
+
"loss": 0.3716,
|
| 932 |
+
"step": 15400
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 4.95,
|
| 936 |
+
"learning_rate": 3.8948507754001205e-05,
|
| 937 |
+
"loss": 0.3728,
|
| 938 |
+
"step": 15500
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 4.98,
|
| 942 |
+
"learning_rate": 3.891302033429149e-05,
|
| 943 |
+
"loss": 0.384,
|
| 944 |
+
"step": 15600
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 5.01,
|
| 948 |
+
"learning_rate": 3.887753291458178e-05,
|
| 949 |
+
"loss": 0.3415,
|
| 950 |
+
"step": 15700
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 5.05,
|
| 954 |
+
"learning_rate": 3.884204549487207e-05,
|
| 955 |
+
"loss": 0.3064,
|
| 956 |
+
"step": 15800
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 5.08,
|
| 960 |
+
"learning_rate": 3.880655807516236e-05,
|
| 961 |
+
"loss": 0.3164,
|
| 962 |
+
"step": 15900
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 5.11,
|
| 966 |
+
"learning_rate": 3.8771070655452645e-05,
|
| 967 |
+
"loss": 0.3215,
|
| 968 |
+
"step": 16000
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 5.14,
|
| 972 |
+
"learning_rate": 3.873558323574293e-05,
|
| 973 |
+
"loss": 0.3234,
|
| 974 |
+
"step": 16100
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 5.17,
|
| 978 |
+
"learning_rate": 3.870009581603322e-05,
|
| 979 |
+
"loss": 0.3119,
|
| 980 |
+
"step": 16200
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 5.21,
|
| 984 |
+
"learning_rate": 3.866460839632351e-05,
|
| 985 |
+
"loss": 0.3145,
|
| 986 |
+
"step": 16300
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 5.24,
|
| 990 |
+
"learning_rate": 3.8629120976613796e-05,
|
| 991 |
+
"loss": 0.3333,
|
| 992 |
+
"step": 16400
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 5.27,
|
| 996 |
+
"learning_rate": 3.8593633556904084e-05,
|
| 997 |
+
"loss": 0.3195,
|
| 998 |
+
"step": 16500
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 5.3,
|
| 1002 |
+
"learning_rate": 3.855814613719437e-05,
|
| 1003 |
+
"loss": 0.3159,
|
| 1004 |
+
"step": 16600
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 5.33,
|
| 1008 |
+
"learning_rate": 3.852265871748465e-05,
|
| 1009 |
+
"loss": 0.3189,
|
| 1010 |
+
"step": 16700
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 5.37,
|
| 1014 |
+
"learning_rate": 3.848717129777494e-05,
|
| 1015 |
+
"loss": 0.317,
|
| 1016 |
+
"step": 16800
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 5.4,
|
| 1020 |
+
"learning_rate": 3.845168387806523e-05,
|
| 1021 |
+
"loss": 0.3231,
|
| 1022 |
+
"step": 16900
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 5.43,
|
| 1026 |
+
"learning_rate": 3.841619645835552e-05,
|
| 1027 |
+
"loss": 0.3328,
|
| 1028 |
+
"step": 17000
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 5.46,
|
| 1032 |
+
"learning_rate": 3.8380709038645805e-05,
|
| 1033 |
+
"loss": 0.3282,
|
| 1034 |
+
"step": 17100
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 5.49,
|
| 1038 |
+
"learning_rate": 3.8345221618936086e-05,
|
| 1039 |
+
"loss": 0.3169,
|
| 1040 |
+
"step": 17200
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 5.53,
|
| 1044 |
+
"learning_rate": 3.8309734199226374e-05,
|
| 1045 |
+
"loss": 0.3156,
|
| 1046 |
+
"step": 17300
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 5.56,
|
| 1050 |
+
"learning_rate": 3.827424677951666e-05,
|
| 1051 |
+
"loss": 0.3313,
|
| 1052 |
+
"step": 17400
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 5.59,
|
| 1056 |
+
"learning_rate": 3.823875935980695e-05,
|
| 1057 |
+
"loss": 0.3181,
|
| 1058 |
+
"step": 17500
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 5.62,
|
| 1062 |
+
"learning_rate": 3.820327194009724e-05,
|
| 1063 |
+
"loss": 0.334,
|
| 1064 |
+
"step": 17600
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 5.65,
|
| 1068 |
+
"learning_rate": 3.8167784520387526e-05,
|
| 1069 |
+
"loss": 0.3313,
|
| 1070 |
+
"step": 17700
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 5.69,
|
| 1074 |
+
"learning_rate": 3.8132297100677814e-05,
|
| 1075 |
+
"loss": 0.3331,
|
| 1076 |
+
"step": 17800
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 5.72,
|
| 1080 |
+
"learning_rate": 3.80968096809681e-05,
|
| 1081 |
+
"loss": 0.3405,
|
| 1082 |
+
"step": 17900
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 5.75,
|
| 1086 |
+
"learning_rate": 3.806132226125839e-05,
|
| 1087 |
+
"loss": 0.3345,
|
| 1088 |
+
"step": 18000
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 5.78,
|
| 1092 |
+
"learning_rate": 3.802583484154868e-05,
|
| 1093 |
+
"loss": 0.3295,
|
| 1094 |
+
"step": 18100
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 5.81,
|
| 1098 |
+
"learning_rate": 3.7990347421838965e-05,
|
| 1099 |
+
"loss": 0.3244,
|
| 1100 |
+
"step": 18200
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 5.84,
|
| 1104 |
+
"learning_rate": 3.795486000212925e-05,
|
| 1105 |
+
"loss": 0.3346,
|
| 1106 |
+
"step": 18300
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 5.88,
|
| 1110 |
+
"learning_rate": 3.791972745661663e-05,
|
| 1111 |
+
"loss": 0.3317,
|
| 1112 |
+
"step": 18400
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 5.91,
|
| 1116 |
+
"learning_rate": 3.788424003690692e-05,
|
| 1117 |
+
"loss": 0.3219,
|
| 1118 |
+
"step": 18500
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 5.94,
|
| 1122 |
+
"learning_rate": 3.784875261719721e-05,
|
| 1123 |
+
"loss": 0.3167,
|
| 1124 |
+
"step": 18600
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 5.97,
|
| 1128 |
+
"learning_rate": 3.781326519748749e-05,
|
| 1129 |
+
"loss": 0.33,
|
| 1130 |
+
"step": 18700
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 6.0,
|
| 1134 |
+
"learning_rate": 3.777777777777778e-05,
|
| 1135 |
+
"loss": 0.3211,
|
| 1136 |
+
"step": 18800
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 6.04,
|
| 1140 |
+
"learning_rate": 3.7742290358068066e-05,
|
| 1141 |
+
"loss": 0.2794,
|
| 1142 |
+
"step": 18900
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 6.07,
|
| 1146 |
+
"learning_rate": 3.7706802938358354e-05,
|
| 1147 |
+
"loss": 0.2675,
|
| 1148 |
+
"step": 19000
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 6.1,
|
| 1152 |
+
"learning_rate": 3.767131551864864e-05,
|
| 1153 |
+
"loss": 0.2676,
|
| 1154 |
+
"step": 19100
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 6.13,
|
| 1158 |
+
"learning_rate": 3.763582809893893e-05,
|
| 1159 |
+
"loss": 0.2693,
|
| 1160 |
+
"step": 19200
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 6.16,
|
| 1164 |
+
"learning_rate": 3.760034067922922e-05,
|
| 1165 |
+
"loss": 0.2669,
|
| 1166 |
+
"step": 19300
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 6.2,
|
| 1170 |
+
"learning_rate": 3.7564853259519506e-05,
|
| 1171 |
+
"loss": 0.2744,
|
| 1172 |
+
"step": 19400
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 6.23,
|
| 1176 |
+
"learning_rate": 3.7529365839809794e-05,
|
| 1177 |
+
"loss": 0.282,
|
| 1178 |
+
"step": 19500
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 6.26,
|
| 1182 |
+
"learning_rate": 3.749387842010008e-05,
|
| 1183 |
+
"loss": 0.2848,
|
| 1184 |
+
"step": 19600
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 6.29,
|
| 1188 |
+
"learning_rate": 3.745839100039037e-05,
|
| 1189 |
+
"loss": 0.2783,
|
| 1190 |
+
"step": 19700
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 6.32,
|
| 1194 |
+
"learning_rate": 3.742290358068066e-05,
|
| 1195 |
+
"loss": 0.2843,
|
| 1196 |
+
"step": 19800
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 6.36,
|
| 1200 |
+
"learning_rate": 3.738741616097094e-05,
|
| 1201 |
+
"loss": 0.2696,
|
| 1202 |
+
"step": 19900
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 6.39,
|
| 1206 |
+
"learning_rate": 3.7351928741261226e-05,
|
| 1207 |
+
"loss": 0.2932,
|
| 1208 |
+
"step": 20000
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 6.42,
|
| 1212 |
+
"learning_rate": 3.7316441321551514e-05,
|
| 1213 |
+
"loss": 0.2775,
|
| 1214 |
+
"step": 20100
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 6.45,
|
| 1218 |
+
"learning_rate": 3.72809539018418e-05,
|
| 1219 |
+
"loss": 0.2893,
|
| 1220 |
+
"step": 20200
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 6.48,
|
| 1224 |
+
"learning_rate": 3.724546648213209e-05,
|
| 1225 |
+
"loss": 0.286,
|
| 1226 |
+
"step": 20300
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 6.52,
|
| 1230 |
+
"learning_rate": 3.720997906242237e-05,
|
| 1231 |
+
"loss": 0.2799,
|
| 1232 |
+
"step": 20400
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 6.55,
|
| 1236 |
+
"learning_rate": 3.717449164271266e-05,
|
| 1237 |
+
"loss": 0.2827,
|
| 1238 |
+
"step": 20500
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 6.58,
|
| 1242 |
+
"learning_rate": 3.713900422300295e-05,
|
| 1243 |
+
"loss": 0.2783,
|
| 1244 |
+
"step": 20600
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 6.61,
|
| 1248 |
+
"learning_rate": 3.7103516803293235e-05,
|
| 1249 |
+
"loss": 0.2729,
|
| 1250 |
+
"step": 20700
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 6.64,
|
| 1254 |
+
"learning_rate": 3.706802938358352e-05,
|
| 1255 |
+
"loss": 0.2747,
|
| 1256 |
+
"step": 20800
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 6.68,
|
| 1260 |
+
"learning_rate": 3.703254196387381e-05,
|
| 1261 |
+
"loss": 0.2764,
|
| 1262 |
+
"step": 20900
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 6.71,
|
| 1266 |
+
"learning_rate": 3.699705454416409e-05,
|
| 1267 |
+
"loss": 0.2804,
|
| 1268 |
+
"step": 21000
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 6.74,
|
| 1272 |
+
"learning_rate": 3.696156712445438e-05,
|
| 1273 |
+
"loss": 0.2841,
|
| 1274 |
+
"step": 21100
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 6.77,
|
| 1278 |
+
"learning_rate": 3.692607970474467e-05,
|
| 1279 |
+
"loss": 0.277,
|
| 1280 |
+
"step": 21200
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 6.8,
|
| 1284 |
+
"learning_rate": 3.6890592285034956e-05,
|
| 1285 |
+
"loss": 0.284,
|
| 1286 |
+
"step": 21300
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 6.83,
|
| 1290 |
+
"learning_rate": 3.6855104865325244e-05,
|
| 1291 |
+
"loss": 0.2778,
|
| 1292 |
+
"step": 21400
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 6.87,
|
| 1296 |
+
"learning_rate": 3.681961744561553e-05,
|
| 1297 |
+
"loss": 0.2797,
|
| 1298 |
+
"step": 21500
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 6.9,
|
| 1302 |
+
"learning_rate": 3.678413002590582e-05,
|
| 1303 |
+
"loss": 0.3036,
|
| 1304 |
+
"step": 21600
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 6.93,
|
| 1308 |
+
"learning_rate": 3.674864260619611e-05,
|
| 1309 |
+
"loss": 0.2869,
|
| 1310 |
+
"step": 21700
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 6.96,
|
| 1314 |
+
"learning_rate": 3.6713155186486395e-05,
|
| 1315 |
+
"loss": 0.2921,
|
| 1316 |
+
"step": 21800
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 6.99,
|
| 1320 |
+
"learning_rate": 3.667766776677668e-05,
|
| 1321 |
+
"loss": 0.2844,
|
| 1322 |
+
"step": 21900
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 7.03,
|
| 1326 |
+
"learning_rate": 3.664218034706697e-05,
|
| 1327 |
+
"loss": 0.2427,
|
| 1328 |
+
"step": 22000
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 7.06,
|
| 1332 |
+
"learning_rate": 3.660669292735726e-05,
|
| 1333 |
+
"loss": 0.243,
|
| 1334 |
+
"step": 22100
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 7.09,
|
| 1338 |
+
"learning_rate": 3.657120550764754e-05,
|
| 1339 |
+
"loss": 0.2311,
|
| 1340 |
+
"step": 22200
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 7.12,
|
| 1344 |
+
"learning_rate": 3.653571808793783e-05,
|
| 1345 |
+
"loss": 0.2358,
|
| 1346 |
+
"step": 22300
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 7.15,
|
| 1350 |
+
"learning_rate": 3.6500230668228116e-05,
|
| 1351 |
+
"loss": 0.2287,
|
| 1352 |
+
"step": 22400
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 7.19,
|
| 1356 |
+
"learning_rate": 3.6464743248518404e-05,
|
| 1357 |
+
"loss": 0.2519,
|
| 1358 |
+
"step": 22500
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 7.22,
|
| 1362 |
+
"learning_rate": 3.642925582880869e-05,
|
| 1363 |
+
"loss": 0.2528,
|
| 1364 |
+
"step": 22600
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 7.25,
|
| 1368 |
+
"learning_rate": 3.639376840909898e-05,
|
| 1369 |
+
"loss": 0.2507,
|
| 1370 |
+
"step": 22700
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 7.28,
|
| 1374 |
+
"learning_rate": 3.635828098938926e-05,
|
| 1375 |
+
"loss": 0.2496,
|
| 1376 |
+
"step": 22800
|
| 1377 |
+
},
|
| 1378 |
+
{
|
| 1379 |
+
"epoch": 7.31,
|
| 1380 |
+
"learning_rate": 3.632279356967955e-05,
|
| 1381 |
+
"loss": 0.2503,
|
| 1382 |
+
"step": 22900
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 7.35,
|
| 1386 |
+
"learning_rate": 3.628730614996984e-05,
|
| 1387 |
+
"loss": 0.2383,
|
| 1388 |
+
"step": 23000
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 7.38,
|
| 1392 |
+
"learning_rate": 3.6251818730260125e-05,
|
| 1393 |
+
"loss": 0.2513,
|
| 1394 |
+
"step": 23100
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 7.41,
|
| 1398 |
+
"learning_rate": 3.621633131055041e-05,
|
| 1399 |
+
"loss": 0.2492,
|
| 1400 |
+
"step": 23200
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 7.44,
|
| 1404 |
+
"learning_rate": 3.61808438908407e-05,
|
| 1405 |
+
"loss": 0.2473,
|
| 1406 |
+
"step": 23300
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 7.47,
|
| 1410 |
+
"learning_rate": 3.614535647113099e-05,
|
| 1411 |
+
"loss": 0.2502,
|
| 1412 |
+
"step": 23400
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 7.51,
|
| 1416 |
+
"learning_rate": 3.6109869051421276e-05,
|
| 1417 |
+
"loss": 0.2492,
|
| 1418 |
+
"step": 23500
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 7.54,
|
| 1422 |
+
"learning_rate": 3.6074381631711564e-05,
|
| 1423 |
+
"loss": 0.2366,
|
| 1424 |
+
"step": 23600
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 7.57,
|
| 1428 |
+
"learning_rate": 3.603889421200185e-05,
|
| 1429 |
+
"loss": 0.2393,
|
| 1430 |
+
"step": 23700
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 7.6,
|
| 1434 |
+
"learning_rate": 3.600340679229214e-05,
|
| 1435 |
+
"loss": 0.2483,
|
| 1436 |
+
"step": 23800
|
| 1437 |
+
},
|
| 1438 |
+
{
|
| 1439 |
+
"epoch": 7.63,
|
| 1440 |
+
"learning_rate": 3.596791937258242e-05,
|
| 1441 |
+
"loss": 0.2428,
|
| 1442 |
+
"step": 23900
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 7.67,
|
| 1446 |
+
"learning_rate": 3.593243195287271e-05,
|
| 1447 |
+
"loss": 0.2507,
|
| 1448 |
+
"step": 24000
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 7.7,
|
| 1452 |
+
"learning_rate": 3.5896944533163e-05,
|
| 1453 |
+
"loss": 0.2491,
|
| 1454 |
+
"step": 24100
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 7.73,
|
| 1458 |
+
"learning_rate": 3.5861457113453285e-05,
|
| 1459 |
+
"loss": 0.2549,
|
| 1460 |
+
"step": 24200
|
| 1461 |
+
},
|
| 1462 |
+
{
|
| 1463 |
+
"epoch": 7.76,
|
| 1464 |
+
"learning_rate": 3.582596969374357e-05,
|
| 1465 |
+
"loss": 0.251,
|
| 1466 |
+
"step": 24300
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 7.79,
|
| 1470 |
+
"learning_rate": 3.579048227403386e-05,
|
| 1471 |
+
"loss": 0.2357,
|
| 1472 |
+
"step": 24400
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 7.82,
|
| 1476 |
+
"learning_rate": 3.575499485432414e-05,
|
| 1477 |
+
"loss": 0.2441,
|
| 1478 |
+
"step": 24500
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 7.86,
|
| 1482 |
+
"learning_rate": 3.571950743461443e-05,
|
| 1483 |
+
"loss": 0.2561,
|
| 1484 |
+
"step": 24600
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 7.89,
|
| 1488 |
+
"learning_rate": 3.568402001490472e-05,
|
| 1489 |
+
"loss": 0.2524,
|
| 1490 |
+
"step": 24700
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 7.92,
|
| 1494 |
+
"learning_rate": 3.5648532595195006e-05,
|
| 1495 |
+
"loss": 0.2467,
|
| 1496 |
+
"step": 24800
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 7.95,
|
| 1500 |
+
"learning_rate": 3.5613045175485293e-05,
|
| 1501 |
+
"loss": 0.2416,
|
| 1502 |
+
"step": 24900
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 7.98,
|
| 1506 |
+
"learning_rate": 3.557755775577558e-05,
|
| 1507 |
+
"loss": 0.2511,
|
| 1508 |
+
"step": 25000
|
| 1509 |
+
}
|
| 1510 |
+
],
|
| 1511 |
+
"max_steps": 125240,
|
| 1512 |
+
"num_train_epochs": 40,
|
| 1513 |
+
"total_flos": 0.0,
|
| 1514 |
+
"trial_name": null,
|
| 1515 |
+
"trial_params": null
|
| 1516 |
+
}
|
checkpoint-30000/config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "Davlan/afro-xlmr-base",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"XLMRobertaModel"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 768,
|
| 13 |
+
"id2label": {
|
| 14 |
+
"0": "LABEL_0"
|
| 15 |
+
},
|
| 16 |
+
"initializer_range": 0.02,
|
| 17 |
+
"intermediate_size": 3072,
|
| 18 |
+
"label2id": {
|
| 19 |
+
"LABEL_0": 0
|
| 20 |
+
},
|
| 21 |
+
"layer_norm_eps": 1e-05,
|
| 22 |
+
"max_position_embeddings": 514,
|
| 23 |
+
"model_type": "xlm-roberta",
|
| 24 |
+
"num_attention_heads": 12,
|
| 25 |
+
"num_hidden_layers": 12,
|
| 26 |
+
"output_past": true,
|
| 27 |
+
"pad_token_id": 1,
|
| 28 |
+
"position_embedding_type": "absolute",
|
| 29 |
+
"torch_dtype": "float32",
|
| 30 |
+
"transformers_version": "4.29.0",
|
| 31 |
+
"type_vocab_size": 1,
|
| 32 |
+
"use_cache": true,
|
| 33 |
+
"vocab_size": 250002
|
| 34 |
+
}
|
checkpoint-30000/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a52ac2fb990c957eb21f774d392ec7dfb4c87b988a281f1d377bf89caec2e57b
|
| 3 |
+
size 1112246250
|
checkpoint-30000/trainer_state.json
ADDED
|
@@ -0,0 +1,1816 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 9.581603321622485,
|
| 5 |
+
"global_step": 30000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.03,
|
| 12 |
+
"learning_rate": 2.8744809964867454e-07,
|
| 13 |
+
"loss": 91.176,
|
| 14 |
+
"step": 100
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.06,
|
| 18 |
+
"learning_rate": 6.068348770360907e-07,
|
| 19 |
+
"loss": 64.0684,
|
| 20 |
+
"step": 200
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.1,
|
| 24 |
+
"learning_rate": 9.26221654423507e-07,
|
| 25 |
+
"loss": 27.1061,
|
| 26 |
+
"step": 300
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.13,
|
| 30 |
+
"learning_rate": 1.242414564037049e-06,
|
| 31 |
+
"loss": 8.0198,
|
| 32 |
+
"step": 400
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.16,
|
| 36 |
+
"learning_rate": 1.561801341424465e-06,
|
| 37 |
+
"loss": 4.7691,
|
| 38 |
+
"step": 500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.19,
|
| 42 |
+
"learning_rate": 1.8811881188118813e-06,
|
| 43 |
+
"loss": 3.4387,
|
| 44 |
+
"step": 600
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.22,
|
| 48 |
+
"learning_rate": 2.2005748961992977e-06,
|
| 49 |
+
"loss": 2.7027,
|
| 50 |
+
"step": 700
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.26,
|
| 54 |
+
"learning_rate": 2.5199616735867137e-06,
|
| 55 |
+
"loss": 2.4431,
|
| 56 |
+
"step": 800
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.29,
|
| 60 |
+
"learning_rate": 2.8393484509741297e-06,
|
| 61 |
+
"loss": 2.0534,
|
| 62 |
+
"step": 900
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.32,
|
| 66 |
+
"learning_rate": 3.1587352283615465e-06,
|
| 67 |
+
"loss": 1.8537,
|
| 68 |
+
"step": 1000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.35,
|
| 72 |
+
"learning_rate": 3.4781220057489625e-06,
|
| 73 |
+
"loss": 1.642,
|
| 74 |
+
"step": 1100
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.38,
|
| 78 |
+
"learning_rate": 3.797508783136378e-06,
|
| 79 |
+
"loss": 1.5092,
|
| 80 |
+
"step": 1200
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.42,
|
| 84 |
+
"learning_rate": 4.1168955605237945e-06,
|
| 85 |
+
"loss": 1.4302,
|
| 86 |
+
"step": 1300
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.45,
|
| 90 |
+
"learning_rate": 4.436282337911211e-06,
|
| 91 |
+
"loss": 1.3527,
|
| 92 |
+
"step": 1400
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.48,
|
| 96 |
+
"learning_rate": 4.755669115298627e-06,
|
| 97 |
+
"loss": 1.2495,
|
| 98 |
+
"step": 1500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.51,
|
| 102 |
+
"learning_rate": 5.075055892686043e-06,
|
| 103 |
+
"loss": 1.2039,
|
| 104 |
+
"step": 1600
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.54,
|
| 108 |
+
"learning_rate": 5.39444267007346e-06,
|
| 109 |
+
"loss": 1.1438,
|
| 110 |
+
"step": 1700
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.57,
|
| 114 |
+
"learning_rate": 5.713829447460876e-06,
|
| 115 |
+
"loss": 1.1056,
|
| 116 |
+
"step": 1800
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.61,
|
| 120 |
+
"learning_rate": 6.033216224848291e-06,
|
| 121 |
+
"loss": 1.0436,
|
| 122 |
+
"step": 1900
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.64,
|
| 126 |
+
"learning_rate": 6.352603002235708e-06,
|
| 127 |
+
"loss": 1.0225,
|
| 128 |
+
"step": 2000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.67,
|
| 132 |
+
"learning_rate": 6.671989779623123e-06,
|
| 133 |
+
"loss": 0.9856,
|
| 134 |
+
"step": 2100
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.7,
|
| 138 |
+
"learning_rate": 6.9913765570105406e-06,
|
| 139 |
+
"loss": 0.9412,
|
| 140 |
+
"step": 2200
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.73,
|
| 144 |
+
"learning_rate": 7.310763334397957e-06,
|
| 145 |
+
"loss": 0.9075,
|
| 146 |
+
"step": 2300
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.77,
|
| 150 |
+
"learning_rate": 7.630150111785373e-06,
|
| 151 |
+
"loss": 0.8752,
|
| 152 |
+
"step": 2400
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.8,
|
| 156 |
+
"learning_rate": 7.949536889172788e-06,
|
| 157 |
+
"loss": 0.854,
|
| 158 |
+
"step": 2500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.83,
|
| 162 |
+
"learning_rate": 8.268923666560205e-06,
|
| 163 |
+
"loss": 0.8282,
|
| 164 |
+
"step": 2600
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.86,
|
| 168 |
+
"learning_rate": 8.588310443947621e-06,
|
| 169 |
+
"loss": 0.8074,
|
| 170 |
+
"step": 2700
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.89,
|
| 174 |
+
"learning_rate": 8.907697221335038e-06,
|
| 175 |
+
"loss": 0.7976,
|
| 176 |
+
"step": 2800
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.93,
|
| 180 |
+
"learning_rate": 9.227083998722454e-06,
|
| 181 |
+
"loss": 0.7955,
|
| 182 |
+
"step": 2900
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.96,
|
| 186 |
+
"learning_rate": 9.54647077610987e-06,
|
| 187 |
+
"loss": 0.7502,
|
| 188 |
+
"step": 3000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.99,
|
| 192 |
+
"learning_rate": 9.865857553497285e-06,
|
| 193 |
+
"loss": 0.7455,
|
| 194 |
+
"step": 3100
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 1.02,
|
| 198 |
+
"learning_rate": 1.0185244330884702e-05,
|
| 199 |
+
"loss": 0.7177,
|
| 200 |
+
"step": 3200
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 1.05,
|
| 204 |
+
"learning_rate": 1.050463110827212e-05,
|
| 205 |
+
"loss": 0.724,
|
| 206 |
+
"step": 3300
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 1.09,
|
| 210 |
+
"learning_rate": 1.0824017885659535e-05,
|
| 211 |
+
"loss": 0.6975,
|
| 212 |
+
"step": 3400
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 1.12,
|
| 216 |
+
"learning_rate": 1.114340466304695e-05,
|
| 217 |
+
"loss": 0.7054,
|
| 218 |
+
"step": 3500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 1.15,
|
| 222 |
+
"learning_rate": 1.1462791440434366e-05,
|
| 223 |
+
"loss": 0.71,
|
| 224 |
+
"step": 3600
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 1.18,
|
| 228 |
+
"learning_rate": 1.1782178217821782e-05,
|
| 229 |
+
"loss": 0.6825,
|
| 230 |
+
"step": 3700
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 1.21,
|
| 234 |
+
"learning_rate": 1.2101564995209199e-05,
|
| 235 |
+
"loss": 0.6792,
|
| 236 |
+
"step": 3800
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 1.25,
|
| 240 |
+
"learning_rate": 1.2420951772596615e-05,
|
| 241 |
+
"loss": 0.6794,
|
| 242 |
+
"step": 3900
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 1.28,
|
| 246 |
+
"learning_rate": 1.2740338549984032e-05,
|
| 247 |
+
"loss": 0.6672,
|
| 248 |
+
"step": 4000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 1.31,
|
| 252 |
+
"learning_rate": 1.3059725327371449e-05,
|
| 253 |
+
"loss": 0.6356,
|
| 254 |
+
"step": 4100
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 1.34,
|
| 258 |
+
"learning_rate": 1.3379112104758865e-05,
|
| 259 |
+
"loss": 0.6503,
|
| 260 |
+
"step": 4200
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 1.37,
|
| 264 |
+
"learning_rate": 1.369849888214628e-05,
|
| 265 |
+
"loss": 0.6495,
|
| 266 |
+
"step": 4300
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 1.41,
|
| 270 |
+
"learning_rate": 1.4017885659533696e-05,
|
| 271 |
+
"loss": 0.6428,
|
| 272 |
+
"step": 4400
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 1.44,
|
| 276 |
+
"learning_rate": 1.4337272436921113e-05,
|
| 277 |
+
"loss": 0.6171,
|
| 278 |
+
"step": 4500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 1.47,
|
| 282 |
+
"learning_rate": 1.4656659214308529e-05,
|
| 283 |
+
"loss": 0.6084,
|
| 284 |
+
"step": 4600
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 1.5,
|
| 288 |
+
"learning_rate": 1.4976045991695944e-05,
|
| 289 |
+
"loss": 0.6222,
|
| 290 |
+
"step": 4700
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 1.53,
|
| 294 |
+
"learning_rate": 1.529543276908336e-05,
|
| 295 |
+
"loss": 0.5971,
|
| 296 |
+
"step": 4800
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 1.56,
|
| 300 |
+
"learning_rate": 1.5614819546470777e-05,
|
| 301 |
+
"loss": 0.5745,
|
| 302 |
+
"step": 4900
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 1.6,
|
| 306 |
+
"learning_rate": 1.5934206323858196e-05,
|
| 307 |
+
"loss": 0.5753,
|
| 308 |
+
"step": 5000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 1.63,
|
| 312 |
+
"learning_rate": 1.625359310124561e-05,
|
| 313 |
+
"loss": 0.6045,
|
| 314 |
+
"step": 5100
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 1.66,
|
| 318 |
+
"learning_rate": 1.6572979878633027e-05,
|
| 319 |
+
"loss": 0.6006,
|
| 320 |
+
"step": 5200
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 1.69,
|
| 324 |
+
"learning_rate": 1.6892366656020443e-05,
|
| 325 |
+
"loss": 0.6003,
|
| 326 |
+
"step": 5300
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 1.72,
|
| 330 |
+
"learning_rate": 1.7211753433407858e-05,
|
| 331 |
+
"loss": 0.5688,
|
| 332 |
+
"step": 5400
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 1.76,
|
| 336 |
+
"learning_rate": 1.7531140210795274e-05,
|
| 337 |
+
"loss": 0.5662,
|
| 338 |
+
"step": 5500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 1.79,
|
| 342 |
+
"learning_rate": 1.785052698818269e-05,
|
| 343 |
+
"loss": 0.5788,
|
| 344 |
+
"step": 5600
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 1.82,
|
| 348 |
+
"learning_rate": 1.8169913765570105e-05,
|
| 349 |
+
"loss": 0.5541,
|
| 350 |
+
"step": 5700
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 1.85,
|
| 354 |
+
"learning_rate": 1.8489300542957524e-05,
|
| 355 |
+
"loss": 0.5656,
|
| 356 |
+
"step": 5800
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 1.88,
|
| 360 |
+
"learning_rate": 1.880868732034494e-05,
|
| 361 |
+
"loss": 0.5466,
|
| 362 |
+
"step": 5900
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 1.92,
|
| 366 |
+
"learning_rate": 1.9128074097732355e-05,
|
| 367 |
+
"loss": 0.563,
|
| 368 |
+
"step": 6000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 1.95,
|
| 372 |
+
"learning_rate": 1.9447460875119774e-05,
|
| 373 |
+
"loss": 0.5603,
|
| 374 |
+
"step": 6100
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 1.98,
|
| 378 |
+
"learning_rate": 1.976684765250719e-05,
|
| 379 |
+
"loss": 0.5579,
|
| 380 |
+
"step": 6200
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 2.01,
|
| 384 |
+
"learning_rate": 2.0086234429894605e-05,
|
| 385 |
+
"loss": 0.5346,
|
| 386 |
+
"step": 6300
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 2.04,
|
| 390 |
+
"learning_rate": 2.0405621207282017e-05,
|
| 391 |
+
"loss": 0.5038,
|
| 392 |
+
"step": 6400
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 2.08,
|
| 396 |
+
"learning_rate": 2.0725007984669436e-05,
|
| 397 |
+
"loss": 0.5121,
|
| 398 |
+
"step": 6500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 2.11,
|
| 402 |
+
"learning_rate": 2.1044394762056855e-05,
|
| 403 |
+
"loss": 0.5135,
|
| 404 |
+
"step": 6600
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 2.14,
|
| 408 |
+
"learning_rate": 2.1363781539444267e-05,
|
| 409 |
+
"loss": 0.504,
|
| 410 |
+
"step": 6700
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 2.17,
|
| 414 |
+
"learning_rate": 2.1683168316831686e-05,
|
| 415 |
+
"loss": 0.5119,
|
| 416 |
+
"step": 6800
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 2.2,
|
| 420 |
+
"learning_rate": 2.20025550942191e-05,
|
| 421 |
+
"loss": 0.4986,
|
| 422 |
+
"step": 6900
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 2.24,
|
| 426 |
+
"learning_rate": 2.2321941871606518e-05,
|
| 427 |
+
"loss": 0.504,
|
| 428 |
+
"step": 7000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 2.27,
|
| 432 |
+
"learning_rate": 2.2641328648993933e-05,
|
| 433 |
+
"loss": 0.5024,
|
| 434 |
+
"step": 7100
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 2.3,
|
| 438 |
+
"learning_rate": 2.296071542638135e-05,
|
| 439 |
+
"loss": 0.479,
|
| 440 |
+
"step": 7200
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 2.33,
|
| 444 |
+
"learning_rate": 2.3280102203768764e-05,
|
| 445 |
+
"loss": 0.492,
|
| 446 |
+
"step": 7300
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 2.36,
|
| 450 |
+
"learning_rate": 2.3599488981156183e-05,
|
| 451 |
+
"loss": 0.4928,
|
| 452 |
+
"step": 7400
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 2.4,
|
| 456 |
+
"learning_rate": 2.39188757585436e-05,
|
| 457 |
+
"loss": 0.4862,
|
| 458 |
+
"step": 7500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 2.43,
|
| 462 |
+
"learning_rate": 2.4238262535931014e-05,
|
| 463 |
+
"loss": 0.5148,
|
| 464 |
+
"step": 7600
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 2.46,
|
| 468 |
+
"learning_rate": 2.4557649313318433e-05,
|
| 469 |
+
"loss": 0.4922,
|
| 470 |
+
"step": 7700
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 2.49,
|
| 474 |
+
"learning_rate": 2.4877036090705846e-05,
|
| 475 |
+
"loss": 0.4882,
|
| 476 |
+
"step": 7800
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 2.52,
|
| 480 |
+
"learning_rate": 2.5196422868093264e-05,
|
| 481 |
+
"loss": 0.4871,
|
| 482 |
+
"step": 7900
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 2.56,
|
| 486 |
+
"learning_rate": 2.5515809645480677e-05,
|
| 487 |
+
"loss": 0.4967,
|
| 488 |
+
"step": 8000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 2.59,
|
| 492 |
+
"learning_rate": 2.5835196422868096e-05,
|
| 493 |
+
"loss": 0.4834,
|
| 494 |
+
"step": 8100
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 2.62,
|
| 498 |
+
"learning_rate": 2.615458320025551e-05,
|
| 499 |
+
"loss": 0.4769,
|
| 500 |
+
"step": 8200
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 2.65,
|
| 504 |
+
"learning_rate": 2.6473969977642927e-05,
|
| 505 |
+
"loss": 0.4776,
|
| 506 |
+
"step": 8300
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 2.68,
|
| 510 |
+
"learning_rate": 2.6793356755030342e-05,
|
| 511 |
+
"loss": 0.4861,
|
| 512 |
+
"step": 8400
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 2.71,
|
| 516 |
+
"learning_rate": 2.711274353241776e-05,
|
| 517 |
+
"loss": 0.486,
|
| 518 |
+
"step": 8500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 2.75,
|
| 522 |
+
"learning_rate": 2.7432130309805177e-05,
|
| 523 |
+
"loss": 0.4745,
|
| 524 |
+
"step": 8600
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 2.78,
|
| 528 |
+
"learning_rate": 2.7751517087192592e-05,
|
| 529 |
+
"loss": 0.4805,
|
| 530 |
+
"step": 8700
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 2.81,
|
| 534 |
+
"learning_rate": 2.807090386458001e-05,
|
| 535 |
+
"loss": 0.5003,
|
| 536 |
+
"step": 8800
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 2.84,
|
| 540 |
+
"learning_rate": 2.8390290641967424e-05,
|
| 541 |
+
"loss": 0.4639,
|
| 542 |
+
"step": 8900
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 2.87,
|
| 546 |
+
"learning_rate": 2.8709677419354843e-05,
|
| 547 |
+
"loss": 0.4763,
|
| 548 |
+
"step": 9000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 2.91,
|
| 552 |
+
"learning_rate": 2.9029064196742255e-05,
|
| 553 |
+
"loss": 0.4727,
|
| 554 |
+
"step": 9100
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 2.94,
|
| 558 |
+
"learning_rate": 2.9348450974129674e-05,
|
| 559 |
+
"loss": 0.4839,
|
| 560 |
+
"step": 9200
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 2.97,
|
| 564 |
+
"learning_rate": 2.966783775151709e-05,
|
| 565 |
+
"loss": 0.4769,
|
| 566 |
+
"step": 9300
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 3.0,
|
| 570 |
+
"learning_rate": 2.9987224528904505e-05,
|
| 571 |
+
"loss": 0.4641,
|
| 572 |
+
"step": 9400
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 3.03,
|
| 576 |
+
"learning_rate": 3.030661130629192e-05,
|
| 577 |
+
"loss": 0.4282,
|
| 578 |
+
"step": 9500
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 3.07,
|
| 582 |
+
"learning_rate": 3.0625998083679336e-05,
|
| 583 |
+
"loss": 0.4057,
|
| 584 |
+
"step": 9600
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 3.1,
|
| 588 |
+
"learning_rate": 3.094538486106676e-05,
|
| 589 |
+
"loss": 0.4192,
|
| 590 |
+
"step": 9700
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 3.13,
|
| 594 |
+
"learning_rate": 3.126477163845417e-05,
|
| 595 |
+
"loss": 0.4186,
|
| 596 |
+
"step": 9800
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 3.16,
|
| 600 |
+
"learning_rate": 3.158415841584159e-05,
|
| 601 |
+
"loss": 0.4086,
|
| 602 |
+
"step": 9900
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 3.19,
|
| 606 |
+
"learning_rate": 3.1903545193229005e-05,
|
| 607 |
+
"loss": 0.4194,
|
| 608 |
+
"step": 10000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 3.23,
|
| 612 |
+
"learning_rate": 3.221973810284254e-05,
|
| 613 |
+
"loss": 0.4443,
|
| 614 |
+
"step": 10100
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 3.26,
|
| 618 |
+
"learning_rate": 3.2539124880229965e-05,
|
| 619 |
+
"loss": 0.4217,
|
| 620 |
+
"step": 10200
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 3.29,
|
| 624 |
+
"learning_rate": 3.2858511657617374e-05,
|
| 625 |
+
"loss": 0.4275,
|
| 626 |
+
"step": 10300
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 3.32,
|
| 630 |
+
"learning_rate": 3.3177898435004796e-05,
|
| 631 |
+
"loss": 0.4422,
|
| 632 |
+
"step": 10400
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 3.35,
|
| 636 |
+
"learning_rate": 3.3497285212392205e-05,
|
| 637 |
+
"loss": 0.4227,
|
| 638 |
+
"step": 10500
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 3.39,
|
| 642 |
+
"learning_rate": 3.381667198977963e-05,
|
| 643 |
+
"loss": 0.4191,
|
| 644 |
+
"step": 10600
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 3.42,
|
| 648 |
+
"learning_rate": 3.413605876716704e-05,
|
| 649 |
+
"loss": 0.4309,
|
| 650 |
+
"step": 10700
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 3.45,
|
| 654 |
+
"learning_rate": 3.445544554455446e-05,
|
| 655 |
+
"loss": 0.4079,
|
| 656 |
+
"step": 10800
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 3.48,
|
| 660 |
+
"learning_rate": 3.4774832321941874e-05,
|
| 661 |
+
"loss": 0.4281,
|
| 662 |
+
"step": 10900
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 3.51,
|
| 666 |
+
"learning_rate": 3.509421909932929e-05,
|
| 667 |
+
"loss": 0.4137,
|
| 668 |
+
"step": 11000
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 3.55,
|
| 672 |
+
"learning_rate": 3.5413605876716705e-05,
|
| 673 |
+
"loss": 0.4168,
|
| 674 |
+
"step": 11100
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 3.58,
|
| 678 |
+
"learning_rate": 3.573299265410412e-05,
|
| 679 |
+
"loss": 0.4186,
|
| 680 |
+
"step": 11200
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 3.61,
|
| 684 |
+
"learning_rate": 3.605237943149154e-05,
|
| 685 |
+
"loss": 0.4325,
|
| 686 |
+
"step": 11300
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 3.64,
|
| 690 |
+
"learning_rate": 3.637176620887895e-05,
|
| 691 |
+
"loss": 0.4422,
|
| 692 |
+
"step": 11400
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 3.67,
|
| 696 |
+
"learning_rate": 3.6691152986266374e-05,
|
| 697 |
+
"loss": 0.4269,
|
| 698 |
+
"step": 11500
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 3.7,
|
| 702 |
+
"learning_rate": 3.701053976365379e-05,
|
| 703 |
+
"loss": 0.4368,
|
| 704 |
+
"step": 11600
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 3.74,
|
| 708 |
+
"learning_rate": 3.7329926541041205e-05,
|
| 709 |
+
"loss": 0.4207,
|
| 710 |
+
"step": 11700
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 3.77,
|
| 714 |
+
"learning_rate": 3.764931331842862e-05,
|
| 715 |
+
"loss": 0.4365,
|
| 716 |
+
"step": 11800
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 3.8,
|
| 720 |
+
"learning_rate": 3.7968700095816036e-05,
|
| 721 |
+
"loss": 0.4394,
|
| 722 |
+
"step": 11900
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 3.83,
|
| 726 |
+
"learning_rate": 3.828808687320345e-05,
|
| 727 |
+
"loss": 0.4351,
|
| 728 |
+
"step": 12000
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 3.86,
|
| 732 |
+
"learning_rate": 3.860747365059087e-05,
|
| 733 |
+
"loss": 0.4113,
|
| 734 |
+
"step": 12100
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 3.9,
|
| 738 |
+
"learning_rate": 3.892686042797828e-05,
|
| 739 |
+
"loss": 0.4261,
|
| 740 |
+
"step": 12200
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 3.93,
|
| 744 |
+
"learning_rate": 3.92462472053657e-05,
|
| 745 |
+
"loss": 0.4311,
|
| 746 |
+
"step": 12300
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 3.96,
|
| 750 |
+
"learning_rate": 3.956563398275312e-05,
|
| 751 |
+
"loss": 0.4246,
|
| 752 |
+
"step": 12400
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 3.99,
|
| 756 |
+
"learning_rate": 3.988502076014053e-05,
|
| 757 |
+
"loss": 0.4294,
|
| 758 |
+
"step": 12500
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 4.02,
|
| 762 |
+
"learning_rate": 3.997728805138579e-05,
|
| 763 |
+
"loss": 0.3854,
|
| 764 |
+
"step": 12600
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 4.06,
|
| 768 |
+
"learning_rate": 3.994180063167608e-05,
|
| 769 |
+
"loss": 0.3656,
|
| 770 |
+
"step": 12700
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 4.09,
|
| 774 |
+
"learning_rate": 3.9906313211966366e-05,
|
| 775 |
+
"loss": 0.3836,
|
| 776 |
+
"step": 12800
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 4.12,
|
| 780 |
+
"learning_rate": 3.987082579225665e-05,
|
| 781 |
+
"loss": 0.3664,
|
| 782 |
+
"step": 12900
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 4.15,
|
| 786 |
+
"learning_rate": 3.9835338372546935e-05,
|
| 787 |
+
"loss": 0.3884,
|
| 788 |
+
"step": 13000
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 4.18,
|
| 792 |
+
"learning_rate": 3.979985095283722e-05,
|
| 793 |
+
"loss": 0.3764,
|
| 794 |
+
"step": 13100
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 4.22,
|
| 798 |
+
"learning_rate": 3.976436353312751e-05,
|
| 799 |
+
"loss": 0.3733,
|
| 800 |
+
"step": 13200
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 4.25,
|
| 804 |
+
"learning_rate": 3.97288761134178e-05,
|
| 805 |
+
"loss": 0.3682,
|
| 806 |
+
"step": 13300
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 4.28,
|
| 810 |
+
"learning_rate": 3.969338869370809e-05,
|
| 811 |
+
"loss": 0.371,
|
| 812 |
+
"step": 13400
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 4.31,
|
| 816 |
+
"learning_rate": 3.965790127399837e-05,
|
| 817 |
+
"loss": 0.3741,
|
| 818 |
+
"step": 13500
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 4.34,
|
| 822 |
+
"learning_rate": 3.9622413854288656e-05,
|
| 823 |
+
"loss": 0.3757,
|
| 824 |
+
"step": 13600
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 4.38,
|
| 828 |
+
"learning_rate": 3.9586926434578944e-05,
|
| 829 |
+
"loss": 0.3652,
|
| 830 |
+
"step": 13700
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 4.41,
|
| 834 |
+
"learning_rate": 3.955143901486923e-05,
|
| 835 |
+
"loss": 0.3892,
|
| 836 |
+
"step": 13800
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 4.44,
|
| 840 |
+
"learning_rate": 3.951630646935662e-05,
|
| 841 |
+
"loss": 0.3893,
|
| 842 |
+
"step": 13900
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 4.47,
|
| 846 |
+
"learning_rate": 3.948081904964691e-05,
|
| 847 |
+
"loss": 0.3823,
|
| 848 |
+
"step": 14000
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 4.5,
|
| 852 |
+
"learning_rate": 3.944533162993719e-05,
|
| 853 |
+
"loss": 0.3654,
|
| 854 |
+
"step": 14100
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 4.54,
|
| 858 |
+
"learning_rate": 3.9409844210227476e-05,
|
| 859 |
+
"loss": 0.3791,
|
| 860 |
+
"step": 14200
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 4.57,
|
| 864 |
+
"learning_rate": 3.9374356790517764e-05,
|
| 865 |
+
"loss": 0.3918,
|
| 866 |
+
"step": 14300
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 4.6,
|
| 870 |
+
"learning_rate": 3.933886937080805e-05,
|
| 871 |
+
"loss": 0.3713,
|
| 872 |
+
"step": 14400
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 4.63,
|
| 876 |
+
"learning_rate": 3.930338195109834e-05,
|
| 877 |
+
"loss": 0.376,
|
| 878 |
+
"step": 14500
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 4.66,
|
| 882 |
+
"learning_rate": 3.926789453138863e-05,
|
| 883 |
+
"loss": 0.3682,
|
| 884 |
+
"step": 14600
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 4.69,
|
| 888 |
+
"learning_rate": 3.9232407111678915e-05,
|
| 889 |
+
"loss": 0.374,
|
| 890 |
+
"step": 14700
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 4.73,
|
| 894 |
+
"learning_rate": 3.91969196919692e-05,
|
| 895 |
+
"loss": 0.3774,
|
| 896 |
+
"step": 14800
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 4.76,
|
| 900 |
+
"learning_rate": 3.9161432272259484e-05,
|
| 901 |
+
"loss": 0.3806,
|
| 902 |
+
"step": 14900
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 4.79,
|
| 906 |
+
"learning_rate": 3.912594485254977e-05,
|
| 907 |
+
"loss": 0.3886,
|
| 908 |
+
"step": 15000
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 4.82,
|
| 912 |
+
"learning_rate": 3.909045743284006e-05,
|
| 913 |
+
"loss": 0.3742,
|
| 914 |
+
"step": 15100
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 4.85,
|
| 918 |
+
"learning_rate": 3.905497001313035e-05,
|
| 919 |
+
"loss": 0.3714,
|
| 920 |
+
"step": 15200
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 4.89,
|
| 924 |
+
"learning_rate": 3.9019482593420636e-05,
|
| 925 |
+
"loss": 0.3721,
|
| 926 |
+
"step": 15300
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 4.92,
|
| 930 |
+
"learning_rate": 3.8983995173710924e-05,
|
| 931 |
+
"loss": 0.3716,
|
| 932 |
+
"step": 15400
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 4.95,
|
| 936 |
+
"learning_rate": 3.8948507754001205e-05,
|
| 937 |
+
"loss": 0.3728,
|
| 938 |
+
"step": 15500
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 4.98,
|
| 942 |
+
"learning_rate": 3.891302033429149e-05,
|
| 943 |
+
"loss": 0.384,
|
| 944 |
+
"step": 15600
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 5.01,
|
| 948 |
+
"learning_rate": 3.887753291458178e-05,
|
| 949 |
+
"loss": 0.3415,
|
| 950 |
+
"step": 15700
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 5.05,
|
| 954 |
+
"learning_rate": 3.884204549487207e-05,
|
| 955 |
+
"loss": 0.3064,
|
| 956 |
+
"step": 15800
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 5.08,
|
| 960 |
+
"learning_rate": 3.880655807516236e-05,
|
| 961 |
+
"loss": 0.3164,
|
| 962 |
+
"step": 15900
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 5.11,
|
| 966 |
+
"learning_rate": 3.8771070655452645e-05,
|
| 967 |
+
"loss": 0.3215,
|
| 968 |
+
"step": 16000
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 5.14,
|
| 972 |
+
"learning_rate": 3.873558323574293e-05,
|
| 973 |
+
"loss": 0.3234,
|
| 974 |
+
"step": 16100
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 5.17,
|
| 978 |
+
"learning_rate": 3.870009581603322e-05,
|
| 979 |
+
"loss": 0.3119,
|
| 980 |
+
"step": 16200
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 5.21,
|
| 984 |
+
"learning_rate": 3.866460839632351e-05,
|
| 985 |
+
"loss": 0.3145,
|
| 986 |
+
"step": 16300
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 5.24,
|
| 990 |
+
"learning_rate": 3.8629120976613796e-05,
|
| 991 |
+
"loss": 0.3333,
|
| 992 |
+
"step": 16400
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 5.27,
|
| 996 |
+
"learning_rate": 3.8593633556904084e-05,
|
| 997 |
+
"loss": 0.3195,
|
| 998 |
+
"step": 16500
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 5.3,
|
| 1002 |
+
"learning_rate": 3.855814613719437e-05,
|
| 1003 |
+
"loss": 0.3159,
|
| 1004 |
+
"step": 16600
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 5.33,
|
| 1008 |
+
"learning_rate": 3.852265871748465e-05,
|
| 1009 |
+
"loss": 0.3189,
|
| 1010 |
+
"step": 16700
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 5.37,
|
| 1014 |
+
"learning_rate": 3.848717129777494e-05,
|
| 1015 |
+
"loss": 0.317,
|
| 1016 |
+
"step": 16800
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 5.4,
|
| 1020 |
+
"learning_rate": 3.845168387806523e-05,
|
| 1021 |
+
"loss": 0.3231,
|
| 1022 |
+
"step": 16900
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 5.43,
|
| 1026 |
+
"learning_rate": 3.841619645835552e-05,
|
| 1027 |
+
"loss": 0.3328,
|
| 1028 |
+
"step": 17000
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 5.46,
|
| 1032 |
+
"learning_rate": 3.8380709038645805e-05,
|
| 1033 |
+
"loss": 0.3282,
|
| 1034 |
+
"step": 17100
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 5.49,
|
| 1038 |
+
"learning_rate": 3.8345221618936086e-05,
|
| 1039 |
+
"loss": 0.3169,
|
| 1040 |
+
"step": 17200
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 5.53,
|
| 1044 |
+
"learning_rate": 3.8309734199226374e-05,
|
| 1045 |
+
"loss": 0.3156,
|
| 1046 |
+
"step": 17300
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 5.56,
|
| 1050 |
+
"learning_rate": 3.827424677951666e-05,
|
| 1051 |
+
"loss": 0.3313,
|
| 1052 |
+
"step": 17400
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 5.59,
|
| 1056 |
+
"learning_rate": 3.823875935980695e-05,
|
| 1057 |
+
"loss": 0.3181,
|
| 1058 |
+
"step": 17500
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 5.62,
|
| 1062 |
+
"learning_rate": 3.820327194009724e-05,
|
| 1063 |
+
"loss": 0.334,
|
| 1064 |
+
"step": 17600
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 5.65,
|
| 1068 |
+
"learning_rate": 3.8167784520387526e-05,
|
| 1069 |
+
"loss": 0.3313,
|
| 1070 |
+
"step": 17700
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 5.69,
|
| 1074 |
+
"learning_rate": 3.8132297100677814e-05,
|
| 1075 |
+
"loss": 0.3331,
|
| 1076 |
+
"step": 17800
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 5.72,
|
| 1080 |
+
"learning_rate": 3.80968096809681e-05,
|
| 1081 |
+
"loss": 0.3405,
|
| 1082 |
+
"step": 17900
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 5.75,
|
| 1086 |
+
"learning_rate": 3.806132226125839e-05,
|
| 1087 |
+
"loss": 0.3345,
|
| 1088 |
+
"step": 18000
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 5.78,
|
| 1092 |
+
"learning_rate": 3.802583484154868e-05,
|
| 1093 |
+
"loss": 0.3295,
|
| 1094 |
+
"step": 18100
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 5.81,
|
| 1098 |
+
"learning_rate": 3.7990347421838965e-05,
|
| 1099 |
+
"loss": 0.3244,
|
| 1100 |
+
"step": 18200
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 5.84,
|
| 1104 |
+
"learning_rate": 3.795486000212925e-05,
|
| 1105 |
+
"loss": 0.3346,
|
| 1106 |
+
"step": 18300
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 5.88,
|
| 1110 |
+
"learning_rate": 3.791972745661663e-05,
|
| 1111 |
+
"loss": 0.3317,
|
| 1112 |
+
"step": 18400
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 5.91,
|
| 1116 |
+
"learning_rate": 3.788424003690692e-05,
|
| 1117 |
+
"loss": 0.3219,
|
| 1118 |
+
"step": 18500
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 5.94,
|
| 1122 |
+
"learning_rate": 3.784875261719721e-05,
|
| 1123 |
+
"loss": 0.3167,
|
| 1124 |
+
"step": 18600
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 5.97,
|
| 1128 |
+
"learning_rate": 3.781326519748749e-05,
|
| 1129 |
+
"loss": 0.33,
|
| 1130 |
+
"step": 18700
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 6.0,
|
| 1134 |
+
"learning_rate": 3.777777777777778e-05,
|
| 1135 |
+
"loss": 0.3211,
|
| 1136 |
+
"step": 18800
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 6.04,
|
| 1140 |
+
"learning_rate": 3.7742290358068066e-05,
|
| 1141 |
+
"loss": 0.2794,
|
| 1142 |
+
"step": 18900
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 6.07,
|
| 1146 |
+
"learning_rate": 3.7706802938358354e-05,
|
| 1147 |
+
"loss": 0.2675,
|
| 1148 |
+
"step": 19000
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 6.1,
|
| 1152 |
+
"learning_rate": 3.767131551864864e-05,
|
| 1153 |
+
"loss": 0.2676,
|
| 1154 |
+
"step": 19100
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 6.13,
|
| 1158 |
+
"learning_rate": 3.763582809893893e-05,
|
| 1159 |
+
"loss": 0.2693,
|
| 1160 |
+
"step": 19200
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 6.16,
|
| 1164 |
+
"learning_rate": 3.760034067922922e-05,
|
| 1165 |
+
"loss": 0.2669,
|
| 1166 |
+
"step": 19300
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 6.2,
|
| 1170 |
+
"learning_rate": 3.7564853259519506e-05,
|
| 1171 |
+
"loss": 0.2744,
|
| 1172 |
+
"step": 19400
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 6.23,
|
| 1176 |
+
"learning_rate": 3.7529365839809794e-05,
|
| 1177 |
+
"loss": 0.282,
|
| 1178 |
+
"step": 19500
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 6.26,
|
| 1182 |
+
"learning_rate": 3.749387842010008e-05,
|
| 1183 |
+
"loss": 0.2848,
|
| 1184 |
+
"step": 19600
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 6.29,
|
| 1188 |
+
"learning_rate": 3.745839100039037e-05,
|
| 1189 |
+
"loss": 0.2783,
|
| 1190 |
+
"step": 19700
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 6.32,
|
| 1194 |
+
"learning_rate": 3.742290358068066e-05,
|
| 1195 |
+
"loss": 0.2843,
|
| 1196 |
+
"step": 19800
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 6.36,
|
| 1200 |
+
"learning_rate": 3.738741616097094e-05,
|
| 1201 |
+
"loss": 0.2696,
|
| 1202 |
+
"step": 19900
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 6.39,
|
| 1206 |
+
"learning_rate": 3.7351928741261226e-05,
|
| 1207 |
+
"loss": 0.2932,
|
| 1208 |
+
"step": 20000
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 6.42,
|
| 1212 |
+
"learning_rate": 3.7316441321551514e-05,
|
| 1213 |
+
"loss": 0.2775,
|
| 1214 |
+
"step": 20100
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 6.45,
|
| 1218 |
+
"learning_rate": 3.72809539018418e-05,
|
| 1219 |
+
"loss": 0.2893,
|
| 1220 |
+
"step": 20200
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 6.48,
|
| 1224 |
+
"learning_rate": 3.724546648213209e-05,
|
| 1225 |
+
"loss": 0.286,
|
| 1226 |
+
"step": 20300
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 6.52,
|
| 1230 |
+
"learning_rate": 3.720997906242237e-05,
|
| 1231 |
+
"loss": 0.2799,
|
| 1232 |
+
"step": 20400
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 6.55,
|
| 1236 |
+
"learning_rate": 3.717449164271266e-05,
|
| 1237 |
+
"loss": 0.2827,
|
| 1238 |
+
"step": 20500
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 6.58,
|
| 1242 |
+
"learning_rate": 3.713900422300295e-05,
|
| 1243 |
+
"loss": 0.2783,
|
| 1244 |
+
"step": 20600
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 6.61,
|
| 1248 |
+
"learning_rate": 3.7103516803293235e-05,
|
| 1249 |
+
"loss": 0.2729,
|
| 1250 |
+
"step": 20700
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 6.64,
|
| 1254 |
+
"learning_rate": 3.706802938358352e-05,
|
| 1255 |
+
"loss": 0.2747,
|
| 1256 |
+
"step": 20800
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 6.68,
|
| 1260 |
+
"learning_rate": 3.703254196387381e-05,
|
| 1261 |
+
"loss": 0.2764,
|
| 1262 |
+
"step": 20900
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 6.71,
|
| 1266 |
+
"learning_rate": 3.699705454416409e-05,
|
| 1267 |
+
"loss": 0.2804,
|
| 1268 |
+
"step": 21000
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 6.74,
|
| 1272 |
+
"learning_rate": 3.696156712445438e-05,
|
| 1273 |
+
"loss": 0.2841,
|
| 1274 |
+
"step": 21100
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 6.77,
|
| 1278 |
+
"learning_rate": 3.692607970474467e-05,
|
| 1279 |
+
"loss": 0.277,
|
| 1280 |
+
"step": 21200
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 6.8,
|
| 1284 |
+
"learning_rate": 3.6890592285034956e-05,
|
| 1285 |
+
"loss": 0.284,
|
| 1286 |
+
"step": 21300
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 6.83,
|
| 1290 |
+
"learning_rate": 3.6855104865325244e-05,
|
| 1291 |
+
"loss": 0.2778,
|
| 1292 |
+
"step": 21400
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 6.87,
|
| 1296 |
+
"learning_rate": 3.681961744561553e-05,
|
| 1297 |
+
"loss": 0.2797,
|
| 1298 |
+
"step": 21500
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 6.9,
|
| 1302 |
+
"learning_rate": 3.678413002590582e-05,
|
| 1303 |
+
"loss": 0.3036,
|
| 1304 |
+
"step": 21600
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 6.93,
|
| 1308 |
+
"learning_rate": 3.674864260619611e-05,
|
| 1309 |
+
"loss": 0.2869,
|
| 1310 |
+
"step": 21700
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 6.96,
|
| 1314 |
+
"learning_rate": 3.6713155186486395e-05,
|
| 1315 |
+
"loss": 0.2921,
|
| 1316 |
+
"step": 21800
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 6.99,
|
| 1320 |
+
"learning_rate": 3.667766776677668e-05,
|
| 1321 |
+
"loss": 0.2844,
|
| 1322 |
+
"step": 21900
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 7.03,
|
| 1326 |
+
"learning_rate": 3.664218034706697e-05,
|
| 1327 |
+
"loss": 0.2427,
|
| 1328 |
+
"step": 22000
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 7.06,
|
| 1332 |
+
"learning_rate": 3.660669292735726e-05,
|
| 1333 |
+
"loss": 0.243,
|
| 1334 |
+
"step": 22100
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 7.09,
|
| 1338 |
+
"learning_rate": 3.657120550764754e-05,
|
| 1339 |
+
"loss": 0.2311,
|
| 1340 |
+
"step": 22200
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 7.12,
|
| 1344 |
+
"learning_rate": 3.653571808793783e-05,
|
| 1345 |
+
"loss": 0.2358,
|
| 1346 |
+
"step": 22300
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 7.15,
|
| 1350 |
+
"learning_rate": 3.6500230668228116e-05,
|
| 1351 |
+
"loss": 0.2287,
|
| 1352 |
+
"step": 22400
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 7.19,
|
| 1356 |
+
"learning_rate": 3.6464743248518404e-05,
|
| 1357 |
+
"loss": 0.2519,
|
| 1358 |
+
"step": 22500
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 7.22,
|
| 1362 |
+
"learning_rate": 3.642925582880869e-05,
|
| 1363 |
+
"loss": 0.2528,
|
| 1364 |
+
"step": 22600
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 7.25,
|
| 1368 |
+
"learning_rate": 3.639376840909898e-05,
|
| 1369 |
+
"loss": 0.2507,
|
| 1370 |
+
"step": 22700
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 7.28,
|
| 1374 |
+
"learning_rate": 3.635828098938926e-05,
|
| 1375 |
+
"loss": 0.2496,
|
| 1376 |
+
"step": 22800
|
| 1377 |
+
},
|
| 1378 |
+
{
|
| 1379 |
+
"epoch": 7.31,
|
| 1380 |
+
"learning_rate": 3.632279356967955e-05,
|
| 1381 |
+
"loss": 0.2503,
|
| 1382 |
+
"step": 22900
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 7.35,
|
| 1386 |
+
"learning_rate": 3.628730614996984e-05,
|
| 1387 |
+
"loss": 0.2383,
|
| 1388 |
+
"step": 23000
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 7.38,
|
| 1392 |
+
"learning_rate": 3.6251818730260125e-05,
|
| 1393 |
+
"loss": 0.2513,
|
| 1394 |
+
"step": 23100
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 7.41,
|
| 1398 |
+
"learning_rate": 3.621633131055041e-05,
|
| 1399 |
+
"loss": 0.2492,
|
| 1400 |
+
"step": 23200
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 7.44,
|
| 1404 |
+
"learning_rate": 3.61808438908407e-05,
|
| 1405 |
+
"loss": 0.2473,
|
| 1406 |
+
"step": 23300
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 7.47,
|
| 1410 |
+
"learning_rate": 3.614535647113099e-05,
|
| 1411 |
+
"loss": 0.2502,
|
| 1412 |
+
"step": 23400
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 7.51,
|
| 1416 |
+
"learning_rate": 3.6109869051421276e-05,
|
| 1417 |
+
"loss": 0.2492,
|
| 1418 |
+
"step": 23500
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 7.54,
|
| 1422 |
+
"learning_rate": 3.6074381631711564e-05,
|
| 1423 |
+
"loss": 0.2366,
|
| 1424 |
+
"step": 23600
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 7.57,
|
| 1428 |
+
"learning_rate": 3.603889421200185e-05,
|
| 1429 |
+
"loss": 0.2393,
|
| 1430 |
+
"step": 23700
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 7.6,
|
| 1434 |
+
"learning_rate": 3.600340679229214e-05,
|
| 1435 |
+
"loss": 0.2483,
|
| 1436 |
+
"step": 23800
|
| 1437 |
+
},
|
| 1438 |
+
{
|
| 1439 |
+
"epoch": 7.63,
|
| 1440 |
+
"learning_rate": 3.596791937258242e-05,
|
| 1441 |
+
"loss": 0.2428,
|
| 1442 |
+
"step": 23900
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 7.67,
|
| 1446 |
+
"learning_rate": 3.593243195287271e-05,
|
| 1447 |
+
"loss": 0.2507,
|
| 1448 |
+
"step": 24000
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 7.7,
|
| 1452 |
+
"learning_rate": 3.5896944533163e-05,
|
| 1453 |
+
"loss": 0.2491,
|
| 1454 |
+
"step": 24100
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 7.73,
|
| 1458 |
+
"learning_rate": 3.5861457113453285e-05,
|
| 1459 |
+
"loss": 0.2549,
|
| 1460 |
+
"step": 24200
|
| 1461 |
+
},
|
| 1462 |
+
{
|
| 1463 |
+
"epoch": 7.76,
|
| 1464 |
+
"learning_rate": 3.582596969374357e-05,
|
| 1465 |
+
"loss": 0.251,
|
| 1466 |
+
"step": 24300
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 7.79,
|
| 1470 |
+
"learning_rate": 3.579048227403386e-05,
|
| 1471 |
+
"loss": 0.2357,
|
| 1472 |
+
"step": 24400
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 7.82,
|
| 1476 |
+
"learning_rate": 3.575499485432414e-05,
|
| 1477 |
+
"loss": 0.2441,
|
| 1478 |
+
"step": 24500
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 7.86,
|
| 1482 |
+
"learning_rate": 3.571950743461443e-05,
|
| 1483 |
+
"loss": 0.2561,
|
| 1484 |
+
"step": 24600
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 7.89,
|
| 1488 |
+
"learning_rate": 3.568402001490472e-05,
|
| 1489 |
+
"loss": 0.2524,
|
| 1490 |
+
"step": 24700
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 7.92,
|
| 1494 |
+
"learning_rate": 3.5648532595195006e-05,
|
| 1495 |
+
"loss": 0.2467,
|
| 1496 |
+
"step": 24800
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 7.95,
|
| 1500 |
+
"learning_rate": 3.5613045175485293e-05,
|
| 1501 |
+
"loss": 0.2416,
|
| 1502 |
+
"step": 24900
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 7.98,
|
| 1506 |
+
"learning_rate": 3.557755775577558e-05,
|
| 1507 |
+
"loss": 0.2511,
|
| 1508 |
+
"step": 25000
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 8.02,
|
| 1512 |
+
"learning_rate": 3.554207033606587e-05,
|
| 1513 |
+
"loss": 0.2323,
|
| 1514 |
+
"step": 25100
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 8.05,
|
| 1518 |
+
"learning_rate": 3.550658291635615e-05,
|
| 1519 |
+
"loss": 0.2158,
|
| 1520 |
+
"step": 25200
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 8.08,
|
| 1524 |
+
"learning_rate": 3.547109549664644e-05,
|
| 1525 |
+
"loss": 0.21,
|
| 1526 |
+
"step": 25300
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 8.11,
|
| 1530 |
+
"learning_rate": 3.5435608076936726e-05,
|
| 1531 |
+
"loss": 0.2102,
|
| 1532 |
+
"step": 25400
|
| 1533 |
+
},
|
| 1534 |
+
{
|
| 1535 |
+
"epoch": 8.14,
|
| 1536 |
+
"learning_rate": 3.5400120657227014e-05,
|
| 1537 |
+
"loss": 0.2071,
|
| 1538 |
+
"step": 25500
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 8.18,
|
| 1542 |
+
"learning_rate": 3.53646332375173e-05,
|
| 1543 |
+
"loss": 0.2198,
|
| 1544 |
+
"step": 25600
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 8.21,
|
| 1548 |
+
"learning_rate": 3.532914581780759e-05,
|
| 1549 |
+
"loss": 0.2193,
|
| 1550 |
+
"step": 25700
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 8.24,
|
| 1554 |
+
"learning_rate": 3.529365839809788e-05,
|
| 1555 |
+
"loss": 0.2166,
|
| 1556 |
+
"step": 25800
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 8.27,
|
| 1560 |
+
"learning_rate": 3.5258170978388166e-05,
|
| 1561 |
+
"loss": 0.2237,
|
| 1562 |
+
"step": 25900
|
| 1563 |
+
},
|
| 1564 |
+
{
|
| 1565 |
+
"epoch": 8.3,
|
| 1566 |
+
"learning_rate": 3.5222683558678454e-05,
|
| 1567 |
+
"loss": 0.2241,
|
| 1568 |
+
"step": 26000
|
| 1569 |
+
},
|
| 1570 |
+
{
|
| 1571 |
+
"epoch": 8.34,
|
| 1572 |
+
"learning_rate": 3.518719613896874e-05,
|
| 1573 |
+
"loss": 0.2132,
|
| 1574 |
+
"step": 26100
|
| 1575 |
+
},
|
| 1576 |
+
{
|
| 1577 |
+
"epoch": 8.37,
|
| 1578 |
+
"learning_rate": 3.515170871925902e-05,
|
| 1579 |
+
"loss": 0.2101,
|
| 1580 |
+
"step": 26200
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 8.4,
|
| 1584 |
+
"learning_rate": 3.511622129954931e-05,
|
| 1585 |
+
"loss": 0.227,
|
| 1586 |
+
"step": 26300
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 8.43,
|
| 1590 |
+
"learning_rate": 3.50807338798396e-05,
|
| 1591 |
+
"loss": 0.2221,
|
| 1592 |
+
"step": 26400
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 8.46,
|
| 1596 |
+
"learning_rate": 3.504560133432698e-05,
|
| 1597 |
+
"loss": 0.2181,
|
| 1598 |
+
"step": 26500
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 8.5,
|
| 1602 |
+
"learning_rate": 3.501011391461727e-05,
|
| 1603 |
+
"loss": 0.2208,
|
| 1604 |
+
"step": 26600
|
| 1605 |
+
},
|
| 1606 |
+
{
|
| 1607 |
+
"epoch": 8.53,
|
| 1608 |
+
"learning_rate": 3.4974626494907555e-05,
|
| 1609 |
+
"loss": 0.2192,
|
| 1610 |
+
"step": 26700
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 8.56,
|
| 1614 |
+
"learning_rate": 3.493913907519784e-05,
|
| 1615 |
+
"loss": 0.2227,
|
| 1616 |
+
"step": 26800
|
| 1617 |
+
},
|
| 1618 |
+
{
|
| 1619 |
+
"epoch": 8.59,
|
| 1620 |
+
"learning_rate": 3.490365165548813e-05,
|
| 1621 |
+
"loss": 0.2308,
|
| 1622 |
+
"step": 26900
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 8.62,
|
| 1626 |
+
"learning_rate": 3.486816423577842e-05,
|
| 1627 |
+
"loss": 0.22,
|
| 1628 |
+
"step": 27000
|
| 1629 |
+
},
|
| 1630 |
+
{
|
| 1631 |
+
"epoch": 8.66,
|
| 1632 |
+
"learning_rate": 3.4832676816068706e-05,
|
| 1633 |
+
"loss": 0.2247,
|
| 1634 |
+
"step": 27100
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 8.69,
|
| 1638 |
+
"learning_rate": 3.4797189396358994e-05,
|
| 1639 |
+
"loss": 0.2148,
|
| 1640 |
+
"step": 27200
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 8.72,
|
| 1644 |
+
"learning_rate": 3.476170197664928e-05,
|
| 1645 |
+
"loss": 0.2136,
|
| 1646 |
+
"step": 27300
|
| 1647 |
+
},
|
| 1648 |
+
{
|
| 1649 |
+
"epoch": 8.75,
|
| 1650 |
+
"learning_rate": 3.472621455693957e-05,
|
| 1651 |
+
"loss": 0.2282,
|
| 1652 |
+
"step": 27400
|
| 1653 |
+
},
|
| 1654 |
+
{
|
| 1655 |
+
"epoch": 8.78,
|
| 1656 |
+
"learning_rate": 3.469072713722986e-05,
|
| 1657 |
+
"loss": 0.2266,
|
| 1658 |
+
"step": 27500
|
| 1659 |
+
},
|
| 1660 |
+
{
|
| 1661 |
+
"epoch": 8.82,
|
| 1662 |
+
"learning_rate": 3.4655239717520146e-05,
|
| 1663 |
+
"loss": 0.2198,
|
| 1664 |
+
"step": 27600
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 8.85,
|
| 1668 |
+
"learning_rate": 3.461975229781043e-05,
|
| 1669 |
+
"loss": 0.2243,
|
| 1670 |
+
"step": 27700
|
| 1671 |
+
},
|
| 1672 |
+
{
|
| 1673 |
+
"epoch": 8.88,
|
| 1674 |
+
"learning_rate": 3.4584264878100715e-05,
|
| 1675 |
+
"loss": 0.2358,
|
| 1676 |
+
"step": 27800
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 8.91,
|
| 1680 |
+
"learning_rate": 3.4548777458391e-05,
|
| 1681 |
+
"loss": 0.2255,
|
| 1682 |
+
"step": 27900
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 8.94,
|
| 1686 |
+
"learning_rate": 3.451329003868129e-05,
|
| 1687 |
+
"loss": 0.2258,
|
| 1688 |
+
"step": 28000
|
| 1689 |
+
},
|
| 1690 |
+
{
|
| 1691 |
+
"epoch": 8.97,
|
| 1692 |
+
"learning_rate": 3.447780261897158e-05,
|
| 1693 |
+
"loss": 0.2228,
|
| 1694 |
+
"step": 28100
|
| 1695 |
+
},
|
| 1696 |
+
{
|
| 1697 |
+
"epoch": 9.01,
|
| 1698 |
+
"learning_rate": 3.4442315199261866e-05,
|
| 1699 |
+
"loss": 0.2183,
|
| 1700 |
+
"step": 28200
|
| 1701 |
+
},
|
| 1702 |
+
{
|
| 1703 |
+
"epoch": 9.04,
|
| 1704 |
+
"learning_rate": 3.440682777955215e-05,
|
| 1705 |
+
"loss": 0.1988,
|
| 1706 |
+
"step": 28300
|
| 1707 |
+
},
|
| 1708 |
+
{
|
| 1709 |
+
"epoch": 9.07,
|
| 1710 |
+
"learning_rate": 3.4371340359842436e-05,
|
| 1711 |
+
"loss": 0.188,
|
| 1712 |
+
"step": 28400
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"epoch": 9.1,
|
| 1716 |
+
"learning_rate": 3.4335852940132723e-05,
|
| 1717 |
+
"loss": 0.1963,
|
| 1718 |
+
"step": 28500
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 9.13,
|
| 1722 |
+
"learning_rate": 3.430036552042301e-05,
|
| 1723 |
+
"loss": 0.1919,
|
| 1724 |
+
"step": 28600
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 9.17,
|
| 1728 |
+
"learning_rate": 3.42648781007133e-05,
|
| 1729 |
+
"loss": 0.1906,
|
| 1730 |
+
"step": 28700
|
| 1731 |
+
},
|
| 1732 |
+
{
|
| 1733 |
+
"epoch": 9.2,
|
| 1734 |
+
"learning_rate": 3.422939068100359e-05,
|
| 1735 |
+
"loss": 0.1911,
|
| 1736 |
+
"step": 28800
|
| 1737 |
+
},
|
| 1738 |
+
{
|
| 1739 |
+
"epoch": 9.23,
|
| 1740 |
+
"learning_rate": 3.4193903261293875e-05,
|
| 1741 |
+
"loss": 0.1971,
|
| 1742 |
+
"step": 28900
|
| 1743 |
+
},
|
| 1744 |
+
{
|
| 1745 |
+
"epoch": 9.26,
|
| 1746 |
+
"learning_rate": 3.415841584158416e-05,
|
| 1747 |
+
"loss": 0.1931,
|
| 1748 |
+
"step": 29000
|
| 1749 |
+
},
|
| 1750 |
+
{
|
| 1751 |
+
"epoch": 9.29,
|
| 1752 |
+
"learning_rate": 3.412292842187445e-05,
|
| 1753 |
+
"loss": 0.1951,
|
| 1754 |
+
"step": 29100
|
| 1755 |
+
},
|
| 1756 |
+
{
|
| 1757 |
+
"epoch": 9.33,
|
| 1758 |
+
"learning_rate": 3.408744100216474e-05,
|
| 1759 |
+
"loss": 0.2008,
|
| 1760 |
+
"step": 29200
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 9.36,
|
| 1764 |
+
"learning_rate": 3.405195358245503e-05,
|
| 1765 |
+
"loss": 0.2053,
|
| 1766 |
+
"step": 29300
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 9.39,
|
| 1770 |
+
"learning_rate": 3.401646616274531e-05,
|
| 1771 |
+
"loss": 0.2015,
|
| 1772 |
+
"step": 29400
|
| 1773 |
+
},
|
| 1774 |
+
{
|
| 1775 |
+
"epoch": 9.42,
|
| 1776 |
+
"learning_rate": 3.3980978743035596e-05,
|
| 1777 |
+
"loss": 0.193,
|
| 1778 |
+
"step": 29500
|
| 1779 |
+
},
|
| 1780 |
+
{
|
| 1781 |
+
"epoch": 9.45,
|
| 1782 |
+
"learning_rate": 3.3945491323325884e-05,
|
| 1783 |
+
"loss": 0.1955,
|
| 1784 |
+
"step": 29600
|
| 1785 |
+
},
|
| 1786 |
+
{
|
| 1787 |
+
"epoch": 9.49,
|
| 1788 |
+
"learning_rate": 3.391000390361617e-05,
|
| 1789 |
+
"loss": 0.1993,
|
| 1790 |
+
"step": 29700
|
| 1791 |
+
},
|
| 1792 |
+
{
|
| 1793 |
+
"epoch": 9.52,
|
| 1794 |
+
"learning_rate": 3.387451648390646e-05,
|
| 1795 |
+
"loss": 0.2043,
|
| 1796 |
+
"step": 29800
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"epoch": 9.55,
|
| 1800 |
+
"learning_rate": 3.383902906419675e-05,
|
| 1801 |
+
"loss": 0.1955,
|
| 1802 |
+
"step": 29900
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 9.58,
|
| 1806 |
+
"learning_rate": 3.380354164448703e-05,
|
| 1807 |
+
"loss": 0.1961,
|
| 1808 |
+
"step": 30000
|
| 1809 |
+
}
|
| 1810 |
+
],
|
| 1811 |
+
"max_steps": 125240,
|
| 1812 |
+
"num_train_epochs": 40,
|
| 1813 |
+
"total_flos": 0.0,
|
| 1814 |
+
"trial_name": null,
|
| 1815 |
+
"trial_params": null
|
| 1816 |
+
}
|
checkpoint-35000/config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "Davlan/afro-xlmr-base",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"XLMRobertaModel"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 768,
|
| 13 |
+
"id2label": {
|
| 14 |
+
"0": "LABEL_0"
|
| 15 |
+
},
|
| 16 |
+
"initializer_range": 0.02,
|
| 17 |
+
"intermediate_size": 3072,
|
| 18 |
+
"label2id": {
|
| 19 |
+
"LABEL_0": 0
|
| 20 |
+
},
|
| 21 |
+
"layer_norm_eps": 1e-05,
|
| 22 |
+
"max_position_embeddings": 514,
|
| 23 |
+
"model_type": "xlm-roberta",
|
| 24 |
+
"num_attention_heads": 12,
|
| 25 |
+
"num_hidden_layers": 12,
|
| 26 |
+
"output_past": true,
|
| 27 |
+
"pad_token_id": 1,
|
| 28 |
+
"position_embedding_type": "absolute",
|
| 29 |
+
"torch_dtype": "float32",
|
| 30 |
+
"transformers_version": "4.29.0",
|
| 31 |
+
"type_vocab_size": 1,
|
| 32 |
+
"use_cache": true,
|
| 33 |
+
"vocab_size": 250002
|
| 34 |
+
}
|
checkpoint-35000/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:766ac23dc09c1a6c0ec6df682ae9d4833a29ae4638f6c74046b452063013d170
|
| 3 |
+
size 1112246250
|
checkpoint-35000/trainer_state.json
ADDED
|
@@ -0,0 +1,2116 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 11.178537208559566,
|
| 5 |
+
"global_step": 35000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.03,
|
| 12 |
+
"learning_rate": 2.8744809964867454e-07,
|
| 13 |
+
"loss": 91.176,
|
| 14 |
+
"step": 100
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.06,
|
| 18 |
+
"learning_rate": 6.068348770360907e-07,
|
| 19 |
+
"loss": 64.0684,
|
| 20 |
+
"step": 200
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.1,
|
| 24 |
+
"learning_rate": 9.26221654423507e-07,
|
| 25 |
+
"loss": 27.1061,
|
| 26 |
+
"step": 300
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.13,
|
| 30 |
+
"learning_rate": 1.242414564037049e-06,
|
| 31 |
+
"loss": 8.0198,
|
| 32 |
+
"step": 400
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.16,
|
| 36 |
+
"learning_rate": 1.561801341424465e-06,
|
| 37 |
+
"loss": 4.7691,
|
| 38 |
+
"step": 500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.19,
|
| 42 |
+
"learning_rate": 1.8811881188118813e-06,
|
| 43 |
+
"loss": 3.4387,
|
| 44 |
+
"step": 600
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.22,
|
| 48 |
+
"learning_rate": 2.2005748961992977e-06,
|
| 49 |
+
"loss": 2.7027,
|
| 50 |
+
"step": 700
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.26,
|
| 54 |
+
"learning_rate": 2.5199616735867137e-06,
|
| 55 |
+
"loss": 2.4431,
|
| 56 |
+
"step": 800
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.29,
|
| 60 |
+
"learning_rate": 2.8393484509741297e-06,
|
| 61 |
+
"loss": 2.0534,
|
| 62 |
+
"step": 900
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.32,
|
| 66 |
+
"learning_rate": 3.1587352283615465e-06,
|
| 67 |
+
"loss": 1.8537,
|
| 68 |
+
"step": 1000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.35,
|
| 72 |
+
"learning_rate": 3.4781220057489625e-06,
|
| 73 |
+
"loss": 1.642,
|
| 74 |
+
"step": 1100
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.38,
|
| 78 |
+
"learning_rate": 3.797508783136378e-06,
|
| 79 |
+
"loss": 1.5092,
|
| 80 |
+
"step": 1200
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.42,
|
| 84 |
+
"learning_rate": 4.1168955605237945e-06,
|
| 85 |
+
"loss": 1.4302,
|
| 86 |
+
"step": 1300
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.45,
|
| 90 |
+
"learning_rate": 4.436282337911211e-06,
|
| 91 |
+
"loss": 1.3527,
|
| 92 |
+
"step": 1400
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.48,
|
| 96 |
+
"learning_rate": 4.755669115298627e-06,
|
| 97 |
+
"loss": 1.2495,
|
| 98 |
+
"step": 1500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.51,
|
| 102 |
+
"learning_rate": 5.075055892686043e-06,
|
| 103 |
+
"loss": 1.2039,
|
| 104 |
+
"step": 1600
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.54,
|
| 108 |
+
"learning_rate": 5.39444267007346e-06,
|
| 109 |
+
"loss": 1.1438,
|
| 110 |
+
"step": 1700
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.57,
|
| 114 |
+
"learning_rate": 5.713829447460876e-06,
|
| 115 |
+
"loss": 1.1056,
|
| 116 |
+
"step": 1800
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.61,
|
| 120 |
+
"learning_rate": 6.033216224848291e-06,
|
| 121 |
+
"loss": 1.0436,
|
| 122 |
+
"step": 1900
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.64,
|
| 126 |
+
"learning_rate": 6.352603002235708e-06,
|
| 127 |
+
"loss": 1.0225,
|
| 128 |
+
"step": 2000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.67,
|
| 132 |
+
"learning_rate": 6.671989779623123e-06,
|
| 133 |
+
"loss": 0.9856,
|
| 134 |
+
"step": 2100
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.7,
|
| 138 |
+
"learning_rate": 6.9913765570105406e-06,
|
| 139 |
+
"loss": 0.9412,
|
| 140 |
+
"step": 2200
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.73,
|
| 144 |
+
"learning_rate": 7.310763334397957e-06,
|
| 145 |
+
"loss": 0.9075,
|
| 146 |
+
"step": 2300
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.77,
|
| 150 |
+
"learning_rate": 7.630150111785373e-06,
|
| 151 |
+
"loss": 0.8752,
|
| 152 |
+
"step": 2400
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.8,
|
| 156 |
+
"learning_rate": 7.949536889172788e-06,
|
| 157 |
+
"loss": 0.854,
|
| 158 |
+
"step": 2500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.83,
|
| 162 |
+
"learning_rate": 8.268923666560205e-06,
|
| 163 |
+
"loss": 0.8282,
|
| 164 |
+
"step": 2600
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.86,
|
| 168 |
+
"learning_rate": 8.588310443947621e-06,
|
| 169 |
+
"loss": 0.8074,
|
| 170 |
+
"step": 2700
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.89,
|
| 174 |
+
"learning_rate": 8.907697221335038e-06,
|
| 175 |
+
"loss": 0.7976,
|
| 176 |
+
"step": 2800
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.93,
|
| 180 |
+
"learning_rate": 9.227083998722454e-06,
|
| 181 |
+
"loss": 0.7955,
|
| 182 |
+
"step": 2900
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.96,
|
| 186 |
+
"learning_rate": 9.54647077610987e-06,
|
| 187 |
+
"loss": 0.7502,
|
| 188 |
+
"step": 3000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.99,
|
| 192 |
+
"learning_rate": 9.865857553497285e-06,
|
| 193 |
+
"loss": 0.7455,
|
| 194 |
+
"step": 3100
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 1.02,
|
| 198 |
+
"learning_rate": 1.0185244330884702e-05,
|
| 199 |
+
"loss": 0.7177,
|
| 200 |
+
"step": 3200
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 1.05,
|
| 204 |
+
"learning_rate": 1.050463110827212e-05,
|
| 205 |
+
"loss": 0.724,
|
| 206 |
+
"step": 3300
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 1.09,
|
| 210 |
+
"learning_rate": 1.0824017885659535e-05,
|
| 211 |
+
"loss": 0.6975,
|
| 212 |
+
"step": 3400
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 1.12,
|
| 216 |
+
"learning_rate": 1.114340466304695e-05,
|
| 217 |
+
"loss": 0.7054,
|
| 218 |
+
"step": 3500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 1.15,
|
| 222 |
+
"learning_rate": 1.1462791440434366e-05,
|
| 223 |
+
"loss": 0.71,
|
| 224 |
+
"step": 3600
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 1.18,
|
| 228 |
+
"learning_rate": 1.1782178217821782e-05,
|
| 229 |
+
"loss": 0.6825,
|
| 230 |
+
"step": 3700
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 1.21,
|
| 234 |
+
"learning_rate": 1.2101564995209199e-05,
|
| 235 |
+
"loss": 0.6792,
|
| 236 |
+
"step": 3800
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 1.25,
|
| 240 |
+
"learning_rate": 1.2420951772596615e-05,
|
| 241 |
+
"loss": 0.6794,
|
| 242 |
+
"step": 3900
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 1.28,
|
| 246 |
+
"learning_rate": 1.2740338549984032e-05,
|
| 247 |
+
"loss": 0.6672,
|
| 248 |
+
"step": 4000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 1.31,
|
| 252 |
+
"learning_rate": 1.3059725327371449e-05,
|
| 253 |
+
"loss": 0.6356,
|
| 254 |
+
"step": 4100
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 1.34,
|
| 258 |
+
"learning_rate": 1.3379112104758865e-05,
|
| 259 |
+
"loss": 0.6503,
|
| 260 |
+
"step": 4200
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 1.37,
|
| 264 |
+
"learning_rate": 1.369849888214628e-05,
|
| 265 |
+
"loss": 0.6495,
|
| 266 |
+
"step": 4300
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 1.41,
|
| 270 |
+
"learning_rate": 1.4017885659533696e-05,
|
| 271 |
+
"loss": 0.6428,
|
| 272 |
+
"step": 4400
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 1.44,
|
| 276 |
+
"learning_rate": 1.4337272436921113e-05,
|
| 277 |
+
"loss": 0.6171,
|
| 278 |
+
"step": 4500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 1.47,
|
| 282 |
+
"learning_rate": 1.4656659214308529e-05,
|
| 283 |
+
"loss": 0.6084,
|
| 284 |
+
"step": 4600
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 1.5,
|
| 288 |
+
"learning_rate": 1.4976045991695944e-05,
|
| 289 |
+
"loss": 0.6222,
|
| 290 |
+
"step": 4700
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 1.53,
|
| 294 |
+
"learning_rate": 1.529543276908336e-05,
|
| 295 |
+
"loss": 0.5971,
|
| 296 |
+
"step": 4800
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 1.56,
|
| 300 |
+
"learning_rate": 1.5614819546470777e-05,
|
| 301 |
+
"loss": 0.5745,
|
| 302 |
+
"step": 4900
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 1.6,
|
| 306 |
+
"learning_rate": 1.5934206323858196e-05,
|
| 307 |
+
"loss": 0.5753,
|
| 308 |
+
"step": 5000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 1.63,
|
| 312 |
+
"learning_rate": 1.625359310124561e-05,
|
| 313 |
+
"loss": 0.6045,
|
| 314 |
+
"step": 5100
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 1.66,
|
| 318 |
+
"learning_rate": 1.6572979878633027e-05,
|
| 319 |
+
"loss": 0.6006,
|
| 320 |
+
"step": 5200
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 1.69,
|
| 324 |
+
"learning_rate": 1.6892366656020443e-05,
|
| 325 |
+
"loss": 0.6003,
|
| 326 |
+
"step": 5300
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 1.72,
|
| 330 |
+
"learning_rate": 1.7211753433407858e-05,
|
| 331 |
+
"loss": 0.5688,
|
| 332 |
+
"step": 5400
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 1.76,
|
| 336 |
+
"learning_rate": 1.7531140210795274e-05,
|
| 337 |
+
"loss": 0.5662,
|
| 338 |
+
"step": 5500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 1.79,
|
| 342 |
+
"learning_rate": 1.785052698818269e-05,
|
| 343 |
+
"loss": 0.5788,
|
| 344 |
+
"step": 5600
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 1.82,
|
| 348 |
+
"learning_rate": 1.8169913765570105e-05,
|
| 349 |
+
"loss": 0.5541,
|
| 350 |
+
"step": 5700
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 1.85,
|
| 354 |
+
"learning_rate": 1.8489300542957524e-05,
|
| 355 |
+
"loss": 0.5656,
|
| 356 |
+
"step": 5800
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 1.88,
|
| 360 |
+
"learning_rate": 1.880868732034494e-05,
|
| 361 |
+
"loss": 0.5466,
|
| 362 |
+
"step": 5900
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 1.92,
|
| 366 |
+
"learning_rate": 1.9128074097732355e-05,
|
| 367 |
+
"loss": 0.563,
|
| 368 |
+
"step": 6000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 1.95,
|
| 372 |
+
"learning_rate": 1.9447460875119774e-05,
|
| 373 |
+
"loss": 0.5603,
|
| 374 |
+
"step": 6100
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 1.98,
|
| 378 |
+
"learning_rate": 1.976684765250719e-05,
|
| 379 |
+
"loss": 0.5579,
|
| 380 |
+
"step": 6200
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 2.01,
|
| 384 |
+
"learning_rate": 2.0086234429894605e-05,
|
| 385 |
+
"loss": 0.5346,
|
| 386 |
+
"step": 6300
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 2.04,
|
| 390 |
+
"learning_rate": 2.0405621207282017e-05,
|
| 391 |
+
"loss": 0.5038,
|
| 392 |
+
"step": 6400
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 2.08,
|
| 396 |
+
"learning_rate": 2.0725007984669436e-05,
|
| 397 |
+
"loss": 0.5121,
|
| 398 |
+
"step": 6500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 2.11,
|
| 402 |
+
"learning_rate": 2.1044394762056855e-05,
|
| 403 |
+
"loss": 0.5135,
|
| 404 |
+
"step": 6600
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 2.14,
|
| 408 |
+
"learning_rate": 2.1363781539444267e-05,
|
| 409 |
+
"loss": 0.504,
|
| 410 |
+
"step": 6700
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 2.17,
|
| 414 |
+
"learning_rate": 2.1683168316831686e-05,
|
| 415 |
+
"loss": 0.5119,
|
| 416 |
+
"step": 6800
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 2.2,
|
| 420 |
+
"learning_rate": 2.20025550942191e-05,
|
| 421 |
+
"loss": 0.4986,
|
| 422 |
+
"step": 6900
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 2.24,
|
| 426 |
+
"learning_rate": 2.2321941871606518e-05,
|
| 427 |
+
"loss": 0.504,
|
| 428 |
+
"step": 7000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 2.27,
|
| 432 |
+
"learning_rate": 2.2641328648993933e-05,
|
| 433 |
+
"loss": 0.5024,
|
| 434 |
+
"step": 7100
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 2.3,
|
| 438 |
+
"learning_rate": 2.296071542638135e-05,
|
| 439 |
+
"loss": 0.479,
|
| 440 |
+
"step": 7200
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 2.33,
|
| 444 |
+
"learning_rate": 2.3280102203768764e-05,
|
| 445 |
+
"loss": 0.492,
|
| 446 |
+
"step": 7300
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 2.36,
|
| 450 |
+
"learning_rate": 2.3599488981156183e-05,
|
| 451 |
+
"loss": 0.4928,
|
| 452 |
+
"step": 7400
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 2.4,
|
| 456 |
+
"learning_rate": 2.39188757585436e-05,
|
| 457 |
+
"loss": 0.4862,
|
| 458 |
+
"step": 7500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 2.43,
|
| 462 |
+
"learning_rate": 2.4238262535931014e-05,
|
| 463 |
+
"loss": 0.5148,
|
| 464 |
+
"step": 7600
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 2.46,
|
| 468 |
+
"learning_rate": 2.4557649313318433e-05,
|
| 469 |
+
"loss": 0.4922,
|
| 470 |
+
"step": 7700
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 2.49,
|
| 474 |
+
"learning_rate": 2.4877036090705846e-05,
|
| 475 |
+
"loss": 0.4882,
|
| 476 |
+
"step": 7800
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 2.52,
|
| 480 |
+
"learning_rate": 2.5196422868093264e-05,
|
| 481 |
+
"loss": 0.4871,
|
| 482 |
+
"step": 7900
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 2.56,
|
| 486 |
+
"learning_rate": 2.5515809645480677e-05,
|
| 487 |
+
"loss": 0.4967,
|
| 488 |
+
"step": 8000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 2.59,
|
| 492 |
+
"learning_rate": 2.5835196422868096e-05,
|
| 493 |
+
"loss": 0.4834,
|
| 494 |
+
"step": 8100
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 2.62,
|
| 498 |
+
"learning_rate": 2.615458320025551e-05,
|
| 499 |
+
"loss": 0.4769,
|
| 500 |
+
"step": 8200
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 2.65,
|
| 504 |
+
"learning_rate": 2.6473969977642927e-05,
|
| 505 |
+
"loss": 0.4776,
|
| 506 |
+
"step": 8300
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 2.68,
|
| 510 |
+
"learning_rate": 2.6793356755030342e-05,
|
| 511 |
+
"loss": 0.4861,
|
| 512 |
+
"step": 8400
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 2.71,
|
| 516 |
+
"learning_rate": 2.711274353241776e-05,
|
| 517 |
+
"loss": 0.486,
|
| 518 |
+
"step": 8500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 2.75,
|
| 522 |
+
"learning_rate": 2.7432130309805177e-05,
|
| 523 |
+
"loss": 0.4745,
|
| 524 |
+
"step": 8600
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 2.78,
|
| 528 |
+
"learning_rate": 2.7751517087192592e-05,
|
| 529 |
+
"loss": 0.4805,
|
| 530 |
+
"step": 8700
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 2.81,
|
| 534 |
+
"learning_rate": 2.807090386458001e-05,
|
| 535 |
+
"loss": 0.5003,
|
| 536 |
+
"step": 8800
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 2.84,
|
| 540 |
+
"learning_rate": 2.8390290641967424e-05,
|
| 541 |
+
"loss": 0.4639,
|
| 542 |
+
"step": 8900
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 2.87,
|
| 546 |
+
"learning_rate": 2.8709677419354843e-05,
|
| 547 |
+
"loss": 0.4763,
|
| 548 |
+
"step": 9000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 2.91,
|
| 552 |
+
"learning_rate": 2.9029064196742255e-05,
|
| 553 |
+
"loss": 0.4727,
|
| 554 |
+
"step": 9100
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 2.94,
|
| 558 |
+
"learning_rate": 2.9348450974129674e-05,
|
| 559 |
+
"loss": 0.4839,
|
| 560 |
+
"step": 9200
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 2.97,
|
| 564 |
+
"learning_rate": 2.966783775151709e-05,
|
| 565 |
+
"loss": 0.4769,
|
| 566 |
+
"step": 9300
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 3.0,
|
| 570 |
+
"learning_rate": 2.9987224528904505e-05,
|
| 571 |
+
"loss": 0.4641,
|
| 572 |
+
"step": 9400
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 3.03,
|
| 576 |
+
"learning_rate": 3.030661130629192e-05,
|
| 577 |
+
"loss": 0.4282,
|
| 578 |
+
"step": 9500
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 3.07,
|
| 582 |
+
"learning_rate": 3.0625998083679336e-05,
|
| 583 |
+
"loss": 0.4057,
|
| 584 |
+
"step": 9600
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 3.1,
|
| 588 |
+
"learning_rate": 3.094538486106676e-05,
|
| 589 |
+
"loss": 0.4192,
|
| 590 |
+
"step": 9700
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 3.13,
|
| 594 |
+
"learning_rate": 3.126477163845417e-05,
|
| 595 |
+
"loss": 0.4186,
|
| 596 |
+
"step": 9800
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 3.16,
|
| 600 |
+
"learning_rate": 3.158415841584159e-05,
|
| 601 |
+
"loss": 0.4086,
|
| 602 |
+
"step": 9900
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 3.19,
|
| 606 |
+
"learning_rate": 3.1903545193229005e-05,
|
| 607 |
+
"loss": 0.4194,
|
| 608 |
+
"step": 10000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 3.23,
|
| 612 |
+
"learning_rate": 3.221973810284254e-05,
|
| 613 |
+
"loss": 0.4443,
|
| 614 |
+
"step": 10100
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 3.26,
|
| 618 |
+
"learning_rate": 3.2539124880229965e-05,
|
| 619 |
+
"loss": 0.4217,
|
| 620 |
+
"step": 10200
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 3.29,
|
| 624 |
+
"learning_rate": 3.2858511657617374e-05,
|
| 625 |
+
"loss": 0.4275,
|
| 626 |
+
"step": 10300
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 3.32,
|
| 630 |
+
"learning_rate": 3.3177898435004796e-05,
|
| 631 |
+
"loss": 0.4422,
|
| 632 |
+
"step": 10400
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 3.35,
|
| 636 |
+
"learning_rate": 3.3497285212392205e-05,
|
| 637 |
+
"loss": 0.4227,
|
| 638 |
+
"step": 10500
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 3.39,
|
| 642 |
+
"learning_rate": 3.381667198977963e-05,
|
| 643 |
+
"loss": 0.4191,
|
| 644 |
+
"step": 10600
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 3.42,
|
| 648 |
+
"learning_rate": 3.413605876716704e-05,
|
| 649 |
+
"loss": 0.4309,
|
| 650 |
+
"step": 10700
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 3.45,
|
| 654 |
+
"learning_rate": 3.445544554455446e-05,
|
| 655 |
+
"loss": 0.4079,
|
| 656 |
+
"step": 10800
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 3.48,
|
| 660 |
+
"learning_rate": 3.4774832321941874e-05,
|
| 661 |
+
"loss": 0.4281,
|
| 662 |
+
"step": 10900
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 3.51,
|
| 666 |
+
"learning_rate": 3.509421909932929e-05,
|
| 667 |
+
"loss": 0.4137,
|
| 668 |
+
"step": 11000
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 3.55,
|
| 672 |
+
"learning_rate": 3.5413605876716705e-05,
|
| 673 |
+
"loss": 0.4168,
|
| 674 |
+
"step": 11100
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 3.58,
|
| 678 |
+
"learning_rate": 3.573299265410412e-05,
|
| 679 |
+
"loss": 0.4186,
|
| 680 |
+
"step": 11200
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 3.61,
|
| 684 |
+
"learning_rate": 3.605237943149154e-05,
|
| 685 |
+
"loss": 0.4325,
|
| 686 |
+
"step": 11300
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 3.64,
|
| 690 |
+
"learning_rate": 3.637176620887895e-05,
|
| 691 |
+
"loss": 0.4422,
|
| 692 |
+
"step": 11400
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 3.67,
|
| 696 |
+
"learning_rate": 3.6691152986266374e-05,
|
| 697 |
+
"loss": 0.4269,
|
| 698 |
+
"step": 11500
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 3.7,
|
| 702 |
+
"learning_rate": 3.701053976365379e-05,
|
| 703 |
+
"loss": 0.4368,
|
| 704 |
+
"step": 11600
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 3.74,
|
| 708 |
+
"learning_rate": 3.7329926541041205e-05,
|
| 709 |
+
"loss": 0.4207,
|
| 710 |
+
"step": 11700
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 3.77,
|
| 714 |
+
"learning_rate": 3.764931331842862e-05,
|
| 715 |
+
"loss": 0.4365,
|
| 716 |
+
"step": 11800
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 3.8,
|
| 720 |
+
"learning_rate": 3.7968700095816036e-05,
|
| 721 |
+
"loss": 0.4394,
|
| 722 |
+
"step": 11900
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 3.83,
|
| 726 |
+
"learning_rate": 3.828808687320345e-05,
|
| 727 |
+
"loss": 0.4351,
|
| 728 |
+
"step": 12000
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 3.86,
|
| 732 |
+
"learning_rate": 3.860747365059087e-05,
|
| 733 |
+
"loss": 0.4113,
|
| 734 |
+
"step": 12100
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 3.9,
|
| 738 |
+
"learning_rate": 3.892686042797828e-05,
|
| 739 |
+
"loss": 0.4261,
|
| 740 |
+
"step": 12200
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 3.93,
|
| 744 |
+
"learning_rate": 3.92462472053657e-05,
|
| 745 |
+
"loss": 0.4311,
|
| 746 |
+
"step": 12300
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 3.96,
|
| 750 |
+
"learning_rate": 3.956563398275312e-05,
|
| 751 |
+
"loss": 0.4246,
|
| 752 |
+
"step": 12400
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 3.99,
|
| 756 |
+
"learning_rate": 3.988502076014053e-05,
|
| 757 |
+
"loss": 0.4294,
|
| 758 |
+
"step": 12500
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 4.02,
|
| 762 |
+
"learning_rate": 3.997728805138579e-05,
|
| 763 |
+
"loss": 0.3854,
|
| 764 |
+
"step": 12600
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 4.06,
|
| 768 |
+
"learning_rate": 3.994180063167608e-05,
|
| 769 |
+
"loss": 0.3656,
|
| 770 |
+
"step": 12700
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 4.09,
|
| 774 |
+
"learning_rate": 3.9906313211966366e-05,
|
| 775 |
+
"loss": 0.3836,
|
| 776 |
+
"step": 12800
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 4.12,
|
| 780 |
+
"learning_rate": 3.987082579225665e-05,
|
| 781 |
+
"loss": 0.3664,
|
| 782 |
+
"step": 12900
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 4.15,
|
| 786 |
+
"learning_rate": 3.9835338372546935e-05,
|
| 787 |
+
"loss": 0.3884,
|
| 788 |
+
"step": 13000
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 4.18,
|
| 792 |
+
"learning_rate": 3.979985095283722e-05,
|
| 793 |
+
"loss": 0.3764,
|
| 794 |
+
"step": 13100
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 4.22,
|
| 798 |
+
"learning_rate": 3.976436353312751e-05,
|
| 799 |
+
"loss": 0.3733,
|
| 800 |
+
"step": 13200
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 4.25,
|
| 804 |
+
"learning_rate": 3.97288761134178e-05,
|
| 805 |
+
"loss": 0.3682,
|
| 806 |
+
"step": 13300
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 4.28,
|
| 810 |
+
"learning_rate": 3.969338869370809e-05,
|
| 811 |
+
"loss": 0.371,
|
| 812 |
+
"step": 13400
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 4.31,
|
| 816 |
+
"learning_rate": 3.965790127399837e-05,
|
| 817 |
+
"loss": 0.3741,
|
| 818 |
+
"step": 13500
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 4.34,
|
| 822 |
+
"learning_rate": 3.9622413854288656e-05,
|
| 823 |
+
"loss": 0.3757,
|
| 824 |
+
"step": 13600
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 4.38,
|
| 828 |
+
"learning_rate": 3.9586926434578944e-05,
|
| 829 |
+
"loss": 0.3652,
|
| 830 |
+
"step": 13700
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 4.41,
|
| 834 |
+
"learning_rate": 3.955143901486923e-05,
|
| 835 |
+
"loss": 0.3892,
|
| 836 |
+
"step": 13800
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 4.44,
|
| 840 |
+
"learning_rate": 3.951630646935662e-05,
|
| 841 |
+
"loss": 0.3893,
|
| 842 |
+
"step": 13900
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 4.47,
|
| 846 |
+
"learning_rate": 3.948081904964691e-05,
|
| 847 |
+
"loss": 0.3823,
|
| 848 |
+
"step": 14000
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 4.5,
|
| 852 |
+
"learning_rate": 3.944533162993719e-05,
|
| 853 |
+
"loss": 0.3654,
|
| 854 |
+
"step": 14100
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 4.54,
|
| 858 |
+
"learning_rate": 3.9409844210227476e-05,
|
| 859 |
+
"loss": 0.3791,
|
| 860 |
+
"step": 14200
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 4.57,
|
| 864 |
+
"learning_rate": 3.9374356790517764e-05,
|
| 865 |
+
"loss": 0.3918,
|
| 866 |
+
"step": 14300
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 4.6,
|
| 870 |
+
"learning_rate": 3.933886937080805e-05,
|
| 871 |
+
"loss": 0.3713,
|
| 872 |
+
"step": 14400
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 4.63,
|
| 876 |
+
"learning_rate": 3.930338195109834e-05,
|
| 877 |
+
"loss": 0.376,
|
| 878 |
+
"step": 14500
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 4.66,
|
| 882 |
+
"learning_rate": 3.926789453138863e-05,
|
| 883 |
+
"loss": 0.3682,
|
| 884 |
+
"step": 14600
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 4.69,
|
| 888 |
+
"learning_rate": 3.9232407111678915e-05,
|
| 889 |
+
"loss": 0.374,
|
| 890 |
+
"step": 14700
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 4.73,
|
| 894 |
+
"learning_rate": 3.91969196919692e-05,
|
| 895 |
+
"loss": 0.3774,
|
| 896 |
+
"step": 14800
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 4.76,
|
| 900 |
+
"learning_rate": 3.9161432272259484e-05,
|
| 901 |
+
"loss": 0.3806,
|
| 902 |
+
"step": 14900
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 4.79,
|
| 906 |
+
"learning_rate": 3.912594485254977e-05,
|
| 907 |
+
"loss": 0.3886,
|
| 908 |
+
"step": 15000
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 4.82,
|
| 912 |
+
"learning_rate": 3.909045743284006e-05,
|
| 913 |
+
"loss": 0.3742,
|
| 914 |
+
"step": 15100
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 4.85,
|
| 918 |
+
"learning_rate": 3.905497001313035e-05,
|
| 919 |
+
"loss": 0.3714,
|
| 920 |
+
"step": 15200
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 4.89,
|
| 924 |
+
"learning_rate": 3.9019482593420636e-05,
|
| 925 |
+
"loss": 0.3721,
|
| 926 |
+
"step": 15300
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 4.92,
|
| 930 |
+
"learning_rate": 3.8983995173710924e-05,
|
| 931 |
+
"loss": 0.3716,
|
| 932 |
+
"step": 15400
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 4.95,
|
| 936 |
+
"learning_rate": 3.8948507754001205e-05,
|
| 937 |
+
"loss": 0.3728,
|
| 938 |
+
"step": 15500
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 4.98,
|
| 942 |
+
"learning_rate": 3.891302033429149e-05,
|
| 943 |
+
"loss": 0.384,
|
| 944 |
+
"step": 15600
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 5.01,
|
| 948 |
+
"learning_rate": 3.887753291458178e-05,
|
| 949 |
+
"loss": 0.3415,
|
| 950 |
+
"step": 15700
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 5.05,
|
| 954 |
+
"learning_rate": 3.884204549487207e-05,
|
| 955 |
+
"loss": 0.3064,
|
| 956 |
+
"step": 15800
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 5.08,
|
| 960 |
+
"learning_rate": 3.880655807516236e-05,
|
| 961 |
+
"loss": 0.3164,
|
| 962 |
+
"step": 15900
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 5.11,
|
| 966 |
+
"learning_rate": 3.8771070655452645e-05,
|
| 967 |
+
"loss": 0.3215,
|
| 968 |
+
"step": 16000
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 5.14,
|
| 972 |
+
"learning_rate": 3.873558323574293e-05,
|
| 973 |
+
"loss": 0.3234,
|
| 974 |
+
"step": 16100
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 5.17,
|
| 978 |
+
"learning_rate": 3.870009581603322e-05,
|
| 979 |
+
"loss": 0.3119,
|
| 980 |
+
"step": 16200
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 5.21,
|
| 984 |
+
"learning_rate": 3.866460839632351e-05,
|
| 985 |
+
"loss": 0.3145,
|
| 986 |
+
"step": 16300
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 5.24,
|
| 990 |
+
"learning_rate": 3.8629120976613796e-05,
|
| 991 |
+
"loss": 0.3333,
|
| 992 |
+
"step": 16400
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 5.27,
|
| 996 |
+
"learning_rate": 3.8593633556904084e-05,
|
| 997 |
+
"loss": 0.3195,
|
| 998 |
+
"step": 16500
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 5.3,
|
| 1002 |
+
"learning_rate": 3.855814613719437e-05,
|
| 1003 |
+
"loss": 0.3159,
|
| 1004 |
+
"step": 16600
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 5.33,
|
| 1008 |
+
"learning_rate": 3.852265871748465e-05,
|
| 1009 |
+
"loss": 0.3189,
|
| 1010 |
+
"step": 16700
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 5.37,
|
| 1014 |
+
"learning_rate": 3.848717129777494e-05,
|
| 1015 |
+
"loss": 0.317,
|
| 1016 |
+
"step": 16800
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 5.4,
|
| 1020 |
+
"learning_rate": 3.845168387806523e-05,
|
| 1021 |
+
"loss": 0.3231,
|
| 1022 |
+
"step": 16900
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 5.43,
|
| 1026 |
+
"learning_rate": 3.841619645835552e-05,
|
| 1027 |
+
"loss": 0.3328,
|
| 1028 |
+
"step": 17000
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 5.46,
|
| 1032 |
+
"learning_rate": 3.8380709038645805e-05,
|
| 1033 |
+
"loss": 0.3282,
|
| 1034 |
+
"step": 17100
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 5.49,
|
| 1038 |
+
"learning_rate": 3.8345221618936086e-05,
|
| 1039 |
+
"loss": 0.3169,
|
| 1040 |
+
"step": 17200
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 5.53,
|
| 1044 |
+
"learning_rate": 3.8309734199226374e-05,
|
| 1045 |
+
"loss": 0.3156,
|
| 1046 |
+
"step": 17300
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 5.56,
|
| 1050 |
+
"learning_rate": 3.827424677951666e-05,
|
| 1051 |
+
"loss": 0.3313,
|
| 1052 |
+
"step": 17400
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 5.59,
|
| 1056 |
+
"learning_rate": 3.823875935980695e-05,
|
| 1057 |
+
"loss": 0.3181,
|
| 1058 |
+
"step": 17500
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 5.62,
|
| 1062 |
+
"learning_rate": 3.820327194009724e-05,
|
| 1063 |
+
"loss": 0.334,
|
| 1064 |
+
"step": 17600
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 5.65,
|
| 1068 |
+
"learning_rate": 3.8167784520387526e-05,
|
| 1069 |
+
"loss": 0.3313,
|
| 1070 |
+
"step": 17700
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 5.69,
|
| 1074 |
+
"learning_rate": 3.8132297100677814e-05,
|
| 1075 |
+
"loss": 0.3331,
|
| 1076 |
+
"step": 17800
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 5.72,
|
| 1080 |
+
"learning_rate": 3.80968096809681e-05,
|
| 1081 |
+
"loss": 0.3405,
|
| 1082 |
+
"step": 17900
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 5.75,
|
| 1086 |
+
"learning_rate": 3.806132226125839e-05,
|
| 1087 |
+
"loss": 0.3345,
|
| 1088 |
+
"step": 18000
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 5.78,
|
| 1092 |
+
"learning_rate": 3.802583484154868e-05,
|
| 1093 |
+
"loss": 0.3295,
|
| 1094 |
+
"step": 18100
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 5.81,
|
| 1098 |
+
"learning_rate": 3.7990347421838965e-05,
|
| 1099 |
+
"loss": 0.3244,
|
| 1100 |
+
"step": 18200
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 5.84,
|
| 1104 |
+
"learning_rate": 3.795486000212925e-05,
|
| 1105 |
+
"loss": 0.3346,
|
| 1106 |
+
"step": 18300
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 5.88,
|
| 1110 |
+
"learning_rate": 3.791972745661663e-05,
|
| 1111 |
+
"loss": 0.3317,
|
| 1112 |
+
"step": 18400
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 5.91,
|
| 1116 |
+
"learning_rate": 3.788424003690692e-05,
|
| 1117 |
+
"loss": 0.3219,
|
| 1118 |
+
"step": 18500
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 5.94,
|
| 1122 |
+
"learning_rate": 3.784875261719721e-05,
|
| 1123 |
+
"loss": 0.3167,
|
| 1124 |
+
"step": 18600
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 5.97,
|
| 1128 |
+
"learning_rate": 3.781326519748749e-05,
|
| 1129 |
+
"loss": 0.33,
|
| 1130 |
+
"step": 18700
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 6.0,
|
| 1134 |
+
"learning_rate": 3.777777777777778e-05,
|
| 1135 |
+
"loss": 0.3211,
|
| 1136 |
+
"step": 18800
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 6.04,
|
| 1140 |
+
"learning_rate": 3.7742290358068066e-05,
|
| 1141 |
+
"loss": 0.2794,
|
| 1142 |
+
"step": 18900
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 6.07,
|
| 1146 |
+
"learning_rate": 3.7706802938358354e-05,
|
| 1147 |
+
"loss": 0.2675,
|
| 1148 |
+
"step": 19000
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 6.1,
|
| 1152 |
+
"learning_rate": 3.767131551864864e-05,
|
| 1153 |
+
"loss": 0.2676,
|
| 1154 |
+
"step": 19100
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 6.13,
|
| 1158 |
+
"learning_rate": 3.763582809893893e-05,
|
| 1159 |
+
"loss": 0.2693,
|
| 1160 |
+
"step": 19200
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 6.16,
|
| 1164 |
+
"learning_rate": 3.760034067922922e-05,
|
| 1165 |
+
"loss": 0.2669,
|
| 1166 |
+
"step": 19300
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 6.2,
|
| 1170 |
+
"learning_rate": 3.7564853259519506e-05,
|
| 1171 |
+
"loss": 0.2744,
|
| 1172 |
+
"step": 19400
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 6.23,
|
| 1176 |
+
"learning_rate": 3.7529365839809794e-05,
|
| 1177 |
+
"loss": 0.282,
|
| 1178 |
+
"step": 19500
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 6.26,
|
| 1182 |
+
"learning_rate": 3.749387842010008e-05,
|
| 1183 |
+
"loss": 0.2848,
|
| 1184 |
+
"step": 19600
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 6.29,
|
| 1188 |
+
"learning_rate": 3.745839100039037e-05,
|
| 1189 |
+
"loss": 0.2783,
|
| 1190 |
+
"step": 19700
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 6.32,
|
| 1194 |
+
"learning_rate": 3.742290358068066e-05,
|
| 1195 |
+
"loss": 0.2843,
|
| 1196 |
+
"step": 19800
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 6.36,
|
| 1200 |
+
"learning_rate": 3.738741616097094e-05,
|
| 1201 |
+
"loss": 0.2696,
|
| 1202 |
+
"step": 19900
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 6.39,
|
| 1206 |
+
"learning_rate": 3.7351928741261226e-05,
|
| 1207 |
+
"loss": 0.2932,
|
| 1208 |
+
"step": 20000
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 6.42,
|
| 1212 |
+
"learning_rate": 3.7316441321551514e-05,
|
| 1213 |
+
"loss": 0.2775,
|
| 1214 |
+
"step": 20100
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 6.45,
|
| 1218 |
+
"learning_rate": 3.72809539018418e-05,
|
| 1219 |
+
"loss": 0.2893,
|
| 1220 |
+
"step": 20200
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 6.48,
|
| 1224 |
+
"learning_rate": 3.724546648213209e-05,
|
| 1225 |
+
"loss": 0.286,
|
| 1226 |
+
"step": 20300
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 6.52,
|
| 1230 |
+
"learning_rate": 3.720997906242237e-05,
|
| 1231 |
+
"loss": 0.2799,
|
| 1232 |
+
"step": 20400
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 6.55,
|
| 1236 |
+
"learning_rate": 3.717449164271266e-05,
|
| 1237 |
+
"loss": 0.2827,
|
| 1238 |
+
"step": 20500
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 6.58,
|
| 1242 |
+
"learning_rate": 3.713900422300295e-05,
|
| 1243 |
+
"loss": 0.2783,
|
| 1244 |
+
"step": 20600
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 6.61,
|
| 1248 |
+
"learning_rate": 3.7103516803293235e-05,
|
| 1249 |
+
"loss": 0.2729,
|
| 1250 |
+
"step": 20700
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 6.64,
|
| 1254 |
+
"learning_rate": 3.706802938358352e-05,
|
| 1255 |
+
"loss": 0.2747,
|
| 1256 |
+
"step": 20800
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 6.68,
|
| 1260 |
+
"learning_rate": 3.703254196387381e-05,
|
| 1261 |
+
"loss": 0.2764,
|
| 1262 |
+
"step": 20900
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 6.71,
|
| 1266 |
+
"learning_rate": 3.699705454416409e-05,
|
| 1267 |
+
"loss": 0.2804,
|
| 1268 |
+
"step": 21000
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 6.74,
|
| 1272 |
+
"learning_rate": 3.696156712445438e-05,
|
| 1273 |
+
"loss": 0.2841,
|
| 1274 |
+
"step": 21100
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 6.77,
|
| 1278 |
+
"learning_rate": 3.692607970474467e-05,
|
| 1279 |
+
"loss": 0.277,
|
| 1280 |
+
"step": 21200
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 6.8,
|
| 1284 |
+
"learning_rate": 3.6890592285034956e-05,
|
| 1285 |
+
"loss": 0.284,
|
| 1286 |
+
"step": 21300
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 6.83,
|
| 1290 |
+
"learning_rate": 3.6855104865325244e-05,
|
| 1291 |
+
"loss": 0.2778,
|
| 1292 |
+
"step": 21400
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 6.87,
|
| 1296 |
+
"learning_rate": 3.681961744561553e-05,
|
| 1297 |
+
"loss": 0.2797,
|
| 1298 |
+
"step": 21500
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 6.9,
|
| 1302 |
+
"learning_rate": 3.678413002590582e-05,
|
| 1303 |
+
"loss": 0.3036,
|
| 1304 |
+
"step": 21600
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 6.93,
|
| 1308 |
+
"learning_rate": 3.674864260619611e-05,
|
| 1309 |
+
"loss": 0.2869,
|
| 1310 |
+
"step": 21700
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 6.96,
|
| 1314 |
+
"learning_rate": 3.6713155186486395e-05,
|
| 1315 |
+
"loss": 0.2921,
|
| 1316 |
+
"step": 21800
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 6.99,
|
| 1320 |
+
"learning_rate": 3.667766776677668e-05,
|
| 1321 |
+
"loss": 0.2844,
|
| 1322 |
+
"step": 21900
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 7.03,
|
| 1326 |
+
"learning_rate": 3.664218034706697e-05,
|
| 1327 |
+
"loss": 0.2427,
|
| 1328 |
+
"step": 22000
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 7.06,
|
| 1332 |
+
"learning_rate": 3.660669292735726e-05,
|
| 1333 |
+
"loss": 0.243,
|
| 1334 |
+
"step": 22100
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 7.09,
|
| 1338 |
+
"learning_rate": 3.657120550764754e-05,
|
| 1339 |
+
"loss": 0.2311,
|
| 1340 |
+
"step": 22200
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 7.12,
|
| 1344 |
+
"learning_rate": 3.653571808793783e-05,
|
| 1345 |
+
"loss": 0.2358,
|
| 1346 |
+
"step": 22300
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 7.15,
|
| 1350 |
+
"learning_rate": 3.6500230668228116e-05,
|
| 1351 |
+
"loss": 0.2287,
|
| 1352 |
+
"step": 22400
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 7.19,
|
| 1356 |
+
"learning_rate": 3.6464743248518404e-05,
|
| 1357 |
+
"loss": 0.2519,
|
| 1358 |
+
"step": 22500
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 7.22,
|
| 1362 |
+
"learning_rate": 3.642925582880869e-05,
|
| 1363 |
+
"loss": 0.2528,
|
| 1364 |
+
"step": 22600
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 7.25,
|
| 1368 |
+
"learning_rate": 3.639376840909898e-05,
|
| 1369 |
+
"loss": 0.2507,
|
| 1370 |
+
"step": 22700
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 7.28,
|
| 1374 |
+
"learning_rate": 3.635828098938926e-05,
|
| 1375 |
+
"loss": 0.2496,
|
| 1376 |
+
"step": 22800
|
| 1377 |
+
},
|
| 1378 |
+
{
|
| 1379 |
+
"epoch": 7.31,
|
| 1380 |
+
"learning_rate": 3.632279356967955e-05,
|
| 1381 |
+
"loss": 0.2503,
|
| 1382 |
+
"step": 22900
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 7.35,
|
| 1386 |
+
"learning_rate": 3.628730614996984e-05,
|
| 1387 |
+
"loss": 0.2383,
|
| 1388 |
+
"step": 23000
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 7.38,
|
| 1392 |
+
"learning_rate": 3.6251818730260125e-05,
|
| 1393 |
+
"loss": 0.2513,
|
| 1394 |
+
"step": 23100
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 7.41,
|
| 1398 |
+
"learning_rate": 3.621633131055041e-05,
|
| 1399 |
+
"loss": 0.2492,
|
| 1400 |
+
"step": 23200
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 7.44,
|
| 1404 |
+
"learning_rate": 3.61808438908407e-05,
|
| 1405 |
+
"loss": 0.2473,
|
| 1406 |
+
"step": 23300
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 7.47,
|
| 1410 |
+
"learning_rate": 3.614535647113099e-05,
|
| 1411 |
+
"loss": 0.2502,
|
| 1412 |
+
"step": 23400
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 7.51,
|
| 1416 |
+
"learning_rate": 3.6109869051421276e-05,
|
| 1417 |
+
"loss": 0.2492,
|
| 1418 |
+
"step": 23500
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 7.54,
|
| 1422 |
+
"learning_rate": 3.6074381631711564e-05,
|
| 1423 |
+
"loss": 0.2366,
|
| 1424 |
+
"step": 23600
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 7.57,
|
| 1428 |
+
"learning_rate": 3.603889421200185e-05,
|
| 1429 |
+
"loss": 0.2393,
|
| 1430 |
+
"step": 23700
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 7.6,
|
| 1434 |
+
"learning_rate": 3.600340679229214e-05,
|
| 1435 |
+
"loss": 0.2483,
|
| 1436 |
+
"step": 23800
|
| 1437 |
+
},
|
| 1438 |
+
{
|
| 1439 |
+
"epoch": 7.63,
|
| 1440 |
+
"learning_rate": 3.596791937258242e-05,
|
| 1441 |
+
"loss": 0.2428,
|
| 1442 |
+
"step": 23900
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 7.67,
|
| 1446 |
+
"learning_rate": 3.593243195287271e-05,
|
| 1447 |
+
"loss": 0.2507,
|
| 1448 |
+
"step": 24000
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 7.7,
|
| 1452 |
+
"learning_rate": 3.5896944533163e-05,
|
| 1453 |
+
"loss": 0.2491,
|
| 1454 |
+
"step": 24100
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 7.73,
|
| 1458 |
+
"learning_rate": 3.5861457113453285e-05,
|
| 1459 |
+
"loss": 0.2549,
|
| 1460 |
+
"step": 24200
|
| 1461 |
+
},
|
| 1462 |
+
{
|
| 1463 |
+
"epoch": 7.76,
|
| 1464 |
+
"learning_rate": 3.582596969374357e-05,
|
| 1465 |
+
"loss": 0.251,
|
| 1466 |
+
"step": 24300
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 7.79,
|
| 1470 |
+
"learning_rate": 3.579048227403386e-05,
|
| 1471 |
+
"loss": 0.2357,
|
| 1472 |
+
"step": 24400
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 7.82,
|
| 1476 |
+
"learning_rate": 3.575499485432414e-05,
|
| 1477 |
+
"loss": 0.2441,
|
| 1478 |
+
"step": 24500
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 7.86,
|
| 1482 |
+
"learning_rate": 3.571950743461443e-05,
|
| 1483 |
+
"loss": 0.2561,
|
| 1484 |
+
"step": 24600
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 7.89,
|
| 1488 |
+
"learning_rate": 3.568402001490472e-05,
|
| 1489 |
+
"loss": 0.2524,
|
| 1490 |
+
"step": 24700
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 7.92,
|
| 1494 |
+
"learning_rate": 3.5648532595195006e-05,
|
| 1495 |
+
"loss": 0.2467,
|
| 1496 |
+
"step": 24800
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 7.95,
|
| 1500 |
+
"learning_rate": 3.5613045175485293e-05,
|
| 1501 |
+
"loss": 0.2416,
|
| 1502 |
+
"step": 24900
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 7.98,
|
| 1506 |
+
"learning_rate": 3.557755775577558e-05,
|
| 1507 |
+
"loss": 0.2511,
|
| 1508 |
+
"step": 25000
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 8.02,
|
| 1512 |
+
"learning_rate": 3.554207033606587e-05,
|
| 1513 |
+
"loss": 0.2323,
|
| 1514 |
+
"step": 25100
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 8.05,
|
| 1518 |
+
"learning_rate": 3.550658291635615e-05,
|
| 1519 |
+
"loss": 0.2158,
|
| 1520 |
+
"step": 25200
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 8.08,
|
| 1524 |
+
"learning_rate": 3.547109549664644e-05,
|
| 1525 |
+
"loss": 0.21,
|
| 1526 |
+
"step": 25300
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 8.11,
|
| 1530 |
+
"learning_rate": 3.5435608076936726e-05,
|
| 1531 |
+
"loss": 0.2102,
|
| 1532 |
+
"step": 25400
|
| 1533 |
+
},
|
| 1534 |
+
{
|
| 1535 |
+
"epoch": 8.14,
|
| 1536 |
+
"learning_rate": 3.5400120657227014e-05,
|
| 1537 |
+
"loss": 0.2071,
|
| 1538 |
+
"step": 25500
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 8.18,
|
| 1542 |
+
"learning_rate": 3.53646332375173e-05,
|
| 1543 |
+
"loss": 0.2198,
|
| 1544 |
+
"step": 25600
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 8.21,
|
| 1548 |
+
"learning_rate": 3.532914581780759e-05,
|
| 1549 |
+
"loss": 0.2193,
|
| 1550 |
+
"step": 25700
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 8.24,
|
| 1554 |
+
"learning_rate": 3.529365839809788e-05,
|
| 1555 |
+
"loss": 0.2166,
|
| 1556 |
+
"step": 25800
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 8.27,
|
| 1560 |
+
"learning_rate": 3.5258170978388166e-05,
|
| 1561 |
+
"loss": 0.2237,
|
| 1562 |
+
"step": 25900
|
| 1563 |
+
},
|
| 1564 |
+
{
|
| 1565 |
+
"epoch": 8.3,
|
| 1566 |
+
"learning_rate": 3.5222683558678454e-05,
|
| 1567 |
+
"loss": 0.2241,
|
| 1568 |
+
"step": 26000
|
| 1569 |
+
},
|
| 1570 |
+
{
|
| 1571 |
+
"epoch": 8.34,
|
| 1572 |
+
"learning_rate": 3.518719613896874e-05,
|
| 1573 |
+
"loss": 0.2132,
|
| 1574 |
+
"step": 26100
|
| 1575 |
+
},
|
| 1576 |
+
{
|
| 1577 |
+
"epoch": 8.37,
|
| 1578 |
+
"learning_rate": 3.515170871925902e-05,
|
| 1579 |
+
"loss": 0.2101,
|
| 1580 |
+
"step": 26200
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 8.4,
|
| 1584 |
+
"learning_rate": 3.511622129954931e-05,
|
| 1585 |
+
"loss": 0.227,
|
| 1586 |
+
"step": 26300
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 8.43,
|
| 1590 |
+
"learning_rate": 3.50807338798396e-05,
|
| 1591 |
+
"loss": 0.2221,
|
| 1592 |
+
"step": 26400
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 8.46,
|
| 1596 |
+
"learning_rate": 3.504560133432698e-05,
|
| 1597 |
+
"loss": 0.2181,
|
| 1598 |
+
"step": 26500
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 8.5,
|
| 1602 |
+
"learning_rate": 3.501011391461727e-05,
|
| 1603 |
+
"loss": 0.2208,
|
| 1604 |
+
"step": 26600
|
| 1605 |
+
},
|
| 1606 |
+
{
|
| 1607 |
+
"epoch": 8.53,
|
| 1608 |
+
"learning_rate": 3.4974626494907555e-05,
|
| 1609 |
+
"loss": 0.2192,
|
| 1610 |
+
"step": 26700
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 8.56,
|
| 1614 |
+
"learning_rate": 3.493913907519784e-05,
|
| 1615 |
+
"loss": 0.2227,
|
| 1616 |
+
"step": 26800
|
| 1617 |
+
},
|
| 1618 |
+
{
|
| 1619 |
+
"epoch": 8.59,
|
| 1620 |
+
"learning_rate": 3.490365165548813e-05,
|
| 1621 |
+
"loss": 0.2308,
|
| 1622 |
+
"step": 26900
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 8.62,
|
| 1626 |
+
"learning_rate": 3.486816423577842e-05,
|
| 1627 |
+
"loss": 0.22,
|
| 1628 |
+
"step": 27000
|
| 1629 |
+
},
|
| 1630 |
+
{
|
| 1631 |
+
"epoch": 8.66,
|
| 1632 |
+
"learning_rate": 3.4832676816068706e-05,
|
| 1633 |
+
"loss": 0.2247,
|
| 1634 |
+
"step": 27100
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 8.69,
|
| 1638 |
+
"learning_rate": 3.4797189396358994e-05,
|
| 1639 |
+
"loss": 0.2148,
|
| 1640 |
+
"step": 27200
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 8.72,
|
| 1644 |
+
"learning_rate": 3.476170197664928e-05,
|
| 1645 |
+
"loss": 0.2136,
|
| 1646 |
+
"step": 27300
|
| 1647 |
+
},
|
| 1648 |
+
{
|
| 1649 |
+
"epoch": 8.75,
|
| 1650 |
+
"learning_rate": 3.472621455693957e-05,
|
| 1651 |
+
"loss": 0.2282,
|
| 1652 |
+
"step": 27400
|
| 1653 |
+
},
|
| 1654 |
+
{
|
| 1655 |
+
"epoch": 8.78,
|
| 1656 |
+
"learning_rate": 3.469072713722986e-05,
|
| 1657 |
+
"loss": 0.2266,
|
| 1658 |
+
"step": 27500
|
| 1659 |
+
},
|
| 1660 |
+
{
|
| 1661 |
+
"epoch": 8.82,
|
| 1662 |
+
"learning_rate": 3.4655239717520146e-05,
|
| 1663 |
+
"loss": 0.2198,
|
| 1664 |
+
"step": 27600
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 8.85,
|
| 1668 |
+
"learning_rate": 3.461975229781043e-05,
|
| 1669 |
+
"loss": 0.2243,
|
| 1670 |
+
"step": 27700
|
| 1671 |
+
},
|
| 1672 |
+
{
|
| 1673 |
+
"epoch": 8.88,
|
| 1674 |
+
"learning_rate": 3.4584264878100715e-05,
|
| 1675 |
+
"loss": 0.2358,
|
| 1676 |
+
"step": 27800
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 8.91,
|
| 1680 |
+
"learning_rate": 3.4548777458391e-05,
|
| 1681 |
+
"loss": 0.2255,
|
| 1682 |
+
"step": 27900
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 8.94,
|
| 1686 |
+
"learning_rate": 3.451329003868129e-05,
|
| 1687 |
+
"loss": 0.2258,
|
| 1688 |
+
"step": 28000
|
| 1689 |
+
},
|
| 1690 |
+
{
|
| 1691 |
+
"epoch": 8.97,
|
| 1692 |
+
"learning_rate": 3.447780261897158e-05,
|
| 1693 |
+
"loss": 0.2228,
|
| 1694 |
+
"step": 28100
|
| 1695 |
+
},
|
| 1696 |
+
{
|
| 1697 |
+
"epoch": 9.01,
|
| 1698 |
+
"learning_rate": 3.4442315199261866e-05,
|
| 1699 |
+
"loss": 0.2183,
|
| 1700 |
+
"step": 28200
|
| 1701 |
+
},
|
| 1702 |
+
{
|
| 1703 |
+
"epoch": 9.04,
|
| 1704 |
+
"learning_rate": 3.440682777955215e-05,
|
| 1705 |
+
"loss": 0.1988,
|
| 1706 |
+
"step": 28300
|
| 1707 |
+
},
|
| 1708 |
+
{
|
| 1709 |
+
"epoch": 9.07,
|
| 1710 |
+
"learning_rate": 3.4371340359842436e-05,
|
| 1711 |
+
"loss": 0.188,
|
| 1712 |
+
"step": 28400
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"epoch": 9.1,
|
| 1716 |
+
"learning_rate": 3.4335852940132723e-05,
|
| 1717 |
+
"loss": 0.1963,
|
| 1718 |
+
"step": 28500
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 9.13,
|
| 1722 |
+
"learning_rate": 3.430036552042301e-05,
|
| 1723 |
+
"loss": 0.1919,
|
| 1724 |
+
"step": 28600
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 9.17,
|
| 1728 |
+
"learning_rate": 3.42648781007133e-05,
|
| 1729 |
+
"loss": 0.1906,
|
| 1730 |
+
"step": 28700
|
| 1731 |
+
},
|
| 1732 |
+
{
|
| 1733 |
+
"epoch": 9.2,
|
| 1734 |
+
"learning_rate": 3.422939068100359e-05,
|
| 1735 |
+
"loss": 0.1911,
|
| 1736 |
+
"step": 28800
|
| 1737 |
+
},
|
| 1738 |
+
{
|
| 1739 |
+
"epoch": 9.23,
|
| 1740 |
+
"learning_rate": 3.4193903261293875e-05,
|
| 1741 |
+
"loss": 0.1971,
|
| 1742 |
+
"step": 28900
|
| 1743 |
+
},
|
| 1744 |
+
{
|
| 1745 |
+
"epoch": 9.26,
|
| 1746 |
+
"learning_rate": 3.415841584158416e-05,
|
| 1747 |
+
"loss": 0.1931,
|
| 1748 |
+
"step": 29000
|
| 1749 |
+
},
|
| 1750 |
+
{
|
| 1751 |
+
"epoch": 9.29,
|
| 1752 |
+
"learning_rate": 3.412292842187445e-05,
|
| 1753 |
+
"loss": 0.1951,
|
| 1754 |
+
"step": 29100
|
| 1755 |
+
},
|
| 1756 |
+
{
|
| 1757 |
+
"epoch": 9.33,
|
| 1758 |
+
"learning_rate": 3.408744100216474e-05,
|
| 1759 |
+
"loss": 0.2008,
|
| 1760 |
+
"step": 29200
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 9.36,
|
| 1764 |
+
"learning_rate": 3.405195358245503e-05,
|
| 1765 |
+
"loss": 0.2053,
|
| 1766 |
+
"step": 29300
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 9.39,
|
| 1770 |
+
"learning_rate": 3.401646616274531e-05,
|
| 1771 |
+
"loss": 0.2015,
|
| 1772 |
+
"step": 29400
|
| 1773 |
+
},
|
| 1774 |
+
{
|
| 1775 |
+
"epoch": 9.42,
|
| 1776 |
+
"learning_rate": 3.3980978743035596e-05,
|
| 1777 |
+
"loss": 0.193,
|
| 1778 |
+
"step": 29500
|
| 1779 |
+
},
|
| 1780 |
+
{
|
| 1781 |
+
"epoch": 9.45,
|
| 1782 |
+
"learning_rate": 3.3945491323325884e-05,
|
| 1783 |
+
"loss": 0.1955,
|
| 1784 |
+
"step": 29600
|
| 1785 |
+
},
|
| 1786 |
+
{
|
| 1787 |
+
"epoch": 9.49,
|
| 1788 |
+
"learning_rate": 3.391000390361617e-05,
|
| 1789 |
+
"loss": 0.1993,
|
| 1790 |
+
"step": 29700
|
| 1791 |
+
},
|
| 1792 |
+
{
|
| 1793 |
+
"epoch": 9.52,
|
| 1794 |
+
"learning_rate": 3.387451648390646e-05,
|
| 1795 |
+
"loss": 0.2043,
|
| 1796 |
+
"step": 29800
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"epoch": 9.55,
|
| 1800 |
+
"learning_rate": 3.383902906419675e-05,
|
| 1801 |
+
"loss": 0.1955,
|
| 1802 |
+
"step": 29900
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 9.58,
|
| 1806 |
+
"learning_rate": 3.380354164448703e-05,
|
| 1807 |
+
"loss": 0.1961,
|
| 1808 |
+
"step": 30000
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 9.61,
|
| 1812 |
+
"learning_rate": 3.3768054224777316e-05,
|
| 1813 |
+
"loss": 0.1961,
|
| 1814 |
+
"step": 30100
|
| 1815 |
+
},
|
| 1816 |
+
{
|
| 1817 |
+
"epoch": 9.65,
|
| 1818 |
+
"learning_rate": 3.3732566805067604e-05,
|
| 1819 |
+
"loss": 0.196,
|
| 1820 |
+
"step": 30200
|
| 1821 |
+
},
|
| 1822 |
+
{
|
| 1823 |
+
"epoch": 9.68,
|
| 1824 |
+
"learning_rate": 3.369707938535789e-05,
|
| 1825 |
+
"loss": 0.2088,
|
| 1826 |
+
"step": 30300
|
| 1827 |
+
},
|
| 1828 |
+
{
|
| 1829 |
+
"epoch": 9.71,
|
| 1830 |
+
"learning_rate": 3.366159196564818e-05,
|
| 1831 |
+
"loss": 0.2031,
|
| 1832 |
+
"step": 30400
|
| 1833 |
+
},
|
| 1834 |
+
{
|
| 1835 |
+
"epoch": 9.74,
|
| 1836 |
+
"learning_rate": 3.362610454593847e-05,
|
| 1837 |
+
"loss": 0.1949,
|
| 1838 |
+
"step": 30500
|
| 1839 |
+
},
|
| 1840 |
+
{
|
| 1841 |
+
"epoch": 9.77,
|
| 1842 |
+
"learning_rate": 3.3590617126228756e-05,
|
| 1843 |
+
"loss": 0.1988,
|
| 1844 |
+
"step": 30600
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 9.81,
|
| 1848 |
+
"learning_rate": 3.3555129706519044e-05,
|
| 1849 |
+
"loss": 0.2061,
|
| 1850 |
+
"step": 30700
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 9.84,
|
| 1854 |
+
"learning_rate": 3.351964228680933e-05,
|
| 1855 |
+
"loss": 0.2022,
|
| 1856 |
+
"step": 30800
|
| 1857 |
+
},
|
| 1858 |
+
{
|
| 1859 |
+
"epoch": 9.87,
|
| 1860 |
+
"learning_rate": 3.348450974129671e-05,
|
| 1861 |
+
"loss": 0.1965,
|
| 1862 |
+
"step": 30900
|
| 1863 |
+
},
|
| 1864 |
+
{
|
| 1865 |
+
"epoch": 9.9,
|
| 1866 |
+
"learning_rate": 3.3449022321587e-05,
|
| 1867 |
+
"loss": 0.1999,
|
| 1868 |
+
"step": 31000
|
| 1869 |
+
},
|
| 1870 |
+
{
|
| 1871 |
+
"epoch": 9.93,
|
| 1872 |
+
"learning_rate": 3.341353490187729e-05,
|
| 1873 |
+
"loss": 0.2065,
|
| 1874 |
+
"step": 31100
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 9.96,
|
| 1878 |
+
"learning_rate": 3.3378047482167576e-05,
|
| 1879 |
+
"loss": 0.2031,
|
| 1880 |
+
"step": 31200
|
| 1881 |
+
},
|
| 1882 |
+
{
|
| 1883 |
+
"epoch": 10.0,
|
| 1884 |
+
"learning_rate": 3.3342560062457864e-05,
|
| 1885 |
+
"loss": 0.2078,
|
| 1886 |
+
"step": 31300
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 10.03,
|
| 1890 |
+
"learning_rate": 3.330707264274815e-05,
|
| 1891 |
+
"loss": 0.1733,
|
| 1892 |
+
"step": 31400
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 10.06,
|
| 1896 |
+
"learning_rate": 3.327158522303843e-05,
|
| 1897 |
+
"loss": 0.1654,
|
| 1898 |
+
"step": 31500
|
| 1899 |
+
},
|
| 1900 |
+
{
|
| 1901 |
+
"epoch": 10.09,
|
| 1902 |
+
"learning_rate": 3.323609780332872e-05,
|
| 1903 |
+
"loss": 0.1674,
|
| 1904 |
+
"step": 31600
|
| 1905 |
+
},
|
| 1906 |
+
{
|
| 1907 |
+
"epoch": 10.12,
|
| 1908 |
+
"learning_rate": 3.320061038361901e-05,
|
| 1909 |
+
"loss": 0.1725,
|
| 1910 |
+
"step": 31700
|
| 1911 |
+
},
|
| 1912 |
+
{
|
| 1913 |
+
"epoch": 10.16,
|
| 1914 |
+
"learning_rate": 3.3165122963909297e-05,
|
| 1915 |
+
"loss": 0.1761,
|
| 1916 |
+
"step": 31800
|
| 1917 |
+
},
|
| 1918 |
+
{
|
| 1919 |
+
"epoch": 10.19,
|
| 1920 |
+
"learning_rate": 3.3129635544199584e-05,
|
| 1921 |
+
"loss": 0.1757,
|
| 1922 |
+
"step": 31900
|
| 1923 |
+
},
|
| 1924 |
+
{
|
| 1925 |
+
"epoch": 10.22,
|
| 1926 |
+
"learning_rate": 3.309414812448987e-05,
|
| 1927 |
+
"loss": 0.1781,
|
| 1928 |
+
"step": 32000
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 10.25,
|
| 1932 |
+
"learning_rate": 3.3058660704780153e-05,
|
| 1933 |
+
"loss": 0.1716,
|
| 1934 |
+
"step": 32100
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 10.28,
|
| 1938 |
+
"learning_rate": 3.302317328507044e-05,
|
| 1939 |
+
"loss": 0.1717,
|
| 1940 |
+
"step": 32200
|
| 1941 |
+
},
|
| 1942 |
+
{
|
| 1943 |
+
"epoch": 10.32,
|
| 1944 |
+
"learning_rate": 3.298768586536073e-05,
|
| 1945 |
+
"loss": 0.179,
|
| 1946 |
+
"step": 32300
|
| 1947 |
+
},
|
| 1948 |
+
{
|
| 1949 |
+
"epoch": 10.35,
|
| 1950 |
+
"learning_rate": 3.295219844565102e-05,
|
| 1951 |
+
"loss": 0.174,
|
| 1952 |
+
"step": 32400
|
| 1953 |
+
},
|
| 1954 |
+
{
|
| 1955 |
+
"epoch": 10.38,
|
| 1956 |
+
"learning_rate": 3.2916711025941305e-05,
|
| 1957 |
+
"loss": 0.178,
|
| 1958 |
+
"step": 32500
|
| 1959 |
+
},
|
| 1960 |
+
{
|
| 1961 |
+
"epoch": 10.41,
|
| 1962 |
+
"learning_rate": 3.288122360623159e-05,
|
| 1963 |
+
"loss": 0.1765,
|
| 1964 |
+
"step": 32600
|
| 1965 |
+
},
|
| 1966 |
+
{
|
| 1967 |
+
"epoch": 10.44,
|
| 1968 |
+
"learning_rate": 3.284573618652188e-05,
|
| 1969 |
+
"loss": 0.1824,
|
| 1970 |
+
"step": 32700
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"epoch": 10.48,
|
| 1974 |
+
"learning_rate": 3.281024876681217e-05,
|
| 1975 |
+
"loss": 0.1806,
|
| 1976 |
+
"step": 32800
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 10.51,
|
| 1980 |
+
"learning_rate": 3.277511622129955e-05,
|
| 1981 |
+
"loss": 0.1785,
|
| 1982 |
+
"step": 32900
|
| 1983 |
+
},
|
| 1984 |
+
{
|
| 1985 |
+
"epoch": 10.54,
|
| 1986 |
+
"learning_rate": 3.2739983675786936e-05,
|
| 1987 |
+
"loss": 0.1698,
|
| 1988 |
+
"step": 33000
|
| 1989 |
+
},
|
| 1990 |
+
{
|
| 1991 |
+
"epoch": 10.57,
|
| 1992 |
+
"learning_rate": 3.2704496256077224e-05,
|
| 1993 |
+
"loss": 0.1734,
|
| 1994 |
+
"step": 33100
|
| 1995 |
+
},
|
| 1996 |
+
{
|
| 1997 |
+
"epoch": 10.6,
|
| 1998 |
+
"learning_rate": 3.2669008836367505e-05,
|
| 1999 |
+
"loss": 0.1761,
|
| 2000 |
+
"step": 33200
|
| 2001 |
+
},
|
| 2002 |
+
{
|
| 2003 |
+
"epoch": 10.64,
|
| 2004 |
+
"learning_rate": 3.263352141665779e-05,
|
| 2005 |
+
"loss": 0.1786,
|
| 2006 |
+
"step": 33300
|
| 2007 |
+
},
|
| 2008 |
+
{
|
| 2009 |
+
"epoch": 10.67,
|
| 2010 |
+
"learning_rate": 3.259803399694808e-05,
|
| 2011 |
+
"loss": 0.1767,
|
| 2012 |
+
"step": 33400
|
| 2013 |
+
},
|
| 2014 |
+
{
|
| 2015 |
+
"epoch": 10.7,
|
| 2016 |
+
"learning_rate": 3.256254657723837e-05,
|
| 2017 |
+
"loss": 0.1881,
|
| 2018 |
+
"step": 33500
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 10.73,
|
| 2022 |
+
"learning_rate": 3.252705915752866e-05,
|
| 2023 |
+
"loss": 0.1792,
|
| 2024 |
+
"step": 33600
|
| 2025 |
+
},
|
| 2026 |
+
{
|
| 2027 |
+
"epoch": 10.76,
|
| 2028 |
+
"learning_rate": 3.2491571737818945e-05,
|
| 2029 |
+
"loss": 0.1832,
|
| 2030 |
+
"step": 33700
|
| 2031 |
+
},
|
| 2032 |
+
{
|
| 2033 |
+
"epoch": 10.8,
|
| 2034 |
+
"learning_rate": 3.245608431810923e-05,
|
| 2035 |
+
"loss": 0.1889,
|
| 2036 |
+
"step": 33800
|
| 2037 |
+
},
|
| 2038 |
+
{
|
| 2039 |
+
"epoch": 10.83,
|
| 2040 |
+
"learning_rate": 3.242059689839952e-05,
|
| 2041 |
+
"loss": 0.1824,
|
| 2042 |
+
"step": 33900
|
| 2043 |
+
},
|
| 2044 |
+
{
|
| 2045 |
+
"epoch": 10.86,
|
| 2046 |
+
"learning_rate": 3.238510947868981e-05,
|
| 2047 |
+
"loss": 0.1746,
|
| 2048 |
+
"step": 34000
|
| 2049 |
+
},
|
| 2050 |
+
{
|
| 2051 |
+
"epoch": 10.89,
|
| 2052 |
+
"learning_rate": 3.2349622058980096e-05,
|
| 2053 |
+
"loss": 0.1735,
|
| 2054 |
+
"step": 34100
|
| 2055 |
+
},
|
| 2056 |
+
{
|
| 2057 |
+
"epoch": 10.92,
|
| 2058 |
+
"learning_rate": 3.2314134639270384e-05,
|
| 2059 |
+
"loss": 0.1822,
|
| 2060 |
+
"step": 34200
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 10.95,
|
| 2064 |
+
"learning_rate": 3.227864721956067e-05,
|
| 2065 |
+
"loss": 0.1768,
|
| 2066 |
+
"step": 34300
|
| 2067 |
+
},
|
| 2068 |
+
{
|
| 2069 |
+
"epoch": 10.99,
|
| 2070 |
+
"learning_rate": 3.224315979985095e-05,
|
| 2071 |
+
"loss": 0.1812,
|
| 2072 |
+
"step": 34400
|
| 2073 |
+
},
|
| 2074 |
+
{
|
| 2075 |
+
"epoch": 11.02,
|
| 2076 |
+
"learning_rate": 3.220767238014124e-05,
|
| 2077 |
+
"loss": 0.1645,
|
| 2078 |
+
"step": 34500
|
| 2079 |
+
},
|
| 2080 |
+
{
|
| 2081 |
+
"epoch": 11.05,
|
| 2082 |
+
"learning_rate": 3.217218496043153e-05,
|
| 2083 |
+
"loss": 0.1483,
|
| 2084 |
+
"step": 34600
|
| 2085 |
+
},
|
| 2086 |
+
{
|
| 2087 |
+
"epoch": 11.08,
|
| 2088 |
+
"learning_rate": 3.213669754072182e-05,
|
| 2089 |
+
"loss": 0.1615,
|
| 2090 |
+
"step": 34700
|
| 2091 |
+
},
|
| 2092 |
+
{
|
| 2093 |
+
"epoch": 11.11,
|
| 2094 |
+
"learning_rate": 3.2101210121012105e-05,
|
| 2095 |
+
"loss": 0.1637,
|
| 2096 |
+
"step": 34800
|
| 2097 |
+
},
|
| 2098 |
+
{
|
| 2099 |
+
"epoch": 11.15,
|
| 2100 |
+
"learning_rate": 3.206572270130239e-05,
|
| 2101 |
+
"loss": 0.1497,
|
| 2102 |
+
"step": 34900
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 11.18,
|
| 2106 |
+
"learning_rate": 3.2030235281592674e-05,
|
| 2107 |
+
"loss": 0.1524,
|
| 2108 |
+
"step": 35000
|
| 2109 |
+
}
|
| 2110 |
+
],
|
| 2111 |
+
"max_steps": 125240,
|
| 2112 |
+
"num_train_epochs": 40,
|
| 2113 |
+
"total_flos": 0.0,
|
| 2114 |
+
"trial_name": null,
|
| 2115 |
+
"trial_params": null
|
| 2116 |
+
}
|
checkpoint-40000/config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "Davlan/afro-xlmr-base",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"XLMRobertaModel"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 768,
|
| 13 |
+
"id2label": {
|
| 14 |
+
"0": "LABEL_0"
|
| 15 |
+
},
|
| 16 |
+
"initializer_range": 0.02,
|
| 17 |
+
"intermediate_size": 3072,
|
| 18 |
+
"label2id": {
|
| 19 |
+
"LABEL_0": 0
|
| 20 |
+
},
|
| 21 |
+
"layer_norm_eps": 1e-05,
|
| 22 |
+
"max_position_embeddings": 514,
|
| 23 |
+
"model_type": "xlm-roberta",
|
| 24 |
+
"num_attention_heads": 12,
|
| 25 |
+
"num_hidden_layers": 12,
|
| 26 |
+
"output_past": true,
|
| 27 |
+
"pad_token_id": 1,
|
| 28 |
+
"position_embedding_type": "absolute",
|
| 29 |
+
"torch_dtype": "float32",
|
| 30 |
+
"transformers_version": "4.29.0",
|
| 31 |
+
"type_vocab_size": 1,
|
| 32 |
+
"use_cache": true,
|
| 33 |
+
"vocab_size": 250002
|
| 34 |
+
}
|
checkpoint-40000/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0773eef0ea94d82d271457982c0e6ed28174dd7b790cd8ab7d2e9514eb4860e5
|
| 3 |
+
size 1112246250
|
checkpoint-40000/trainer_state.json
ADDED
|
@@ -0,0 +1,2416 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 12.775471095496647,
|
| 5 |
+
"global_step": 40000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.03,
|
| 12 |
+
"learning_rate": 2.8744809964867454e-07,
|
| 13 |
+
"loss": 91.176,
|
| 14 |
+
"step": 100
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.06,
|
| 18 |
+
"learning_rate": 6.068348770360907e-07,
|
| 19 |
+
"loss": 64.0684,
|
| 20 |
+
"step": 200
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.1,
|
| 24 |
+
"learning_rate": 9.26221654423507e-07,
|
| 25 |
+
"loss": 27.1061,
|
| 26 |
+
"step": 300
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.13,
|
| 30 |
+
"learning_rate": 1.242414564037049e-06,
|
| 31 |
+
"loss": 8.0198,
|
| 32 |
+
"step": 400
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.16,
|
| 36 |
+
"learning_rate": 1.561801341424465e-06,
|
| 37 |
+
"loss": 4.7691,
|
| 38 |
+
"step": 500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.19,
|
| 42 |
+
"learning_rate": 1.8811881188118813e-06,
|
| 43 |
+
"loss": 3.4387,
|
| 44 |
+
"step": 600
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.22,
|
| 48 |
+
"learning_rate": 2.2005748961992977e-06,
|
| 49 |
+
"loss": 2.7027,
|
| 50 |
+
"step": 700
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.26,
|
| 54 |
+
"learning_rate": 2.5199616735867137e-06,
|
| 55 |
+
"loss": 2.4431,
|
| 56 |
+
"step": 800
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.29,
|
| 60 |
+
"learning_rate": 2.8393484509741297e-06,
|
| 61 |
+
"loss": 2.0534,
|
| 62 |
+
"step": 900
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.32,
|
| 66 |
+
"learning_rate": 3.1587352283615465e-06,
|
| 67 |
+
"loss": 1.8537,
|
| 68 |
+
"step": 1000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.35,
|
| 72 |
+
"learning_rate": 3.4781220057489625e-06,
|
| 73 |
+
"loss": 1.642,
|
| 74 |
+
"step": 1100
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.38,
|
| 78 |
+
"learning_rate": 3.797508783136378e-06,
|
| 79 |
+
"loss": 1.5092,
|
| 80 |
+
"step": 1200
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.42,
|
| 84 |
+
"learning_rate": 4.1168955605237945e-06,
|
| 85 |
+
"loss": 1.4302,
|
| 86 |
+
"step": 1300
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.45,
|
| 90 |
+
"learning_rate": 4.436282337911211e-06,
|
| 91 |
+
"loss": 1.3527,
|
| 92 |
+
"step": 1400
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.48,
|
| 96 |
+
"learning_rate": 4.755669115298627e-06,
|
| 97 |
+
"loss": 1.2495,
|
| 98 |
+
"step": 1500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.51,
|
| 102 |
+
"learning_rate": 5.075055892686043e-06,
|
| 103 |
+
"loss": 1.2039,
|
| 104 |
+
"step": 1600
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.54,
|
| 108 |
+
"learning_rate": 5.39444267007346e-06,
|
| 109 |
+
"loss": 1.1438,
|
| 110 |
+
"step": 1700
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.57,
|
| 114 |
+
"learning_rate": 5.713829447460876e-06,
|
| 115 |
+
"loss": 1.1056,
|
| 116 |
+
"step": 1800
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.61,
|
| 120 |
+
"learning_rate": 6.033216224848291e-06,
|
| 121 |
+
"loss": 1.0436,
|
| 122 |
+
"step": 1900
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.64,
|
| 126 |
+
"learning_rate": 6.352603002235708e-06,
|
| 127 |
+
"loss": 1.0225,
|
| 128 |
+
"step": 2000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.67,
|
| 132 |
+
"learning_rate": 6.671989779623123e-06,
|
| 133 |
+
"loss": 0.9856,
|
| 134 |
+
"step": 2100
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.7,
|
| 138 |
+
"learning_rate": 6.9913765570105406e-06,
|
| 139 |
+
"loss": 0.9412,
|
| 140 |
+
"step": 2200
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.73,
|
| 144 |
+
"learning_rate": 7.310763334397957e-06,
|
| 145 |
+
"loss": 0.9075,
|
| 146 |
+
"step": 2300
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.77,
|
| 150 |
+
"learning_rate": 7.630150111785373e-06,
|
| 151 |
+
"loss": 0.8752,
|
| 152 |
+
"step": 2400
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.8,
|
| 156 |
+
"learning_rate": 7.949536889172788e-06,
|
| 157 |
+
"loss": 0.854,
|
| 158 |
+
"step": 2500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.83,
|
| 162 |
+
"learning_rate": 8.268923666560205e-06,
|
| 163 |
+
"loss": 0.8282,
|
| 164 |
+
"step": 2600
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.86,
|
| 168 |
+
"learning_rate": 8.588310443947621e-06,
|
| 169 |
+
"loss": 0.8074,
|
| 170 |
+
"step": 2700
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.89,
|
| 174 |
+
"learning_rate": 8.907697221335038e-06,
|
| 175 |
+
"loss": 0.7976,
|
| 176 |
+
"step": 2800
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.93,
|
| 180 |
+
"learning_rate": 9.227083998722454e-06,
|
| 181 |
+
"loss": 0.7955,
|
| 182 |
+
"step": 2900
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.96,
|
| 186 |
+
"learning_rate": 9.54647077610987e-06,
|
| 187 |
+
"loss": 0.7502,
|
| 188 |
+
"step": 3000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.99,
|
| 192 |
+
"learning_rate": 9.865857553497285e-06,
|
| 193 |
+
"loss": 0.7455,
|
| 194 |
+
"step": 3100
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 1.02,
|
| 198 |
+
"learning_rate": 1.0185244330884702e-05,
|
| 199 |
+
"loss": 0.7177,
|
| 200 |
+
"step": 3200
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 1.05,
|
| 204 |
+
"learning_rate": 1.050463110827212e-05,
|
| 205 |
+
"loss": 0.724,
|
| 206 |
+
"step": 3300
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 1.09,
|
| 210 |
+
"learning_rate": 1.0824017885659535e-05,
|
| 211 |
+
"loss": 0.6975,
|
| 212 |
+
"step": 3400
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 1.12,
|
| 216 |
+
"learning_rate": 1.114340466304695e-05,
|
| 217 |
+
"loss": 0.7054,
|
| 218 |
+
"step": 3500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 1.15,
|
| 222 |
+
"learning_rate": 1.1462791440434366e-05,
|
| 223 |
+
"loss": 0.71,
|
| 224 |
+
"step": 3600
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 1.18,
|
| 228 |
+
"learning_rate": 1.1782178217821782e-05,
|
| 229 |
+
"loss": 0.6825,
|
| 230 |
+
"step": 3700
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 1.21,
|
| 234 |
+
"learning_rate": 1.2101564995209199e-05,
|
| 235 |
+
"loss": 0.6792,
|
| 236 |
+
"step": 3800
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 1.25,
|
| 240 |
+
"learning_rate": 1.2420951772596615e-05,
|
| 241 |
+
"loss": 0.6794,
|
| 242 |
+
"step": 3900
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 1.28,
|
| 246 |
+
"learning_rate": 1.2740338549984032e-05,
|
| 247 |
+
"loss": 0.6672,
|
| 248 |
+
"step": 4000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 1.31,
|
| 252 |
+
"learning_rate": 1.3059725327371449e-05,
|
| 253 |
+
"loss": 0.6356,
|
| 254 |
+
"step": 4100
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 1.34,
|
| 258 |
+
"learning_rate": 1.3379112104758865e-05,
|
| 259 |
+
"loss": 0.6503,
|
| 260 |
+
"step": 4200
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 1.37,
|
| 264 |
+
"learning_rate": 1.369849888214628e-05,
|
| 265 |
+
"loss": 0.6495,
|
| 266 |
+
"step": 4300
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 1.41,
|
| 270 |
+
"learning_rate": 1.4017885659533696e-05,
|
| 271 |
+
"loss": 0.6428,
|
| 272 |
+
"step": 4400
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 1.44,
|
| 276 |
+
"learning_rate": 1.4337272436921113e-05,
|
| 277 |
+
"loss": 0.6171,
|
| 278 |
+
"step": 4500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 1.47,
|
| 282 |
+
"learning_rate": 1.4656659214308529e-05,
|
| 283 |
+
"loss": 0.6084,
|
| 284 |
+
"step": 4600
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 1.5,
|
| 288 |
+
"learning_rate": 1.4976045991695944e-05,
|
| 289 |
+
"loss": 0.6222,
|
| 290 |
+
"step": 4700
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 1.53,
|
| 294 |
+
"learning_rate": 1.529543276908336e-05,
|
| 295 |
+
"loss": 0.5971,
|
| 296 |
+
"step": 4800
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 1.56,
|
| 300 |
+
"learning_rate": 1.5614819546470777e-05,
|
| 301 |
+
"loss": 0.5745,
|
| 302 |
+
"step": 4900
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 1.6,
|
| 306 |
+
"learning_rate": 1.5934206323858196e-05,
|
| 307 |
+
"loss": 0.5753,
|
| 308 |
+
"step": 5000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 1.63,
|
| 312 |
+
"learning_rate": 1.625359310124561e-05,
|
| 313 |
+
"loss": 0.6045,
|
| 314 |
+
"step": 5100
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 1.66,
|
| 318 |
+
"learning_rate": 1.6572979878633027e-05,
|
| 319 |
+
"loss": 0.6006,
|
| 320 |
+
"step": 5200
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 1.69,
|
| 324 |
+
"learning_rate": 1.6892366656020443e-05,
|
| 325 |
+
"loss": 0.6003,
|
| 326 |
+
"step": 5300
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 1.72,
|
| 330 |
+
"learning_rate": 1.7211753433407858e-05,
|
| 331 |
+
"loss": 0.5688,
|
| 332 |
+
"step": 5400
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 1.76,
|
| 336 |
+
"learning_rate": 1.7531140210795274e-05,
|
| 337 |
+
"loss": 0.5662,
|
| 338 |
+
"step": 5500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 1.79,
|
| 342 |
+
"learning_rate": 1.785052698818269e-05,
|
| 343 |
+
"loss": 0.5788,
|
| 344 |
+
"step": 5600
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 1.82,
|
| 348 |
+
"learning_rate": 1.8169913765570105e-05,
|
| 349 |
+
"loss": 0.5541,
|
| 350 |
+
"step": 5700
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 1.85,
|
| 354 |
+
"learning_rate": 1.8489300542957524e-05,
|
| 355 |
+
"loss": 0.5656,
|
| 356 |
+
"step": 5800
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 1.88,
|
| 360 |
+
"learning_rate": 1.880868732034494e-05,
|
| 361 |
+
"loss": 0.5466,
|
| 362 |
+
"step": 5900
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 1.92,
|
| 366 |
+
"learning_rate": 1.9128074097732355e-05,
|
| 367 |
+
"loss": 0.563,
|
| 368 |
+
"step": 6000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 1.95,
|
| 372 |
+
"learning_rate": 1.9447460875119774e-05,
|
| 373 |
+
"loss": 0.5603,
|
| 374 |
+
"step": 6100
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 1.98,
|
| 378 |
+
"learning_rate": 1.976684765250719e-05,
|
| 379 |
+
"loss": 0.5579,
|
| 380 |
+
"step": 6200
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 2.01,
|
| 384 |
+
"learning_rate": 2.0086234429894605e-05,
|
| 385 |
+
"loss": 0.5346,
|
| 386 |
+
"step": 6300
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 2.04,
|
| 390 |
+
"learning_rate": 2.0405621207282017e-05,
|
| 391 |
+
"loss": 0.5038,
|
| 392 |
+
"step": 6400
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 2.08,
|
| 396 |
+
"learning_rate": 2.0725007984669436e-05,
|
| 397 |
+
"loss": 0.5121,
|
| 398 |
+
"step": 6500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 2.11,
|
| 402 |
+
"learning_rate": 2.1044394762056855e-05,
|
| 403 |
+
"loss": 0.5135,
|
| 404 |
+
"step": 6600
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 2.14,
|
| 408 |
+
"learning_rate": 2.1363781539444267e-05,
|
| 409 |
+
"loss": 0.504,
|
| 410 |
+
"step": 6700
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 2.17,
|
| 414 |
+
"learning_rate": 2.1683168316831686e-05,
|
| 415 |
+
"loss": 0.5119,
|
| 416 |
+
"step": 6800
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 2.2,
|
| 420 |
+
"learning_rate": 2.20025550942191e-05,
|
| 421 |
+
"loss": 0.4986,
|
| 422 |
+
"step": 6900
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 2.24,
|
| 426 |
+
"learning_rate": 2.2321941871606518e-05,
|
| 427 |
+
"loss": 0.504,
|
| 428 |
+
"step": 7000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 2.27,
|
| 432 |
+
"learning_rate": 2.2641328648993933e-05,
|
| 433 |
+
"loss": 0.5024,
|
| 434 |
+
"step": 7100
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 2.3,
|
| 438 |
+
"learning_rate": 2.296071542638135e-05,
|
| 439 |
+
"loss": 0.479,
|
| 440 |
+
"step": 7200
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 2.33,
|
| 444 |
+
"learning_rate": 2.3280102203768764e-05,
|
| 445 |
+
"loss": 0.492,
|
| 446 |
+
"step": 7300
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 2.36,
|
| 450 |
+
"learning_rate": 2.3599488981156183e-05,
|
| 451 |
+
"loss": 0.4928,
|
| 452 |
+
"step": 7400
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 2.4,
|
| 456 |
+
"learning_rate": 2.39188757585436e-05,
|
| 457 |
+
"loss": 0.4862,
|
| 458 |
+
"step": 7500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 2.43,
|
| 462 |
+
"learning_rate": 2.4238262535931014e-05,
|
| 463 |
+
"loss": 0.5148,
|
| 464 |
+
"step": 7600
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 2.46,
|
| 468 |
+
"learning_rate": 2.4557649313318433e-05,
|
| 469 |
+
"loss": 0.4922,
|
| 470 |
+
"step": 7700
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 2.49,
|
| 474 |
+
"learning_rate": 2.4877036090705846e-05,
|
| 475 |
+
"loss": 0.4882,
|
| 476 |
+
"step": 7800
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 2.52,
|
| 480 |
+
"learning_rate": 2.5196422868093264e-05,
|
| 481 |
+
"loss": 0.4871,
|
| 482 |
+
"step": 7900
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 2.56,
|
| 486 |
+
"learning_rate": 2.5515809645480677e-05,
|
| 487 |
+
"loss": 0.4967,
|
| 488 |
+
"step": 8000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 2.59,
|
| 492 |
+
"learning_rate": 2.5835196422868096e-05,
|
| 493 |
+
"loss": 0.4834,
|
| 494 |
+
"step": 8100
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 2.62,
|
| 498 |
+
"learning_rate": 2.615458320025551e-05,
|
| 499 |
+
"loss": 0.4769,
|
| 500 |
+
"step": 8200
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 2.65,
|
| 504 |
+
"learning_rate": 2.6473969977642927e-05,
|
| 505 |
+
"loss": 0.4776,
|
| 506 |
+
"step": 8300
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 2.68,
|
| 510 |
+
"learning_rate": 2.6793356755030342e-05,
|
| 511 |
+
"loss": 0.4861,
|
| 512 |
+
"step": 8400
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 2.71,
|
| 516 |
+
"learning_rate": 2.711274353241776e-05,
|
| 517 |
+
"loss": 0.486,
|
| 518 |
+
"step": 8500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 2.75,
|
| 522 |
+
"learning_rate": 2.7432130309805177e-05,
|
| 523 |
+
"loss": 0.4745,
|
| 524 |
+
"step": 8600
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 2.78,
|
| 528 |
+
"learning_rate": 2.7751517087192592e-05,
|
| 529 |
+
"loss": 0.4805,
|
| 530 |
+
"step": 8700
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 2.81,
|
| 534 |
+
"learning_rate": 2.807090386458001e-05,
|
| 535 |
+
"loss": 0.5003,
|
| 536 |
+
"step": 8800
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 2.84,
|
| 540 |
+
"learning_rate": 2.8390290641967424e-05,
|
| 541 |
+
"loss": 0.4639,
|
| 542 |
+
"step": 8900
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 2.87,
|
| 546 |
+
"learning_rate": 2.8709677419354843e-05,
|
| 547 |
+
"loss": 0.4763,
|
| 548 |
+
"step": 9000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 2.91,
|
| 552 |
+
"learning_rate": 2.9029064196742255e-05,
|
| 553 |
+
"loss": 0.4727,
|
| 554 |
+
"step": 9100
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 2.94,
|
| 558 |
+
"learning_rate": 2.9348450974129674e-05,
|
| 559 |
+
"loss": 0.4839,
|
| 560 |
+
"step": 9200
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 2.97,
|
| 564 |
+
"learning_rate": 2.966783775151709e-05,
|
| 565 |
+
"loss": 0.4769,
|
| 566 |
+
"step": 9300
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 3.0,
|
| 570 |
+
"learning_rate": 2.9987224528904505e-05,
|
| 571 |
+
"loss": 0.4641,
|
| 572 |
+
"step": 9400
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 3.03,
|
| 576 |
+
"learning_rate": 3.030661130629192e-05,
|
| 577 |
+
"loss": 0.4282,
|
| 578 |
+
"step": 9500
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 3.07,
|
| 582 |
+
"learning_rate": 3.0625998083679336e-05,
|
| 583 |
+
"loss": 0.4057,
|
| 584 |
+
"step": 9600
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 3.1,
|
| 588 |
+
"learning_rate": 3.094538486106676e-05,
|
| 589 |
+
"loss": 0.4192,
|
| 590 |
+
"step": 9700
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 3.13,
|
| 594 |
+
"learning_rate": 3.126477163845417e-05,
|
| 595 |
+
"loss": 0.4186,
|
| 596 |
+
"step": 9800
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 3.16,
|
| 600 |
+
"learning_rate": 3.158415841584159e-05,
|
| 601 |
+
"loss": 0.4086,
|
| 602 |
+
"step": 9900
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 3.19,
|
| 606 |
+
"learning_rate": 3.1903545193229005e-05,
|
| 607 |
+
"loss": 0.4194,
|
| 608 |
+
"step": 10000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 3.23,
|
| 612 |
+
"learning_rate": 3.221973810284254e-05,
|
| 613 |
+
"loss": 0.4443,
|
| 614 |
+
"step": 10100
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 3.26,
|
| 618 |
+
"learning_rate": 3.2539124880229965e-05,
|
| 619 |
+
"loss": 0.4217,
|
| 620 |
+
"step": 10200
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 3.29,
|
| 624 |
+
"learning_rate": 3.2858511657617374e-05,
|
| 625 |
+
"loss": 0.4275,
|
| 626 |
+
"step": 10300
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 3.32,
|
| 630 |
+
"learning_rate": 3.3177898435004796e-05,
|
| 631 |
+
"loss": 0.4422,
|
| 632 |
+
"step": 10400
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 3.35,
|
| 636 |
+
"learning_rate": 3.3497285212392205e-05,
|
| 637 |
+
"loss": 0.4227,
|
| 638 |
+
"step": 10500
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 3.39,
|
| 642 |
+
"learning_rate": 3.381667198977963e-05,
|
| 643 |
+
"loss": 0.4191,
|
| 644 |
+
"step": 10600
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 3.42,
|
| 648 |
+
"learning_rate": 3.413605876716704e-05,
|
| 649 |
+
"loss": 0.4309,
|
| 650 |
+
"step": 10700
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 3.45,
|
| 654 |
+
"learning_rate": 3.445544554455446e-05,
|
| 655 |
+
"loss": 0.4079,
|
| 656 |
+
"step": 10800
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 3.48,
|
| 660 |
+
"learning_rate": 3.4774832321941874e-05,
|
| 661 |
+
"loss": 0.4281,
|
| 662 |
+
"step": 10900
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 3.51,
|
| 666 |
+
"learning_rate": 3.509421909932929e-05,
|
| 667 |
+
"loss": 0.4137,
|
| 668 |
+
"step": 11000
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 3.55,
|
| 672 |
+
"learning_rate": 3.5413605876716705e-05,
|
| 673 |
+
"loss": 0.4168,
|
| 674 |
+
"step": 11100
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 3.58,
|
| 678 |
+
"learning_rate": 3.573299265410412e-05,
|
| 679 |
+
"loss": 0.4186,
|
| 680 |
+
"step": 11200
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 3.61,
|
| 684 |
+
"learning_rate": 3.605237943149154e-05,
|
| 685 |
+
"loss": 0.4325,
|
| 686 |
+
"step": 11300
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 3.64,
|
| 690 |
+
"learning_rate": 3.637176620887895e-05,
|
| 691 |
+
"loss": 0.4422,
|
| 692 |
+
"step": 11400
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 3.67,
|
| 696 |
+
"learning_rate": 3.6691152986266374e-05,
|
| 697 |
+
"loss": 0.4269,
|
| 698 |
+
"step": 11500
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 3.7,
|
| 702 |
+
"learning_rate": 3.701053976365379e-05,
|
| 703 |
+
"loss": 0.4368,
|
| 704 |
+
"step": 11600
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 3.74,
|
| 708 |
+
"learning_rate": 3.7329926541041205e-05,
|
| 709 |
+
"loss": 0.4207,
|
| 710 |
+
"step": 11700
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 3.77,
|
| 714 |
+
"learning_rate": 3.764931331842862e-05,
|
| 715 |
+
"loss": 0.4365,
|
| 716 |
+
"step": 11800
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 3.8,
|
| 720 |
+
"learning_rate": 3.7968700095816036e-05,
|
| 721 |
+
"loss": 0.4394,
|
| 722 |
+
"step": 11900
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 3.83,
|
| 726 |
+
"learning_rate": 3.828808687320345e-05,
|
| 727 |
+
"loss": 0.4351,
|
| 728 |
+
"step": 12000
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 3.86,
|
| 732 |
+
"learning_rate": 3.860747365059087e-05,
|
| 733 |
+
"loss": 0.4113,
|
| 734 |
+
"step": 12100
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 3.9,
|
| 738 |
+
"learning_rate": 3.892686042797828e-05,
|
| 739 |
+
"loss": 0.4261,
|
| 740 |
+
"step": 12200
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 3.93,
|
| 744 |
+
"learning_rate": 3.92462472053657e-05,
|
| 745 |
+
"loss": 0.4311,
|
| 746 |
+
"step": 12300
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 3.96,
|
| 750 |
+
"learning_rate": 3.956563398275312e-05,
|
| 751 |
+
"loss": 0.4246,
|
| 752 |
+
"step": 12400
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 3.99,
|
| 756 |
+
"learning_rate": 3.988502076014053e-05,
|
| 757 |
+
"loss": 0.4294,
|
| 758 |
+
"step": 12500
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 4.02,
|
| 762 |
+
"learning_rate": 3.997728805138579e-05,
|
| 763 |
+
"loss": 0.3854,
|
| 764 |
+
"step": 12600
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 4.06,
|
| 768 |
+
"learning_rate": 3.994180063167608e-05,
|
| 769 |
+
"loss": 0.3656,
|
| 770 |
+
"step": 12700
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 4.09,
|
| 774 |
+
"learning_rate": 3.9906313211966366e-05,
|
| 775 |
+
"loss": 0.3836,
|
| 776 |
+
"step": 12800
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 4.12,
|
| 780 |
+
"learning_rate": 3.987082579225665e-05,
|
| 781 |
+
"loss": 0.3664,
|
| 782 |
+
"step": 12900
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 4.15,
|
| 786 |
+
"learning_rate": 3.9835338372546935e-05,
|
| 787 |
+
"loss": 0.3884,
|
| 788 |
+
"step": 13000
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 4.18,
|
| 792 |
+
"learning_rate": 3.979985095283722e-05,
|
| 793 |
+
"loss": 0.3764,
|
| 794 |
+
"step": 13100
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 4.22,
|
| 798 |
+
"learning_rate": 3.976436353312751e-05,
|
| 799 |
+
"loss": 0.3733,
|
| 800 |
+
"step": 13200
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 4.25,
|
| 804 |
+
"learning_rate": 3.97288761134178e-05,
|
| 805 |
+
"loss": 0.3682,
|
| 806 |
+
"step": 13300
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 4.28,
|
| 810 |
+
"learning_rate": 3.969338869370809e-05,
|
| 811 |
+
"loss": 0.371,
|
| 812 |
+
"step": 13400
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 4.31,
|
| 816 |
+
"learning_rate": 3.965790127399837e-05,
|
| 817 |
+
"loss": 0.3741,
|
| 818 |
+
"step": 13500
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 4.34,
|
| 822 |
+
"learning_rate": 3.9622413854288656e-05,
|
| 823 |
+
"loss": 0.3757,
|
| 824 |
+
"step": 13600
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 4.38,
|
| 828 |
+
"learning_rate": 3.9586926434578944e-05,
|
| 829 |
+
"loss": 0.3652,
|
| 830 |
+
"step": 13700
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 4.41,
|
| 834 |
+
"learning_rate": 3.955143901486923e-05,
|
| 835 |
+
"loss": 0.3892,
|
| 836 |
+
"step": 13800
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 4.44,
|
| 840 |
+
"learning_rate": 3.951630646935662e-05,
|
| 841 |
+
"loss": 0.3893,
|
| 842 |
+
"step": 13900
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 4.47,
|
| 846 |
+
"learning_rate": 3.948081904964691e-05,
|
| 847 |
+
"loss": 0.3823,
|
| 848 |
+
"step": 14000
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 4.5,
|
| 852 |
+
"learning_rate": 3.944533162993719e-05,
|
| 853 |
+
"loss": 0.3654,
|
| 854 |
+
"step": 14100
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 4.54,
|
| 858 |
+
"learning_rate": 3.9409844210227476e-05,
|
| 859 |
+
"loss": 0.3791,
|
| 860 |
+
"step": 14200
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 4.57,
|
| 864 |
+
"learning_rate": 3.9374356790517764e-05,
|
| 865 |
+
"loss": 0.3918,
|
| 866 |
+
"step": 14300
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 4.6,
|
| 870 |
+
"learning_rate": 3.933886937080805e-05,
|
| 871 |
+
"loss": 0.3713,
|
| 872 |
+
"step": 14400
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 4.63,
|
| 876 |
+
"learning_rate": 3.930338195109834e-05,
|
| 877 |
+
"loss": 0.376,
|
| 878 |
+
"step": 14500
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 4.66,
|
| 882 |
+
"learning_rate": 3.926789453138863e-05,
|
| 883 |
+
"loss": 0.3682,
|
| 884 |
+
"step": 14600
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 4.69,
|
| 888 |
+
"learning_rate": 3.9232407111678915e-05,
|
| 889 |
+
"loss": 0.374,
|
| 890 |
+
"step": 14700
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 4.73,
|
| 894 |
+
"learning_rate": 3.91969196919692e-05,
|
| 895 |
+
"loss": 0.3774,
|
| 896 |
+
"step": 14800
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 4.76,
|
| 900 |
+
"learning_rate": 3.9161432272259484e-05,
|
| 901 |
+
"loss": 0.3806,
|
| 902 |
+
"step": 14900
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 4.79,
|
| 906 |
+
"learning_rate": 3.912594485254977e-05,
|
| 907 |
+
"loss": 0.3886,
|
| 908 |
+
"step": 15000
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 4.82,
|
| 912 |
+
"learning_rate": 3.909045743284006e-05,
|
| 913 |
+
"loss": 0.3742,
|
| 914 |
+
"step": 15100
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 4.85,
|
| 918 |
+
"learning_rate": 3.905497001313035e-05,
|
| 919 |
+
"loss": 0.3714,
|
| 920 |
+
"step": 15200
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 4.89,
|
| 924 |
+
"learning_rate": 3.9019482593420636e-05,
|
| 925 |
+
"loss": 0.3721,
|
| 926 |
+
"step": 15300
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 4.92,
|
| 930 |
+
"learning_rate": 3.8983995173710924e-05,
|
| 931 |
+
"loss": 0.3716,
|
| 932 |
+
"step": 15400
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 4.95,
|
| 936 |
+
"learning_rate": 3.8948507754001205e-05,
|
| 937 |
+
"loss": 0.3728,
|
| 938 |
+
"step": 15500
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 4.98,
|
| 942 |
+
"learning_rate": 3.891302033429149e-05,
|
| 943 |
+
"loss": 0.384,
|
| 944 |
+
"step": 15600
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 5.01,
|
| 948 |
+
"learning_rate": 3.887753291458178e-05,
|
| 949 |
+
"loss": 0.3415,
|
| 950 |
+
"step": 15700
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 5.05,
|
| 954 |
+
"learning_rate": 3.884204549487207e-05,
|
| 955 |
+
"loss": 0.3064,
|
| 956 |
+
"step": 15800
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 5.08,
|
| 960 |
+
"learning_rate": 3.880655807516236e-05,
|
| 961 |
+
"loss": 0.3164,
|
| 962 |
+
"step": 15900
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 5.11,
|
| 966 |
+
"learning_rate": 3.8771070655452645e-05,
|
| 967 |
+
"loss": 0.3215,
|
| 968 |
+
"step": 16000
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 5.14,
|
| 972 |
+
"learning_rate": 3.873558323574293e-05,
|
| 973 |
+
"loss": 0.3234,
|
| 974 |
+
"step": 16100
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 5.17,
|
| 978 |
+
"learning_rate": 3.870009581603322e-05,
|
| 979 |
+
"loss": 0.3119,
|
| 980 |
+
"step": 16200
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 5.21,
|
| 984 |
+
"learning_rate": 3.866460839632351e-05,
|
| 985 |
+
"loss": 0.3145,
|
| 986 |
+
"step": 16300
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 5.24,
|
| 990 |
+
"learning_rate": 3.8629120976613796e-05,
|
| 991 |
+
"loss": 0.3333,
|
| 992 |
+
"step": 16400
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 5.27,
|
| 996 |
+
"learning_rate": 3.8593633556904084e-05,
|
| 997 |
+
"loss": 0.3195,
|
| 998 |
+
"step": 16500
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 5.3,
|
| 1002 |
+
"learning_rate": 3.855814613719437e-05,
|
| 1003 |
+
"loss": 0.3159,
|
| 1004 |
+
"step": 16600
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 5.33,
|
| 1008 |
+
"learning_rate": 3.852265871748465e-05,
|
| 1009 |
+
"loss": 0.3189,
|
| 1010 |
+
"step": 16700
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 5.37,
|
| 1014 |
+
"learning_rate": 3.848717129777494e-05,
|
| 1015 |
+
"loss": 0.317,
|
| 1016 |
+
"step": 16800
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 5.4,
|
| 1020 |
+
"learning_rate": 3.845168387806523e-05,
|
| 1021 |
+
"loss": 0.3231,
|
| 1022 |
+
"step": 16900
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 5.43,
|
| 1026 |
+
"learning_rate": 3.841619645835552e-05,
|
| 1027 |
+
"loss": 0.3328,
|
| 1028 |
+
"step": 17000
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 5.46,
|
| 1032 |
+
"learning_rate": 3.8380709038645805e-05,
|
| 1033 |
+
"loss": 0.3282,
|
| 1034 |
+
"step": 17100
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 5.49,
|
| 1038 |
+
"learning_rate": 3.8345221618936086e-05,
|
| 1039 |
+
"loss": 0.3169,
|
| 1040 |
+
"step": 17200
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 5.53,
|
| 1044 |
+
"learning_rate": 3.8309734199226374e-05,
|
| 1045 |
+
"loss": 0.3156,
|
| 1046 |
+
"step": 17300
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 5.56,
|
| 1050 |
+
"learning_rate": 3.827424677951666e-05,
|
| 1051 |
+
"loss": 0.3313,
|
| 1052 |
+
"step": 17400
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 5.59,
|
| 1056 |
+
"learning_rate": 3.823875935980695e-05,
|
| 1057 |
+
"loss": 0.3181,
|
| 1058 |
+
"step": 17500
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 5.62,
|
| 1062 |
+
"learning_rate": 3.820327194009724e-05,
|
| 1063 |
+
"loss": 0.334,
|
| 1064 |
+
"step": 17600
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 5.65,
|
| 1068 |
+
"learning_rate": 3.8167784520387526e-05,
|
| 1069 |
+
"loss": 0.3313,
|
| 1070 |
+
"step": 17700
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 5.69,
|
| 1074 |
+
"learning_rate": 3.8132297100677814e-05,
|
| 1075 |
+
"loss": 0.3331,
|
| 1076 |
+
"step": 17800
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 5.72,
|
| 1080 |
+
"learning_rate": 3.80968096809681e-05,
|
| 1081 |
+
"loss": 0.3405,
|
| 1082 |
+
"step": 17900
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 5.75,
|
| 1086 |
+
"learning_rate": 3.806132226125839e-05,
|
| 1087 |
+
"loss": 0.3345,
|
| 1088 |
+
"step": 18000
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 5.78,
|
| 1092 |
+
"learning_rate": 3.802583484154868e-05,
|
| 1093 |
+
"loss": 0.3295,
|
| 1094 |
+
"step": 18100
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 5.81,
|
| 1098 |
+
"learning_rate": 3.7990347421838965e-05,
|
| 1099 |
+
"loss": 0.3244,
|
| 1100 |
+
"step": 18200
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 5.84,
|
| 1104 |
+
"learning_rate": 3.795486000212925e-05,
|
| 1105 |
+
"loss": 0.3346,
|
| 1106 |
+
"step": 18300
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 5.88,
|
| 1110 |
+
"learning_rate": 3.791972745661663e-05,
|
| 1111 |
+
"loss": 0.3317,
|
| 1112 |
+
"step": 18400
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 5.91,
|
| 1116 |
+
"learning_rate": 3.788424003690692e-05,
|
| 1117 |
+
"loss": 0.3219,
|
| 1118 |
+
"step": 18500
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 5.94,
|
| 1122 |
+
"learning_rate": 3.784875261719721e-05,
|
| 1123 |
+
"loss": 0.3167,
|
| 1124 |
+
"step": 18600
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 5.97,
|
| 1128 |
+
"learning_rate": 3.781326519748749e-05,
|
| 1129 |
+
"loss": 0.33,
|
| 1130 |
+
"step": 18700
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 6.0,
|
| 1134 |
+
"learning_rate": 3.777777777777778e-05,
|
| 1135 |
+
"loss": 0.3211,
|
| 1136 |
+
"step": 18800
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 6.04,
|
| 1140 |
+
"learning_rate": 3.7742290358068066e-05,
|
| 1141 |
+
"loss": 0.2794,
|
| 1142 |
+
"step": 18900
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 6.07,
|
| 1146 |
+
"learning_rate": 3.7706802938358354e-05,
|
| 1147 |
+
"loss": 0.2675,
|
| 1148 |
+
"step": 19000
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 6.1,
|
| 1152 |
+
"learning_rate": 3.767131551864864e-05,
|
| 1153 |
+
"loss": 0.2676,
|
| 1154 |
+
"step": 19100
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 6.13,
|
| 1158 |
+
"learning_rate": 3.763582809893893e-05,
|
| 1159 |
+
"loss": 0.2693,
|
| 1160 |
+
"step": 19200
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 6.16,
|
| 1164 |
+
"learning_rate": 3.760034067922922e-05,
|
| 1165 |
+
"loss": 0.2669,
|
| 1166 |
+
"step": 19300
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 6.2,
|
| 1170 |
+
"learning_rate": 3.7564853259519506e-05,
|
| 1171 |
+
"loss": 0.2744,
|
| 1172 |
+
"step": 19400
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 6.23,
|
| 1176 |
+
"learning_rate": 3.7529365839809794e-05,
|
| 1177 |
+
"loss": 0.282,
|
| 1178 |
+
"step": 19500
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 6.26,
|
| 1182 |
+
"learning_rate": 3.749387842010008e-05,
|
| 1183 |
+
"loss": 0.2848,
|
| 1184 |
+
"step": 19600
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 6.29,
|
| 1188 |
+
"learning_rate": 3.745839100039037e-05,
|
| 1189 |
+
"loss": 0.2783,
|
| 1190 |
+
"step": 19700
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 6.32,
|
| 1194 |
+
"learning_rate": 3.742290358068066e-05,
|
| 1195 |
+
"loss": 0.2843,
|
| 1196 |
+
"step": 19800
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 6.36,
|
| 1200 |
+
"learning_rate": 3.738741616097094e-05,
|
| 1201 |
+
"loss": 0.2696,
|
| 1202 |
+
"step": 19900
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 6.39,
|
| 1206 |
+
"learning_rate": 3.7351928741261226e-05,
|
| 1207 |
+
"loss": 0.2932,
|
| 1208 |
+
"step": 20000
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 6.42,
|
| 1212 |
+
"learning_rate": 3.7316441321551514e-05,
|
| 1213 |
+
"loss": 0.2775,
|
| 1214 |
+
"step": 20100
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 6.45,
|
| 1218 |
+
"learning_rate": 3.72809539018418e-05,
|
| 1219 |
+
"loss": 0.2893,
|
| 1220 |
+
"step": 20200
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 6.48,
|
| 1224 |
+
"learning_rate": 3.724546648213209e-05,
|
| 1225 |
+
"loss": 0.286,
|
| 1226 |
+
"step": 20300
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 6.52,
|
| 1230 |
+
"learning_rate": 3.720997906242237e-05,
|
| 1231 |
+
"loss": 0.2799,
|
| 1232 |
+
"step": 20400
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 6.55,
|
| 1236 |
+
"learning_rate": 3.717449164271266e-05,
|
| 1237 |
+
"loss": 0.2827,
|
| 1238 |
+
"step": 20500
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 6.58,
|
| 1242 |
+
"learning_rate": 3.713900422300295e-05,
|
| 1243 |
+
"loss": 0.2783,
|
| 1244 |
+
"step": 20600
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 6.61,
|
| 1248 |
+
"learning_rate": 3.7103516803293235e-05,
|
| 1249 |
+
"loss": 0.2729,
|
| 1250 |
+
"step": 20700
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 6.64,
|
| 1254 |
+
"learning_rate": 3.706802938358352e-05,
|
| 1255 |
+
"loss": 0.2747,
|
| 1256 |
+
"step": 20800
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 6.68,
|
| 1260 |
+
"learning_rate": 3.703254196387381e-05,
|
| 1261 |
+
"loss": 0.2764,
|
| 1262 |
+
"step": 20900
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 6.71,
|
| 1266 |
+
"learning_rate": 3.699705454416409e-05,
|
| 1267 |
+
"loss": 0.2804,
|
| 1268 |
+
"step": 21000
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 6.74,
|
| 1272 |
+
"learning_rate": 3.696156712445438e-05,
|
| 1273 |
+
"loss": 0.2841,
|
| 1274 |
+
"step": 21100
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 6.77,
|
| 1278 |
+
"learning_rate": 3.692607970474467e-05,
|
| 1279 |
+
"loss": 0.277,
|
| 1280 |
+
"step": 21200
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 6.8,
|
| 1284 |
+
"learning_rate": 3.6890592285034956e-05,
|
| 1285 |
+
"loss": 0.284,
|
| 1286 |
+
"step": 21300
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 6.83,
|
| 1290 |
+
"learning_rate": 3.6855104865325244e-05,
|
| 1291 |
+
"loss": 0.2778,
|
| 1292 |
+
"step": 21400
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 6.87,
|
| 1296 |
+
"learning_rate": 3.681961744561553e-05,
|
| 1297 |
+
"loss": 0.2797,
|
| 1298 |
+
"step": 21500
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 6.9,
|
| 1302 |
+
"learning_rate": 3.678413002590582e-05,
|
| 1303 |
+
"loss": 0.3036,
|
| 1304 |
+
"step": 21600
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 6.93,
|
| 1308 |
+
"learning_rate": 3.674864260619611e-05,
|
| 1309 |
+
"loss": 0.2869,
|
| 1310 |
+
"step": 21700
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 6.96,
|
| 1314 |
+
"learning_rate": 3.6713155186486395e-05,
|
| 1315 |
+
"loss": 0.2921,
|
| 1316 |
+
"step": 21800
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 6.99,
|
| 1320 |
+
"learning_rate": 3.667766776677668e-05,
|
| 1321 |
+
"loss": 0.2844,
|
| 1322 |
+
"step": 21900
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 7.03,
|
| 1326 |
+
"learning_rate": 3.664218034706697e-05,
|
| 1327 |
+
"loss": 0.2427,
|
| 1328 |
+
"step": 22000
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 7.06,
|
| 1332 |
+
"learning_rate": 3.660669292735726e-05,
|
| 1333 |
+
"loss": 0.243,
|
| 1334 |
+
"step": 22100
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 7.09,
|
| 1338 |
+
"learning_rate": 3.657120550764754e-05,
|
| 1339 |
+
"loss": 0.2311,
|
| 1340 |
+
"step": 22200
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 7.12,
|
| 1344 |
+
"learning_rate": 3.653571808793783e-05,
|
| 1345 |
+
"loss": 0.2358,
|
| 1346 |
+
"step": 22300
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 7.15,
|
| 1350 |
+
"learning_rate": 3.6500230668228116e-05,
|
| 1351 |
+
"loss": 0.2287,
|
| 1352 |
+
"step": 22400
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 7.19,
|
| 1356 |
+
"learning_rate": 3.6464743248518404e-05,
|
| 1357 |
+
"loss": 0.2519,
|
| 1358 |
+
"step": 22500
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 7.22,
|
| 1362 |
+
"learning_rate": 3.642925582880869e-05,
|
| 1363 |
+
"loss": 0.2528,
|
| 1364 |
+
"step": 22600
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 7.25,
|
| 1368 |
+
"learning_rate": 3.639376840909898e-05,
|
| 1369 |
+
"loss": 0.2507,
|
| 1370 |
+
"step": 22700
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 7.28,
|
| 1374 |
+
"learning_rate": 3.635828098938926e-05,
|
| 1375 |
+
"loss": 0.2496,
|
| 1376 |
+
"step": 22800
|
| 1377 |
+
},
|
| 1378 |
+
{
|
| 1379 |
+
"epoch": 7.31,
|
| 1380 |
+
"learning_rate": 3.632279356967955e-05,
|
| 1381 |
+
"loss": 0.2503,
|
| 1382 |
+
"step": 22900
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 7.35,
|
| 1386 |
+
"learning_rate": 3.628730614996984e-05,
|
| 1387 |
+
"loss": 0.2383,
|
| 1388 |
+
"step": 23000
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 7.38,
|
| 1392 |
+
"learning_rate": 3.6251818730260125e-05,
|
| 1393 |
+
"loss": 0.2513,
|
| 1394 |
+
"step": 23100
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 7.41,
|
| 1398 |
+
"learning_rate": 3.621633131055041e-05,
|
| 1399 |
+
"loss": 0.2492,
|
| 1400 |
+
"step": 23200
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 7.44,
|
| 1404 |
+
"learning_rate": 3.61808438908407e-05,
|
| 1405 |
+
"loss": 0.2473,
|
| 1406 |
+
"step": 23300
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 7.47,
|
| 1410 |
+
"learning_rate": 3.614535647113099e-05,
|
| 1411 |
+
"loss": 0.2502,
|
| 1412 |
+
"step": 23400
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 7.51,
|
| 1416 |
+
"learning_rate": 3.6109869051421276e-05,
|
| 1417 |
+
"loss": 0.2492,
|
| 1418 |
+
"step": 23500
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 7.54,
|
| 1422 |
+
"learning_rate": 3.6074381631711564e-05,
|
| 1423 |
+
"loss": 0.2366,
|
| 1424 |
+
"step": 23600
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 7.57,
|
| 1428 |
+
"learning_rate": 3.603889421200185e-05,
|
| 1429 |
+
"loss": 0.2393,
|
| 1430 |
+
"step": 23700
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 7.6,
|
| 1434 |
+
"learning_rate": 3.600340679229214e-05,
|
| 1435 |
+
"loss": 0.2483,
|
| 1436 |
+
"step": 23800
|
| 1437 |
+
},
|
| 1438 |
+
{
|
| 1439 |
+
"epoch": 7.63,
|
| 1440 |
+
"learning_rate": 3.596791937258242e-05,
|
| 1441 |
+
"loss": 0.2428,
|
| 1442 |
+
"step": 23900
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 7.67,
|
| 1446 |
+
"learning_rate": 3.593243195287271e-05,
|
| 1447 |
+
"loss": 0.2507,
|
| 1448 |
+
"step": 24000
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 7.7,
|
| 1452 |
+
"learning_rate": 3.5896944533163e-05,
|
| 1453 |
+
"loss": 0.2491,
|
| 1454 |
+
"step": 24100
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 7.73,
|
| 1458 |
+
"learning_rate": 3.5861457113453285e-05,
|
| 1459 |
+
"loss": 0.2549,
|
| 1460 |
+
"step": 24200
|
| 1461 |
+
},
|
| 1462 |
+
{
|
| 1463 |
+
"epoch": 7.76,
|
| 1464 |
+
"learning_rate": 3.582596969374357e-05,
|
| 1465 |
+
"loss": 0.251,
|
| 1466 |
+
"step": 24300
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 7.79,
|
| 1470 |
+
"learning_rate": 3.579048227403386e-05,
|
| 1471 |
+
"loss": 0.2357,
|
| 1472 |
+
"step": 24400
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 7.82,
|
| 1476 |
+
"learning_rate": 3.575499485432414e-05,
|
| 1477 |
+
"loss": 0.2441,
|
| 1478 |
+
"step": 24500
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 7.86,
|
| 1482 |
+
"learning_rate": 3.571950743461443e-05,
|
| 1483 |
+
"loss": 0.2561,
|
| 1484 |
+
"step": 24600
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 7.89,
|
| 1488 |
+
"learning_rate": 3.568402001490472e-05,
|
| 1489 |
+
"loss": 0.2524,
|
| 1490 |
+
"step": 24700
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 7.92,
|
| 1494 |
+
"learning_rate": 3.5648532595195006e-05,
|
| 1495 |
+
"loss": 0.2467,
|
| 1496 |
+
"step": 24800
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 7.95,
|
| 1500 |
+
"learning_rate": 3.5613045175485293e-05,
|
| 1501 |
+
"loss": 0.2416,
|
| 1502 |
+
"step": 24900
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 7.98,
|
| 1506 |
+
"learning_rate": 3.557755775577558e-05,
|
| 1507 |
+
"loss": 0.2511,
|
| 1508 |
+
"step": 25000
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 8.02,
|
| 1512 |
+
"learning_rate": 3.554207033606587e-05,
|
| 1513 |
+
"loss": 0.2323,
|
| 1514 |
+
"step": 25100
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 8.05,
|
| 1518 |
+
"learning_rate": 3.550658291635615e-05,
|
| 1519 |
+
"loss": 0.2158,
|
| 1520 |
+
"step": 25200
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 8.08,
|
| 1524 |
+
"learning_rate": 3.547109549664644e-05,
|
| 1525 |
+
"loss": 0.21,
|
| 1526 |
+
"step": 25300
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 8.11,
|
| 1530 |
+
"learning_rate": 3.5435608076936726e-05,
|
| 1531 |
+
"loss": 0.2102,
|
| 1532 |
+
"step": 25400
|
| 1533 |
+
},
|
| 1534 |
+
{
|
| 1535 |
+
"epoch": 8.14,
|
| 1536 |
+
"learning_rate": 3.5400120657227014e-05,
|
| 1537 |
+
"loss": 0.2071,
|
| 1538 |
+
"step": 25500
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 8.18,
|
| 1542 |
+
"learning_rate": 3.53646332375173e-05,
|
| 1543 |
+
"loss": 0.2198,
|
| 1544 |
+
"step": 25600
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 8.21,
|
| 1548 |
+
"learning_rate": 3.532914581780759e-05,
|
| 1549 |
+
"loss": 0.2193,
|
| 1550 |
+
"step": 25700
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 8.24,
|
| 1554 |
+
"learning_rate": 3.529365839809788e-05,
|
| 1555 |
+
"loss": 0.2166,
|
| 1556 |
+
"step": 25800
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 8.27,
|
| 1560 |
+
"learning_rate": 3.5258170978388166e-05,
|
| 1561 |
+
"loss": 0.2237,
|
| 1562 |
+
"step": 25900
|
| 1563 |
+
},
|
| 1564 |
+
{
|
| 1565 |
+
"epoch": 8.3,
|
| 1566 |
+
"learning_rate": 3.5222683558678454e-05,
|
| 1567 |
+
"loss": 0.2241,
|
| 1568 |
+
"step": 26000
|
| 1569 |
+
},
|
| 1570 |
+
{
|
| 1571 |
+
"epoch": 8.34,
|
| 1572 |
+
"learning_rate": 3.518719613896874e-05,
|
| 1573 |
+
"loss": 0.2132,
|
| 1574 |
+
"step": 26100
|
| 1575 |
+
},
|
| 1576 |
+
{
|
| 1577 |
+
"epoch": 8.37,
|
| 1578 |
+
"learning_rate": 3.515170871925902e-05,
|
| 1579 |
+
"loss": 0.2101,
|
| 1580 |
+
"step": 26200
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 8.4,
|
| 1584 |
+
"learning_rate": 3.511622129954931e-05,
|
| 1585 |
+
"loss": 0.227,
|
| 1586 |
+
"step": 26300
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 8.43,
|
| 1590 |
+
"learning_rate": 3.50807338798396e-05,
|
| 1591 |
+
"loss": 0.2221,
|
| 1592 |
+
"step": 26400
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 8.46,
|
| 1596 |
+
"learning_rate": 3.504560133432698e-05,
|
| 1597 |
+
"loss": 0.2181,
|
| 1598 |
+
"step": 26500
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 8.5,
|
| 1602 |
+
"learning_rate": 3.501011391461727e-05,
|
| 1603 |
+
"loss": 0.2208,
|
| 1604 |
+
"step": 26600
|
| 1605 |
+
},
|
| 1606 |
+
{
|
| 1607 |
+
"epoch": 8.53,
|
| 1608 |
+
"learning_rate": 3.4974626494907555e-05,
|
| 1609 |
+
"loss": 0.2192,
|
| 1610 |
+
"step": 26700
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 8.56,
|
| 1614 |
+
"learning_rate": 3.493913907519784e-05,
|
| 1615 |
+
"loss": 0.2227,
|
| 1616 |
+
"step": 26800
|
| 1617 |
+
},
|
| 1618 |
+
{
|
| 1619 |
+
"epoch": 8.59,
|
| 1620 |
+
"learning_rate": 3.490365165548813e-05,
|
| 1621 |
+
"loss": 0.2308,
|
| 1622 |
+
"step": 26900
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 8.62,
|
| 1626 |
+
"learning_rate": 3.486816423577842e-05,
|
| 1627 |
+
"loss": 0.22,
|
| 1628 |
+
"step": 27000
|
| 1629 |
+
},
|
| 1630 |
+
{
|
| 1631 |
+
"epoch": 8.66,
|
| 1632 |
+
"learning_rate": 3.4832676816068706e-05,
|
| 1633 |
+
"loss": 0.2247,
|
| 1634 |
+
"step": 27100
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 8.69,
|
| 1638 |
+
"learning_rate": 3.4797189396358994e-05,
|
| 1639 |
+
"loss": 0.2148,
|
| 1640 |
+
"step": 27200
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 8.72,
|
| 1644 |
+
"learning_rate": 3.476170197664928e-05,
|
| 1645 |
+
"loss": 0.2136,
|
| 1646 |
+
"step": 27300
|
| 1647 |
+
},
|
| 1648 |
+
{
|
| 1649 |
+
"epoch": 8.75,
|
| 1650 |
+
"learning_rate": 3.472621455693957e-05,
|
| 1651 |
+
"loss": 0.2282,
|
| 1652 |
+
"step": 27400
|
| 1653 |
+
},
|
| 1654 |
+
{
|
| 1655 |
+
"epoch": 8.78,
|
| 1656 |
+
"learning_rate": 3.469072713722986e-05,
|
| 1657 |
+
"loss": 0.2266,
|
| 1658 |
+
"step": 27500
|
| 1659 |
+
},
|
| 1660 |
+
{
|
| 1661 |
+
"epoch": 8.82,
|
| 1662 |
+
"learning_rate": 3.4655239717520146e-05,
|
| 1663 |
+
"loss": 0.2198,
|
| 1664 |
+
"step": 27600
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 8.85,
|
| 1668 |
+
"learning_rate": 3.461975229781043e-05,
|
| 1669 |
+
"loss": 0.2243,
|
| 1670 |
+
"step": 27700
|
| 1671 |
+
},
|
| 1672 |
+
{
|
| 1673 |
+
"epoch": 8.88,
|
| 1674 |
+
"learning_rate": 3.4584264878100715e-05,
|
| 1675 |
+
"loss": 0.2358,
|
| 1676 |
+
"step": 27800
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 8.91,
|
| 1680 |
+
"learning_rate": 3.4548777458391e-05,
|
| 1681 |
+
"loss": 0.2255,
|
| 1682 |
+
"step": 27900
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 8.94,
|
| 1686 |
+
"learning_rate": 3.451329003868129e-05,
|
| 1687 |
+
"loss": 0.2258,
|
| 1688 |
+
"step": 28000
|
| 1689 |
+
},
|
| 1690 |
+
{
|
| 1691 |
+
"epoch": 8.97,
|
| 1692 |
+
"learning_rate": 3.447780261897158e-05,
|
| 1693 |
+
"loss": 0.2228,
|
| 1694 |
+
"step": 28100
|
| 1695 |
+
},
|
| 1696 |
+
{
|
| 1697 |
+
"epoch": 9.01,
|
| 1698 |
+
"learning_rate": 3.4442315199261866e-05,
|
| 1699 |
+
"loss": 0.2183,
|
| 1700 |
+
"step": 28200
|
| 1701 |
+
},
|
| 1702 |
+
{
|
| 1703 |
+
"epoch": 9.04,
|
| 1704 |
+
"learning_rate": 3.440682777955215e-05,
|
| 1705 |
+
"loss": 0.1988,
|
| 1706 |
+
"step": 28300
|
| 1707 |
+
},
|
| 1708 |
+
{
|
| 1709 |
+
"epoch": 9.07,
|
| 1710 |
+
"learning_rate": 3.4371340359842436e-05,
|
| 1711 |
+
"loss": 0.188,
|
| 1712 |
+
"step": 28400
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"epoch": 9.1,
|
| 1716 |
+
"learning_rate": 3.4335852940132723e-05,
|
| 1717 |
+
"loss": 0.1963,
|
| 1718 |
+
"step": 28500
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 9.13,
|
| 1722 |
+
"learning_rate": 3.430036552042301e-05,
|
| 1723 |
+
"loss": 0.1919,
|
| 1724 |
+
"step": 28600
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 9.17,
|
| 1728 |
+
"learning_rate": 3.42648781007133e-05,
|
| 1729 |
+
"loss": 0.1906,
|
| 1730 |
+
"step": 28700
|
| 1731 |
+
},
|
| 1732 |
+
{
|
| 1733 |
+
"epoch": 9.2,
|
| 1734 |
+
"learning_rate": 3.422939068100359e-05,
|
| 1735 |
+
"loss": 0.1911,
|
| 1736 |
+
"step": 28800
|
| 1737 |
+
},
|
| 1738 |
+
{
|
| 1739 |
+
"epoch": 9.23,
|
| 1740 |
+
"learning_rate": 3.4193903261293875e-05,
|
| 1741 |
+
"loss": 0.1971,
|
| 1742 |
+
"step": 28900
|
| 1743 |
+
},
|
| 1744 |
+
{
|
| 1745 |
+
"epoch": 9.26,
|
| 1746 |
+
"learning_rate": 3.415841584158416e-05,
|
| 1747 |
+
"loss": 0.1931,
|
| 1748 |
+
"step": 29000
|
| 1749 |
+
},
|
| 1750 |
+
{
|
| 1751 |
+
"epoch": 9.29,
|
| 1752 |
+
"learning_rate": 3.412292842187445e-05,
|
| 1753 |
+
"loss": 0.1951,
|
| 1754 |
+
"step": 29100
|
| 1755 |
+
},
|
| 1756 |
+
{
|
| 1757 |
+
"epoch": 9.33,
|
| 1758 |
+
"learning_rate": 3.408744100216474e-05,
|
| 1759 |
+
"loss": 0.2008,
|
| 1760 |
+
"step": 29200
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 9.36,
|
| 1764 |
+
"learning_rate": 3.405195358245503e-05,
|
| 1765 |
+
"loss": 0.2053,
|
| 1766 |
+
"step": 29300
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 9.39,
|
| 1770 |
+
"learning_rate": 3.401646616274531e-05,
|
| 1771 |
+
"loss": 0.2015,
|
| 1772 |
+
"step": 29400
|
| 1773 |
+
},
|
| 1774 |
+
{
|
| 1775 |
+
"epoch": 9.42,
|
| 1776 |
+
"learning_rate": 3.3980978743035596e-05,
|
| 1777 |
+
"loss": 0.193,
|
| 1778 |
+
"step": 29500
|
| 1779 |
+
},
|
| 1780 |
+
{
|
| 1781 |
+
"epoch": 9.45,
|
| 1782 |
+
"learning_rate": 3.3945491323325884e-05,
|
| 1783 |
+
"loss": 0.1955,
|
| 1784 |
+
"step": 29600
|
| 1785 |
+
},
|
| 1786 |
+
{
|
| 1787 |
+
"epoch": 9.49,
|
| 1788 |
+
"learning_rate": 3.391000390361617e-05,
|
| 1789 |
+
"loss": 0.1993,
|
| 1790 |
+
"step": 29700
|
| 1791 |
+
},
|
| 1792 |
+
{
|
| 1793 |
+
"epoch": 9.52,
|
| 1794 |
+
"learning_rate": 3.387451648390646e-05,
|
| 1795 |
+
"loss": 0.2043,
|
| 1796 |
+
"step": 29800
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"epoch": 9.55,
|
| 1800 |
+
"learning_rate": 3.383902906419675e-05,
|
| 1801 |
+
"loss": 0.1955,
|
| 1802 |
+
"step": 29900
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 9.58,
|
| 1806 |
+
"learning_rate": 3.380354164448703e-05,
|
| 1807 |
+
"loss": 0.1961,
|
| 1808 |
+
"step": 30000
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 9.61,
|
| 1812 |
+
"learning_rate": 3.3768054224777316e-05,
|
| 1813 |
+
"loss": 0.1961,
|
| 1814 |
+
"step": 30100
|
| 1815 |
+
},
|
| 1816 |
+
{
|
| 1817 |
+
"epoch": 9.65,
|
| 1818 |
+
"learning_rate": 3.3732566805067604e-05,
|
| 1819 |
+
"loss": 0.196,
|
| 1820 |
+
"step": 30200
|
| 1821 |
+
},
|
| 1822 |
+
{
|
| 1823 |
+
"epoch": 9.68,
|
| 1824 |
+
"learning_rate": 3.369707938535789e-05,
|
| 1825 |
+
"loss": 0.2088,
|
| 1826 |
+
"step": 30300
|
| 1827 |
+
},
|
| 1828 |
+
{
|
| 1829 |
+
"epoch": 9.71,
|
| 1830 |
+
"learning_rate": 3.366159196564818e-05,
|
| 1831 |
+
"loss": 0.2031,
|
| 1832 |
+
"step": 30400
|
| 1833 |
+
},
|
| 1834 |
+
{
|
| 1835 |
+
"epoch": 9.74,
|
| 1836 |
+
"learning_rate": 3.362610454593847e-05,
|
| 1837 |
+
"loss": 0.1949,
|
| 1838 |
+
"step": 30500
|
| 1839 |
+
},
|
| 1840 |
+
{
|
| 1841 |
+
"epoch": 9.77,
|
| 1842 |
+
"learning_rate": 3.3590617126228756e-05,
|
| 1843 |
+
"loss": 0.1988,
|
| 1844 |
+
"step": 30600
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 9.81,
|
| 1848 |
+
"learning_rate": 3.3555129706519044e-05,
|
| 1849 |
+
"loss": 0.2061,
|
| 1850 |
+
"step": 30700
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 9.84,
|
| 1854 |
+
"learning_rate": 3.351964228680933e-05,
|
| 1855 |
+
"loss": 0.2022,
|
| 1856 |
+
"step": 30800
|
| 1857 |
+
},
|
| 1858 |
+
{
|
| 1859 |
+
"epoch": 9.87,
|
| 1860 |
+
"learning_rate": 3.348450974129671e-05,
|
| 1861 |
+
"loss": 0.1965,
|
| 1862 |
+
"step": 30900
|
| 1863 |
+
},
|
| 1864 |
+
{
|
| 1865 |
+
"epoch": 9.9,
|
| 1866 |
+
"learning_rate": 3.3449022321587e-05,
|
| 1867 |
+
"loss": 0.1999,
|
| 1868 |
+
"step": 31000
|
| 1869 |
+
},
|
| 1870 |
+
{
|
| 1871 |
+
"epoch": 9.93,
|
| 1872 |
+
"learning_rate": 3.341353490187729e-05,
|
| 1873 |
+
"loss": 0.2065,
|
| 1874 |
+
"step": 31100
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 9.96,
|
| 1878 |
+
"learning_rate": 3.3378047482167576e-05,
|
| 1879 |
+
"loss": 0.2031,
|
| 1880 |
+
"step": 31200
|
| 1881 |
+
},
|
| 1882 |
+
{
|
| 1883 |
+
"epoch": 10.0,
|
| 1884 |
+
"learning_rate": 3.3342560062457864e-05,
|
| 1885 |
+
"loss": 0.2078,
|
| 1886 |
+
"step": 31300
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 10.03,
|
| 1890 |
+
"learning_rate": 3.330707264274815e-05,
|
| 1891 |
+
"loss": 0.1733,
|
| 1892 |
+
"step": 31400
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 10.06,
|
| 1896 |
+
"learning_rate": 3.327158522303843e-05,
|
| 1897 |
+
"loss": 0.1654,
|
| 1898 |
+
"step": 31500
|
| 1899 |
+
},
|
| 1900 |
+
{
|
| 1901 |
+
"epoch": 10.09,
|
| 1902 |
+
"learning_rate": 3.323609780332872e-05,
|
| 1903 |
+
"loss": 0.1674,
|
| 1904 |
+
"step": 31600
|
| 1905 |
+
},
|
| 1906 |
+
{
|
| 1907 |
+
"epoch": 10.12,
|
| 1908 |
+
"learning_rate": 3.320061038361901e-05,
|
| 1909 |
+
"loss": 0.1725,
|
| 1910 |
+
"step": 31700
|
| 1911 |
+
},
|
| 1912 |
+
{
|
| 1913 |
+
"epoch": 10.16,
|
| 1914 |
+
"learning_rate": 3.3165122963909297e-05,
|
| 1915 |
+
"loss": 0.1761,
|
| 1916 |
+
"step": 31800
|
| 1917 |
+
},
|
| 1918 |
+
{
|
| 1919 |
+
"epoch": 10.19,
|
| 1920 |
+
"learning_rate": 3.3129635544199584e-05,
|
| 1921 |
+
"loss": 0.1757,
|
| 1922 |
+
"step": 31900
|
| 1923 |
+
},
|
| 1924 |
+
{
|
| 1925 |
+
"epoch": 10.22,
|
| 1926 |
+
"learning_rate": 3.309414812448987e-05,
|
| 1927 |
+
"loss": 0.1781,
|
| 1928 |
+
"step": 32000
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 10.25,
|
| 1932 |
+
"learning_rate": 3.3058660704780153e-05,
|
| 1933 |
+
"loss": 0.1716,
|
| 1934 |
+
"step": 32100
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 10.28,
|
| 1938 |
+
"learning_rate": 3.302317328507044e-05,
|
| 1939 |
+
"loss": 0.1717,
|
| 1940 |
+
"step": 32200
|
| 1941 |
+
},
|
| 1942 |
+
{
|
| 1943 |
+
"epoch": 10.32,
|
| 1944 |
+
"learning_rate": 3.298768586536073e-05,
|
| 1945 |
+
"loss": 0.179,
|
| 1946 |
+
"step": 32300
|
| 1947 |
+
},
|
| 1948 |
+
{
|
| 1949 |
+
"epoch": 10.35,
|
| 1950 |
+
"learning_rate": 3.295219844565102e-05,
|
| 1951 |
+
"loss": 0.174,
|
| 1952 |
+
"step": 32400
|
| 1953 |
+
},
|
| 1954 |
+
{
|
| 1955 |
+
"epoch": 10.38,
|
| 1956 |
+
"learning_rate": 3.2916711025941305e-05,
|
| 1957 |
+
"loss": 0.178,
|
| 1958 |
+
"step": 32500
|
| 1959 |
+
},
|
| 1960 |
+
{
|
| 1961 |
+
"epoch": 10.41,
|
| 1962 |
+
"learning_rate": 3.288122360623159e-05,
|
| 1963 |
+
"loss": 0.1765,
|
| 1964 |
+
"step": 32600
|
| 1965 |
+
},
|
| 1966 |
+
{
|
| 1967 |
+
"epoch": 10.44,
|
| 1968 |
+
"learning_rate": 3.284573618652188e-05,
|
| 1969 |
+
"loss": 0.1824,
|
| 1970 |
+
"step": 32700
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"epoch": 10.48,
|
| 1974 |
+
"learning_rate": 3.281024876681217e-05,
|
| 1975 |
+
"loss": 0.1806,
|
| 1976 |
+
"step": 32800
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 10.51,
|
| 1980 |
+
"learning_rate": 3.277511622129955e-05,
|
| 1981 |
+
"loss": 0.1785,
|
| 1982 |
+
"step": 32900
|
| 1983 |
+
},
|
| 1984 |
+
{
|
| 1985 |
+
"epoch": 10.54,
|
| 1986 |
+
"learning_rate": 3.2739983675786936e-05,
|
| 1987 |
+
"loss": 0.1698,
|
| 1988 |
+
"step": 33000
|
| 1989 |
+
},
|
| 1990 |
+
{
|
| 1991 |
+
"epoch": 10.57,
|
| 1992 |
+
"learning_rate": 3.2704496256077224e-05,
|
| 1993 |
+
"loss": 0.1734,
|
| 1994 |
+
"step": 33100
|
| 1995 |
+
},
|
| 1996 |
+
{
|
| 1997 |
+
"epoch": 10.6,
|
| 1998 |
+
"learning_rate": 3.2669008836367505e-05,
|
| 1999 |
+
"loss": 0.1761,
|
| 2000 |
+
"step": 33200
|
| 2001 |
+
},
|
| 2002 |
+
{
|
| 2003 |
+
"epoch": 10.64,
|
| 2004 |
+
"learning_rate": 3.263352141665779e-05,
|
| 2005 |
+
"loss": 0.1786,
|
| 2006 |
+
"step": 33300
|
| 2007 |
+
},
|
| 2008 |
+
{
|
| 2009 |
+
"epoch": 10.67,
|
| 2010 |
+
"learning_rate": 3.259803399694808e-05,
|
| 2011 |
+
"loss": 0.1767,
|
| 2012 |
+
"step": 33400
|
| 2013 |
+
},
|
| 2014 |
+
{
|
| 2015 |
+
"epoch": 10.7,
|
| 2016 |
+
"learning_rate": 3.256254657723837e-05,
|
| 2017 |
+
"loss": 0.1881,
|
| 2018 |
+
"step": 33500
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 10.73,
|
| 2022 |
+
"learning_rate": 3.252705915752866e-05,
|
| 2023 |
+
"loss": 0.1792,
|
| 2024 |
+
"step": 33600
|
| 2025 |
+
},
|
| 2026 |
+
{
|
| 2027 |
+
"epoch": 10.76,
|
| 2028 |
+
"learning_rate": 3.2491571737818945e-05,
|
| 2029 |
+
"loss": 0.1832,
|
| 2030 |
+
"step": 33700
|
| 2031 |
+
},
|
| 2032 |
+
{
|
| 2033 |
+
"epoch": 10.8,
|
| 2034 |
+
"learning_rate": 3.245608431810923e-05,
|
| 2035 |
+
"loss": 0.1889,
|
| 2036 |
+
"step": 33800
|
| 2037 |
+
},
|
| 2038 |
+
{
|
| 2039 |
+
"epoch": 10.83,
|
| 2040 |
+
"learning_rate": 3.242059689839952e-05,
|
| 2041 |
+
"loss": 0.1824,
|
| 2042 |
+
"step": 33900
|
| 2043 |
+
},
|
| 2044 |
+
{
|
| 2045 |
+
"epoch": 10.86,
|
| 2046 |
+
"learning_rate": 3.238510947868981e-05,
|
| 2047 |
+
"loss": 0.1746,
|
| 2048 |
+
"step": 34000
|
| 2049 |
+
},
|
| 2050 |
+
{
|
| 2051 |
+
"epoch": 10.89,
|
| 2052 |
+
"learning_rate": 3.2349622058980096e-05,
|
| 2053 |
+
"loss": 0.1735,
|
| 2054 |
+
"step": 34100
|
| 2055 |
+
},
|
| 2056 |
+
{
|
| 2057 |
+
"epoch": 10.92,
|
| 2058 |
+
"learning_rate": 3.2314134639270384e-05,
|
| 2059 |
+
"loss": 0.1822,
|
| 2060 |
+
"step": 34200
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 10.95,
|
| 2064 |
+
"learning_rate": 3.227864721956067e-05,
|
| 2065 |
+
"loss": 0.1768,
|
| 2066 |
+
"step": 34300
|
| 2067 |
+
},
|
| 2068 |
+
{
|
| 2069 |
+
"epoch": 10.99,
|
| 2070 |
+
"learning_rate": 3.224315979985095e-05,
|
| 2071 |
+
"loss": 0.1812,
|
| 2072 |
+
"step": 34400
|
| 2073 |
+
},
|
| 2074 |
+
{
|
| 2075 |
+
"epoch": 11.02,
|
| 2076 |
+
"learning_rate": 3.220767238014124e-05,
|
| 2077 |
+
"loss": 0.1645,
|
| 2078 |
+
"step": 34500
|
| 2079 |
+
},
|
| 2080 |
+
{
|
| 2081 |
+
"epoch": 11.05,
|
| 2082 |
+
"learning_rate": 3.217218496043153e-05,
|
| 2083 |
+
"loss": 0.1483,
|
| 2084 |
+
"step": 34600
|
| 2085 |
+
},
|
| 2086 |
+
{
|
| 2087 |
+
"epoch": 11.08,
|
| 2088 |
+
"learning_rate": 3.213669754072182e-05,
|
| 2089 |
+
"loss": 0.1615,
|
| 2090 |
+
"step": 34700
|
| 2091 |
+
},
|
| 2092 |
+
{
|
| 2093 |
+
"epoch": 11.11,
|
| 2094 |
+
"learning_rate": 3.2101210121012105e-05,
|
| 2095 |
+
"loss": 0.1637,
|
| 2096 |
+
"step": 34800
|
| 2097 |
+
},
|
| 2098 |
+
{
|
| 2099 |
+
"epoch": 11.15,
|
| 2100 |
+
"learning_rate": 3.206572270130239e-05,
|
| 2101 |
+
"loss": 0.1497,
|
| 2102 |
+
"step": 34900
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 11.18,
|
| 2106 |
+
"learning_rate": 3.2030235281592674e-05,
|
| 2107 |
+
"loss": 0.1524,
|
| 2108 |
+
"step": 35000
|
| 2109 |
+
},
|
| 2110 |
+
{
|
| 2111 |
+
"epoch": 11.21,
|
| 2112 |
+
"learning_rate": 3.199474786188296e-05,
|
| 2113 |
+
"loss": 0.1545,
|
| 2114 |
+
"step": 35100
|
| 2115 |
+
},
|
| 2116 |
+
{
|
| 2117 |
+
"epoch": 11.24,
|
| 2118 |
+
"learning_rate": 3.195926044217325e-05,
|
| 2119 |
+
"loss": 0.1526,
|
| 2120 |
+
"step": 35200
|
| 2121 |
+
},
|
| 2122 |
+
{
|
| 2123 |
+
"epoch": 11.27,
|
| 2124 |
+
"learning_rate": 3.192377302246354e-05,
|
| 2125 |
+
"loss": 0.1499,
|
| 2126 |
+
"step": 35300
|
| 2127 |
+
},
|
| 2128 |
+
{
|
| 2129 |
+
"epoch": 11.31,
|
| 2130 |
+
"learning_rate": 3.1888285602753826e-05,
|
| 2131 |
+
"loss": 0.1553,
|
| 2132 |
+
"step": 35400
|
| 2133 |
+
},
|
| 2134 |
+
{
|
| 2135 |
+
"epoch": 11.34,
|
| 2136 |
+
"learning_rate": 3.1852798183044114e-05,
|
| 2137 |
+
"loss": 0.1583,
|
| 2138 |
+
"step": 35500
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 11.37,
|
| 2142 |
+
"learning_rate": 3.18173107633344e-05,
|
| 2143 |
+
"loss": 0.156,
|
| 2144 |
+
"step": 35600
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 11.4,
|
| 2148 |
+
"learning_rate": 3.178182334362469e-05,
|
| 2149 |
+
"loss": 0.1582,
|
| 2150 |
+
"step": 35700
|
| 2151 |
+
},
|
| 2152 |
+
{
|
| 2153 |
+
"epoch": 11.43,
|
| 2154 |
+
"learning_rate": 3.174633592391498e-05,
|
| 2155 |
+
"loss": 0.155,
|
| 2156 |
+
"step": 35800
|
| 2157 |
+
},
|
| 2158 |
+
{
|
| 2159 |
+
"epoch": 11.47,
|
| 2160 |
+
"learning_rate": 3.1710848504205265e-05,
|
| 2161 |
+
"loss": 0.1546,
|
| 2162 |
+
"step": 35900
|
| 2163 |
+
},
|
| 2164 |
+
{
|
| 2165 |
+
"epoch": 11.5,
|
| 2166 |
+
"learning_rate": 3.167536108449555e-05,
|
| 2167 |
+
"loss": 0.157,
|
| 2168 |
+
"step": 36000
|
| 2169 |
+
},
|
| 2170 |
+
{
|
| 2171 |
+
"epoch": 11.53,
|
| 2172 |
+
"learning_rate": 3.1639873664785834e-05,
|
| 2173 |
+
"loss": 0.1644,
|
| 2174 |
+
"step": 36100
|
| 2175 |
+
},
|
| 2176 |
+
{
|
| 2177 |
+
"epoch": 11.56,
|
| 2178 |
+
"learning_rate": 3.160438624507612e-05,
|
| 2179 |
+
"loss": 0.1627,
|
| 2180 |
+
"step": 36200
|
| 2181 |
+
},
|
| 2182 |
+
{
|
| 2183 |
+
"epoch": 11.59,
|
| 2184 |
+
"learning_rate": 3.156889882536641e-05,
|
| 2185 |
+
"loss": 0.165,
|
| 2186 |
+
"step": 36300
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 11.63,
|
| 2190 |
+
"learning_rate": 3.15334114056567e-05,
|
| 2191 |
+
"loss": 0.1602,
|
| 2192 |
+
"step": 36400
|
| 2193 |
+
},
|
| 2194 |
+
{
|
| 2195 |
+
"epoch": 11.66,
|
| 2196 |
+
"learning_rate": 3.1497923985946986e-05,
|
| 2197 |
+
"loss": 0.1694,
|
| 2198 |
+
"step": 36500
|
| 2199 |
+
},
|
| 2200 |
+
{
|
| 2201 |
+
"epoch": 11.69,
|
| 2202 |
+
"learning_rate": 3.1462436566237274e-05,
|
| 2203 |
+
"loss": 0.1604,
|
| 2204 |
+
"step": 36600
|
| 2205 |
+
},
|
| 2206 |
+
{
|
| 2207 |
+
"epoch": 11.72,
|
| 2208 |
+
"learning_rate": 3.1426949146527555e-05,
|
| 2209 |
+
"loss": 0.1725,
|
| 2210 |
+
"step": 36700
|
| 2211 |
+
},
|
| 2212 |
+
{
|
| 2213 |
+
"epoch": 11.75,
|
| 2214 |
+
"learning_rate": 3.139146172681784e-05,
|
| 2215 |
+
"loss": 0.1577,
|
| 2216 |
+
"step": 36800
|
| 2217 |
+
},
|
| 2218 |
+
{
|
| 2219 |
+
"epoch": 11.79,
|
| 2220 |
+
"learning_rate": 3.135597430710813e-05,
|
| 2221 |
+
"loss": 0.1741,
|
| 2222 |
+
"step": 36900
|
| 2223 |
+
},
|
| 2224 |
+
{
|
| 2225 |
+
"epoch": 11.82,
|
| 2226 |
+
"learning_rate": 3.132048688739842e-05,
|
| 2227 |
+
"loss": 0.1547,
|
| 2228 |
+
"step": 37000
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 11.85,
|
| 2232 |
+
"learning_rate": 3.1284999467688707e-05,
|
| 2233 |
+
"loss": 0.1621,
|
| 2234 |
+
"step": 37100
|
| 2235 |
+
},
|
| 2236 |
+
{
|
| 2237 |
+
"epoch": 11.88,
|
| 2238 |
+
"learning_rate": 3.1249512047978994e-05,
|
| 2239 |
+
"loss": 0.1663,
|
| 2240 |
+
"step": 37200
|
| 2241 |
+
},
|
| 2242 |
+
{
|
| 2243 |
+
"epoch": 11.91,
|
| 2244 |
+
"learning_rate": 3.121437950246638e-05,
|
| 2245 |
+
"loss": 0.1594,
|
| 2246 |
+
"step": 37300
|
| 2247 |
+
},
|
| 2248 |
+
{
|
| 2249 |
+
"epoch": 11.95,
|
| 2250 |
+
"learning_rate": 3.117889208275667e-05,
|
| 2251 |
+
"loss": 0.174,
|
| 2252 |
+
"step": 37400
|
| 2253 |
+
},
|
| 2254 |
+
{
|
| 2255 |
+
"epoch": 11.98,
|
| 2256 |
+
"learning_rate": 3.114340466304696e-05,
|
| 2257 |
+
"loss": 0.1755,
|
| 2258 |
+
"step": 37500
|
| 2259 |
+
},
|
| 2260 |
+
{
|
| 2261 |
+
"epoch": 12.01,
|
| 2262 |
+
"learning_rate": 3.110791724333724e-05,
|
| 2263 |
+
"loss": 0.1627,
|
| 2264 |
+
"step": 37600
|
| 2265 |
+
},
|
| 2266 |
+
{
|
| 2267 |
+
"epoch": 12.04,
|
| 2268 |
+
"learning_rate": 3.1072429823627526e-05,
|
| 2269 |
+
"loss": 0.1445,
|
| 2270 |
+
"step": 37700
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 12.07,
|
| 2274 |
+
"learning_rate": 3.1036942403917814e-05,
|
| 2275 |
+
"loss": 0.1399,
|
| 2276 |
+
"step": 37800
|
| 2277 |
+
},
|
| 2278 |
+
{
|
| 2279 |
+
"epoch": 12.1,
|
| 2280 |
+
"learning_rate": 3.10014549842081e-05,
|
| 2281 |
+
"loss": 0.1475,
|
| 2282 |
+
"step": 37900
|
| 2283 |
+
},
|
| 2284 |
+
{
|
| 2285 |
+
"epoch": 12.14,
|
| 2286 |
+
"learning_rate": 3.096596756449839e-05,
|
| 2287 |
+
"loss": 0.1373,
|
| 2288 |
+
"step": 38000
|
| 2289 |
+
},
|
| 2290 |
+
{
|
| 2291 |
+
"epoch": 12.17,
|
| 2292 |
+
"learning_rate": 3.093048014478868e-05,
|
| 2293 |
+
"loss": 0.1379,
|
| 2294 |
+
"step": 38100
|
| 2295 |
+
},
|
| 2296 |
+
{
|
| 2297 |
+
"epoch": 12.2,
|
| 2298 |
+
"learning_rate": 3.089499272507896e-05,
|
| 2299 |
+
"loss": 0.1389,
|
| 2300 |
+
"step": 38200
|
| 2301 |
+
},
|
| 2302 |
+
{
|
| 2303 |
+
"epoch": 12.23,
|
| 2304 |
+
"learning_rate": 3.085950530536925e-05,
|
| 2305 |
+
"loss": 0.1394,
|
| 2306 |
+
"step": 38300
|
| 2307 |
+
},
|
| 2308 |
+
{
|
| 2309 |
+
"epoch": 12.26,
|
| 2310 |
+
"learning_rate": 3.0824017885659535e-05,
|
| 2311 |
+
"loss": 0.1434,
|
| 2312 |
+
"step": 38400
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 12.3,
|
| 2316 |
+
"learning_rate": 3.078853046594982e-05,
|
| 2317 |
+
"loss": 0.1467,
|
| 2318 |
+
"step": 38500
|
| 2319 |
+
},
|
| 2320 |
+
{
|
| 2321 |
+
"epoch": 12.33,
|
| 2322 |
+
"learning_rate": 3.075339792043721e-05,
|
| 2323 |
+
"loss": 0.1272,
|
| 2324 |
+
"step": 38600
|
| 2325 |
+
},
|
| 2326 |
+
{
|
| 2327 |
+
"epoch": 12.36,
|
| 2328 |
+
"learning_rate": 3.07179105007275e-05,
|
| 2329 |
+
"loss": 0.1363,
|
| 2330 |
+
"step": 38700
|
| 2331 |
+
},
|
| 2332 |
+
{
|
| 2333 |
+
"epoch": 12.39,
|
| 2334 |
+
"learning_rate": 3.0682423081017786e-05,
|
| 2335 |
+
"loss": 0.1467,
|
| 2336 |
+
"step": 38800
|
| 2337 |
+
},
|
| 2338 |
+
{
|
| 2339 |
+
"epoch": 12.42,
|
| 2340 |
+
"learning_rate": 3.0646935661308074e-05,
|
| 2341 |
+
"loss": 0.1389,
|
| 2342 |
+
"step": 38900
|
| 2343 |
+
},
|
| 2344 |
+
{
|
| 2345 |
+
"epoch": 12.46,
|
| 2346 |
+
"learning_rate": 3.0611448241598355e-05,
|
| 2347 |
+
"loss": 0.1447,
|
| 2348 |
+
"step": 39000
|
| 2349 |
+
},
|
| 2350 |
+
{
|
| 2351 |
+
"epoch": 12.49,
|
| 2352 |
+
"learning_rate": 3.057596082188864e-05,
|
| 2353 |
+
"loss": 0.1524,
|
| 2354 |
+
"step": 39100
|
| 2355 |
+
},
|
| 2356 |
+
{
|
| 2357 |
+
"epoch": 12.52,
|
| 2358 |
+
"learning_rate": 3.054047340217893e-05,
|
| 2359 |
+
"loss": 0.1434,
|
| 2360 |
+
"step": 39200
|
| 2361 |
+
},
|
| 2362 |
+
{
|
| 2363 |
+
"epoch": 12.55,
|
| 2364 |
+
"learning_rate": 3.050498598246922e-05,
|
| 2365 |
+
"loss": 0.142,
|
| 2366 |
+
"step": 39300
|
| 2367 |
+
},
|
| 2368 |
+
{
|
| 2369 |
+
"epoch": 12.58,
|
| 2370 |
+
"learning_rate": 3.0469498562759506e-05,
|
| 2371 |
+
"loss": 0.1495,
|
| 2372 |
+
"step": 39400
|
| 2373 |
+
},
|
| 2374 |
+
{
|
| 2375 |
+
"epoch": 12.62,
|
| 2376 |
+
"learning_rate": 3.0434011143049794e-05,
|
| 2377 |
+
"loss": 0.1451,
|
| 2378 |
+
"step": 39500
|
| 2379 |
+
},
|
| 2380 |
+
{
|
| 2381 |
+
"epoch": 12.65,
|
| 2382 |
+
"learning_rate": 3.0398523723340075e-05,
|
| 2383 |
+
"loss": 0.1464,
|
| 2384 |
+
"step": 39600
|
| 2385 |
+
},
|
| 2386 |
+
{
|
| 2387 |
+
"epoch": 12.68,
|
| 2388 |
+
"learning_rate": 3.0363036303630363e-05,
|
| 2389 |
+
"loss": 0.1419,
|
| 2390 |
+
"step": 39700
|
| 2391 |
+
},
|
| 2392 |
+
{
|
| 2393 |
+
"epoch": 12.71,
|
| 2394 |
+
"learning_rate": 3.032754888392065e-05,
|
| 2395 |
+
"loss": 0.1451,
|
| 2396 |
+
"step": 39800
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 12.74,
|
| 2400 |
+
"learning_rate": 3.029206146421094e-05,
|
| 2401 |
+
"loss": 0.151,
|
| 2402 |
+
"step": 39900
|
| 2403 |
+
},
|
| 2404 |
+
{
|
| 2405 |
+
"epoch": 12.78,
|
| 2406 |
+
"learning_rate": 3.0256574044501227e-05,
|
| 2407 |
+
"loss": 0.1442,
|
| 2408 |
+
"step": 40000
|
| 2409 |
+
}
|
| 2410 |
+
],
|
| 2411 |
+
"max_steps": 125240,
|
| 2412 |
+
"num_train_epochs": 40,
|
| 2413 |
+
"total_flos": 0.0,
|
| 2414 |
+
"trial_name": null,
|
| 2415 |
+
"trial_params": null
|
| 2416 |
+
}
|
checkpoint-45000/config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "Davlan/afro-xlmr-base",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"XLMRobertaModel"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 768,
|
| 13 |
+
"id2label": {
|
| 14 |
+
"0": "LABEL_0"
|
| 15 |
+
},
|
| 16 |
+
"initializer_range": 0.02,
|
| 17 |
+
"intermediate_size": 3072,
|
| 18 |
+
"label2id": {
|
| 19 |
+
"LABEL_0": 0
|
| 20 |
+
},
|
| 21 |
+
"layer_norm_eps": 1e-05,
|
| 22 |
+
"max_position_embeddings": 514,
|
| 23 |
+
"model_type": "xlm-roberta",
|
| 24 |
+
"num_attention_heads": 12,
|
| 25 |
+
"num_hidden_layers": 12,
|
| 26 |
+
"output_past": true,
|
| 27 |
+
"pad_token_id": 1,
|
| 28 |
+
"position_embedding_type": "absolute",
|
| 29 |
+
"torch_dtype": "float32",
|
| 30 |
+
"transformers_version": "4.29.0",
|
| 31 |
+
"type_vocab_size": 1,
|
| 32 |
+
"use_cache": true,
|
| 33 |
+
"vocab_size": 250002
|
| 34 |
+
}
|
checkpoint-45000/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3dc79da63fd55997e655becfcdb1a63e09a8e5bbe27b2fc67ea72d785a285c94
|
| 3 |
+
size 1112246250
|
checkpoint-45000/trainer_state.json
ADDED
|
@@ -0,0 +1,2716 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 14.372404982433727,
|
| 5 |
+
"global_step": 45000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.03,
|
| 12 |
+
"learning_rate": 2.8744809964867454e-07,
|
| 13 |
+
"loss": 91.176,
|
| 14 |
+
"step": 100
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.06,
|
| 18 |
+
"learning_rate": 6.068348770360907e-07,
|
| 19 |
+
"loss": 64.0684,
|
| 20 |
+
"step": 200
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.1,
|
| 24 |
+
"learning_rate": 9.26221654423507e-07,
|
| 25 |
+
"loss": 27.1061,
|
| 26 |
+
"step": 300
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.13,
|
| 30 |
+
"learning_rate": 1.242414564037049e-06,
|
| 31 |
+
"loss": 8.0198,
|
| 32 |
+
"step": 400
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.16,
|
| 36 |
+
"learning_rate": 1.561801341424465e-06,
|
| 37 |
+
"loss": 4.7691,
|
| 38 |
+
"step": 500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.19,
|
| 42 |
+
"learning_rate": 1.8811881188118813e-06,
|
| 43 |
+
"loss": 3.4387,
|
| 44 |
+
"step": 600
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.22,
|
| 48 |
+
"learning_rate": 2.2005748961992977e-06,
|
| 49 |
+
"loss": 2.7027,
|
| 50 |
+
"step": 700
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.26,
|
| 54 |
+
"learning_rate": 2.5199616735867137e-06,
|
| 55 |
+
"loss": 2.4431,
|
| 56 |
+
"step": 800
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.29,
|
| 60 |
+
"learning_rate": 2.8393484509741297e-06,
|
| 61 |
+
"loss": 2.0534,
|
| 62 |
+
"step": 900
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.32,
|
| 66 |
+
"learning_rate": 3.1587352283615465e-06,
|
| 67 |
+
"loss": 1.8537,
|
| 68 |
+
"step": 1000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.35,
|
| 72 |
+
"learning_rate": 3.4781220057489625e-06,
|
| 73 |
+
"loss": 1.642,
|
| 74 |
+
"step": 1100
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.38,
|
| 78 |
+
"learning_rate": 3.797508783136378e-06,
|
| 79 |
+
"loss": 1.5092,
|
| 80 |
+
"step": 1200
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.42,
|
| 84 |
+
"learning_rate": 4.1168955605237945e-06,
|
| 85 |
+
"loss": 1.4302,
|
| 86 |
+
"step": 1300
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.45,
|
| 90 |
+
"learning_rate": 4.436282337911211e-06,
|
| 91 |
+
"loss": 1.3527,
|
| 92 |
+
"step": 1400
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.48,
|
| 96 |
+
"learning_rate": 4.755669115298627e-06,
|
| 97 |
+
"loss": 1.2495,
|
| 98 |
+
"step": 1500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.51,
|
| 102 |
+
"learning_rate": 5.075055892686043e-06,
|
| 103 |
+
"loss": 1.2039,
|
| 104 |
+
"step": 1600
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.54,
|
| 108 |
+
"learning_rate": 5.39444267007346e-06,
|
| 109 |
+
"loss": 1.1438,
|
| 110 |
+
"step": 1700
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.57,
|
| 114 |
+
"learning_rate": 5.713829447460876e-06,
|
| 115 |
+
"loss": 1.1056,
|
| 116 |
+
"step": 1800
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.61,
|
| 120 |
+
"learning_rate": 6.033216224848291e-06,
|
| 121 |
+
"loss": 1.0436,
|
| 122 |
+
"step": 1900
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.64,
|
| 126 |
+
"learning_rate": 6.352603002235708e-06,
|
| 127 |
+
"loss": 1.0225,
|
| 128 |
+
"step": 2000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.67,
|
| 132 |
+
"learning_rate": 6.671989779623123e-06,
|
| 133 |
+
"loss": 0.9856,
|
| 134 |
+
"step": 2100
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.7,
|
| 138 |
+
"learning_rate": 6.9913765570105406e-06,
|
| 139 |
+
"loss": 0.9412,
|
| 140 |
+
"step": 2200
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.73,
|
| 144 |
+
"learning_rate": 7.310763334397957e-06,
|
| 145 |
+
"loss": 0.9075,
|
| 146 |
+
"step": 2300
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.77,
|
| 150 |
+
"learning_rate": 7.630150111785373e-06,
|
| 151 |
+
"loss": 0.8752,
|
| 152 |
+
"step": 2400
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.8,
|
| 156 |
+
"learning_rate": 7.949536889172788e-06,
|
| 157 |
+
"loss": 0.854,
|
| 158 |
+
"step": 2500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.83,
|
| 162 |
+
"learning_rate": 8.268923666560205e-06,
|
| 163 |
+
"loss": 0.8282,
|
| 164 |
+
"step": 2600
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.86,
|
| 168 |
+
"learning_rate": 8.588310443947621e-06,
|
| 169 |
+
"loss": 0.8074,
|
| 170 |
+
"step": 2700
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.89,
|
| 174 |
+
"learning_rate": 8.907697221335038e-06,
|
| 175 |
+
"loss": 0.7976,
|
| 176 |
+
"step": 2800
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.93,
|
| 180 |
+
"learning_rate": 9.227083998722454e-06,
|
| 181 |
+
"loss": 0.7955,
|
| 182 |
+
"step": 2900
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.96,
|
| 186 |
+
"learning_rate": 9.54647077610987e-06,
|
| 187 |
+
"loss": 0.7502,
|
| 188 |
+
"step": 3000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.99,
|
| 192 |
+
"learning_rate": 9.865857553497285e-06,
|
| 193 |
+
"loss": 0.7455,
|
| 194 |
+
"step": 3100
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 1.02,
|
| 198 |
+
"learning_rate": 1.0185244330884702e-05,
|
| 199 |
+
"loss": 0.7177,
|
| 200 |
+
"step": 3200
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 1.05,
|
| 204 |
+
"learning_rate": 1.050463110827212e-05,
|
| 205 |
+
"loss": 0.724,
|
| 206 |
+
"step": 3300
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 1.09,
|
| 210 |
+
"learning_rate": 1.0824017885659535e-05,
|
| 211 |
+
"loss": 0.6975,
|
| 212 |
+
"step": 3400
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 1.12,
|
| 216 |
+
"learning_rate": 1.114340466304695e-05,
|
| 217 |
+
"loss": 0.7054,
|
| 218 |
+
"step": 3500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 1.15,
|
| 222 |
+
"learning_rate": 1.1462791440434366e-05,
|
| 223 |
+
"loss": 0.71,
|
| 224 |
+
"step": 3600
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 1.18,
|
| 228 |
+
"learning_rate": 1.1782178217821782e-05,
|
| 229 |
+
"loss": 0.6825,
|
| 230 |
+
"step": 3700
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 1.21,
|
| 234 |
+
"learning_rate": 1.2101564995209199e-05,
|
| 235 |
+
"loss": 0.6792,
|
| 236 |
+
"step": 3800
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 1.25,
|
| 240 |
+
"learning_rate": 1.2420951772596615e-05,
|
| 241 |
+
"loss": 0.6794,
|
| 242 |
+
"step": 3900
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 1.28,
|
| 246 |
+
"learning_rate": 1.2740338549984032e-05,
|
| 247 |
+
"loss": 0.6672,
|
| 248 |
+
"step": 4000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 1.31,
|
| 252 |
+
"learning_rate": 1.3059725327371449e-05,
|
| 253 |
+
"loss": 0.6356,
|
| 254 |
+
"step": 4100
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 1.34,
|
| 258 |
+
"learning_rate": 1.3379112104758865e-05,
|
| 259 |
+
"loss": 0.6503,
|
| 260 |
+
"step": 4200
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 1.37,
|
| 264 |
+
"learning_rate": 1.369849888214628e-05,
|
| 265 |
+
"loss": 0.6495,
|
| 266 |
+
"step": 4300
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 1.41,
|
| 270 |
+
"learning_rate": 1.4017885659533696e-05,
|
| 271 |
+
"loss": 0.6428,
|
| 272 |
+
"step": 4400
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 1.44,
|
| 276 |
+
"learning_rate": 1.4337272436921113e-05,
|
| 277 |
+
"loss": 0.6171,
|
| 278 |
+
"step": 4500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 1.47,
|
| 282 |
+
"learning_rate": 1.4656659214308529e-05,
|
| 283 |
+
"loss": 0.6084,
|
| 284 |
+
"step": 4600
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 1.5,
|
| 288 |
+
"learning_rate": 1.4976045991695944e-05,
|
| 289 |
+
"loss": 0.6222,
|
| 290 |
+
"step": 4700
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 1.53,
|
| 294 |
+
"learning_rate": 1.529543276908336e-05,
|
| 295 |
+
"loss": 0.5971,
|
| 296 |
+
"step": 4800
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 1.56,
|
| 300 |
+
"learning_rate": 1.5614819546470777e-05,
|
| 301 |
+
"loss": 0.5745,
|
| 302 |
+
"step": 4900
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 1.6,
|
| 306 |
+
"learning_rate": 1.5934206323858196e-05,
|
| 307 |
+
"loss": 0.5753,
|
| 308 |
+
"step": 5000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 1.63,
|
| 312 |
+
"learning_rate": 1.625359310124561e-05,
|
| 313 |
+
"loss": 0.6045,
|
| 314 |
+
"step": 5100
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 1.66,
|
| 318 |
+
"learning_rate": 1.6572979878633027e-05,
|
| 319 |
+
"loss": 0.6006,
|
| 320 |
+
"step": 5200
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 1.69,
|
| 324 |
+
"learning_rate": 1.6892366656020443e-05,
|
| 325 |
+
"loss": 0.6003,
|
| 326 |
+
"step": 5300
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 1.72,
|
| 330 |
+
"learning_rate": 1.7211753433407858e-05,
|
| 331 |
+
"loss": 0.5688,
|
| 332 |
+
"step": 5400
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 1.76,
|
| 336 |
+
"learning_rate": 1.7531140210795274e-05,
|
| 337 |
+
"loss": 0.5662,
|
| 338 |
+
"step": 5500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 1.79,
|
| 342 |
+
"learning_rate": 1.785052698818269e-05,
|
| 343 |
+
"loss": 0.5788,
|
| 344 |
+
"step": 5600
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 1.82,
|
| 348 |
+
"learning_rate": 1.8169913765570105e-05,
|
| 349 |
+
"loss": 0.5541,
|
| 350 |
+
"step": 5700
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 1.85,
|
| 354 |
+
"learning_rate": 1.8489300542957524e-05,
|
| 355 |
+
"loss": 0.5656,
|
| 356 |
+
"step": 5800
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 1.88,
|
| 360 |
+
"learning_rate": 1.880868732034494e-05,
|
| 361 |
+
"loss": 0.5466,
|
| 362 |
+
"step": 5900
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 1.92,
|
| 366 |
+
"learning_rate": 1.9128074097732355e-05,
|
| 367 |
+
"loss": 0.563,
|
| 368 |
+
"step": 6000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 1.95,
|
| 372 |
+
"learning_rate": 1.9447460875119774e-05,
|
| 373 |
+
"loss": 0.5603,
|
| 374 |
+
"step": 6100
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 1.98,
|
| 378 |
+
"learning_rate": 1.976684765250719e-05,
|
| 379 |
+
"loss": 0.5579,
|
| 380 |
+
"step": 6200
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 2.01,
|
| 384 |
+
"learning_rate": 2.0086234429894605e-05,
|
| 385 |
+
"loss": 0.5346,
|
| 386 |
+
"step": 6300
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 2.04,
|
| 390 |
+
"learning_rate": 2.0405621207282017e-05,
|
| 391 |
+
"loss": 0.5038,
|
| 392 |
+
"step": 6400
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 2.08,
|
| 396 |
+
"learning_rate": 2.0725007984669436e-05,
|
| 397 |
+
"loss": 0.5121,
|
| 398 |
+
"step": 6500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 2.11,
|
| 402 |
+
"learning_rate": 2.1044394762056855e-05,
|
| 403 |
+
"loss": 0.5135,
|
| 404 |
+
"step": 6600
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 2.14,
|
| 408 |
+
"learning_rate": 2.1363781539444267e-05,
|
| 409 |
+
"loss": 0.504,
|
| 410 |
+
"step": 6700
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 2.17,
|
| 414 |
+
"learning_rate": 2.1683168316831686e-05,
|
| 415 |
+
"loss": 0.5119,
|
| 416 |
+
"step": 6800
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 2.2,
|
| 420 |
+
"learning_rate": 2.20025550942191e-05,
|
| 421 |
+
"loss": 0.4986,
|
| 422 |
+
"step": 6900
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 2.24,
|
| 426 |
+
"learning_rate": 2.2321941871606518e-05,
|
| 427 |
+
"loss": 0.504,
|
| 428 |
+
"step": 7000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 2.27,
|
| 432 |
+
"learning_rate": 2.2641328648993933e-05,
|
| 433 |
+
"loss": 0.5024,
|
| 434 |
+
"step": 7100
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 2.3,
|
| 438 |
+
"learning_rate": 2.296071542638135e-05,
|
| 439 |
+
"loss": 0.479,
|
| 440 |
+
"step": 7200
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 2.33,
|
| 444 |
+
"learning_rate": 2.3280102203768764e-05,
|
| 445 |
+
"loss": 0.492,
|
| 446 |
+
"step": 7300
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 2.36,
|
| 450 |
+
"learning_rate": 2.3599488981156183e-05,
|
| 451 |
+
"loss": 0.4928,
|
| 452 |
+
"step": 7400
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 2.4,
|
| 456 |
+
"learning_rate": 2.39188757585436e-05,
|
| 457 |
+
"loss": 0.4862,
|
| 458 |
+
"step": 7500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 2.43,
|
| 462 |
+
"learning_rate": 2.4238262535931014e-05,
|
| 463 |
+
"loss": 0.5148,
|
| 464 |
+
"step": 7600
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 2.46,
|
| 468 |
+
"learning_rate": 2.4557649313318433e-05,
|
| 469 |
+
"loss": 0.4922,
|
| 470 |
+
"step": 7700
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 2.49,
|
| 474 |
+
"learning_rate": 2.4877036090705846e-05,
|
| 475 |
+
"loss": 0.4882,
|
| 476 |
+
"step": 7800
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 2.52,
|
| 480 |
+
"learning_rate": 2.5196422868093264e-05,
|
| 481 |
+
"loss": 0.4871,
|
| 482 |
+
"step": 7900
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 2.56,
|
| 486 |
+
"learning_rate": 2.5515809645480677e-05,
|
| 487 |
+
"loss": 0.4967,
|
| 488 |
+
"step": 8000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 2.59,
|
| 492 |
+
"learning_rate": 2.5835196422868096e-05,
|
| 493 |
+
"loss": 0.4834,
|
| 494 |
+
"step": 8100
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 2.62,
|
| 498 |
+
"learning_rate": 2.615458320025551e-05,
|
| 499 |
+
"loss": 0.4769,
|
| 500 |
+
"step": 8200
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 2.65,
|
| 504 |
+
"learning_rate": 2.6473969977642927e-05,
|
| 505 |
+
"loss": 0.4776,
|
| 506 |
+
"step": 8300
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 2.68,
|
| 510 |
+
"learning_rate": 2.6793356755030342e-05,
|
| 511 |
+
"loss": 0.4861,
|
| 512 |
+
"step": 8400
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 2.71,
|
| 516 |
+
"learning_rate": 2.711274353241776e-05,
|
| 517 |
+
"loss": 0.486,
|
| 518 |
+
"step": 8500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 2.75,
|
| 522 |
+
"learning_rate": 2.7432130309805177e-05,
|
| 523 |
+
"loss": 0.4745,
|
| 524 |
+
"step": 8600
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 2.78,
|
| 528 |
+
"learning_rate": 2.7751517087192592e-05,
|
| 529 |
+
"loss": 0.4805,
|
| 530 |
+
"step": 8700
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 2.81,
|
| 534 |
+
"learning_rate": 2.807090386458001e-05,
|
| 535 |
+
"loss": 0.5003,
|
| 536 |
+
"step": 8800
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 2.84,
|
| 540 |
+
"learning_rate": 2.8390290641967424e-05,
|
| 541 |
+
"loss": 0.4639,
|
| 542 |
+
"step": 8900
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 2.87,
|
| 546 |
+
"learning_rate": 2.8709677419354843e-05,
|
| 547 |
+
"loss": 0.4763,
|
| 548 |
+
"step": 9000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 2.91,
|
| 552 |
+
"learning_rate": 2.9029064196742255e-05,
|
| 553 |
+
"loss": 0.4727,
|
| 554 |
+
"step": 9100
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 2.94,
|
| 558 |
+
"learning_rate": 2.9348450974129674e-05,
|
| 559 |
+
"loss": 0.4839,
|
| 560 |
+
"step": 9200
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 2.97,
|
| 564 |
+
"learning_rate": 2.966783775151709e-05,
|
| 565 |
+
"loss": 0.4769,
|
| 566 |
+
"step": 9300
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 3.0,
|
| 570 |
+
"learning_rate": 2.9987224528904505e-05,
|
| 571 |
+
"loss": 0.4641,
|
| 572 |
+
"step": 9400
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 3.03,
|
| 576 |
+
"learning_rate": 3.030661130629192e-05,
|
| 577 |
+
"loss": 0.4282,
|
| 578 |
+
"step": 9500
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 3.07,
|
| 582 |
+
"learning_rate": 3.0625998083679336e-05,
|
| 583 |
+
"loss": 0.4057,
|
| 584 |
+
"step": 9600
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 3.1,
|
| 588 |
+
"learning_rate": 3.094538486106676e-05,
|
| 589 |
+
"loss": 0.4192,
|
| 590 |
+
"step": 9700
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 3.13,
|
| 594 |
+
"learning_rate": 3.126477163845417e-05,
|
| 595 |
+
"loss": 0.4186,
|
| 596 |
+
"step": 9800
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 3.16,
|
| 600 |
+
"learning_rate": 3.158415841584159e-05,
|
| 601 |
+
"loss": 0.4086,
|
| 602 |
+
"step": 9900
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 3.19,
|
| 606 |
+
"learning_rate": 3.1903545193229005e-05,
|
| 607 |
+
"loss": 0.4194,
|
| 608 |
+
"step": 10000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 3.23,
|
| 612 |
+
"learning_rate": 3.221973810284254e-05,
|
| 613 |
+
"loss": 0.4443,
|
| 614 |
+
"step": 10100
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 3.26,
|
| 618 |
+
"learning_rate": 3.2539124880229965e-05,
|
| 619 |
+
"loss": 0.4217,
|
| 620 |
+
"step": 10200
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 3.29,
|
| 624 |
+
"learning_rate": 3.2858511657617374e-05,
|
| 625 |
+
"loss": 0.4275,
|
| 626 |
+
"step": 10300
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 3.32,
|
| 630 |
+
"learning_rate": 3.3177898435004796e-05,
|
| 631 |
+
"loss": 0.4422,
|
| 632 |
+
"step": 10400
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 3.35,
|
| 636 |
+
"learning_rate": 3.3497285212392205e-05,
|
| 637 |
+
"loss": 0.4227,
|
| 638 |
+
"step": 10500
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 3.39,
|
| 642 |
+
"learning_rate": 3.381667198977963e-05,
|
| 643 |
+
"loss": 0.4191,
|
| 644 |
+
"step": 10600
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 3.42,
|
| 648 |
+
"learning_rate": 3.413605876716704e-05,
|
| 649 |
+
"loss": 0.4309,
|
| 650 |
+
"step": 10700
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 3.45,
|
| 654 |
+
"learning_rate": 3.445544554455446e-05,
|
| 655 |
+
"loss": 0.4079,
|
| 656 |
+
"step": 10800
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 3.48,
|
| 660 |
+
"learning_rate": 3.4774832321941874e-05,
|
| 661 |
+
"loss": 0.4281,
|
| 662 |
+
"step": 10900
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 3.51,
|
| 666 |
+
"learning_rate": 3.509421909932929e-05,
|
| 667 |
+
"loss": 0.4137,
|
| 668 |
+
"step": 11000
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 3.55,
|
| 672 |
+
"learning_rate": 3.5413605876716705e-05,
|
| 673 |
+
"loss": 0.4168,
|
| 674 |
+
"step": 11100
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 3.58,
|
| 678 |
+
"learning_rate": 3.573299265410412e-05,
|
| 679 |
+
"loss": 0.4186,
|
| 680 |
+
"step": 11200
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 3.61,
|
| 684 |
+
"learning_rate": 3.605237943149154e-05,
|
| 685 |
+
"loss": 0.4325,
|
| 686 |
+
"step": 11300
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 3.64,
|
| 690 |
+
"learning_rate": 3.637176620887895e-05,
|
| 691 |
+
"loss": 0.4422,
|
| 692 |
+
"step": 11400
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 3.67,
|
| 696 |
+
"learning_rate": 3.6691152986266374e-05,
|
| 697 |
+
"loss": 0.4269,
|
| 698 |
+
"step": 11500
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 3.7,
|
| 702 |
+
"learning_rate": 3.701053976365379e-05,
|
| 703 |
+
"loss": 0.4368,
|
| 704 |
+
"step": 11600
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 3.74,
|
| 708 |
+
"learning_rate": 3.7329926541041205e-05,
|
| 709 |
+
"loss": 0.4207,
|
| 710 |
+
"step": 11700
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 3.77,
|
| 714 |
+
"learning_rate": 3.764931331842862e-05,
|
| 715 |
+
"loss": 0.4365,
|
| 716 |
+
"step": 11800
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 3.8,
|
| 720 |
+
"learning_rate": 3.7968700095816036e-05,
|
| 721 |
+
"loss": 0.4394,
|
| 722 |
+
"step": 11900
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 3.83,
|
| 726 |
+
"learning_rate": 3.828808687320345e-05,
|
| 727 |
+
"loss": 0.4351,
|
| 728 |
+
"step": 12000
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 3.86,
|
| 732 |
+
"learning_rate": 3.860747365059087e-05,
|
| 733 |
+
"loss": 0.4113,
|
| 734 |
+
"step": 12100
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 3.9,
|
| 738 |
+
"learning_rate": 3.892686042797828e-05,
|
| 739 |
+
"loss": 0.4261,
|
| 740 |
+
"step": 12200
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 3.93,
|
| 744 |
+
"learning_rate": 3.92462472053657e-05,
|
| 745 |
+
"loss": 0.4311,
|
| 746 |
+
"step": 12300
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 3.96,
|
| 750 |
+
"learning_rate": 3.956563398275312e-05,
|
| 751 |
+
"loss": 0.4246,
|
| 752 |
+
"step": 12400
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 3.99,
|
| 756 |
+
"learning_rate": 3.988502076014053e-05,
|
| 757 |
+
"loss": 0.4294,
|
| 758 |
+
"step": 12500
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 4.02,
|
| 762 |
+
"learning_rate": 3.997728805138579e-05,
|
| 763 |
+
"loss": 0.3854,
|
| 764 |
+
"step": 12600
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 4.06,
|
| 768 |
+
"learning_rate": 3.994180063167608e-05,
|
| 769 |
+
"loss": 0.3656,
|
| 770 |
+
"step": 12700
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 4.09,
|
| 774 |
+
"learning_rate": 3.9906313211966366e-05,
|
| 775 |
+
"loss": 0.3836,
|
| 776 |
+
"step": 12800
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 4.12,
|
| 780 |
+
"learning_rate": 3.987082579225665e-05,
|
| 781 |
+
"loss": 0.3664,
|
| 782 |
+
"step": 12900
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 4.15,
|
| 786 |
+
"learning_rate": 3.9835338372546935e-05,
|
| 787 |
+
"loss": 0.3884,
|
| 788 |
+
"step": 13000
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 4.18,
|
| 792 |
+
"learning_rate": 3.979985095283722e-05,
|
| 793 |
+
"loss": 0.3764,
|
| 794 |
+
"step": 13100
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 4.22,
|
| 798 |
+
"learning_rate": 3.976436353312751e-05,
|
| 799 |
+
"loss": 0.3733,
|
| 800 |
+
"step": 13200
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 4.25,
|
| 804 |
+
"learning_rate": 3.97288761134178e-05,
|
| 805 |
+
"loss": 0.3682,
|
| 806 |
+
"step": 13300
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 4.28,
|
| 810 |
+
"learning_rate": 3.969338869370809e-05,
|
| 811 |
+
"loss": 0.371,
|
| 812 |
+
"step": 13400
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 4.31,
|
| 816 |
+
"learning_rate": 3.965790127399837e-05,
|
| 817 |
+
"loss": 0.3741,
|
| 818 |
+
"step": 13500
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 4.34,
|
| 822 |
+
"learning_rate": 3.9622413854288656e-05,
|
| 823 |
+
"loss": 0.3757,
|
| 824 |
+
"step": 13600
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 4.38,
|
| 828 |
+
"learning_rate": 3.9586926434578944e-05,
|
| 829 |
+
"loss": 0.3652,
|
| 830 |
+
"step": 13700
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 4.41,
|
| 834 |
+
"learning_rate": 3.955143901486923e-05,
|
| 835 |
+
"loss": 0.3892,
|
| 836 |
+
"step": 13800
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 4.44,
|
| 840 |
+
"learning_rate": 3.951630646935662e-05,
|
| 841 |
+
"loss": 0.3893,
|
| 842 |
+
"step": 13900
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 4.47,
|
| 846 |
+
"learning_rate": 3.948081904964691e-05,
|
| 847 |
+
"loss": 0.3823,
|
| 848 |
+
"step": 14000
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 4.5,
|
| 852 |
+
"learning_rate": 3.944533162993719e-05,
|
| 853 |
+
"loss": 0.3654,
|
| 854 |
+
"step": 14100
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 4.54,
|
| 858 |
+
"learning_rate": 3.9409844210227476e-05,
|
| 859 |
+
"loss": 0.3791,
|
| 860 |
+
"step": 14200
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 4.57,
|
| 864 |
+
"learning_rate": 3.9374356790517764e-05,
|
| 865 |
+
"loss": 0.3918,
|
| 866 |
+
"step": 14300
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 4.6,
|
| 870 |
+
"learning_rate": 3.933886937080805e-05,
|
| 871 |
+
"loss": 0.3713,
|
| 872 |
+
"step": 14400
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 4.63,
|
| 876 |
+
"learning_rate": 3.930338195109834e-05,
|
| 877 |
+
"loss": 0.376,
|
| 878 |
+
"step": 14500
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 4.66,
|
| 882 |
+
"learning_rate": 3.926789453138863e-05,
|
| 883 |
+
"loss": 0.3682,
|
| 884 |
+
"step": 14600
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 4.69,
|
| 888 |
+
"learning_rate": 3.9232407111678915e-05,
|
| 889 |
+
"loss": 0.374,
|
| 890 |
+
"step": 14700
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 4.73,
|
| 894 |
+
"learning_rate": 3.91969196919692e-05,
|
| 895 |
+
"loss": 0.3774,
|
| 896 |
+
"step": 14800
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 4.76,
|
| 900 |
+
"learning_rate": 3.9161432272259484e-05,
|
| 901 |
+
"loss": 0.3806,
|
| 902 |
+
"step": 14900
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 4.79,
|
| 906 |
+
"learning_rate": 3.912594485254977e-05,
|
| 907 |
+
"loss": 0.3886,
|
| 908 |
+
"step": 15000
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 4.82,
|
| 912 |
+
"learning_rate": 3.909045743284006e-05,
|
| 913 |
+
"loss": 0.3742,
|
| 914 |
+
"step": 15100
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 4.85,
|
| 918 |
+
"learning_rate": 3.905497001313035e-05,
|
| 919 |
+
"loss": 0.3714,
|
| 920 |
+
"step": 15200
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 4.89,
|
| 924 |
+
"learning_rate": 3.9019482593420636e-05,
|
| 925 |
+
"loss": 0.3721,
|
| 926 |
+
"step": 15300
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 4.92,
|
| 930 |
+
"learning_rate": 3.8983995173710924e-05,
|
| 931 |
+
"loss": 0.3716,
|
| 932 |
+
"step": 15400
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 4.95,
|
| 936 |
+
"learning_rate": 3.8948507754001205e-05,
|
| 937 |
+
"loss": 0.3728,
|
| 938 |
+
"step": 15500
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 4.98,
|
| 942 |
+
"learning_rate": 3.891302033429149e-05,
|
| 943 |
+
"loss": 0.384,
|
| 944 |
+
"step": 15600
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 5.01,
|
| 948 |
+
"learning_rate": 3.887753291458178e-05,
|
| 949 |
+
"loss": 0.3415,
|
| 950 |
+
"step": 15700
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 5.05,
|
| 954 |
+
"learning_rate": 3.884204549487207e-05,
|
| 955 |
+
"loss": 0.3064,
|
| 956 |
+
"step": 15800
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 5.08,
|
| 960 |
+
"learning_rate": 3.880655807516236e-05,
|
| 961 |
+
"loss": 0.3164,
|
| 962 |
+
"step": 15900
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 5.11,
|
| 966 |
+
"learning_rate": 3.8771070655452645e-05,
|
| 967 |
+
"loss": 0.3215,
|
| 968 |
+
"step": 16000
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 5.14,
|
| 972 |
+
"learning_rate": 3.873558323574293e-05,
|
| 973 |
+
"loss": 0.3234,
|
| 974 |
+
"step": 16100
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 5.17,
|
| 978 |
+
"learning_rate": 3.870009581603322e-05,
|
| 979 |
+
"loss": 0.3119,
|
| 980 |
+
"step": 16200
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 5.21,
|
| 984 |
+
"learning_rate": 3.866460839632351e-05,
|
| 985 |
+
"loss": 0.3145,
|
| 986 |
+
"step": 16300
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 5.24,
|
| 990 |
+
"learning_rate": 3.8629120976613796e-05,
|
| 991 |
+
"loss": 0.3333,
|
| 992 |
+
"step": 16400
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 5.27,
|
| 996 |
+
"learning_rate": 3.8593633556904084e-05,
|
| 997 |
+
"loss": 0.3195,
|
| 998 |
+
"step": 16500
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 5.3,
|
| 1002 |
+
"learning_rate": 3.855814613719437e-05,
|
| 1003 |
+
"loss": 0.3159,
|
| 1004 |
+
"step": 16600
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 5.33,
|
| 1008 |
+
"learning_rate": 3.852265871748465e-05,
|
| 1009 |
+
"loss": 0.3189,
|
| 1010 |
+
"step": 16700
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 5.37,
|
| 1014 |
+
"learning_rate": 3.848717129777494e-05,
|
| 1015 |
+
"loss": 0.317,
|
| 1016 |
+
"step": 16800
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 5.4,
|
| 1020 |
+
"learning_rate": 3.845168387806523e-05,
|
| 1021 |
+
"loss": 0.3231,
|
| 1022 |
+
"step": 16900
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 5.43,
|
| 1026 |
+
"learning_rate": 3.841619645835552e-05,
|
| 1027 |
+
"loss": 0.3328,
|
| 1028 |
+
"step": 17000
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 5.46,
|
| 1032 |
+
"learning_rate": 3.8380709038645805e-05,
|
| 1033 |
+
"loss": 0.3282,
|
| 1034 |
+
"step": 17100
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 5.49,
|
| 1038 |
+
"learning_rate": 3.8345221618936086e-05,
|
| 1039 |
+
"loss": 0.3169,
|
| 1040 |
+
"step": 17200
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 5.53,
|
| 1044 |
+
"learning_rate": 3.8309734199226374e-05,
|
| 1045 |
+
"loss": 0.3156,
|
| 1046 |
+
"step": 17300
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 5.56,
|
| 1050 |
+
"learning_rate": 3.827424677951666e-05,
|
| 1051 |
+
"loss": 0.3313,
|
| 1052 |
+
"step": 17400
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 5.59,
|
| 1056 |
+
"learning_rate": 3.823875935980695e-05,
|
| 1057 |
+
"loss": 0.3181,
|
| 1058 |
+
"step": 17500
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 5.62,
|
| 1062 |
+
"learning_rate": 3.820327194009724e-05,
|
| 1063 |
+
"loss": 0.334,
|
| 1064 |
+
"step": 17600
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 5.65,
|
| 1068 |
+
"learning_rate": 3.8167784520387526e-05,
|
| 1069 |
+
"loss": 0.3313,
|
| 1070 |
+
"step": 17700
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 5.69,
|
| 1074 |
+
"learning_rate": 3.8132297100677814e-05,
|
| 1075 |
+
"loss": 0.3331,
|
| 1076 |
+
"step": 17800
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 5.72,
|
| 1080 |
+
"learning_rate": 3.80968096809681e-05,
|
| 1081 |
+
"loss": 0.3405,
|
| 1082 |
+
"step": 17900
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 5.75,
|
| 1086 |
+
"learning_rate": 3.806132226125839e-05,
|
| 1087 |
+
"loss": 0.3345,
|
| 1088 |
+
"step": 18000
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 5.78,
|
| 1092 |
+
"learning_rate": 3.802583484154868e-05,
|
| 1093 |
+
"loss": 0.3295,
|
| 1094 |
+
"step": 18100
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 5.81,
|
| 1098 |
+
"learning_rate": 3.7990347421838965e-05,
|
| 1099 |
+
"loss": 0.3244,
|
| 1100 |
+
"step": 18200
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 5.84,
|
| 1104 |
+
"learning_rate": 3.795486000212925e-05,
|
| 1105 |
+
"loss": 0.3346,
|
| 1106 |
+
"step": 18300
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 5.88,
|
| 1110 |
+
"learning_rate": 3.791972745661663e-05,
|
| 1111 |
+
"loss": 0.3317,
|
| 1112 |
+
"step": 18400
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 5.91,
|
| 1116 |
+
"learning_rate": 3.788424003690692e-05,
|
| 1117 |
+
"loss": 0.3219,
|
| 1118 |
+
"step": 18500
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 5.94,
|
| 1122 |
+
"learning_rate": 3.784875261719721e-05,
|
| 1123 |
+
"loss": 0.3167,
|
| 1124 |
+
"step": 18600
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 5.97,
|
| 1128 |
+
"learning_rate": 3.781326519748749e-05,
|
| 1129 |
+
"loss": 0.33,
|
| 1130 |
+
"step": 18700
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 6.0,
|
| 1134 |
+
"learning_rate": 3.777777777777778e-05,
|
| 1135 |
+
"loss": 0.3211,
|
| 1136 |
+
"step": 18800
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 6.04,
|
| 1140 |
+
"learning_rate": 3.7742290358068066e-05,
|
| 1141 |
+
"loss": 0.2794,
|
| 1142 |
+
"step": 18900
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 6.07,
|
| 1146 |
+
"learning_rate": 3.7706802938358354e-05,
|
| 1147 |
+
"loss": 0.2675,
|
| 1148 |
+
"step": 19000
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 6.1,
|
| 1152 |
+
"learning_rate": 3.767131551864864e-05,
|
| 1153 |
+
"loss": 0.2676,
|
| 1154 |
+
"step": 19100
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 6.13,
|
| 1158 |
+
"learning_rate": 3.763582809893893e-05,
|
| 1159 |
+
"loss": 0.2693,
|
| 1160 |
+
"step": 19200
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 6.16,
|
| 1164 |
+
"learning_rate": 3.760034067922922e-05,
|
| 1165 |
+
"loss": 0.2669,
|
| 1166 |
+
"step": 19300
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 6.2,
|
| 1170 |
+
"learning_rate": 3.7564853259519506e-05,
|
| 1171 |
+
"loss": 0.2744,
|
| 1172 |
+
"step": 19400
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 6.23,
|
| 1176 |
+
"learning_rate": 3.7529365839809794e-05,
|
| 1177 |
+
"loss": 0.282,
|
| 1178 |
+
"step": 19500
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 6.26,
|
| 1182 |
+
"learning_rate": 3.749387842010008e-05,
|
| 1183 |
+
"loss": 0.2848,
|
| 1184 |
+
"step": 19600
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 6.29,
|
| 1188 |
+
"learning_rate": 3.745839100039037e-05,
|
| 1189 |
+
"loss": 0.2783,
|
| 1190 |
+
"step": 19700
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 6.32,
|
| 1194 |
+
"learning_rate": 3.742290358068066e-05,
|
| 1195 |
+
"loss": 0.2843,
|
| 1196 |
+
"step": 19800
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 6.36,
|
| 1200 |
+
"learning_rate": 3.738741616097094e-05,
|
| 1201 |
+
"loss": 0.2696,
|
| 1202 |
+
"step": 19900
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 6.39,
|
| 1206 |
+
"learning_rate": 3.7351928741261226e-05,
|
| 1207 |
+
"loss": 0.2932,
|
| 1208 |
+
"step": 20000
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 6.42,
|
| 1212 |
+
"learning_rate": 3.7316441321551514e-05,
|
| 1213 |
+
"loss": 0.2775,
|
| 1214 |
+
"step": 20100
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 6.45,
|
| 1218 |
+
"learning_rate": 3.72809539018418e-05,
|
| 1219 |
+
"loss": 0.2893,
|
| 1220 |
+
"step": 20200
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 6.48,
|
| 1224 |
+
"learning_rate": 3.724546648213209e-05,
|
| 1225 |
+
"loss": 0.286,
|
| 1226 |
+
"step": 20300
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 6.52,
|
| 1230 |
+
"learning_rate": 3.720997906242237e-05,
|
| 1231 |
+
"loss": 0.2799,
|
| 1232 |
+
"step": 20400
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 6.55,
|
| 1236 |
+
"learning_rate": 3.717449164271266e-05,
|
| 1237 |
+
"loss": 0.2827,
|
| 1238 |
+
"step": 20500
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 6.58,
|
| 1242 |
+
"learning_rate": 3.713900422300295e-05,
|
| 1243 |
+
"loss": 0.2783,
|
| 1244 |
+
"step": 20600
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 6.61,
|
| 1248 |
+
"learning_rate": 3.7103516803293235e-05,
|
| 1249 |
+
"loss": 0.2729,
|
| 1250 |
+
"step": 20700
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 6.64,
|
| 1254 |
+
"learning_rate": 3.706802938358352e-05,
|
| 1255 |
+
"loss": 0.2747,
|
| 1256 |
+
"step": 20800
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 6.68,
|
| 1260 |
+
"learning_rate": 3.703254196387381e-05,
|
| 1261 |
+
"loss": 0.2764,
|
| 1262 |
+
"step": 20900
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 6.71,
|
| 1266 |
+
"learning_rate": 3.699705454416409e-05,
|
| 1267 |
+
"loss": 0.2804,
|
| 1268 |
+
"step": 21000
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 6.74,
|
| 1272 |
+
"learning_rate": 3.696156712445438e-05,
|
| 1273 |
+
"loss": 0.2841,
|
| 1274 |
+
"step": 21100
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 6.77,
|
| 1278 |
+
"learning_rate": 3.692607970474467e-05,
|
| 1279 |
+
"loss": 0.277,
|
| 1280 |
+
"step": 21200
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 6.8,
|
| 1284 |
+
"learning_rate": 3.6890592285034956e-05,
|
| 1285 |
+
"loss": 0.284,
|
| 1286 |
+
"step": 21300
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 6.83,
|
| 1290 |
+
"learning_rate": 3.6855104865325244e-05,
|
| 1291 |
+
"loss": 0.2778,
|
| 1292 |
+
"step": 21400
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 6.87,
|
| 1296 |
+
"learning_rate": 3.681961744561553e-05,
|
| 1297 |
+
"loss": 0.2797,
|
| 1298 |
+
"step": 21500
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 6.9,
|
| 1302 |
+
"learning_rate": 3.678413002590582e-05,
|
| 1303 |
+
"loss": 0.3036,
|
| 1304 |
+
"step": 21600
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 6.93,
|
| 1308 |
+
"learning_rate": 3.674864260619611e-05,
|
| 1309 |
+
"loss": 0.2869,
|
| 1310 |
+
"step": 21700
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 6.96,
|
| 1314 |
+
"learning_rate": 3.6713155186486395e-05,
|
| 1315 |
+
"loss": 0.2921,
|
| 1316 |
+
"step": 21800
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 6.99,
|
| 1320 |
+
"learning_rate": 3.667766776677668e-05,
|
| 1321 |
+
"loss": 0.2844,
|
| 1322 |
+
"step": 21900
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 7.03,
|
| 1326 |
+
"learning_rate": 3.664218034706697e-05,
|
| 1327 |
+
"loss": 0.2427,
|
| 1328 |
+
"step": 22000
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 7.06,
|
| 1332 |
+
"learning_rate": 3.660669292735726e-05,
|
| 1333 |
+
"loss": 0.243,
|
| 1334 |
+
"step": 22100
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 7.09,
|
| 1338 |
+
"learning_rate": 3.657120550764754e-05,
|
| 1339 |
+
"loss": 0.2311,
|
| 1340 |
+
"step": 22200
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 7.12,
|
| 1344 |
+
"learning_rate": 3.653571808793783e-05,
|
| 1345 |
+
"loss": 0.2358,
|
| 1346 |
+
"step": 22300
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 7.15,
|
| 1350 |
+
"learning_rate": 3.6500230668228116e-05,
|
| 1351 |
+
"loss": 0.2287,
|
| 1352 |
+
"step": 22400
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 7.19,
|
| 1356 |
+
"learning_rate": 3.6464743248518404e-05,
|
| 1357 |
+
"loss": 0.2519,
|
| 1358 |
+
"step": 22500
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 7.22,
|
| 1362 |
+
"learning_rate": 3.642925582880869e-05,
|
| 1363 |
+
"loss": 0.2528,
|
| 1364 |
+
"step": 22600
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 7.25,
|
| 1368 |
+
"learning_rate": 3.639376840909898e-05,
|
| 1369 |
+
"loss": 0.2507,
|
| 1370 |
+
"step": 22700
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 7.28,
|
| 1374 |
+
"learning_rate": 3.635828098938926e-05,
|
| 1375 |
+
"loss": 0.2496,
|
| 1376 |
+
"step": 22800
|
| 1377 |
+
},
|
| 1378 |
+
{
|
| 1379 |
+
"epoch": 7.31,
|
| 1380 |
+
"learning_rate": 3.632279356967955e-05,
|
| 1381 |
+
"loss": 0.2503,
|
| 1382 |
+
"step": 22900
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 7.35,
|
| 1386 |
+
"learning_rate": 3.628730614996984e-05,
|
| 1387 |
+
"loss": 0.2383,
|
| 1388 |
+
"step": 23000
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 7.38,
|
| 1392 |
+
"learning_rate": 3.6251818730260125e-05,
|
| 1393 |
+
"loss": 0.2513,
|
| 1394 |
+
"step": 23100
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 7.41,
|
| 1398 |
+
"learning_rate": 3.621633131055041e-05,
|
| 1399 |
+
"loss": 0.2492,
|
| 1400 |
+
"step": 23200
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 7.44,
|
| 1404 |
+
"learning_rate": 3.61808438908407e-05,
|
| 1405 |
+
"loss": 0.2473,
|
| 1406 |
+
"step": 23300
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 7.47,
|
| 1410 |
+
"learning_rate": 3.614535647113099e-05,
|
| 1411 |
+
"loss": 0.2502,
|
| 1412 |
+
"step": 23400
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 7.51,
|
| 1416 |
+
"learning_rate": 3.6109869051421276e-05,
|
| 1417 |
+
"loss": 0.2492,
|
| 1418 |
+
"step": 23500
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 7.54,
|
| 1422 |
+
"learning_rate": 3.6074381631711564e-05,
|
| 1423 |
+
"loss": 0.2366,
|
| 1424 |
+
"step": 23600
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 7.57,
|
| 1428 |
+
"learning_rate": 3.603889421200185e-05,
|
| 1429 |
+
"loss": 0.2393,
|
| 1430 |
+
"step": 23700
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 7.6,
|
| 1434 |
+
"learning_rate": 3.600340679229214e-05,
|
| 1435 |
+
"loss": 0.2483,
|
| 1436 |
+
"step": 23800
|
| 1437 |
+
},
|
| 1438 |
+
{
|
| 1439 |
+
"epoch": 7.63,
|
| 1440 |
+
"learning_rate": 3.596791937258242e-05,
|
| 1441 |
+
"loss": 0.2428,
|
| 1442 |
+
"step": 23900
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 7.67,
|
| 1446 |
+
"learning_rate": 3.593243195287271e-05,
|
| 1447 |
+
"loss": 0.2507,
|
| 1448 |
+
"step": 24000
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 7.7,
|
| 1452 |
+
"learning_rate": 3.5896944533163e-05,
|
| 1453 |
+
"loss": 0.2491,
|
| 1454 |
+
"step": 24100
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 7.73,
|
| 1458 |
+
"learning_rate": 3.5861457113453285e-05,
|
| 1459 |
+
"loss": 0.2549,
|
| 1460 |
+
"step": 24200
|
| 1461 |
+
},
|
| 1462 |
+
{
|
| 1463 |
+
"epoch": 7.76,
|
| 1464 |
+
"learning_rate": 3.582596969374357e-05,
|
| 1465 |
+
"loss": 0.251,
|
| 1466 |
+
"step": 24300
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 7.79,
|
| 1470 |
+
"learning_rate": 3.579048227403386e-05,
|
| 1471 |
+
"loss": 0.2357,
|
| 1472 |
+
"step": 24400
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 7.82,
|
| 1476 |
+
"learning_rate": 3.575499485432414e-05,
|
| 1477 |
+
"loss": 0.2441,
|
| 1478 |
+
"step": 24500
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 7.86,
|
| 1482 |
+
"learning_rate": 3.571950743461443e-05,
|
| 1483 |
+
"loss": 0.2561,
|
| 1484 |
+
"step": 24600
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 7.89,
|
| 1488 |
+
"learning_rate": 3.568402001490472e-05,
|
| 1489 |
+
"loss": 0.2524,
|
| 1490 |
+
"step": 24700
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 7.92,
|
| 1494 |
+
"learning_rate": 3.5648532595195006e-05,
|
| 1495 |
+
"loss": 0.2467,
|
| 1496 |
+
"step": 24800
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 7.95,
|
| 1500 |
+
"learning_rate": 3.5613045175485293e-05,
|
| 1501 |
+
"loss": 0.2416,
|
| 1502 |
+
"step": 24900
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 7.98,
|
| 1506 |
+
"learning_rate": 3.557755775577558e-05,
|
| 1507 |
+
"loss": 0.2511,
|
| 1508 |
+
"step": 25000
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 8.02,
|
| 1512 |
+
"learning_rate": 3.554207033606587e-05,
|
| 1513 |
+
"loss": 0.2323,
|
| 1514 |
+
"step": 25100
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 8.05,
|
| 1518 |
+
"learning_rate": 3.550658291635615e-05,
|
| 1519 |
+
"loss": 0.2158,
|
| 1520 |
+
"step": 25200
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 8.08,
|
| 1524 |
+
"learning_rate": 3.547109549664644e-05,
|
| 1525 |
+
"loss": 0.21,
|
| 1526 |
+
"step": 25300
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 8.11,
|
| 1530 |
+
"learning_rate": 3.5435608076936726e-05,
|
| 1531 |
+
"loss": 0.2102,
|
| 1532 |
+
"step": 25400
|
| 1533 |
+
},
|
| 1534 |
+
{
|
| 1535 |
+
"epoch": 8.14,
|
| 1536 |
+
"learning_rate": 3.5400120657227014e-05,
|
| 1537 |
+
"loss": 0.2071,
|
| 1538 |
+
"step": 25500
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 8.18,
|
| 1542 |
+
"learning_rate": 3.53646332375173e-05,
|
| 1543 |
+
"loss": 0.2198,
|
| 1544 |
+
"step": 25600
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 8.21,
|
| 1548 |
+
"learning_rate": 3.532914581780759e-05,
|
| 1549 |
+
"loss": 0.2193,
|
| 1550 |
+
"step": 25700
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 8.24,
|
| 1554 |
+
"learning_rate": 3.529365839809788e-05,
|
| 1555 |
+
"loss": 0.2166,
|
| 1556 |
+
"step": 25800
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 8.27,
|
| 1560 |
+
"learning_rate": 3.5258170978388166e-05,
|
| 1561 |
+
"loss": 0.2237,
|
| 1562 |
+
"step": 25900
|
| 1563 |
+
},
|
| 1564 |
+
{
|
| 1565 |
+
"epoch": 8.3,
|
| 1566 |
+
"learning_rate": 3.5222683558678454e-05,
|
| 1567 |
+
"loss": 0.2241,
|
| 1568 |
+
"step": 26000
|
| 1569 |
+
},
|
| 1570 |
+
{
|
| 1571 |
+
"epoch": 8.34,
|
| 1572 |
+
"learning_rate": 3.518719613896874e-05,
|
| 1573 |
+
"loss": 0.2132,
|
| 1574 |
+
"step": 26100
|
| 1575 |
+
},
|
| 1576 |
+
{
|
| 1577 |
+
"epoch": 8.37,
|
| 1578 |
+
"learning_rate": 3.515170871925902e-05,
|
| 1579 |
+
"loss": 0.2101,
|
| 1580 |
+
"step": 26200
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 8.4,
|
| 1584 |
+
"learning_rate": 3.511622129954931e-05,
|
| 1585 |
+
"loss": 0.227,
|
| 1586 |
+
"step": 26300
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 8.43,
|
| 1590 |
+
"learning_rate": 3.50807338798396e-05,
|
| 1591 |
+
"loss": 0.2221,
|
| 1592 |
+
"step": 26400
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 8.46,
|
| 1596 |
+
"learning_rate": 3.504560133432698e-05,
|
| 1597 |
+
"loss": 0.2181,
|
| 1598 |
+
"step": 26500
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 8.5,
|
| 1602 |
+
"learning_rate": 3.501011391461727e-05,
|
| 1603 |
+
"loss": 0.2208,
|
| 1604 |
+
"step": 26600
|
| 1605 |
+
},
|
| 1606 |
+
{
|
| 1607 |
+
"epoch": 8.53,
|
| 1608 |
+
"learning_rate": 3.4974626494907555e-05,
|
| 1609 |
+
"loss": 0.2192,
|
| 1610 |
+
"step": 26700
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 8.56,
|
| 1614 |
+
"learning_rate": 3.493913907519784e-05,
|
| 1615 |
+
"loss": 0.2227,
|
| 1616 |
+
"step": 26800
|
| 1617 |
+
},
|
| 1618 |
+
{
|
| 1619 |
+
"epoch": 8.59,
|
| 1620 |
+
"learning_rate": 3.490365165548813e-05,
|
| 1621 |
+
"loss": 0.2308,
|
| 1622 |
+
"step": 26900
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 8.62,
|
| 1626 |
+
"learning_rate": 3.486816423577842e-05,
|
| 1627 |
+
"loss": 0.22,
|
| 1628 |
+
"step": 27000
|
| 1629 |
+
},
|
| 1630 |
+
{
|
| 1631 |
+
"epoch": 8.66,
|
| 1632 |
+
"learning_rate": 3.4832676816068706e-05,
|
| 1633 |
+
"loss": 0.2247,
|
| 1634 |
+
"step": 27100
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 8.69,
|
| 1638 |
+
"learning_rate": 3.4797189396358994e-05,
|
| 1639 |
+
"loss": 0.2148,
|
| 1640 |
+
"step": 27200
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 8.72,
|
| 1644 |
+
"learning_rate": 3.476170197664928e-05,
|
| 1645 |
+
"loss": 0.2136,
|
| 1646 |
+
"step": 27300
|
| 1647 |
+
},
|
| 1648 |
+
{
|
| 1649 |
+
"epoch": 8.75,
|
| 1650 |
+
"learning_rate": 3.472621455693957e-05,
|
| 1651 |
+
"loss": 0.2282,
|
| 1652 |
+
"step": 27400
|
| 1653 |
+
},
|
| 1654 |
+
{
|
| 1655 |
+
"epoch": 8.78,
|
| 1656 |
+
"learning_rate": 3.469072713722986e-05,
|
| 1657 |
+
"loss": 0.2266,
|
| 1658 |
+
"step": 27500
|
| 1659 |
+
},
|
| 1660 |
+
{
|
| 1661 |
+
"epoch": 8.82,
|
| 1662 |
+
"learning_rate": 3.4655239717520146e-05,
|
| 1663 |
+
"loss": 0.2198,
|
| 1664 |
+
"step": 27600
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 8.85,
|
| 1668 |
+
"learning_rate": 3.461975229781043e-05,
|
| 1669 |
+
"loss": 0.2243,
|
| 1670 |
+
"step": 27700
|
| 1671 |
+
},
|
| 1672 |
+
{
|
| 1673 |
+
"epoch": 8.88,
|
| 1674 |
+
"learning_rate": 3.4584264878100715e-05,
|
| 1675 |
+
"loss": 0.2358,
|
| 1676 |
+
"step": 27800
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 8.91,
|
| 1680 |
+
"learning_rate": 3.4548777458391e-05,
|
| 1681 |
+
"loss": 0.2255,
|
| 1682 |
+
"step": 27900
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 8.94,
|
| 1686 |
+
"learning_rate": 3.451329003868129e-05,
|
| 1687 |
+
"loss": 0.2258,
|
| 1688 |
+
"step": 28000
|
| 1689 |
+
},
|
| 1690 |
+
{
|
| 1691 |
+
"epoch": 8.97,
|
| 1692 |
+
"learning_rate": 3.447780261897158e-05,
|
| 1693 |
+
"loss": 0.2228,
|
| 1694 |
+
"step": 28100
|
| 1695 |
+
},
|
| 1696 |
+
{
|
| 1697 |
+
"epoch": 9.01,
|
| 1698 |
+
"learning_rate": 3.4442315199261866e-05,
|
| 1699 |
+
"loss": 0.2183,
|
| 1700 |
+
"step": 28200
|
| 1701 |
+
},
|
| 1702 |
+
{
|
| 1703 |
+
"epoch": 9.04,
|
| 1704 |
+
"learning_rate": 3.440682777955215e-05,
|
| 1705 |
+
"loss": 0.1988,
|
| 1706 |
+
"step": 28300
|
| 1707 |
+
},
|
| 1708 |
+
{
|
| 1709 |
+
"epoch": 9.07,
|
| 1710 |
+
"learning_rate": 3.4371340359842436e-05,
|
| 1711 |
+
"loss": 0.188,
|
| 1712 |
+
"step": 28400
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"epoch": 9.1,
|
| 1716 |
+
"learning_rate": 3.4335852940132723e-05,
|
| 1717 |
+
"loss": 0.1963,
|
| 1718 |
+
"step": 28500
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 9.13,
|
| 1722 |
+
"learning_rate": 3.430036552042301e-05,
|
| 1723 |
+
"loss": 0.1919,
|
| 1724 |
+
"step": 28600
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 9.17,
|
| 1728 |
+
"learning_rate": 3.42648781007133e-05,
|
| 1729 |
+
"loss": 0.1906,
|
| 1730 |
+
"step": 28700
|
| 1731 |
+
},
|
| 1732 |
+
{
|
| 1733 |
+
"epoch": 9.2,
|
| 1734 |
+
"learning_rate": 3.422939068100359e-05,
|
| 1735 |
+
"loss": 0.1911,
|
| 1736 |
+
"step": 28800
|
| 1737 |
+
},
|
| 1738 |
+
{
|
| 1739 |
+
"epoch": 9.23,
|
| 1740 |
+
"learning_rate": 3.4193903261293875e-05,
|
| 1741 |
+
"loss": 0.1971,
|
| 1742 |
+
"step": 28900
|
| 1743 |
+
},
|
| 1744 |
+
{
|
| 1745 |
+
"epoch": 9.26,
|
| 1746 |
+
"learning_rate": 3.415841584158416e-05,
|
| 1747 |
+
"loss": 0.1931,
|
| 1748 |
+
"step": 29000
|
| 1749 |
+
},
|
| 1750 |
+
{
|
| 1751 |
+
"epoch": 9.29,
|
| 1752 |
+
"learning_rate": 3.412292842187445e-05,
|
| 1753 |
+
"loss": 0.1951,
|
| 1754 |
+
"step": 29100
|
| 1755 |
+
},
|
| 1756 |
+
{
|
| 1757 |
+
"epoch": 9.33,
|
| 1758 |
+
"learning_rate": 3.408744100216474e-05,
|
| 1759 |
+
"loss": 0.2008,
|
| 1760 |
+
"step": 29200
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 9.36,
|
| 1764 |
+
"learning_rate": 3.405195358245503e-05,
|
| 1765 |
+
"loss": 0.2053,
|
| 1766 |
+
"step": 29300
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 9.39,
|
| 1770 |
+
"learning_rate": 3.401646616274531e-05,
|
| 1771 |
+
"loss": 0.2015,
|
| 1772 |
+
"step": 29400
|
| 1773 |
+
},
|
| 1774 |
+
{
|
| 1775 |
+
"epoch": 9.42,
|
| 1776 |
+
"learning_rate": 3.3980978743035596e-05,
|
| 1777 |
+
"loss": 0.193,
|
| 1778 |
+
"step": 29500
|
| 1779 |
+
},
|
| 1780 |
+
{
|
| 1781 |
+
"epoch": 9.45,
|
| 1782 |
+
"learning_rate": 3.3945491323325884e-05,
|
| 1783 |
+
"loss": 0.1955,
|
| 1784 |
+
"step": 29600
|
| 1785 |
+
},
|
| 1786 |
+
{
|
| 1787 |
+
"epoch": 9.49,
|
| 1788 |
+
"learning_rate": 3.391000390361617e-05,
|
| 1789 |
+
"loss": 0.1993,
|
| 1790 |
+
"step": 29700
|
| 1791 |
+
},
|
| 1792 |
+
{
|
| 1793 |
+
"epoch": 9.52,
|
| 1794 |
+
"learning_rate": 3.387451648390646e-05,
|
| 1795 |
+
"loss": 0.2043,
|
| 1796 |
+
"step": 29800
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"epoch": 9.55,
|
| 1800 |
+
"learning_rate": 3.383902906419675e-05,
|
| 1801 |
+
"loss": 0.1955,
|
| 1802 |
+
"step": 29900
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 9.58,
|
| 1806 |
+
"learning_rate": 3.380354164448703e-05,
|
| 1807 |
+
"loss": 0.1961,
|
| 1808 |
+
"step": 30000
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 9.61,
|
| 1812 |
+
"learning_rate": 3.3768054224777316e-05,
|
| 1813 |
+
"loss": 0.1961,
|
| 1814 |
+
"step": 30100
|
| 1815 |
+
},
|
| 1816 |
+
{
|
| 1817 |
+
"epoch": 9.65,
|
| 1818 |
+
"learning_rate": 3.3732566805067604e-05,
|
| 1819 |
+
"loss": 0.196,
|
| 1820 |
+
"step": 30200
|
| 1821 |
+
},
|
| 1822 |
+
{
|
| 1823 |
+
"epoch": 9.68,
|
| 1824 |
+
"learning_rate": 3.369707938535789e-05,
|
| 1825 |
+
"loss": 0.2088,
|
| 1826 |
+
"step": 30300
|
| 1827 |
+
},
|
| 1828 |
+
{
|
| 1829 |
+
"epoch": 9.71,
|
| 1830 |
+
"learning_rate": 3.366159196564818e-05,
|
| 1831 |
+
"loss": 0.2031,
|
| 1832 |
+
"step": 30400
|
| 1833 |
+
},
|
| 1834 |
+
{
|
| 1835 |
+
"epoch": 9.74,
|
| 1836 |
+
"learning_rate": 3.362610454593847e-05,
|
| 1837 |
+
"loss": 0.1949,
|
| 1838 |
+
"step": 30500
|
| 1839 |
+
},
|
| 1840 |
+
{
|
| 1841 |
+
"epoch": 9.77,
|
| 1842 |
+
"learning_rate": 3.3590617126228756e-05,
|
| 1843 |
+
"loss": 0.1988,
|
| 1844 |
+
"step": 30600
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 9.81,
|
| 1848 |
+
"learning_rate": 3.3555129706519044e-05,
|
| 1849 |
+
"loss": 0.2061,
|
| 1850 |
+
"step": 30700
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 9.84,
|
| 1854 |
+
"learning_rate": 3.351964228680933e-05,
|
| 1855 |
+
"loss": 0.2022,
|
| 1856 |
+
"step": 30800
|
| 1857 |
+
},
|
| 1858 |
+
{
|
| 1859 |
+
"epoch": 9.87,
|
| 1860 |
+
"learning_rate": 3.348450974129671e-05,
|
| 1861 |
+
"loss": 0.1965,
|
| 1862 |
+
"step": 30900
|
| 1863 |
+
},
|
| 1864 |
+
{
|
| 1865 |
+
"epoch": 9.9,
|
| 1866 |
+
"learning_rate": 3.3449022321587e-05,
|
| 1867 |
+
"loss": 0.1999,
|
| 1868 |
+
"step": 31000
|
| 1869 |
+
},
|
| 1870 |
+
{
|
| 1871 |
+
"epoch": 9.93,
|
| 1872 |
+
"learning_rate": 3.341353490187729e-05,
|
| 1873 |
+
"loss": 0.2065,
|
| 1874 |
+
"step": 31100
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 9.96,
|
| 1878 |
+
"learning_rate": 3.3378047482167576e-05,
|
| 1879 |
+
"loss": 0.2031,
|
| 1880 |
+
"step": 31200
|
| 1881 |
+
},
|
| 1882 |
+
{
|
| 1883 |
+
"epoch": 10.0,
|
| 1884 |
+
"learning_rate": 3.3342560062457864e-05,
|
| 1885 |
+
"loss": 0.2078,
|
| 1886 |
+
"step": 31300
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 10.03,
|
| 1890 |
+
"learning_rate": 3.330707264274815e-05,
|
| 1891 |
+
"loss": 0.1733,
|
| 1892 |
+
"step": 31400
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 10.06,
|
| 1896 |
+
"learning_rate": 3.327158522303843e-05,
|
| 1897 |
+
"loss": 0.1654,
|
| 1898 |
+
"step": 31500
|
| 1899 |
+
},
|
| 1900 |
+
{
|
| 1901 |
+
"epoch": 10.09,
|
| 1902 |
+
"learning_rate": 3.323609780332872e-05,
|
| 1903 |
+
"loss": 0.1674,
|
| 1904 |
+
"step": 31600
|
| 1905 |
+
},
|
| 1906 |
+
{
|
| 1907 |
+
"epoch": 10.12,
|
| 1908 |
+
"learning_rate": 3.320061038361901e-05,
|
| 1909 |
+
"loss": 0.1725,
|
| 1910 |
+
"step": 31700
|
| 1911 |
+
},
|
| 1912 |
+
{
|
| 1913 |
+
"epoch": 10.16,
|
| 1914 |
+
"learning_rate": 3.3165122963909297e-05,
|
| 1915 |
+
"loss": 0.1761,
|
| 1916 |
+
"step": 31800
|
| 1917 |
+
},
|
| 1918 |
+
{
|
| 1919 |
+
"epoch": 10.19,
|
| 1920 |
+
"learning_rate": 3.3129635544199584e-05,
|
| 1921 |
+
"loss": 0.1757,
|
| 1922 |
+
"step": 31900
|
| 1923 |
+
},
|
| 1924 |
+
{
|
| 1925 |
+
"epoch": 10.22,
|
| 1926 |
+
"learning_rate": 3.309414812448987e-05,
|
| 1927 |
+
"loss": 0.1781,
|
| 1928 |
+
"step": 32000
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 10.25,
|
| 1932 |
+
"learning_rate": 3.3058660704780153e-05,
|
| 1933 |
+
"loss": 0.1716,
|
| 1934 |
+
"step": 32100
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 10.28,
|
| 1938 |
+
"learning_rate": 3.302317328507044e-05,
|
| 1939 |
+
"loss": 0.1717,
|
| 1940 |
+
"step": 32200
|
| 1941 |
+
},
|
| 1942 |
+
{
|
| 1943 |
+
"epoch": 10.32,
|
| 1944 |
+
"learning_rate": 3.298768586536073e-05,
|
| 1945 |
+
"loss": 0.179,
|
| 1946 |
+
"step": 32300
|
| 1947 |
+
},
|
| 1948 |
+
{
|
| 1949 |
+
"epoch": 10.35,
|
| 1950 |
+
"learning_rate": 3.295219844565102e-05,
|
| 1951 |
+
"loss": 0.174,
|
| 1952 |
+
"step": 32400
|
| 1953 |
+
},
|
| 1954 |
+
{
|
| 1955 |
+
"epoch": 10.38,
|
| 1956 |
+
"learning_rate": 3.2916711025941305e-05,
|
| 1957 |
+
"loss": 0.178,
|
| 1958 |
+
"step": 32500
|
| 1959 |
+
},
|
| 1960 |
+
{
|
| 1961 |
+
"epoch": 10.41,
|
| 1962 |
+
"learning_rate": 3.288122360623159e-05,
|
| 1963 |
+
"loss": 0.1765,
|
| 1964 |
+
"step": 32600
|
| 1965 |
+
},
|
| 1966 |
+
{
|
| 1967 |
+
"epoch": 10.44,
|
| 1968 |
+
"learning_rate": 3.284573618652188e-05,
|
| 1969 |
+
"loss": 0.1824,
|
| 1970 |
+
"step": 32700
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"epoch": 10.48,
|
| 1974 |
+
"learning_rate": 3.281024876681217e-05,
|
| 1975 |
+
"loss": 0.1806,
|
| 1976 |
+
"step": 32800
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 10.51,
|
| 1980 |
+
"learning_rate": 3.277511622129955e-05,
|
| 1981 |
+
"loss": 0.1785,
|
| 1982 |
+
"step": 32900
|
| 1983 |
+
},
|
| 1984 |
+
{
|
| 1985 |
+
"epoch": 10.54,
|
| 1986 |
+
"learning_rate": 3.2739983675786936e-05,
|
| 1987 |
+
"loss": 0.1698,
|
| 1988 |
+
"step": 33000
|
| 1989 |
+
},
|
| 1990 |
+
{
|
| 1991 |
+
"epoch": 10.57,
|
| 1992 |
+
"learning_rate": 3.2704496256077224e-05,
|
| 1993 |
+
"loss": 0.1734,
|
| 1994 |
+
"step": 33100
|
| 1995 |
+
},
|
| 1996 |
+
{
|
| 1997 |
+
"epoch": 10.6,
|
| 1998 |
+
"learning_rate": 3.2669008836367505e-05,
|
| 1999 |
+
"loss": 0.1761,
|
| 2000 |
+
"step": 33200
|
| 2001 |
+
},
|
| 2002 |
+
{
|
| 2003 |
+
"epoch": 10.64,
|
| 2004 |
+
"learning_rate": 3.263352141665779e-05,
|
| 2005 |
+
"loss": 0.1786,
|
| 2006 |
+
"step": 33300
|
| 2007 |
+
},
|
| 2008 |
+
{
|
| 2009 |
+
"epoch": 10.67,
|
| 2010 |
+
"learning_rate": 3.259803399694808e-05,
|
| 2011 |
+
"loss": 0.1767,
|
| 2012 |
+
"step": 33400
|
| 2013 |
+
},
|
| 2014 |
+
{
|
| 2015 |
+
"epoch": 10.7,
|
| 2016 |
+
"learning_rate": 3.256254657723837e-05,
|
| 2017 |
+
"loss": 0.1881,
|
| 2018 |
+
"step": 33500
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 10.73,
|
| 2022 |
+
"learning_rate": 3.252705915752866e-05,
|
| 2023 |
+
"loss": 0.1792,
|
| 2024 |
+
"step": 33600
|
| 2025 |
+
},
|
| 2026 |
+
{
|
| 2027 |
+
"epoch": 10.76,
|
| 2028 |
+
"learning_rate": 3.2491571737818945e-05,
|
| 2029 |
+
"loss": 0.1832,
|
| 2030 |
+
"step": 33700
|
| 2031 |
+
},
|
| 2032 |
+
{
|
| 2033 |
+
"epoch": 10.8,
|
| 2034 |
+
"learning_rate": 3.245608431810923e-05,
|
| 2035 |
+
"loss": 0.1889,
|
| 2036 |
+
"step": 33800
|
| 2037 |
+
},
|
| 2038 |
+
{
|
| 2039 |
+
"epoch": 10.83,
|
| 2040 |
+
"learning_rate": 3.242059689839952e-05,
|
| 2041 |
+
"loss": 0.1824,
|
| 2042 |
+
"step": 33900
|
| 2043 |
+
},
|
| 2044 |
+
{
|
| 2045 |
+
"epoch": 10.86,
|
| 2046 |
+
"learning_rate": 3.238510947868981e-05,
|
| 2047 |
+
"loss": 0.1746,
|
| 2048 |
+
"step": 34000
|
| 2049 |
+
},
|
| 2050 |
+
{
|
| 2051 |
+
"epoch": 10.89,
|
| 2052 |
+
"learning_rate": 3.2349622058980096e-05,
|
| 2053 |
+
"loss": 0.1735,
|
| 2054 |
+
"step": 34100
|
| 2055 |
+
},
|
| 2056 |
+
{
|
| 2057 |
+
"epoch": 10.92,
|
| 2058 |
+
"learning_rate": 3.2314134639270384e-05,
|
| 2059 |
+
"loss": 0.1822,
|
| 2060 |
+
"step": 34200
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 10.95,
|
| 2064 |
+
"learning_rate": 3.227864721956067e-05,
|
| 2065 |
+
"loss": 0.1768,
|
| 2066 |
+
"step": 34300
|
| 2067 |
+
},
|
| 2068 |
+
{
|
| 2069 |
+
"epoch": 10.99,
|
| 2070 |
+
"learning_rate": 3.224315979985095e-05,
|
| 2071 |
+
"loss": 0.1812,
|
| 2072 |
+
"step": 34400
|
| 2073 |
+
},
|
| 2074 |
+
{
|
| 2075 |
+
"epoch": 11.02,
|
| 2076 |
+
"learning_rate": 3.220767238014124e-05,
|
| 2077 |
+
"loss": 0.1645,
|
| 2078 |
+
"step": 34500
|
| 2079 |
+
},
|
| 2080 |
+
{
|
| 2081 |
+
"epoch": 11.05,
|
| 2082 |
+
"learning_rate": 3.217218496043153e-05,
|
| 2083 |
+
"loss": 0.1483,
|
| 2084 |
+
"step": 34600
|
| 2085 |
+
},
|
| 2086 |
+
{
|
| 2087 |
+
"epoch": 11.08,
|
| 2088 |
+
"learning_rate": 3.213669754072182e-05,
|
| 2089 |
+
"loss": 0.1615,
|
| 2090 |
+
"step": 34700
|
| 2091 |
+
},
|
| 2092 |
+
{
|
| 2093 |
+
"epoch": 11.11,
|
| 2094 |
+
"learning_rate": 3.2101210121012105e-05,
|
| 2095 |
+
"loss": 0.1637,
|
| 2096 |
+
"step": 34800
|
| 2097 |
+
},
|
| 2098 |
+
{
|
| 2099 |
+
"epoch": 11.15,
|
| 2100 |
+
"learning_rate": 3.206572270130239e-05,
|
| 2101 |
+
"loss": 0.1497,
|
| 2102 |
+
"step": 34900
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 11.18,
|
| 2106 |
+
"learning_rate": 3.2030235281592674e-05,
|
| 2107 |
+
"loss": 0.1524,
|
| 2108 |
+
"step": 35000
|
| 2109 |
+
},
|
| 2110 |
+
{
|
| 2111 |
+
"epoch": 11.21,
|
| 2112 |
+
"learning_rate": 3.199474786188296e-05,
|
| 2113 |
+
"loss": 0.1545,
|
| 2114 |
+
"step": 35100
|
| 2115 |
+
},
|
| 2116 |
+
{
|
| 2117 |
+
"epoch": 11.24,
|
| 2118 |
+
"learning_rate": 3.195926044217325e-05,
|
| 2119 |
+
"loss": 0.1526,
|
| 2120 |
+
"step": 35200
|
| 2121 |
+
},
|
| 2122 |
+
{
|
| 2123 |
+
"epoch": 11.27,
|
| 2124 |
+
"learning_rate": 3.192377302246354e-05,
|
| 2125 |
+
"loss": 0.1499,
|
| 2126 |
+
"step": 35300
|
| 2127 |
+
},
|
| 2128 |
+
{
|
| 2129 |
+
"epoch": 11.31,
|
| 2130 |
+
"learning_rate": 3.1888285602753826e-05,
|
| 2131 |
+
"loss": 0.1553,
|
| 2132 |
+
"step": 35400
|
| 2133 |
+
},
|
| 2134 |
+
{
|
| 2135 |
+
"epoch": 11.34,
|
| 2136 |
+
"learning_rate": 3.1852798183044114e-05,
|
| 2137 |
+
"loss": 0.1583,
|
| 2138 |
+
"step": 35500
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 11.37,
|
| 2142 |
+
"learning_rate": 3.18173107633344e-05,
|
| 2143 |
+
"loss": 0.156,
|
| 2144 |
+
"step": 35600
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 11.4,
|
| 2148 |
+
"learning_rate": 3.178182334362469e-05,
|
| 2149 |
+
"loss": 0.1582,
|
| 2150 |
+
"step": 35700
|
| 2151 |
+
},
|
| 2152 |
+
{
|
| 2153 |
+
"epoch": 11.43,
|
| 2154 |
+
"learning_rate": 3.174633592391498e-05,
|
| 2155 |
+
"loss": 0.155,
|
| 2156 |
+
"step": 35800
|
| 2157 |
+
},
|
| 2158 |
+
{
|
| 2159 |
+
"epoch": 11.47,
|
| 2160 |
+
"learning_rate": 3.1710848504205265e-05,
|
| 2161 |
+
"loss": 0.1546,
|
| 2162 |
+
"step": 35900
|
| 2163 |
+
},
|
| 2164 |
+
{
|
| 2165 |
+
"epoch": 11.5,
|
| 2166 |
+
"learning_rate": 3.167536108449555e-05,
|
| 2167 |
+
"loss": 0.157,
|
| 2168 |
+
"step": 36000
|
| 2169 |
+
},
|
| 2170 |
+
{
|
| 2171 |
+
"epoch": 11.53,
|
| 2172 |
+
"learning_rate": 3.1639873664785834e-05,
|
| 2173 |
+
"loss": 0.1644,
|
| 2174 |
+
"step": 36100
|
| 2175 |
+
},
|
| 2176 |
+
{
|
| 2177 |
+
"epoch": 11.56,
|
| 2178 |
+
"learning_rate": 3.160438624507612e-05,
|
| 2179 |
+
"loss": 0.1627,
|
| 2180 |
+
"step": 36200
|
| 2181 |
+
},
|
| 2182 |
+
{
|
| 2183 |
+
"epoch": 11.59,
|
| 2184 |
+
"learning_rate": 3.156889882536641e-05,
|
| 2185 |
+
"loss": 0.165,
|
| 2186 |
+
"step": 36300
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 11.63,
|
| 2190 |
+
"learning_rate": 3.15334114056567e-05,
|
| 2191 |
+
"loss": 0.1602,
|
| 2192 |
+
"step": 36400
|
| 2193 |
+
},
|
| 2194 |
+
{
|
| 2195 |
+
"epoch": 11.66,
|
| 2196 |
+
"learning_rate": 3.1497923985946986e-05,
|
| 2197 |
+
"loss": 0.1694,
|
| 2198 |
+
"step": 36500
|
| 2199 |
+
},
|
| 2200 |
+
{
|
| 2201 |
+
"epoch": 11.69,
|
| 2202 |
+
"learning_rate": 3.1462436566237274e-05,
|
| 2203 |
+
"loss": 0.1604,
|
| 2204 |
+
"step": 36600
|
| 2205 |
+
},
|
| 2206 |
+
{
|
| 2207 |
+
"epoch": 11.72,
|
| 2208 |
+
"learning_rate": 3.1426949146527555e-05,
|
| 2209 |
+
"loss": 0.1725,
|
| 2210 |
+
"step": 36700
|
| 2211 |
+
},
|
| 2212 |
+
{
|
| 2213 |
+
"epoch": 11.75,
|
| 2214 |
+
"learning_rate": 3.139146172681784e-05,
|
| 2215 |
+
"loss": 0.1577,
|
| 2216 |
+
"step": 36800
|
| 2217 |
+
},
|
| 2218 |
+
{
|
| 2219 |
+
"epoch": 11.79,
|
| 2220 |
+
"learning_rate": 3.135597430710813e-05,
|
| 2221 |
+
"loss": 0.1741,
|
| 2222 |
+
"step": 36900
|
| 2223 |
+
},
|
| 2224 |
+
{
|
| 2225 |
+
"epoch": 11.82,
|
| 2226 |
+
"learning_rate": 3.132048688739842e-05,
|
| 2227 |
+
"loss": 0.1547,
|
| 2228 |
+
"step": 37000
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 11.85,
|
| 2232 |
+
"learning_rate": 3.1284999467688707e-05,
|
| 2233 |
+
"loss": 0.1621,
|
| 2234 |
+
"step": 37100
|
| 2235 |
+
},
|
| 2236 |
+
{
|
| 2237 |
+
"epoch": 11.88,
|
| 2238 |
+
"learning_rate": 3.1249512047978994e-05,
|
| 2239 |
+
"loss": 0.1663,
|
| 2240 |
+
"step": 37200
|
| 2241 |
+
},
|
| 2242 |
+
{
|
| 2243 |
+
"epoch": 11.91,
|
| 2244 |
+
"learning_rate": 3.121437950246638e-05,
|
| 2245 |
+
"loss": 0.1594,
|
| 2246 |
+
"step": 37300
|
| 2247 |
+
},
|
| 2248 |
+
{
|
| 2249 |
+
"epoch": 11.95,
|
| 2250 |
+
"learning_rate": 3.117889208275667e-05,
|
| 2251 |
+
"loss": 0.174,
|
| 2252 |
+
"step": 37400
|
| 2253 |
+
},
|
| 2254 |
+
{
|
| 2255 |
+
"epoch": 11.98,
|
| 2256 |
+
"learning_rate": 3.114340466304696e-05,
|
| 2257 |
+
"loss": 0.1755,
|
| 2258 |
+
"step": 37500
|
| 2259 |
+
},
|
| 2260 |
+
{
|
| 2261 |
+
"epoch": 12.01,
|
| 2262 |
+
"learning_rate": 3.110791724333724e-05,
|
| 2263 |
+
"loss": 0.1627,
|
| 2264 |
+
"step": 37600
|
| 2265 |
+
},
|
| 2266 |
+
{
|
| 2267 |
+
"epoch": 12.04,
|
| 2268 |
+
"learning_rate": 3.1072429823627526e-05,
|
| 2269 |
+
"loss": 0.1445,
|
| 2270 |
+
"step": 37700
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 12.07,
|
| 2274 |
+
"learning_rate": 3.1036942403917814e-05,
|
| 2275 |
+
"loss": 0.1399,
|
| 2276 |
+
"step": 37800
|
| 2277 |
+
},
|
| 2278 |
+
{
|
| 2279 |
+
"epoch": 12.1,
|
| 2280 |
+
"learning_rate": 3.10014549842081e-05,
|
| 2281 |
+
"loss": 0.1475,
|
| 2282 |
+
"step": 37900
|
| 2283 |
+
},
|
| 2284 |
+
{
|
| 2285 |
+
"epoch": 12.14,
|
| 2286 |
+
"learning_rate": 3.096596756449839e-05,
|
| 2287 |
+
"loss": 0.1373,
|
| 2288 |
+
"step": 38000
|
| 2289 |
+
},
|
| 2290 |
+
{
|
| 2291 |
+
"epoch": 12.17,
|
| 2292 |
+
"learning_rate": 3.093048014478868e-05,
|
| 2293 |
+
"loss": 0.1379,
|
| 2294 |
+
"step": 38100
|
| 2295 |
+
},
|
| 2296 |
+
{
|
| 2297 |
+
"epoch": 12.2,
|
| 2298 |
+
"learning_rate": 3.089499272507896e-05,
|
| 2299 |
+
"loss": 0.1389,
|
| 2300 |
+
"step": 38200
|
| 2301 |
+
},
|
| 2302 |
+
{
|
| 2303 |
+
"epoch": 12.23,
|
| 2304 |
+
"learning_rate": 3.085950530536925e-05,
|
| 2305 |
+
"loss": 0.1394,
|
| 2306 |
+
"step": 38300
|
| 2307 |
+
},
|
| 2308 |
+
{
|
| 2309 |
+
"epoch": 12.26,
|
| 2310 |
+
"learning_rate": 3.0824017885659535e-05,
|
| 2311 |
+
"loss": 0.1434,
|
| 2312 |
+
"step": 38400
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 12.3,
|
| 2316 |
+
"learning_rate": 3.078853046594982e-05,
|
| 2317 |
+
"loss": 0.1467,
|
| 2318 |
+
"step": 38500
|
| 2319 |
+
},
|
| 2320 |
+
{
|
| 2321 |
+
"epoch": 12.33,
|
| 2322 |
+
"learning_rate": 3.075339792043721e-05,
|
| 2323 |
+
"loss": 0.1272,
|
| 2324 |
+
"step": 38600
|
| 2325 |
+
},
|
| 2326 |
+
{
|
| 2327 |
+
"epoch": 12.36,
|
| 2328 |
+
"learning_rate": 3.07179105007275e-05,
|
| 2329 |
+
"loss": 0.1363,
|
| 2330 |
+
"step": 38700
|
| 2331 |
+
},
|
| 2332 |
+
{
|
| 2333 |
+
"epoch": 12.39,
|
| 2334 |
+
"learning_rate": 3.0682423081017786e-05,
|
| 2335 |
+
"loss": 0.1467,
|
| 2336 |
+
"step": 38800
|
| 2337 |
+
},
|
| 2338 |
+
{
|
| 2339 |
+
"epoch": 12.42,
|
| 2340 |
+
"learning_rate": 3.0646935661308074e-05,
|
| 2341 |
+
"loss": 0.1389,
|
| 2342 |
+
"step": 38900
|
| 2343 |
+
},
|
| 2344 |
+
{
|
| 2345 |
+
"epoch": 12.46,
|
| 2346 |
+
"learning_rate": 3.0611448241598355e-05,
|
| 2347 |
+
"loss": 0.1447,
|
| 2348 |
+
"step": 39000
|
| 2349 |
+
},
|
| 2350 |
+
{
|
| 2351 |
+
"epoch": 12.49,
|
| 2352 |
+
"learning_rate": 3.057596082188864e-05,
|
| 2353 |
+
"loss": 0.1524,
|
| 2354 |
+
"step": 39100
|
| 2355 |
+
},
|
| 2356 |
+
{
|
| 2357 |
+
"epoch": 12.52,
|
| 2358 |
+
"learning_rate": 3.054047340217893e-05,
|
| 2359 |
+
"loss": 0.1434,
|
| 2360 |
+
"step": 39200
|
| 2361 |
+
},
|
| 2362 |
+
{
|
| 2363 |
+
"epoch": 12.55,
|
| 2364 |
+
"learning_rate": 3.050498598246922e-05,
|
| 2365 |
+
"loss": 0.142,
|
| 2366 |
+
"step": 39300
|
| 2367 |
+
},
|
| 2368 |
+
{
|
| 2369 |
+
"epoch": 12.58,
|
| 2370 |
+
"learning_rate": 3.0469498562759506e-05,
|
| 2371 |
+
"loss": 0.1495,
|
| 2372 |
+
"step": 39400
|
| 2373 |
+
},
|
| 2374 |
+
{
|
| 2375 |
+
"epoch": 12.62,
|
| 2376 |
+
"learning_rate": 3.0434011143049794e-05,
|
| 2377 |
+
"loss": 0.1451,
|
| 2378 |
+
"step": 39500
|
| 2379 |
+
},
|
| 2380 |
+
{
|
| 2381 |
+
"epoch": 12.65,
|
| 2382 |
+
"learning_rate": 3.0398523723340075e-05,
|
| 2383 |
+
"loss": 0.1464,
|
| 2384 |
+
"step": 39600
|
| 2385 |
+
},
|
| 2386 |
+
{
|
| 2387 |
+
"epoch": 12.68,
|
| 2388 |
+
"learning_rate": 3.0363036303630363e-05,
|
| 2389 |
+
"loss": 0.1419,
|
| 2390 |
+
"step": 39700
|
| 2391 |
+
},
|
| 2392 |
+
{
|
| 2393 |
+
"epoch": 12.71,
|
| 2394 |
+
"learning_rate": 3.032754888392065e-05,
|
| 2395 |
+
"loss": 0.1451,
|
| 2396 |
+
"step": 39800
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 12.74,
|
| 2400 |
+
"learning_rate": 3.029206146421094e-05,
|
| 2401 |
+
"loss": 0.151,
|
| 2402 |
+
"step": 39900
|
| 2403 |
+
},
|
| 2404 |
+
{
|
| 2405 |
+
"epoch": 12.78,
|
| 2406 |
+
"learning_rate": 3.0256574044501227e-05,
|
| 2407 |
+
"loss": 0.1442,
|
| 2408 |
+
"step": 40000
|
| 2409 |
+
},
|
| 2410 |
+
{
|
| 2411 |
+
"epoch": 12.81,
|
| 2412 |
+
"learning_rate": 3.0221086624791515e-05,
|
| 2413 |
+
"loss": 0.1553,
|
| 2414 |
+
"step": 40100
|
| 2415 |
+
},
|
| 2416 |
+
{
|
| 2417 |
+
"epoch": 12.84,
|
| 2418 |
+
"learning_rate": 3.01855992050818e-05,
|
| 2419 |
+
"loss": 0.1481,
|
| 2420 |
+
"step": 40200
|
| 2421 |
+
},
|
| 2422 |
+
{
|
| 2423 |
+
"epoch": 12.87,
|
| 2424 |
+
"learning_rate": 3.0150111785372087e-05,
|
| 2425 |
+
"loss": 0.1442,
|
| 2426 |
+
"step": 40300
|
| 2427 |
+
},
|
| 2428 |
+
{
|
| 2429 |
+
"epoch": 12.9,
|
| 2430 |
+
"learning_rate": 3.0114624365662375e-05,
|
| 2431 |
+
"loss": 0.1504,
|
| 2432 |
+
"step": 40400
|
| 2433 |
+
},
|
| 2434 |
+
{
|
| 2435 |
+
"epoch": 12.94,
|
| 2436 |
+
"learning_rate": 3.0079136945952663e-05,
|
| 2437 |
+
"loss": 0.1476,
|
| 2438 |
+
"step": 40500
|
| 2439 |
+
},
|
| 2440 |
+
{
|
| 2441 |
+
"epoch": 12.97,
|
| 2442 |
+
"learning_rate": 3.004364952624295e-05,
|
| 2443 |
+
"loss": 0.145,
|
| 2444 |
+
"step": 40600
|
| 2445 |
+
},
|
| 2446 |
+
{
|
| 2447 |
+
"epoch": 13.0,
|
| 2448 |
+
"learning_rate": 3.000816210653324e-05,
|
| 2449 |
+
"loss": 0.1549,
|
| 2450 |
+
"step": 40700
|
| 2451 |
+
},
|
| 2452 |
+
{
|
| 2453 |
+
"epoch": 13.03,
|
| 2454 |
+
"learning_rate": 2.997267468682352e-05,
|
| 2455 |
+
"loss": 0.1253,
|
| 2456 |
+
"step": 40800
|
| 2457 |
+
},
|
| 2458 |
+
{
|
| 2459 |
+
"epoch": 13.06,
|
| 2460 |
+
"learning_rate": 2.9937187267113808e-05,
|
| 2461 |
+
"loss": 0.1259,
|
| 2462 |
+
"step": 40900
|
| 2463 |
+
},
|
| 2464 |
+
{
|
| 2465 |
+
"epoch": 13.09,
|
| 2466 |
+
"learning_rate": 2.9901699847404096e-05,
|
| 2467 |
+
"loss": 0.1334,
|
| 2468 |
+
"step": 41000
|
| 2469 |
+
},
|
| 2470 |
+
{
|
| 2471 |
+
"epoch": 13.13,
|
| 2472 |
+
"learning_rate": 2.9866212427694384e-05,
|
| 2473 |
+
"loss": 0.1184,
|
| 2474 |
+
"step": 41100
|
| 2475 |
+
},
|
| 2476 |
+
{
|
| 2477 |
+
"epoch": 13.16,
|
| 2478 |
+
"learning_rate": 2.9830725007984672e-05,
|
| 2479 |
+
"loss": 0.1323,
|
| 2480 |
+
"step": 41200
|
| 2481 |
+
},
|
| 2482 |
+
{
|
| 2483 |
+
"epoch": 13.19,
|
| 2484 |
+
"learning_rate": 2.979523758827496e-05,
|
| 2485 |
+
"loss": 0.1294,
|
| 2486 |
+
"step": 41300
|
| 2487 |
+
},
|
| 2488 |
+
{
|
| 2489 |
+
"epoch": 13.22,
|
| 2490 |
+
"learning_rate": 2.9759750168565244e-05,
|
| 2491 |
+
"loss": 0.1283,
|
| 2492 |
+
"step": 41400
|
| 2493 |
+
},
|
| 2494 |
+
{
|
| 2495 |
+
"epoch": 13.25,
|
| 2496 |
+
"learning_rate": 2.9724262748855532e-05,
|
| 2497 |
+
"loss": 0.1313,
|
| 2498 |
+
"step": 41500
|
| 2499 |
+
},
|
| 2500 |
+
{
|
| 2501 |
+
"epoch": 13.29,
|
| 2502 |
+
"learning_rate": 2.968877532914582e-05,
|
| 2503 |
+
"loss": 0.1336,
|
| 2504 |
+
"step": 41600
|
| 2505 |
+
},
|
| 2506 |
+
{
|
| 2507 |
+
"epoch": 13.32,
|
| 2508 |
+
"learning_rate": 2.9653287909436108e-05,
|
| 2509 |
+
"loss": 0.1329,
|
| 2510 |
+
"step": 41700
|
| 2511 |
+
},
|
| 2512 |
+
{
|
| 2513 |
+
"epoch": 13.35,
|
| 2514 |
+
"learning_rate": 2.9617800489726396e-05,
|
| 2515 |
+
"loss": 0.1336,
|
| 2516 |
+
"step": 41800
|
| 2517 |
+
},
|
| 2518 |
+
{
|
| 2519 |
+
"epoch": 13.38,
|
| 2520 |
+
"learning_rate": 2.958231307001668e-05,
|
| 2521 |
+
"loss": 0.1305,
|
| 2522 |
+
"step": 41900
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 13.41,
|
| 2526 |
+
"learning_rate": 2.954682565030697e-05,
|
| 2527 |
+
"loss": 0.1308,
|
| 2528 |
+
"step": 42000
|
| 2529 |
+
},
|
| 2530 |
+
{
|
| 2531 |
+
"epoch": 13.45,
|
| 2532 |
+
"learning_rate": 2.9511338230597256e-05,
|
| 2533 |
+
"loss": 0.1381,
|
| 2534 |
+
"step": 42100
|
| 2535 |
+
},
|
| 2536 |
+
{
|
| 2537 |
+
"epoch": 13.48,
|
| 2538 |
+
"learning_rate": 2.9475850810887544e-05,
|
| 2539 |
+
"loss": 0.1379,
|
| 2540 |
+
"step": 42200
|
| 2541 |
+
},
|
| 2542 |
+
{
|
| 2543 |
+
"epoch": 13.51,
|
| 2544 |
+
"learning_rate": 2.9440363391177832e-05,
|
| 2545 |
+
"loss": 0.1359,
|
| 2546 |
+
"step": 42300
|
| 2547 |
+
},
|
| 2548 |
+
{
|
| 2549 |
+
"epoch": 13.54,
|
| 2550 |
+
"learning_rate": 2.940487597146812e-05,
|
| 2551 |
+
"loss": 0.1273,
|
| 2552 |
+
"step": 42400
|
| 2553 |
+
},
|
| 2554 |
+
{
|
| 2555 |
+
"epoch": 13.57,
|
| 2556 |
+
"learning_rate": 2.93693885517584e-05,
|
| 2557 |
+
"loss": 0.1331,
|
| 2558 |
+
"step": 42500
|
| 2559 |
+
},
|
| 2560 |
+
{
|
| 2561 |
+
"epoch": 13.61,
|
| 2562 |
+
"learning_rate": 2.933390113204869e-05,
|
| 2563 |
+
"loss": 0.1315,
|
| 2564 |
+
"step": 42600
|
| 2565 |
+
},
|
| 2566 |
+
{
|
| 2567 |
+
"epoch": 13.64,
|
| 2568 |
+
"learning_rate": 2.9298413712338977e-05,
|
| 2569 |
+
"loss": 0.1335,
|
| 2570 |
+
"step": 42700
|
| 2571 |
+
},
|
| 2572 |
+
{
|
| 2573 |
+
"epoch": 13.67,
|
| 2574 |
+
"learning_rate": 2.9262926292629265e-05,
|
| 2575 |
+
"loss": 0.1344,
|
| 2576 |
+
"step": 42800
|
| 2577 |
+
},
|
| 2578 |
+
{
|
| 2579 |
+
"epoch": 13.7,
|
| 2580 |
+
"learning_rate": 2.9227438872919553e-05,
|
| 2581 |
+
"loss": 0.1411,
|
| 2582 |
+
"step": 42900
|
| 2583 |
+
},
|
| 2584 |
+
{
|
| 2585 |
+
"epoch": 13.73,
|
| 2586 |
+
"learning_rate": 2.919195145320984e-05,
|
| 2587 |
+
"loss": 0.132,
|
| 2588 |
+
"step": 43000
|
| 2589 |
+
},
|
| 2590 |
+
{
|
| 2591 |
+
"epoch": 13.77,
|
| 2592 |
+
"learning_rate": 2.9156464033500125e-05,
|
| 2593 |
+
"loss": 0.1338,
|
| 2594 |
+
"step": 43100
|
| 2595 |
+
},
|
| 2596 |
+
{
|
| 2597 |
+
"epoch": 13.8,
|
| 2598 |
+
"learning_rate": 2.9120976613790413e-05,
|
| 2599 |
+
"loss": 0.125,
|
| 2600 |
+
"step": 43200
|
| 2601 |
+
},
|
| 2602 |
+
{
|
| 2603 |
+
"epoch": 13.83,
|
| 2604 |
+
"learning_rate": 2.90858440682778e-05,
|
| 2605 |
+
"loss": 0.1389,
|
| 2606 |
+
"step": 43300
|
| 2607 |
+
},
|
| 2608 |
+
{
|
| 2609 |
+
"epoch": 13.86,
|
| 2610 |
+
"learning_rate": 2.9050356648568085e-05,
|
| 2611 |
+
"loss": 0.1386,
|
| 2612 |
+
"step": 43400
|
| 2613 |
+
},
|
| 2614 |
+
{
|
| 2615 |
+
"epoch": 13.89,
|
| 2616 |
+
"learning_rate": 2.9014869228858373e-05,
|
| 2617 |
+
"loss": 0.13,
|
| 2618 |
+
"step": 43500
|
| 2619 |
+
},
|
| 2620 |
+
{
|
| 2621 |
+
"epoch": 13.93,
|
| 2622 |
+
"learning_rate": 2.897938180914866e-05,
|
| 2623 |
+
"loss": 0.1351,
|
| 2624 |
+
"step": 43600
|
| 2625 |
+
},
|
| 2626 |
+
{
|
| 2627 |
+
"epoch": 13.96,
|
| 2628 |
+
"learning_rate": 2.894389438943895e-05,
|
| 2629 |
+
"loss": 0.1341,
|
| 2630 |
+
"step": 43700
|
| 2631 |
+
},
|
| 2632 |
+
{
|
| 2633 |
+
"epoch": 13.99,
|
| 2634 |
+
"learning_rate": 2.8908406969729236e-05,
|
| 2635 |
+
"loss": 0.1399,
|
| 2636 |
+
"step": 43800
|
| 2637 |
+
},
|
| 2638 |
+
{
|
| 2639 |
+
"epoch": 14.02,
|
| 2640 |
+
"learning_rate": 2.887291955001952e-05,
|
| 2641 |
+
"loss": 0.1186,
|
| 2642 |
+
"step": 43900
|
| 2643 |
+
},
|
| 2644 |
+
{
|
| 2645 |
+
"epoch": 14.05,
|
| 2646 |
+
"learning_rate": 2.8837432130309805e-05,
|
| 2647 |
+
"loss": 0.1124,
|
| 2648 |
+
"step": 44000
|
| 2649 |
+
},
|
| 2650 |
+
{
|
| 2651 |
+
"epoch": 14.08,
|
| 2652 |
+
"learning_rate": 2.8801944710600093e-05,
|
| 2653 |
+
"loss": 0.1156,
|
| 2654 |
+
"step": 44100
|
| 2655 |
+
},
|
| 2656 |
+
{
|
| 2657 |
+
"epoch": 14.12,
|
| 2658 |
+
"learning_rate": 2.876645729089038e-05,
|
| 2659 |
+
"loss": 0.1161,
|
| 2660 |
+
"step": 44200
|
| 2661 |
+
},
|
| 2662 |
+
{
|
| 2663 |
+
"epoch": 14.15,
|
| 2664 |
+
"learning_rate": 2.873096987118067e-05,
|
| 2665 |
+
"loss": 0.113,
|
| 2666 |
+
"step": 44300
|
| 2667 |
+
},
|
| 2668 |
+
{
|
| 2669 |
+
"epoch": 14.18,
|
| 2670 |
+
"learning_rate": 2.8695482451470957e-05,
|
| 2671 |
+
"loss": 0.1219,
|
| 2672 |
+
"step": 44400
|
| 2673 |
+
},
|
| 2674 |
+
{
|
| 2675 |
+
"epoch": 14.21,
|
| 2676 |
+
"learning_rate": 2.8659995031761245e-05,
|
| 2677 |
+
"loss": 0.1228,
|
| 2678 |
+
"step": 44500
|
| 2679 |
+
},
|
| 2680 |
+
{
|
| 2681 |
+
"epoch": 14.24,
|
| 2682 |
+
"learning_rate": 2.862450761205153e-05,
|
| 2683 |
+
"loss": 0.1208,
|
| 2684 |
+
"step": 44600
|
| 2685 |
+
},
|
| 2686 |
+
{
|
| 2687 |
+
"epoch": 14.28,
|
| 2688 |
+
"learning_rate": 2.8589020192341817e-05,
|
| 2689 |
+
"loss": 0.1213,
|
| 2690 |
+
"step": 44700
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 14.31,
|
| 2694 |
+
"learning_rate": 2.8553532772632105e-05,
|
| 2695 |
+
"loss": 0.1147,
|
| 2696 |
+
"step": 44800
|
| 2697 |
+
},
|
| 2698 |
+
{
|
| 2699 |
+
"epoch": 14.34,
|
| 2700 |
+
"learning_rate": 2.8518045352922393e-05,
|
| 2701 |
+
"loss": 0.1198,
|
| 2702 |
+
"step": 44900
|
| 2703 |
+
},
|
| 2704 |
+
{
|
| 2705 |
+
"epoch": 14.37,
|
| 2706 |
+
"learning_rate": 2.848255793321268e-05,
|
| 2707 |
+
"loss": 0.1144,
|
| 2708 |
+
"step": 45000
|
| 2709 |
+
}
|
| 2710 |
+
],
|
| 2711 |
+
"max_steps": 125240,
|
| 2712 |
+
"num_train_epochs": 40,
|
| 2713 |
+
"total_flos": 0.0,
|
| 2714 |
+
"trial_name": null,
|
| 2715 |
+
"trial_params": null
|
| 2716 |
+
}
|
checkpoint-5000/config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "Davlan/afro-xlmr-base",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"XLMRobertaModel"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 768,
|
| 13 |
+
"id2label": {
|
| 14 |
+
"0": "LABEL_0"
|
| 15 |
+
},
|
| 16 |
+
"initializer_range": 0.02,
|
| 17 |
+
"intermediate_size": 3072,
|
| 18 |
+
"label2id": {
|
| 19 |
+
"LABEL_0": 0
|
| 20 |
+
},
|
| 21 |
+
"layer_norm_eps": 1e-05,
|
| 22 |
+
"max_position_embeddings": 514,
|
| 23 |
+
"model_type": "xlm-roberta",
|
| 24 |
+
"num_attention_heads": 12,
|
| 25 |
+
"num_hidden_layers": 12,
|
| 26 |
+
"output_past": true,
|
| 27 |
+
"pad_token_id": 1,
|
| 28 |
+
"position_embedding_type": "absolute",
|
| 29 |
+
"torch_dtype": "float32",
|
| 30 |
+
"transformers_version": "4.29.0",
|
| 31 |
+
"type_vocab_size": 1,
|
| 32 |
+
"use_cache": true,
|
| 33 |
+
"vocab_size": 250002
|
| 34 |
+
}
|
checkpoint-5000/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f2db7830f63262f422d5f4548086d28536aa5cce56f4933659f1ee26c968b557
|
| 3 |
+
size 1112246250
|
checkpoint-5000/trainer_state.json
ADDED
|
@@ -0,0 +1,316 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.5969338869370808,
|
| 5 |
+
"global_step": 5000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.03,
|
| 12 |
+
"learning_rate": 2.8744809964867454e-07,
|
| 13 |
+
"loss": 91.176,
|
| 14 |
+
"step": 100
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.06,
|
| 18 |
+
"learning_rate": 6.068348770360907e-07,
|
| 19 |
+
"loss": 64.0684,
|
| 20 |
+
"step": 200
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.1,
|
| 24 |
+
"learning_rate": 9.26221654423507e-07,
|
| 25 |
+
"loss": 27.1061,
|
| 26 |
+
"step": 300
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.13,
|
| 30 |
+
"learning_rate": 1.242414564037049e-06,
|
| 31 |
+
"loss": 8.0198,
|
| 32 |
+
"step": 400
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.16,
|
| 36 |
+
"learning_rate": 1.561801341424465e-06,
|
| 37 |
+
"loss": 4.7691,
|
| 38 |
+
"step": 500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.19,
|
| 42 |
+
"learning_rate": 1.8811881188118813e-06,
|
| 43 |
+
"loss": 3.4387,
|
| 44 |
+
"step": 600
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.22,
|
| 48 |
+
"learning_rate": 2.2005748961992977e-06,
|
| 49 |
+
"loss": 2.7027,
|
| 50 |
+
"step": 700
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.26,
|
| 54 |
+
"learning_rate": 2.5199616735867137e-06,
|
| 55 |
+
"loss": 2.4431,
|
| 56 |
+
"step": 800
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.29,
|
| 60 |
+
"learning_rate": 2.8393484509741297e-06,
|
| 61 |
+
"loss": 2.0534,
|
| 62 |
+
"step": 900
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.32,
|
| 66 |
+
"learning_rate": 3.1587352283615465e-06,
|
| 67 |
+
"loss": 1.8537,
|
| 68 |
+
"step": 1000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.35,
|
| 72 |
+
"learning_rate": 3.4781220057489625e-06,
|
| 73 |
+
"loss": 1.642,
|
| 74 |
+
"step": 1100
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.38,
|
| 78 |
+
"learning_rate": 3.797508783136378e-06,
|
| 79 |
+
"loss": 1.5092,
|
| 80 |
+
"step": 1200
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.42,
|
| 84 |
+
"learning_rate": 4.1168955605237945e-06,
|
| 85 |
+
"loss": 1.4302,
|
| 86 |
+
"step": 1300
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.45,
|
| 90 |
+
"learning_rate": 4.436282337911211e-06,
|
| 91 |
+
"loss": 1.3527,
|
| 92 |
+
"step": 1400
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.48,
|
| 96 |
+
"learning_rate": 4.755669115298627e-06,
|
| 97 |
+
"loss": 1.2495,
|
| 98 |
+
"step": 1500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.51,
|
| 102 |
+
"learning_rate": 5.075055892686043e-06,
|
| 103 |
+
"loss": 1.2039,
|
| 104 |
+
"step": 1600
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.54,
|
| 108 |
+
"learning_rate": 5.39444267007346e-06,
|
| 109 |
+
"loss": 1.1438,
|
| 110 |
+
"step": 1700
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.57,
|
| 114 |
+
"learning_rate": 5.713829447460876e-06,
|
| 115 |
+
"loss": 1.1056,
|
| 116 |
+
"step": 1800
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.61,
|
| 120 |
+
"learning_rate": 6.033216224848291e-06,
|
| 121 |
+
"loss": 1.0436,
|
| 122 |
+
"step": 1900
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.64,
|
| 126 |
+
"learning_rate": 6.352603002235708e-06,
|
| 127 |
+
"loss": 1.0225,
|
| 128 |
+
"step": 2000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.67,
|
| 132 |
+
"learning_rate": 6.671989779623123e-06,
|
| 133 |
+
"loss": 0.9856,
|
| 134 |
+
"step": 2100
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.7,
|
| 138 |
+
"learning_rate": 6.9913765570105406e-06,
|
| 139 |
+
"loss": 0.9412,
|
| 140 |
+
"step": 2200
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.73,
|
| 144 |
+
"learning_rate": 7.310763334397957e-06,
|
| 145 |
+
"loss": 0.9075,
|
| 146 |
+
"step": 2300
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.77,
|
| 150 |
+
"learning_rate": 7.630150111785373e-06,
|
| 151 |
+
"loss": 0.8752,
|
| 152 |
+
"step": 2400
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.8,
|
| 156 |
+
"learning_rate": 7.949536889172788e-06,
|
| 157 |
+
"loss": 0.854,
|
| 158 |
+
"step": 2500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.83,
|
| 162 |
+
"learning_rate": 8.268923666560205e-06,
|
| 163 |
+
"loss": 0.8282,
|
| 164 |
+
"step": 2600
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.86,
|
| 168 |
+
"learning_rate": 8.588310443947621e-06,
|
| 169 |
+
"loss": 0.8074,
|
| 170 |
+
"step": 2700
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.89,
|
| 174 |
+
"learning_rate": 8.907697221335038e-06,
|
| 175 |
+
"loss": 0.7976,
|
| 176 |
+
"step": 2800
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.93,
|
| 180 |
+
"learning_rate": 9.227083998722454e-06,
|
| 181 |
+
"loss": 0.7955,
|
| 182 |
+
"step": 2900
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.96,
|
| 186 |
+
"learning_rate": 9.54647077610987e-06,
|
| 187 |
+
"loss": 0.7502,
|
| 188 |
+
"step": 3000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.99,
|
| 192 |
+
"learning_rate": 9.865857553497285e-06,
|
| 193 |
+
"loss": 0.7455,
|
| 194 |
+
"step": 3100
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 1.02,
|
| 198 |
+
"learning_rate": 1.0185244330884702e-05,
|
| 199 |
+
"loss": 0.7177,
|
| 200 |
+
"step": 3200
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 1.05,
|
| 204 |
+
"learning_rate": 1.050463110827212e-05,
|
| 205 |
+
"loss": 0.724,
|
| 206 |
+
"step": 3300
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 1.09,
|
| 210 |
+
"learning_rate": 1.0824017885659535e-05,
|
| 211 |
+
"loss": 0.6975,
|
| 212 |
+
"step": 3400
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 1.12,
|
| 216 |
+
"learning_rate": 1.114340466304695e-05,
|
| 217 |
+
"loss": 0.7054,
|
| 218 |
+
"step": 3500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 1.15,
|
| 222 |
+
"learning_rate": 1.1462791440434366e-05,
|
| 223 |
+
"loss": 0.71,
|
| 224 |
+
"step": 3600
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 1.18,
|
| 228 |
+
"learning_rate": 1.1782178217821782e-05,
|
| 229 |
+
"loss": 0.6825,
|
| 230 |
+
"step": 3700
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 1.21,
|
| 234 |
+
"learning_rate": 1.2101564995209199e-05,
|
| 235 |
+
"loss": 0.6792,
|
| 236 |
+
"step": 3800
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 1.25,
|
| 240 |
+
"learning_rate": 1.2420951772596615e-05,
|
| 241 |
+
"loss": 0.6794,
|
| 242 |
+
"step": 3900
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 1.28,
|
| 246 |
+
"learning_rate": 1.2740338549984032e-05,
|
| 247 |
+
"loss": 0.6672,
|
| 248 |
+
"step": 4000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 1.31,
|
| 252 |
+
"learning_rate": 1.3059725327371449e-05,
|
| 253 |
+
"loss": 0.6356,
|
| 254 |
+
"step": 4100
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 1.34,
|
| 258 |
+
"learning_rate": 1.3379112104758865e-05,
|
| 259 |
+
"loss": 0.6503,
|
| 260 |
+
"step": 4200
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 1.37,
|
| 264 |
+
"learning_rate": 1.369849888214628e-05,
|
| 265 |
+
"loss": 0.6495,
|
| 266 |
+
"step": 4300
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 1.41,
|
| 270 |
+
"learning_rate": 1.4017885659533696e-05,
|
| 271 |
+
"loss": 0.6428,
|
| 272 |
+
"step": 4400
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 1.44,
|
| 276 |
+
"learning_rate": 1.4337272436921113e-05,
|
| 277 |
+
"loss": 0.6171,
|
| 278 |
+
"step": 4500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 1.47,
|
| 282 |
+
"learning_rate": 1.4656659214308529e-05,
|
| 283 |
+
"loss": 0.6084,
|
| 284 |
+
"step": 4600
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 1.5,
|
| 288 |
+
"learning_rate": 1.4976045991695944e-05,
|
| 289 |
+
"loss": 0.6222,
|
| 290 |
+
"step": 4700
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 1.53,
|
| 294 |
+
"learning_rate": 1.529543276908336e-05,
|
| 295 |
+
"loss": 0.5971,
|
| 296 |
+
"step": 4800
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 1.56,
|
| 300 |
+
"learning_rate": 1.5614819546470777e-05,
|
| 301 |
+
"loss": 0.5745,
|
| 302 |
+
"step": 4900
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 1.6,
|
| 306 |
+
"learning_rate": 1.5934206323858196e-05,
|
| 307 |
+
"loss": 0.5753,
|
| 308 |
+
"step": 5000
|
| 309 |
+
}
|
| 310 |
+
],
|
| 311 |
+
"max_steps": 125240,
|
| 312 |
+
"num_train_epochs": 40,
|
| 313 |
+
"total_flos": 0.0,
|
| 314 |
+
"trial_name": null,
|
| 315 |
+
"trial_params": null
|
| 316 |
+
}
|
checkpoint-50000/config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "Davlan/afro-xlmr-base",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"XLMRobertaModel"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 768,
|
| 13 |
+
"id2label": {
|
| 14 |
+
"0": "LABEL_0"
|
| 15 |
+
},
|
| 16 |
+
"initializer_range": 0.02,
|
| 17 |
+
"intermediate_size": 3072,
|
| 18 |
+
"label2id": {
|
| 19 |
+
"LABEL_0": 0
|
| 20 |
+
},
|
| 21 |
+
"layer_norm_eps": 1e-05,
|
| 22 |
+
"max_position_embeddings": 514,
|
| 23 |
+
"model_type": "xlm-roberta",
|
| 24 |
+
"num_attention_heads": 12,
|
| 25 |
+
"num_hidden_layers": 12,
|
| 26 |
+
"output_past": true,
|
| 27 |
+
"pad_token_id": 1,
|
| 28 |
+
"position_embedding_type": "absolute",
|
| 29 |
+
"torch_dtype": "float32",
|
| 30 |
+
"transformers_version": "4.29.0",
|
| 31 |
+
"type_vocab_size": 1,
|
| 32 |
+
"use_cache": true,
|
| 33 |
+
"vocab_size": 250002
|
| 34 |
+
}
|
checkpoint-50000/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:12a24b9d42a0590339c8598ce93a194334c646c90f3a2a703e908202e7a33acf
|
| 3 |
+
size 1112246250
|
checkpoint-50000/trainer_state.json
ADDED
|
@@ -0,0 +1,3016 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 15.969338869370809,
|
| 5 |
+
"global_step": 50000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.03,
|
| 12 |
+
"learning_rate": 2.8744809964867454e-07,
|
| 13 |
+
"loss": 91.176,
|
| 14 |
+
"step": 100
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.06,
|
| 18 |
+
"learning_rate": 6.068348770360907e-07,
|
| 19 |
+
"loss": 64.0684,
|
| 20 |
+
"step": 200
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.1,
|
| 24 |
+
"learning_rate": 9.26221654423507e-07,
|
| 25 |
+
"loss": 27.1061,
|
| 26 |
+
"step": 300
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.13,
|
| 30 |
+
"learning_rate": 1.242414564037049e-06,
|
| 31 |
+
"loss": 8.0198,
|
| 32 |
+
"step": 400
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.16,
|
| 36 |
+
"learning_rate": 1.561801341424465e-06,
|
| 37 |
+
"loss": 4.7691,
|
| 38 |
+
"step": 500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.19,
|
| 42 |
+
"learning_rate": 1.8811881188118813e-06,
|
| 43 |
+
"loss": 3.4387,
|
| 44 |
+
"step": 600
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.22,
|
| 48 |
+
"learning_rate": 2.2005748961992977e-06,
|
| 49 |
+
"loss": 2.7027,
|
| 50 |
+
"step": 700
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.26,
|
| 54 |
+
"learning_rate": 2.5199616735867137e-06,
|
| 55 |
+
"loss": 2.4431,
|
| 56 |
+
"step": 800
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.29,
|
| 60 |
+
"learning_rate": 2.8393484509741297e-06,
|
| 61 |
+
"loss": 2.0534,
|
| 62 |
+
"step": 900
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.32,
|
| 66 |
+
"learning_rate": 3.1587352283615465e-06,
|
| 67 |
+
"loss": 1.8537,
|
| 68 |
+
"step": 1000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.35,
|
| 72 |
+
"learning_rate": 3.4781220057489625e-06,
|
| 73 |
+
"loss": 1.642,
|
| 74 |
+
"step": 1100
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.38,
|
| 78 |
+
"learning_rate": 3.797508783136378e-06,
|
| 79 |
+
"loss": 1.5092,
|
| 80 |
+
"step": 1200
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.42,
|
| 84 |
+
"learning_rate": 4.1168955605237945e-06,
|
| 85 |
+
"loss": 1.4302,
|
| 86 |
+
"step": 1300
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.45,
|
| 90 |
+
"learning_rate": 4.436282337911211e-06,
|
| 91 |
+
"loss": 1.3527,
|
| 92 |
+
"step": 1400
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.48,
|
| 96 |
+
"learning_rate": 4.755669115298627e-06,
|
| 97 |
+
"loss": 1.2495,
|
| 98 |
+
"step": 1500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.51,
|
| 102 |
+
"learning_rate": 5.075055892686043e-06,
|
| 103 |
+
"loss": 1.2039,
|
| 104 |
+
"step": 1600
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.54,
|
| 108 |
+
"learning_rate": 5.39444267007346e-06,
|
| 109 |
+
"loss": 1.1438,
|
| 110 |
+
"step": 1700
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.57,
|
| 114 |
+
"learning_rate": 5.713829447460876e-06,
|
| 115 |
+
"loss": 1.1056,
|
| 116 |
+
"step": 1800
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.61,
|
| 120 |
+
"learning_rate": 6.033216224848291e-06,
|
| 121 |
+
"loss": 1.0436,
|
| 122 |
+
"step": 1900
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.64,
|
| 126 |
+
"learning_rate": 6.352603002235708e-06,
|
| 127 |
+
"loss": 1.0225,
|
| 128 |
+
"step": 2000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.67,
|
| 132 |
+
"learning_rate": 6.671989779623123e-06,
|
| 133 |
+
"loss": 0.9856,
|
| 134 |
+
"step": 2100
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.7,
|
| 138 |
+
"learning_rate": 6.9913765570105406e-06,
|
| 139 |
+
"loss": 0.9412,
|
| 140 |
+
"step": 2200
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.73,
|
| 144 |
+
"learning_rate": 7.310763334397957e-06,
|
| 145 |
+
"loss": 0.9075,
|
| 146 |
+
"step": 2300
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.77,
|
| 150 |
+
"learning_rate": 7.630150111785373e-06,
|
| 151 |
+
"loss": 0.8752,
|
| 152 |
+
"step": 2400
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.8,
|
| 156 |
+
"learning_rate": 7.949536889172788e-06,
|
| 157 |
+
"loss": 0.854,
|
| 158 |
+
"step": 2500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.83,
|
| 162 |
+
"learning_rate": 8.268923666560205e-06,
|
| 163 |
+
"loss": 0.8282,
|
| 164 |
+
"step": 2600
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.86,
|
| 168 |
+
"learning_rate": 8.588310443947621e-06,
|
| 169 |
+
"loss": 0.8074,
|
| 170 |
+
"step": 2700
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.89,
|
| 174 |
+
"learning_rate": 8.907697221335038e-06,
|
| 175 |
+
"loss": 0.7976,
|
| 176 |
+
"step": 2800
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.93,
|
| 180 |
+
"learning_rate": 9.227083998722454e-06,
|
| 181 |
+
"loss": 0.7955,
|
| 182 |
+
"step": 2900
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.96,
|
| 186 |
+
"learning_rate": 9.54647077610987e-06,
|
| 187 |
+
"loss": 0.7502,
|
| 188 |
+
"step": 3000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.99,
|
| 192 |
+
"learning_rate": 9.865857553497285e-06,
|
| 193 |
+
"loss": 0.7455,
|
| 194 |
+
"step": 3100
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 1.02,
|
| 198 |
+
"learning_rate": 1.0185244330884702e-05,
|
| 199 |
+
"loss": 0.7177,
|
| 200 |
+
"step": 3200
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 1.05,
|
| 204 |
+
"learning_rate": 1.050463110827212e-05,
|
| 205 |
+
"loss": 0.724,
|
| 206 |
+
"step": 3300
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 1.09,
|
| 210 |
+
"learning_rate": 1.0824017885659535e-05,
|
| 211 |
+
"loss": 0.6975,
|
| 212 |
+
"step": 3400
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 1.12,
|
| 216 |
+
"learning_rate": 1.114340466304695e-05,
|
| 217 |
+
"loss": 0.7054,
|
| 218 |
+
"step": 3500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 1.15,
|
| 222 |
+
"learning_rate": 1.1462791440434366e-05,
|
| 223 |
+
"loss": 0.71,
|
| 224 |
+
"step": 3600
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 1.18,
|
| 228 |
+
"learning_rate": 1.1782178217821782e-05,
|
| 229 |
+
"loss": 0.6825,
|
| 230 |
+
"step": 3700
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 1.21,
|
| 234 |
+
"learning_rate": 1.2101564995209199e-05,
|
| 235 |
+
"loss": 0.6792,
|
| 236 |
+
"step": 3800
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 1.25,
|
| 240 |
+
"learning_rate": 1.2420951772596615e-05,
|
| 241 |
+
"loss": 0.6794,
|
| 242 |
+
"step": 3900
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 1.28,
|
| 246 |
+
"learning_rate": 1.2740338549984032e-05,
|
| 247 |
+
"loss": 0.6672,
|
| 248 |
+
"step": 4000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 1.31,
|
| 252 |
+
"learning_rate": 1.3059725327371449e-05,
|
| 253 |
+
"loss": 0.6356,
|
| 254 |
+
"step": 4100
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 1.34,
|
| 258 |
+
"learning_rate": 1.3379112104758865e-05,
|
| 259 |
+
"loss": 0.6503,
|
| 260 |
+
"step": 4200
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 1.37,
|
| 264 |
+
"learning_rate": 1.369849888214628e-05,
|
| 265 |
+
"loss": 0.6495,
|
| 266 |
+
"step": 4300
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 1.41,
|
| 270 |
+
"learning_rate": 1.4017885659533696e-05,
|
| 271 |
+
"loss": 0.6428,
|
| 272 |
+
"step": 4400
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 1.44,
|
| 276 |
+
"learning_rate": 1.4337272436921113e-05,
|
| 277 |
+
"loss": 0.6171,
|
| 278 |
+
"step": 4500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 1.47,
|
| 282 |
+
"learning_rate": 1.4656659214308529e-05,
|
| 283 |
+
"loss": 0.6084,
|
| 284 |
+
"step": 4600
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 1.5,
|
| 288 |
+
"learning_rate": 1.4976045991695944e-05,
|
| 289 |
+
"loss": 0.6222,
|
| 290 |
+
"step": 4700
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 1.53,
|
| 294 |
+
"learning_rate": 1.529543276908336e-05,
|
| 295 |
+
"loss": 0.5971,
|
| 296 |
+
"step": 4800
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 1.56,
|
| 300 |
+
"learning_rate": 1.5614819546470777e-05,
|
| 301 |
+
"loss": 0.5745,
|
| 302 |
+
"step": 4900
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 1.6,
|
| 306 |
+
"learning_rate": 1.5934206323858196e-05,
|
| 307 |
+
"loss": 0.5753,
|
| 308 |
+
"step": 5000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 1.63,
|
| 312 |
+
"learning_rate": 1.625359310124561e-05,
|
| 313 |
+
"loss": 0.6045,
|
| 314 |
+
"step": 5100
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 1.66,
|
| 318 |
+
"learning_rate": 1.6572979878633027e-05,
|
| 319 |
+
"loss": 0.6006,
|
| 320 |
+
"step": 5200
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 1.69,
|
| 324 |
+
"learning_rate": 1.6892366656020443e-05,
|
| 325 |
+
"loss": 0.6003,
|
| 326 |
+
"step": 5300
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 1.72,
|
| 330 |
+
"learning_rate": 1.7211753433407858e-05,
|
| 331 |
+
"loss": 0.5688,
|
| 332 |
+
"step": 5400
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 1.76,
|
| 336 |
+
"learning_rate": 1.7531140210795274e-05,
|
| 337 |
+
"loss": 0.5662,
|
| 338 |
+
"step": 5500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 1.79,
|
| 342 |
+
"learning_rate": 1.785052698818269e-05,
|
| 343 |
+
"loss": 0.5788,
|
| 344 |
+
"step": 5600
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 1.82,
|
| 348 |
+
"learning_rate": 1.8169913765570105e-05,
|
| 349 |
+
"loss": 0.5541,
|
| 350 |
+
"step": 5700
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 1.85,
|
| 354 |
+
"learning_rate": 1.8489300542957524e-05,
|
| 355 |
+
"loss": 0.5656,
|
| 356 |
+
"step": 5800
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 1.88,
|
| 360 |
+
"learning_rate": 1.880868732034494e-05,
|
| 361 |
+
"loss": 0.5466,
|
| 362 |
+
"step": 5900
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 1.92,
|
| 366 |
+
"learning_rate": 1.9128074097732355e-05,
|
| 367 |
+
"loss": 0.563,
|
| 368 |
+
"step": 6000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 1.95,
|
| 372 |
+
"learning_rate": 1.9447460875119774e-05,
|
| 373 |
+
"loss": 0.5603,
|
| 374 |
+
"step": 6100
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 1.98,
|
| 378 |
+
"learning_rate": 1.976684765250719e-05,
|
| 379 |
+
"loss": 0.5579,
|
| 380 |
+
"step": 6200
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 2.01,
|
| 384 |
+
"learning_rate": 2.0086234429894605e-05,
|
| 385 |
+
"loss": 0.5346,
|
| 386 |
+
"step": 6300
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 2.04,
|
| 390 |
+
"learning_rate": 2.0405621207282017e-05,
|
| 391 |
+
"loss": 0.5038,
|
| 392 |
+
"step": 6400
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 2.08,
|
| 396 |
+
"learning_rate": 2.0725007984669436e-05,
|
| 397 |
+
"loss": 0.5121,
|
| 398 |
+
"step": 6500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 2.11,
|
| 402 |
+
"learning_rate": 2.1044394762056855e-05,
|
| 403 |
+
"loss": 0.5135,
|
| 404 |
+
"step": 6600
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 2.14,
|
| 408 |
+
"learning_rate": 2.1363781539444267e-05,
|
| 409 |
+
"loss": 0.504,
|
| 410 |
+
"step": 6700
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 2.17,
|
| 414 |
+
"learning_rate": 2.1683168316831686e-05,
|
| 415 |
+
"loss": 0.5119,
|
| 416 |
+
"step": 6800
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 2.2,
|
| 420 |
+
"learning_rate": 2.20025550942191e-05,
|
| 421 |
+
"loss": 0.4986,
|
| 422 |
+
"step": 6900
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 2.24,
|
| 426 |
+
"learning_rate": 2.2321941871606518e-05,
|
| 427 |
+
"loss": 0.504,
|
| 428 |
+
"step": 7000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 2.27,
|
| 432 |
+
"learning_rate": 2.2641328648993933e-05,
|
| 433 |
+
"loss": 0.5024,
|
| 434 |
+
"step": 7100
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 2.3,
|
| 438 |
+
"learning_rate": 2.296071542638135e-05,
|
| 439 |
+
"loss": 0.479,
|
| 440 |
+
"step": 7200
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 2.33,
|
| 444 |
+
"learning_rate": 2.3280102203768764e-05,
|
| 445 |
+
"loss": 0.492,
|
| 446 |
+
"step": 7300
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 2.36,
|
| 450 |
+
"learning_rate": 2.3599488981156183e-05,
|
| 451 |
+
"loss": 0.4928,
|
| 452 |
+
"step": 7400
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 2.4,
|
| 456 |
+
"learning_rate": 2.39188757585436e-05,
|
| 457 |
+
"loss": 0.4862,
|
| 458 |
+
"step": 7500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 2.43,
|
| 462 |
+
"learning_rate": 2.4238262535931014e-05,
|
| 463 |
+
"loss": 0.5148,
|
| 464 |
+
"step": 7600
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 2.46,
|
| 468 |
+
"learning_rate": 2.4557649313318433e-05,
|
| 469 |
+
"loss": 0.4922,
|
| 470 |
+
"step": 7700
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 2.49,
|
| 474 |
+
"learning_rate": 2.4877036090705846e-05,
|
| 475 |
+
"loss": 0.4882,
|
| 476 |
+
"step": 7800
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 2.52,
|
| 480 |
+
"learning_rate": 2.5196422868093264e-05,
|
| 481 |
+
"loss": 0.4871,
|
| 482 |
+
"step": 7900
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 2.56,
|
| 486 |
+
"learning_rate": 2.5515809645480677e-05,
|
| 487 |
+
"loss": 0.4967,
|
| 488 |
+
"step": 8000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 2.59,
|
| 492 |
+
"learning_rate": 2.5835196422868096e-05,
|
| 493 |
+
"loss": 0.4834,
|
| 494 |
+
"step": 8100
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 2.62,
|
| 498 |
+
"learning_rate": 2.615458320025551e-05,
|
| 499 |
+
"loss": 0.4769,
|
| 500 |
+
"step": 8200
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 2.65,
|
| 504 |
+
"learning_rate": 2.6473969977642927e-05,
|
| 505 |
+
"loss": 0.4776,
|
| 506 |
+
"step": 8300
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 2.68,
|
| 510 |
+
"learning_rate": 2.6793356755030342e-05,
|
| 511 |
+
"loss": 0.4861,
|
| 512 |
+
"step": 8400
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 2.71,
|
| 516 |
+
"learning_rate": 2.711274353241776e-05,
|
| 517 |
+
"loss": 0.486,
|
| 518 |
+
"step": 8500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 2.75,
|
| 522 |
+
"learning_rate": 2.7432130309805177e-05,
|
| 523 |
+
"loss": 0.4745,
|
| 524 |
+
"step": 8600
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 2.78,
|
| 528 |
+
"learning_rate": 2.7751517087192592e-05,
|
| 529 |
+
"loss": 0.4805,
|
| 530 |
+
"step": 8700
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 2.81,
|
| 534 |
+
"learning_rate": 2.807090386458001e-05,
|
| 535 |
+
"loss": 0.5003,
|
| 536 |
+
"step": 8800
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 2.84,
|
| 540 |
+
"learning_rate": 2.8390290641967424e-05,
|
| 541 |
+
"loss": 0.4639,
|
| 542 |
+
"step": 8900
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 2.87,
|
| 546 |
+
"learning_rate": 2.8709677419354843e-05,
|
| 547 |
+
"loss": 0.4763,
|
| 548 |
+
"step": 9000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 2.91,
|
| 552 |
+
"learning_rate": 2.9029064196742255e-05,
|
| 553 |
+
"loss": 0.4727,
|
| 554 |
+
"step": 9100
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 2.94,
|
| 558 |
+
"learning_rate": 2.9348450974129674e-05,
|
| 559 |
+
"loss": 0.4839,
|
| 560 |
+
"step": 9200
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 2.97,
|
| 564 |
+
"learning_rate": 2.966783775151709e-05,
|
| 565 |
+
"loss": 0.4769,
|
| 566 |
+
"step": 9300
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 3.0,
|
| 570 |
+
"learning_rate": 2.9987224528904505e-05,
|
| 571 |
+
"loss": 0.4641,
|
| 572 |
+
"step": 9400
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 3.03,
|
| 576 |
+
"learning_rate": 3.030661130629192e-05,
|
| 577 |
+
"loss": 0.4282,
|
| 578 |
+
"step": 9500
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 3.07,
|
| 582 |
+
"learning_rate": 3.0625998083679336e-05,
|
| 583 |
+
"loss": 0.4057,
|
| 584 |
+
"step": 9600
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 3.1,
|
| 588 |
+
"learning_rate": 3.094538486106676e-05,
|
| 589 |
+
"loss": 0.4192,
|
| 590 |
+
"step": 9700
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 3.13,
|
| 594 |
+
"learning_rate": 3.126477163845417e-05,
|
| 595 |
+
"loss": 0.4186,
|
| 596 |
+
"step": 9800
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 3.16,
|
| 600 |
+
"learning_rate": 3.158415841584159e-05,
|
| 601 |
+
"loss": 0.4086,
|
| 602 |
+
"step": 9900
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 3.19,
|
| 606 |
+
"learning_rate": 3.1903545193229005e-05,
|
| 607 |
+
"loss": 0.4194,
|
| 608 |
+
"step": 10000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 3.23,
|
| 612 |
+
"learning_rate": 3.221973810284254e-05,
|
| 613 |
+
"loss": 0.4443,
|
| 614 |
+
"step": 10100
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 3.26,
|
| 618 |
+
"learning_rate": 3.2539124880229965e-05,
|
| 619 |
+
"loss": 0.4217,
|
| 620 |
+
"step": 10200
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 3.29,
|
| 624 |
+
"learning_rate": 3.2858511657617374e-05,
|
| 625 |
+
"loss": 0.4275,
|
| 626 |
+
"step": 10300
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 3.32,
|
| 630 |
+
"learning_rate": 3.3177898435004796e-05,
|
| 631 |
+
"loss": 0.4422,
|
| 632 |
+
"step": 10400
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 3.35,
|
| 636 |
+
"learning_rate": 3.3497285212392205e-05,
|
| 637 |
+
"loss": 0.4227,
|
| 638 |
+
"step": 10500
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 3.39,
|
| 642 |
+
"learning_rate": 3.381667198977963e-05,
|
| 643 |
+
"loss": 0.4191,
|
| 644 |
+
"step": 10600
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 3.42,
|
| 648 |
+
"learning_rate": 3.413605876716704e-05,
|
| 649 |
+
"loss": 0.4309,
|
| 650 |
+
"step": 10700
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 3.45,
|
| 654 |
+
"learning_rate": 3.445544554455446e-05,
|
| 655 |
+
"loss": 0.4079,
|
| 656 |
+
"step": 10800
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 3.48,
|
| 660 |
+
"learning_rate": 3.4774832321941874e-05,
|
| 661 |
+
"loss": 0.4281,
|
| 662 |
+
"step": 10900
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 3.51,
|
| 666 |
+
"learning_rate": 3.509421909932929e-05,
|
| 667 |
+
"loss": 0.4137,
|
| 668 |
+
"step": 11000
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 3.55,
|
| 672 |
+
"learning_rate": 3.5413605876716705e-05,
|
| 673 |
+
"loss": 0.4168,
|
| 674 |
+
"step": 11100
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 3.58,
|
| 678 |
+
"learning_rate": 3.573299265410412e-05,
|
| 679 |
+
"loss": 0.4186,
|
| 680 |
+
"step": 11200
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 3.61,
|
| 684 |
+
"learning_rate": 3.605237943149154e-05,
|
| 685 |
+
"loss": 0.4325,
|
| 686 |
+
"step": 11300
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 3.64,
|
| 690 |
+
"learning_rate": 3.637176620887895e-05,
|
| 691 |
+
"loss": 0.4422,
|
| 692 |
+
"step": 11400
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 3.67,
|
| 696 |
+
"learning_rate": 3.6691152986266374e-05,
|
| 697 |
+
"loss": 0.4269,
|
| 698 |
+
"step": 11500
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 3.7,
|
| 702 |
+
"learning_rate": 3.701053976365379e-05,
|
| 703 |
+
"loss": 0.4368,
|
| 704 |
+
"step": 11600
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 3.74,
|
| 708 |
+
"learning_rate": 3.7329926541041205e-05,
|
| 709 |
+
"loss": 0.4207,
|
| 710 |
+
"step": 11700
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 3.77,
|
| 714 |
+
"learning_rate": 3.764931331842862e-05,
|
| 715 |
+
"loss": 0.4365,
|
| 716 |
+
"step": 11800
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 3.8,
|
| 720 |
+
"learning_rate": 3.7968700095816036e-05,
|
| 721 |
+
"loss": 0.4394,
|
| 722 |
+
"step": 11900
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 3.83,
|
| 726 |
+
"learning_rate": 3.828808687320345e-05,
|
| 727 |
+
"loss": 0.4351,
|
| 728 |
+
"step": 12000
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 3.86,
|
| 732 |
+
"learning_rate": 3.860747365059087e-05,
|
| 733 |
+
"loss": 0.4113,
|
| 734 |
+
"step": 12100
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 3.9,
|
| 738 |
+
"learning_rate": 3.892686042797828e-05,
|
| 739 |
+
"loss": 0.4261,
|
| 740 |
+
"step": 12200
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 3.93,
|
| 744 |
+
"learning_rate": 3.92462472053657e-05,
|
| 745 |
+
"loss": 0.4311,
|
| 746 |
+
"step": 12300
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 3.96,
|
| 750 |
+
"learning_rate": 3.956563398275312e-05,
|
| 751 |
+
"loss": 0.4246,
|
| 752 |
+
"step": 12400
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 3.99,
|
| 756 |
+
"learning_rate": 3.988502076014053e-05,
|
| 757 |
+
"loss": 0.4294,
|
| 758 |
+
"step": 12500
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 4.02,
|
| 762 |
+
"learning_rate": 3.997728805138579e-05,
|
| 763 |
+
"loss": 0.3854,
|
| 764 |
+
"step": 12600
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 4.06,
|
| 768 |
+
"learning_rate": 3.994180063167608e-05,
|
| 769 |
+
"loss": 0.3656,
|
| 770 |
+
"step": 12700
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 4.09,
|
| 774 |
+
"learning_rate": 3.9906313211966366e-05,
|
| 775 |
+
"loss": 0.3836,
|
| 776 |
+
"step": 12800
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 4.12,
|
| 780 |
+
"learning_rate": 3.987082579225665e-05,
|
| 781 |
+
"loss": 0.3664,
|
| 782 |
+
"step": 12900
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 4.15,
|
| 786 |
+
"learning_rate": 3.9835338372546935e-05,
|
| 787 |
+
"loss": 0.3884,
|
| 788 |
+
"step": 13000
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 4.18,
|
| 792 |
+
"learning_rate": 3.979985095283722e-05,
|
| 793 |
+
"loss": 0.3764,
|
| 794 |
+
"step": 13100
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 4.22,
|
| 798 |
+
"learning_rate": 3.976436353312751e-05,
|
| 799 |
+
"loss": 0.3733,
|
| 800 |
+
"step": 13200
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 4.25,
|
| 804 |
+
"learning_rate": 3.97288761134178e-05,
|
| 805 |
+
"loss": 0.3682,
|
| 806 |
+
"step": 13300
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 4.28,
|
| 810 |
+
"learning_rate": 3.969338869370809e-05,
|
| 811 |
+
"loss": 0.371,
|
| 812 |
+
"step": 13400
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 4.31,
|
| 816 |
+
"learning_rate": 3.965790127399837e-05,
|
| 817 |
+
"loss": 0.3741,
|
| 818 |
+
"step": 13500
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 4.34,
|
| 822 |
+
"learning_rate": 3.9622413854288656e-05,
|
| 823 |
+
"loss": 0.3757,
|
| 824 |
+
"step": 13600
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 4.38,
|
| 828 |
+
"learning_rate": 3.9586926434578944e-05,
|
| 829 |
+
"loss": 0.3652,
|
| 830 |
+
"step": 13700
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 4.41,
|
| 834 |
+
"learning_rate": 3.955143901486923e-05,
|
| 835 |
+
"loss": 0.3892,
|
| 836 |
+
"step": 13800
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 4.44,
|
| 840 |
+
"learning_rate": 3.951630646935662e-05,
|
| 841 |
+
"loss": 0.3893,
|
| 842 |
+
"step": 13900
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 4.47,
|
| 846 |
+
"learning_rate": 3.948081904964691e-05,
|
| 847 |
+
"loss": 0.3823,
|
| 848 |
+
"step": 14000
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 4.5,
|
| 852 |
+
"learning_rate": 3.944533162993719e-05,
|
| 853 |
+
"loss": 0.3654,
|
| 854 |
+
"step": 14100
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 4.54,
|
| 858 |
+
"learning_rate": 3.9409844210227476e-05,
|
| 859 |
+
"loss": 0.3791,
|
| 860 |
+
"step": 14200
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 4.57,
|
| 864 |
+
"learning_rate": 3.9374356790517764e-05,
|
| 865 |
+
"loss": 0.3918,
|
| 866 |
+
"step": 14300
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 4.6,
|
| 870 |
+
"learning_rate": 3.933886937080805e-05,
|
| 871 |
+
"loss": 0.3713,
|
| 872 |
+
"step": 14400
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 4.63,
|
| 876 |
+
"learning_rate": 3.930338195109834e-05,
|
| 877 |
+
"loss": 0.376,
|
| 878 |
+
"step": 14500
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 4.66,
|
| 882 |
+
"learning_rate": 3.926789453138863e-05,
|
| 883 |
+
"loss": 0.3682,
|
| 884 |
+
"step": 14600
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 4.69,
|
| 888 |
+
"learning_rate": 3.9232407111678915e-05,
|
| 889 |
+
"loss": 0.374,
|
| 890 |
+
"step": 14700
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 4.73,
|
| 894 |
+
"learning_rate": 3.91969196919692e-05,
|
| 895 |
+
"loss": 0.3774,
|
| 896 |
+
"step": 14800
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 4.76,
|
| 900 |
+
"learning_rate": 3.9161432272259484e-05,
|
| 901 |
+
"loss": 0.3806,
|
| 902 |
+
"step": 14900
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 4.79,
|
| 906 |
+
"learning_rate": 3.912594485254977e-05,
|
| 907 |
+
"loss": 0.3886,
|
| 908 |
+
"step": 15000
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 4.82,
|
| 912 |
+
"learning_rate": 3.909045743284006e-05,
|
| 913 |
+
"loss": 0.3742,
|
| 914 |
+
"step": 15100
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 4.85,
|
| 918 |
+
"learning_rate": 3.905497001313035e-05,
|
| 919 |
+
"loss": 0.3714,
|
| 920 |
+
"step": 15200
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 4.89,
|
| 924 |
+
"learning_rate": 3.9019482593420636e-05,
|
| 925 |
+
"loss": 0.3721,
|
| 926 |
+
"step": 15300
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 4.92,
|
| 930 |
+
"learning_rate": 3.8983995173710924e-05,
|
| 931 |
+
"loss": 0.3716,
|
| 932 |
+
"step": 15400
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 4.95,
|
| 936 |
+
"learning_rate": 3.8948507754001205e-05,
|
| 937 |
+
"loss": 0.3728,
|
| 938 |
+
"step": 15500
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 4.98,
|
| 942 |
+
"learning_rate": 3.891302033429149e-05,
|
| 943 |
+
"loss": 0.384,
|
| 944 |
+
"step": 15600
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 5.01,
|
| 948 |
+
"learning_rate": 3.887753291458178e-05,
|
| 949 |
+
"loss": 0.3415,
|
| 950 |
+
"step": 15700
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 5.05,
|
| 954 |
+
"learning_rate": 3.884204549487207e-05,
|
| 955 |
+
"loss": 0.3064,
|
| 956 |
+
"step": 15800
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 5.08,
|
| 960 |
+
"learning_rate": 3.880655807516236e-05,
|
| 961 |
+
"loss": 0.3164,
|
| 962 |
+
"step": 15900
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 5.11,
|
| 966 |
+
"learning_rate": 3.8771070655452645e-05,
|
| 967 |
+
"loss": 0.3215,
|
| 968 |
+
"step": 16000
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 5.14,
|
| 972 |
+
"learning_rate": 3.873558323574293e-05,
|
| 973 |
+
"loss": 0.3234,
|
| 974 |
+
"step": 16100
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 5.17,
|
| 978 |
+
"learning_rate": 3.870009581603322e-05,
|
| 979 |
+
"loss": 0.3119,
|
| 980 |
+
"step": 16200
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 5.21,
|
| 984 |
+
"learning_rate": 3.866460839632351e-05,
|
| 985 |
+
"loss": 0.3145,
|
| 986 |
+
"step": 16300
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 5.24,
|
| 990 |
+
"learning_rate": 3.8629120976613796e-05,
|
| 991 |
+
"loss": 0.3333,
|
| 992 |
+
"step": 16400
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 5.27,
|
| 996 |
+
"learning_rate": 3.8593633556904084e-05,
|
| 997 |
+
"loss": 0.3195,
|
| 998 |
+
"step": 16500
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 5.3,
|
| 1002 |
+
"learning_rate": 3.855814613719437e-05,
|
| 1003 |
+
"loss": 0.3159,
|
| 1004 |
+
"step": 16600
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 5.33,
|
| 1008 |
+
"learning_rate": 3.852265871748465e-05,
|
| 1009 |
+
"loss": 0.3189,
|
| 1010 |
+
"step": 16700
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 5.37,
|
| 1014 |
+
"learning_rate": 3.848717129777494e-05,
|
| 1015 |
+
"loss": 0.317,
|
| 1016 |
+
"step": 16800
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 5.4,
|
| 1020 |
+
"learning_rate": 3.845168387806523e-05,
|
| 1021 |
+
"loss": 0.3231,
|
| 1022 |
+
"step": 16900
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 5.43,
|
| 1026 |
+
"learning_rate": 3.841619645835552e-05,
|
| 1027 |
+
"loss": 0.3328,
|
| 1028 |
+
"step": 17000
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 5.46,
|
| 1032 |
+
"learning_rate": 3.8380709038645805e-05,
|
| 1033 |
+
"loss": 0.3282,
|
| 1034 |
+
"step": 17100
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 5.49,
|
| 1038 |
+
"learning_rate": 3.8345221618936086e-05,
|
| 1039 |
+
"loss": 0.3169,
|
| 1040 |
+
"step": 17200
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 5.53,
|
| 1044 |
+
"learning_rate": 3.8309734199226374e-05,
|
| 1045 |
+
"loss": 0.3156,
|
| 1046 |
+
"step": 17300
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 5.56,
|
| 1050 |
+
"learning_rate": 3.827424677951666e-05,
|
| 1051 |
+
"loss": 0.3313,
|
| 1052 |
+
"step": 17400
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 5.59,
|
| 1056 |
+
"learning_rate": 3.823875935980695e-05,
|
| 1057 |
+
"loss": 0.3181,
|
| 1058 |
+
"step": 17500
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 5.62,
|
| 1062 |
+
"learning_rate": 3.820327194009724e-05,
|
| 1063 |
+
"loss": 0.334,
|
| 1064 |
+
"step": 17600
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 5.65,
|
| 1068 |
+
"learning_rate": 3.8167784520387526e-05,
|
| 1069 |
+
"loss": 0.3313,
|
| 1070 |
+
"step": 17700
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 5.69,
|
| 1074 |
+
"learning_rate": 3.8132297100677814e-05,
|
| 1075 |
+
"loss": 0.3331,
|
| 1076 |
+
"step": 17800
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 5.72,
|
| 1080 |
+
"learning_rate": 3.80968096809681e-05,
|
| 1081 |
+
"loss": 0.3405,
|
| 1082 |
+
"step": 17900
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 5.75,
|
| 1086 |
+
"learning_rate": 3.806132226125839e-05,
|
| 1087 |
+
"loss": 0.3345,
|
| 1088 |
+
"step": 18000
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 5.78,
|
| 1092 |
+
"learning_rate": 3.802583484154868e-05,
|
| 1093 |
+
"loss": 0.3295,
|
| 1094 |
+
"step": 18100
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 5.81,
|
| 1098 |
+
"learning_rate": 3.7990347421838965e-05,
|
| 1099 |
+
"loss": 0.3244,
|
| 1100 |
+
"step": 18200
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 5.84,
|
| 1104 |
+
"learning_rate": 3.795486000212925e-05,
|
| 1105 |
+
"loss": 0.3346,
|
| 1106 |
+
"step": 18300
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 5.88,
|
| 1110 |
+
"learning_rate": 3.791972745661663e-05,
|
| 1111 |
+
"loss": 0.3317,
|
| 1112 |
+
"step": 18400
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 5.91,
|
| 1116 |
+
"learning_rate": 3.788424003690692e-05,
|
| 1117 |
+
"loss": 0.3219,
|
| 1118 |
+
"step": 18500
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 5.94,
|
| 1122 |
+
"learning_rate": 3.784875261719721e-05,
|
| 1123 |
+
"loss": 0.3167,
|
| 1124 |
+
"step": 18600
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 5.97,
|
| 1128 |
+
"learning_rate": 3.781326519748749e-05,
|
| 1129 |
+
"loss": 0.33,
|
| 1130 |
+
"step": 18700
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 6.0,
|
| 1134 |
+
"learning_rate": 3.777777777777778e-05,
|
| 1135 |
+
"loss": 0.3211,
|
| 1136 |
+
"step": 18800
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 6.04,
|
| 1140 |
+
"learning_rate": 3.7742290358068066e-05,
|
| 1141 |
+
"loss": 0.2794,
|
| 1142 |
+
"step": 18900
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 6.07,
|
| 1146 |
+
"learning_rate": 3.7706802938358354e-05,
|
| 1147 |
+
"loss": 0.2675,
|
| 1148 |
+
"step": 19000
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 6.1,
|
| 1152 |
+
"learning_rate": 3.767131551864864e-05,
|
| 1153 |
+
"loss": 0.2676,
|
| 1154 |
+
"step": 19100
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 6.13,
|
| 1158 |
+
"learning_rate": 3.763582809893893e-05,
|
| 1159 |
+
"loss": 0.2693,
|
| 1160 |
+
"step": 19200
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 6.16,
|
| 1164 |
+
"learning_rate": 3.760034067922922e-05,
|
| 1165 |
+
"loss": 0.2669,
|
| 1166 |
+
"step": 19300
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 6.2,
|
| 1170 |
+
"learning_rate": 3.7564853259519506e-05,
|
| 1171 |
+
"loss": 0.2744,
|
| 1172 |
+
"step": 19400
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 6.23,
|
| 1176 |
+
"learning_rate": 3.7529365839809794e-05,
|
| 1177 |
+
"loss": 0.282,
|
| 1178 |
+
"step": 19500
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 6.26,
|
| 1182 |
+
"learning_rate": 3.749387842010008e-05,
|
| 1183 |
+
"loss": 0.2848,
|
| 1184 |
+
"step": 19600
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 6.29,
|
| 1188 |
+
"learning_rate": 3.745839100039037e-05,
|
| 1189 |
+
"loss": 0.2783,
|
| 1190 |
+
"step": 19700
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 6.32,
|
| 1194 |
+
"learning_rate": 3.742290358068066e-05,
|
| 1195 |
+
"loss": 0.2843,
|
| 1196 |
+
"step": 19800
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 6.36,
|
| 1200 |
+
"learning_rate": 3.738741616097094e-05,
|
| 1201 |
+
"loss": 0.2696,
|
| 1202 |
+
"step": 19900
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 6.39,
|
| 1206 |
+
"learning_rate": 3.7351928741261226e-05,
|
| 1207 |
+
"loss": 0.2932,
|
| 1208 |
+
"step": 20000
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 6.42,
|
| 1212 |
+
"learning_rate": 3.7316441321551514e-05,
|
| 1213 |
+
"loss": 0.2775,
|
| 1214 |
+
"step": 20100
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 6.45,
|
| 1218 |
+
"learning_rate": 3.72809539018418e-05,
|
| 1219 |
+
"loss": 0.2893,
|
| 1220 |
+
"step": 20200
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 6.48,
|
| 1224 |
+
"learning_rate": 3.724546648213209e-05,
|
| 1225 |
+
"loss": 0.286,
|
| 1226 |
+
"step": 20300
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 6.52,
|
| 1230 |
+
"learning_rate": 3.720997906242237e-05,
|
| 1231 |
+
"loss": 0.2799,
|
| 1232 |
+
"step": 20400
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 6.55,
|
| 1236 |
+
"learning_rate": 3.717449164271266e-05,
|
| 1237 |
+
"loss": 0.2827,
|
| 1238 |
+
"step": 20500
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 6.58,
|
| 1242 |
+
"learning_rate": 3.713900422300295e-05,
|
| 1243 |
+
"loss": 0.2783,
|
| 1244 |
+
"step": 20600
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 6.61,
|
| 1248 |
+
"learning_rate": 3.7103516803293235e-05,
|
| 1249 |
+
"loss": 0.2729,
|
| 1250 |
+
"step": 20700
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 6.64,
|
| 1254 |
+
"learning_rate": 3.706802938358352e-05,
|
| 1255 |
+
"loss": 0.2747,
|
| 1256 |
+
"step": 20800
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 6.68,
|
| 1260 |
+
"learning_rate": 3.703254196387381e-05,
|
| 1261 |
+
"loss": 0.2764,
|
| 1262 |
+
"step": 20900
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 6.71,
|
| 1266 |
+
"learning_rate": 3.699705454416409e-05,
|
| 1267 |
+
"loss": 0.2804,
|
| 1268 |
+
"step": 21000
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 6.74,
|
| 1272 |
+
"learning_rate": 3.696156712445438e-05,
|
| 1273 |
+
"loss": 0.2841,
|
| 1274 |
+
"step": 21100
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 6.77,
|
| 1278 |
+
"learning_rate": 3.692607970474467e-05,
|
| 1279 |
+
"loss": 0.277,
|
| 1280 |
+
"step": 21200
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 6.8,
|
| 1284 |
+
"learning_rate": 3.6890592285034956e-05,
|
| 1285 |
+
"loss": 0.284,
|
| 1286 |
+
"step": 21300
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 6.83,
|
| 1290 |
+
"learning_rate": 3.6855104865325244e-05,
|
| 1291 |
+
"loss": 0.2778,
|
| 1292 |
+
"step": 21400
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 6.87,
|
| 1296 |
+
"learning_rate": 3.681961744561553e-05,
|
| 1297 |
+
"loss": 0.2797,
|
| 1298 |
+
"step": 21500
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 6.9,
|
| 1302 |
+
"learning_rate": 3.678413002590582e-05,
|
| 1303 |
+
"loss": 0.3036,
|
| 1304 |
+
"step": 21600
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 6.93,
|
| 1308 |
+
"learning_rate": 3.674864260619611e-05,
|
| 1309 |
+
"loss": 0.2869,
|
| 1310 |
+
"step": 21700
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 6.96,
|
| 1314 |
+
"learning_rate": 3.6713155186486395e-05,
|
| 1315 |
+
"loss": 0.2921,
|
| 1316 |
+
"step": 21800
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 6.99,
|
| 1320 |
+
"learning_rate": 3.667766776677668e-05,
|
| 1321 |
+
"loss": 0.2844,
|
| 1322 |
+
"step": 21900
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 7.03,
|
| 1326 |
+
"learning_rate": 3.664218034706697e-05,
|
| 1327 |
+
"loss": 0.2427,
|
| 1328 |
+
"step": 22000
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 7.06,
|
| 1332 |
+
"learning_rate": 3.660669292735726e-05,
|
| 1333 |
+
"loss": 0.243,
|
| 1334 |
+
"step": 22100
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 7.09,
|
| 1338 |
+
"learning_rate": 3.657120550764754e-05,
|
| 1339 |
+
"loss": 0.2311,
|
| 1340 |
+
"step": 22200
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 7.12,
|
| 1344 |
+
"learning_rate": 3.653571808793783e-05,
|
| 1345 |
+
"loss": 0.2358,
|
| 1346 |
+
"step": 22300
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 7.15,
|
| 1350 |
+
"learning_rate": 3.6500230668228116e-05,
|
| 1351 |
+
"loss": 0.2287,
|
| 1352 |
+
"step": 22400
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 7.19,
|
| 1356 |
+
"learning_rate": 3.6464743248518404e-05,
|
| 1357 |
+
"loss": 0.2519,
|
| 1358 |
+
"step": 22500
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 7.22,
|
| 1362 |
+
"learning_rate": 3.642925582880869e-05,
|
| 1363 |
+
"loss": 0.2528,
|
| 1364 |
+
"step": 22600
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 7.25,
|
| 1368 |
+
"learning_rate": 3.639376840909898e-05,
|
| 1369 |
+
"loss": 0.2507,
|
| 1370 |
+
"step": 22700
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 7.28,
|
| 1374 |
+
"learning_rate": 3.635828098938926e-05,
|
| 1375 |
+
"loss": 0.2496,
|
| 1376 |
+
"step": 22800
|
| 1377 |
+
},
|
| 1378 |
+
{
|
| 1379 |
+
"epoch": 7.31,
|
| 1380 |
+
"learning_rate": 3.632279356967955e-05,
|
| 1381 |
+
"loss": 0.2503,
|
| 1382 |
+
"step": 22900
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 7.35,
|
| 1386 |
+
"learning_rate": 3.628730614996984e-05,
|
| 1387 |
+
"loss": 0.2383,
|
| 1388 |
+
"step": 23000
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 7.38,
|
| 1392 |
+
"learning_rate": 3.6251818730260125e-05,
|
| 1393 |
+
"loss": 0.2513,
|
| 1394 |
+
"step": 23100
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 7.41,
|
| 1398 |
+
"learning_rate": 3.621633131055041e-05,
|
| 1399 |
+
"loss": 0.2492,
|
| 1400 |
+
"step": 23200
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 7.44,
|
| 1404 |
+
"learning_rate": 3.61808438908407e-05,
|
| 1405 |
+
"loss": 0.2473,
|
| 1406 |
+
"step": 23300
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 7.47,
|
| 1410 |
+
"learning_rate": 3.614535647113099e-05,
|
| 1411 |
+
"loss": 0.2502,
|
| 1412 |
+
"step": 23400
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 7.51,
|
| 1416 |
+
"learning_rate": 3.6109869051421276e-05,
|
| 1417 |
+
"loss": 0.2492,
|
| 1418 |
+
"step": 23500
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 7.54,
|
| 1422 |
+
"learning_rate": 3.6074381631711564e-05,
|
| 1423 |
+
"loss": 0.2366,
|
| 1424 |
+
"step": 23600
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 7.57,
|
| 1428 |
+
"learning_rate": 3.603889421200185e-05,
|
| 1429 |
+
"loss": 0.2393,
|
| 1430 |
+
"step": 23700
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 7.6,
|
| 1434 |
+
"learning_rate": 3.600340679229214e-05,
|
| 1435 |
+
"loss": 0.2483,
|
| 1436 |
+
"step": 23800
|
| 1437 |
+
},
|
| 1438 |
+
{
|
| 1439 |
+
"epoch": 7.63,
|
| 1440 |
+
"learning_rate": 3.596791937258242e-05,
|
| 1441 |
+
"loss": 0.2428,
|
| 1442 |
+
"step": 23900
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 7.67,
|
| 1446 |
+
"learning_rate": 3.593243195287271e-05,
|
| 1447 |
+
"loss": 0.2507,
|
| 1448 |
+
"step": 24000
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 7.7,
|
| 1452 |
+
"learning_rate": 3.5896944533163e-05,
|
| 1453 |
+
"loss": 0.2491,
|
| 1454 |
+
"step": 24100
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 7.73,
|
| 1458 |
+
"learning_rate": 3.5861457113453285e-05,
|
| 1459 |
+
"loss": 0.2549,
|
| 1460 |
+
"step": 24200
|
| 1461 |
+
},
|
| 1462 |
+
{
|
| 1463 |
+
"epoch": 7.76,
|
| 1464 |
+
"learning_rate": 3.582596969374357e-05,
|
| 1465 |
+
"loss": 0.251,
|
| 1466 |
+
"step": 24300
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 7.79,
|
| 1470 |
+
"learning_rate": 3.579048227403386e-05,
|
| 1471 |
+
"loss": 0.2357,
|
| 1472 |
+
"step": 24400
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 7.82,
|
| 1476 |
+
"learning_rate": 3.575499485432414e-05,
|
| 1477 |
+
"loss": 0.2441,
|
| 1478 |
+
"step": 24500
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 7.86,
|
| 1482 |
+
"learning_rate": 3.571950743461443e-05,
|
| 1483 |
+
"loss": 0.2561,
|
| 1484 |
+
"step": 24600
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 7.89,
|
| 1488 |
+
"learning_rate": 3.568402001490472e-05,
|
| 1489 |
+
"loss": 0.2524,
|
| 1490 |
+
"step": 24700
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 7.92,
|
| 1494 |
+
"learning_rate": 3.5648532595195006e-05,
|
| 1495 |
+
"loss": 0.2467,
|
| 1496 |
+
"step": 24800
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 7.95,
|
| 1500 |
+
"learning_rate": 3.5613045175485293e-05,
|
| 1501 |
+
"loss": 0.2416,
|
| 1502 |
+
"step": 24900
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 7.98,
|
| 1506 |
+
"learning_rate": 3.557755775577558e-05,
|
| 1507 |
+
"loss": 0.2511,
|
| 1508 |
+
"step": 25000
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 8.02,
|
| 1512 |
+
"learning_rate": 3.554207033606587e-05,
|
| 1513 |
+
"loss": 0.2323,
|
| 1514 |
+
"step": 25100
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 8.05,
|
| 1518 |
+
"learning_rate": 3.550658291635615e-05,
|
| 1519 |
+
"loss": 0.2158,
|
| 1520 |
+
"step": 25200
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 8.08,
|
| 1524 |
+
"learning_rate": 3.547109549664644e-05,
|
| 1525 |
+
"loss": 0.21,
|
| 1526 |
+
"step": 25300
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 8.11,
|
| 1530 |
+
"learning_rate": 3.5435608076936726e-05,
|
| 1531 |
+
"loss": 0.2102,
|
| 1532 |
+
"step": 25400
|
| 1533 |
+
},
|
| 1534 |
+
{
|
| 1535 |
+
"epoch": 8.14,
|
| 1536 |
+
"learning_rate": 3.5400120657227014e-05,
|
| 1537 |
+
"loss": 0.2071,
|
| 1538 |
+
"step": 25500
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 8.18,
|
| 1542 |
+
"learning_rate": 3.53646332375173e-05,
|
| 1543 |
+
"loss": 0.2198,
|
| 1544 |
+
"step": 25600
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 8.21,
|
| 1548 |
+
"learning_rate": 3.532914581780759e-05,
|
| 1549 |
+
"loss": 0.2193,
|
| 1550 |
+
"step": 25700
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 8.24,
|
| 1554 |
+
"learning_rate": 3.529365839809788e-05,
|
| 1555 |
+
"loss": 0.2166,
|
| 1556 |
+
"step": 25800
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 8.27,
|
| 1560 |
+
"learning_rate": 3.5258170978388166e-05,
|
| 1561 |
+
"loss": 0.2237,
|
| 1562 |
+
"step": 25900
|
| 1563 |
+
},
|
| 1564 |
+
{
|
| 1565 |
+
"epoch": 8.3,
|
| 1566 |
+
"learning_rate": 3.5222683558678454e-05,
|
| 1567 |
+
"loss": 0.2241,
|
| 1568 |
+
"step": 26000
|
| 1569 |
+
},
|
| 1570 |
+
{
|
| 1571 |
+
"epoch": 8.34,
|
| 1572 |
+
"learning_rate": 3.518719613896874e-05,
|
| 1573 |
+
"loss": 0.2132,
|
| 1574 |
+
"step": 26100
|
| 1575 |
+
},
|
| 1576 |
+
{
|
| 1577 |
+
"epoch": 8.37,
|
| 1578 |
+
"learning_rate": 3.515170871925902e-05,
|
| 1579 |
+
"loss": 0.2101,
|
| 1580 |
+
"step": 26200
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 8.4,
|
| 1584 |
+
"learning_rate": 3.511622129954931e-05,
|
| 1585 |
+
"loss": 0.227,
|
| 1586 |
+
"step": 26300
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 8.43,
|
| 1590 |
+
"learning_rate": 3.50807338798396e-05,
|
| 1591 |
+
"loss": 0.2221,
|
| 1592 |
+
"step": 26400
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 8.46,
|
| 1596 |
+
"learning_rate": 3.504560133432698e-05,
|
| 1597 |
+
"loss": 0.2181,
|
| 1598 |
+
"step": 26500
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 8.5,
|
| 1602 |
+
"learning_rate": 3.501011391461727e-05,
|
| 1603 |
+
"loss": 0.2208,
|
| 1604 |
+
"step": 26600
|
| 1605 |
+
},
|
| 1606 |
+
{
|
| 1607 |
+
"epoch": 8.53,
|
| 1608 |
+
"learning_rate": 3.4974626494907555e-05,
|
| 1609 |
+
"loss": 0.2192,
|
| 1610 |
+
"step": 26700
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 8.56,
|
| 1614 |
+
"learning_rate": 3.493913907519784e-05,
|
| 1615 |
+
"loss": 0.2227,
|
| 1616 |
+
"step": 26800
|
| 1617 |
+
},
|
| 1618 |
+
{
|
| 1619 |
+
"epoch": 8.59,
|
| 1620 |
+
"learning_rate": 3.490365165548813e-05,
|
| 1621 |
+
"loss": 0.2308,
|
| 1622 |
+
"step": 26900
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 8.62,
|
| 1626 |
+
"learning_rate": 3.486816423577842e-05,
|
| 1627 |
+
"loss": 0.22,
|
| 1628 |
+
"step": 27000
|
| 1629 |
+
},
|
| 1630 |
+
{
|
| 1631 |
+
"epoch": 8.66,
|
| 1632 |
+
"learning_rate": 3.4832676816068706e-05,
|
| 1633 |
+
"loss": 0.2247,
|
| 1634 |
+
"step": 27100
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 8.69,
|
| 1638 |
+
"learning_rate": 3.4797189396358994e-05,
|
| 1639 |
+
"loss": 0.2148,
|
| 1640 |
+
"step": 27200
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 8.72,
|
| 1644 |
+
"learning_rate": 3.476170197664928e-05,
|
| 1645 |
+
"loss": 0.2136,
|
| 1646 |
+
"step": 27300
|
| 1647 |
+
},
|
| 1648 |
+
{
|
| 1649 |
+
"epoch": 8.75,
|
| 1650 |
+
"learning_rate": 3.472621455693957e-05,
|
| 1651 |
+
"loss": 0.2282,
|
| 1652 |
+
"step": 27400
|
| 1653 |
+
},
|
| 1654 |
+
{
|
| 1655 |
+
"epoch": 8.78,
|
| 1656 |
+
"learning_rate": 3.469072713722986e-05,
|
| 1657 |
+
"loss": 0.2266,
|
| 1658 |
+
"step": 27500
|
| 1659 |
+
},
|
| 1660 |
+
{
|
| 1661 |
+
"epoch": 8.82,
|
| 1662 |
+
"learning_rate": 3.4655239717520146e-05,
|
| 1663 |
+
"loss": 0.2198,
|
| 1664 |
+
"step": 27600
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 8.85,
|
| 1668 |
+
"learning_rate": 3.461975229781043e-05,
|
| 1669 |
+
"loss": 0.2243,
|
| 1670 |
+
"step": 27700
|
| 1671 |
+
},
|
| 1672 |
+
{
|
| 1673 |
+
"epoch": 8.88,
|
| 1674 |
+
"learning_rate": 3.4584264878100715e-05,
|
| 1675 |
+
"loss": 0.2358,
|
| 1676 |
+
"step": 27800
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 8.91,
|
| 1680 |
+
"learning_rate": 3.4548777458391e-05,
|
| 1681 |
+
"loss": 0.2255,
|
| 1682 |
+
"step": 27900
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 8.94,
|
| 1686 |
+
"learning_rate": 3.451329003868129e-05,
|
| 1687 |
+
"loss": 0.2258,
|
| 1688 |
+
"step": 28000
|
| 1689 |
+
},
|
| 1690 |
+
{
|
| 1691 |
+
"epoch": 8.97,
|
| 1692 |
+
"learning_rate": 3.447780261897158e-05,
|
| 1693 |
+
"loss": 0.2228,
|
| 1694 |
+
"step": 28100
|
| 1695 |
+
},
|
| 1696 |
+
{
|
| 1697 |
+
"epoch": 9.01,
|
| 1698 |
+
"learning_rate": 3.4442315199261866e-05,
|
| 1699 |
+
"loss": 0.2183,
|
| 1700 |
+
"step": 28200
|
| 1701 |
+
},
|
| 1702 |
+
{
|
| 1703 |
+
"epoch": 9.04,
|
| 1704 |
+
"learning_rate": 3.440682777955215e-05,
|
| 1705 |
+
"loss": 0.1988,
|
| 1706 |
+
"step": 28300
|
| 1707 |
+
},
|
| 1708 |
+
{
|
| 1709 |
+
"epoch": 9.07,
|
| 1710 |
+
"learning_rate": 3.4371340359842436e-05,
|
| 1711 |
+
"loss": 0.188,
|
| 1712 |
+
"step": 28400
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"epoch": 9.1,
|
| 1716 |
+
"learning_rate": 3.4335852940132723e-05,
|
| 1717 |
+
"loss": 0.1963,
|
| 1718 |
+
"step": 28500
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 9.13,
|
| 1722 |
+
"learning_rate": 3.430036552042301e-05,
|
| 1723 |
+
"loss": 0.1919,
|
| 1724 |
+
"step": 28600
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 9.17,
|
| 1728 |
+
"learning_rate": 3.42648781007133e-05,
|
| 1729 |
+
"loss": 0.1906,
|
| 1730 |
+
"step": 28700
|
| 1731 |
+
},
|
| 1732 |
+
{
|
| 1733 |
+
"epoch": 9.2,
|
| 1734 |
+
"learning_rate": 3.422939068100359e-05,
|
| 1735 |
+
"loss": 0.1911,
|
| 1736 |
+
"step": 28800
|
| 1737 |
+
},
|
| 1738 |
+
{
|
| 1739 |
+
"epoch": 9.23,
|
| 1740 |
+
"learning_rate": 3.4193903261293875e-05,
|
| 1741 |
+
"loss": 0.1971,
|
| 1742 |
+
"step": 28900
|
| 1743 |
+
},
|
| 1744 |
+
{
|
| 1745 |
+
"epoch": 9.26,
|
| 1746 |
+
"learning_rate": 3.415841584158416e-05,
|
| 1747 |
+
"loss": 0.1931,
|
| 1748 |
+
"step": 29000
|
| 1749 |
+
},
|
| 1750 |
+
{
|
| 1751 |
+
"epoch": 9.29,
|
| 1752 |
+
"learning_rate": 3.412292842187445e-05,
|
| 1753 |
+
"loss": 0.1951,
|
| 1754 |
+
"step": 29100
|
| 1755 |
+
},
|
| 1756 |
+
{
|
| 1757 |
+
"epoch": 9.33,
|
| 1758 |
+
"learning_rate": 3.408744100216474e-05,
|
| 1759 |
+
"loss": 0.2008,
|
| 1760 |
+
"step": 29200
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 9.36,
|
| 1764 |
+
"learning_rate": 3.405195358245503e-05,
|
| 1765 |
+
"loss": 0.2053,
|
| 1766 |
+
"step": 29300
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 9.39,
|
| 1770 |
+
"learning_rate": 3.401646616274531e-05,
|
| 1771 |
+
"loss": 0.2015,
|
| 1772 |
+
"step": 29400
|
| 1773 |
+
},
|
| 1774 |
+
{
|
| 1775 |
+
"epoch": 9.42,
|
| 1776 |
+
"learning_rate": 3.3980978743035596e-05,
|
| 1777 |
+
"loss": 0.193,
|
| 1778 |
+
"step": 29500
|
| 1779 |
+
},
|
| 1780 |
+
{
|
| 1781 |
+
"epoch": 9.45,
|
| 1782 |
+
"learning_rate": 3.3945491323325884e-05,
|
| 1783 |
+
"loss": 0.1955,
|
| 1784 |
+
"step": 29600
|
| 1785 |
+
},
|
| 1786 |
+
{
|
| 1787 |
+
"epoch": 9.49,
|
| 1788 |
+
"learning_rate": 3.391000390361617e-05,
|
| 1789 |
+
"loss": 0.1993,
|
| 1790 |
+
"step": 29700
|
| 1791 |
+
},
|
| 1792 |
+
{
|
| 1793 |
+
"epoch": 9.52,
|
| 1794 |
+
"learning_rate": 3.387451648390646e-05,
|
| 1795 |
+
"loss": 0.2043,
|
| 1796 |
+
"step": 29800
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"epoch": 9.55,
|
| 1800 |
+
"learning_rate": 3.383902906419675e-05,
|
| 1801 |
+
"loss": 0.1955,
|
| 1802 |
+
"step": 29900
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 9.58,
|
| 1806 |
+
"learning_rate": 3.380354164448703e-05,
|
| 1807 |
+
"loss": 0.1961,
|
| 1808 |
+
"step": 30000
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 9.61,
|
| 1812 |
+
"learning_rate": 3.3768054224777316e-05,
|
| 1813 |
+
"loss": 0.1961,
|
| 1814 |
+
"step": 30100
|
| 1815 |
+
},
|
| 1816 |
+
{
|
| 1817 |
+
"epoch": 9.65,
|
| 1818 |
+
"learning_rate": 3.3732566805067604e-05,
|
| 1819 |
+
"loss": 0.196,
|
| 1820 |
+
"step": 30200
|
| 1821 |
+
},
|
| 1822 |
+
{
|
| 1823 |
+
"epoch": 9.68,
|
| 1824 |
+
"learning_rate": 3.369707938535789e-05,
|
| 1825 |
+
"loss": 0.2088,
|
| 1826 |
+
"step": 30300
|
| 1827 |
+
},
|
| 1828 |
+
{
|
| 1829 |
+
"epoch": 9.71,
|
| 1830 |
+
"learning_rate": 3.366159196564818e-05,
|
| 1831 |
+
"loss": 0.2031,
|
| 1832 |
+
"step": 30400
|
| 1833 |
+
},
|
| 1834 |
+
{
|
| 1835 |
+
"epoch": 9.74,
|
| 1836 |
+
"learning_rate": 3.362610454593847e-05,
|
| 1837 |
+
"loss": 0.1949,
|
| 1838 |
+
"step": 30500
|
| 1839 |
+
},
|
| 1840 |
+
{
|
| 1841 |
+
"epoch": 9.77,
|
| 1842 |
+
"learning_rate": 3.3590617126228756e-05,
|
| 1843 |
+
"loss": 0.1988,
|
| 1844 |
+
"step": 30600
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 9.81,
|
| 1848 |
+
"learning_rate": 3.3555129706519044e-05,
|
| 1849 |
+
"loss": 0.2061,
|
| 1850 |
+
"step": 30700
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 9.84,
|
| 1854 |
+
"learning_rate": 3.351964228680933e-05,
|
| 1855 |
+
"loss": 0.2022,
|
| 1856 |
+
"step": 30800
|
| 1857 |
+
},
|
| 1858 |
+
{
|
| 1859 |
+
"epoch": 9.87,
|
| 1860 |
+
"learning_rate": 3.348450974129671e-05,
|
| 1861 |
+
"loss": 0.1965,
|
| 1862 |
+
"step": 30900
|
| 1863 |
+
},
|
| 1864 |
+
{
|
| 1865 |
+
"epoch": 9.9,
|
| 1866 |
+
"learning_rate": 3.3449022321587e-05,
|
| 1867 |
+
"loss": 0.1999,
|
| 1868 |
+
"step": 31000
|
| 1869 |
+
},
|
| 1870 |
+
{
|
| 1871 |
+
"epoch": 9.93,
|
| 1872 |
+
"learning_rate": 3.341353490187729e-05,
|
| 1873 |
+
"loss": 0.2065,
|
| 1874 |
+
"step": 31100
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 9.96,
|
| 1878 |
+
"learning_rate": 3.3378047482167576e-05,
|
| 1879 |
+
"loss": 0.2031,
|
| 1880 |
+
"step": 31200
|
| 1881 |
+
},
|
| 1882 |
+
{
|
| 1883 |
+
"epoch": 10.0,
|
| 1884 |
+
"learning_rate": 3.3342560062457864e-05,
|
| 1885 |
+
"loss": 0.2078,
|
| 1886 |
+
"step": 31300
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 10.03,
|
| 1890 |
+
"learning_rate": 3.330707264274815e-05,
|
| 1891 |
+
"loss": 0.1733,
|
| 1892 |
+
"step": 31400
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 10.06,
|
| 1896 |
+
"learning_rate": 3.327158522303843e-05,
|
| 1897 |
+
"loss": 0.1654,
|
| 1898 |
+
"step": 31500
|
| 1899 |
+
},
|
| 1900 |
+
{
|
| 1901 |
+
"epoch": 10.09,
|
| 1902 |
+
"learning_rate": 3.323609780332872e-05,
|
| 1903 |
+
"loss": 0.1674,
|
| 1904 |
+
"step": 31600
|
| 1905 |
+
},
|
| 1906 |
+
{
|
| 1907 |
+
"epoch": 10.12,
|
| 1908 |
+
"learning_rate": 3.320061038361901e-05,
|
| 1909 |
+
"loss": 0.1725,
|
| 1910 |
+
"step": 31700
|
| 1911 |
+
},
|
| 1912 |
+
{
|
| 1913 |
+
"epoch": 10.16,
|
| 1914 |
+
"learning_rate": 3.3165122963909297e-05,
|
| 1915 |
+
"loss": 0.1761,
|
| 1916 |
+
"step": 31800
|
| 1917 |
+
},
|
| 1918 |
+
{
|
| 1919 |
+
"epoch": 10.19,
|
| 1920 |
+
"learning_rate": 3.3129635544199584e-05,
|
| 1921 |
+
"loss": 0.1757,
|
| 1922 |
+
"step": 31900
|
| 1923 |
+
},
|
| 1924 |
+
{
|
| 1925 |
+
"epoch": 10.22,
|
| 1926 |
+
"learning_rate": 3.309414812448987e-05,
|
| 1927 |
+
"loss": 0.1781,
|
| 1928 |
+
"step": 32000
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 10.25,
|
| 1932 |
+
"learning_rate": 3.3058660704780153e-05,
|
| 1933 |
+
"loss": 0.1716,
|
| 1934 |
+
"step": 32100
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 10.28,
|
| 1938 |
+
"learning_rate": 3.302317328507044e-05,
|
| 1939 |
+
"loss": 0.1717,
|
| 1940 |
+
"step": 32200
|
| 1941 |
+
},
|
| 1942 |
+
{
|
| 1943 |
+
"epoch": 10.32,
|
| 1944 |
+
"learning_rate": 3.298768586536073e-05,
|
| 1945 |
+
"loss": 0.179,
|
| 1946 |
+
"step": 32300
|
| 1947 |
+
},
|
| 1948 |
+
{
|
| 1949 |
+
"epoch": 10.35,
|
| 1950 |
+
"learning_rate": 3.295219844565102e-05,
|
| 1951 |
+
"loss": 0.174,
|
| 1952 |
+
"step": 32400
|
| 1953 |
+
},
|
| 1954 |
+
{
|
| 1955 |
+
"epoch": 10.38,
|
| 1956 |
+
"learning_rate": 3.2916711025941305e-05,
|
| 1957 |
+
"loss": 0.178,
|
| 1958 |
+
"step": 32500
|
| 1959 |
+
},
|
| 1960 |
+
{
|
| 1961 |
+
"epoch": 10.41,
|
| 1962 |
+
"learning_rate": 3.288122360623159e-05,
|
| 1963 |
+
"loss": 0.1765,
|
| 1964 |
+
"step": 32600
|
| 1965 |
+
},
|
| 1966 |
+
{
|
| 1967 |
+
"epoch": 10.44,
|
| 1968 |
+
"learning_rate": 3.284573618652188e-05,
|
| 1969 |
+
"loss": 0.1824,
|
| 1970 |
+
"step": 32700
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"epoch": 10.48,
|
| 1974 |
+
"learning_rate": 3.281024876681217e-05,
|
| 1975 |
+
"loss": 0.1806,
|
| 1976 |
+
"step": 32800
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 10.51,
|
| 1980 |
+
"learning_rate": 3.277511622129955e-05,
|
| 1981 |
+
"loss": 0.1785,
|
| 1982 |
+
"step": 32900
|
| 1983 |
+
},
|
| 1984 |
+
{
|
| 1985 |
+
"epoch": 10.54,
|
| 1986 |
+
"learning_rate": 3.2739983675786936e-05,
|
| 1987 |
+
"loss": 0.1698,
|
| 1988 |
+
"step": 33000
|
| 1989 |
+
},
|
| 1990 |
+
{
|
| 1991 |
+
"epoch": 10.57,
|
| 1992 |
+
"learning_rate": 3.2704496256077224e-05,
|
| 1993 |
+
"loss": 0.1734,
|
| 1994 |
+
"step": 33100
|
| 1995 |
+
},
|
| 1996 |
+
{
|
| 1997 |
+
"epoch": 10.6,
|
| 1998 |
+
"learning_rate": 3.2669008836367505e-05,
|
| 1999 |
+
"loss": 0.1761,
|
| 2000 |
+
"step": 33200
|
| 2001 |
+
},
|
| 2002 |
+
{
|
| 2003 |
+
"epoch": 10.64,
|
| 2004 |
+
"learning_rate": 3.263352141665779e-05,
|
| 2005 |
+
"loss": 0.1786,
|
| 2006 |
+
"step": 33300
|
| 2007 |
+
},
|
| 2008 |
+
{
|
| 2009 |
+
"epoch": 10.67,
|
| 2010 |
+
"learning_rate": 3.259803399694808e-05,
|
| 2011 |
+
"loss": 0.1767,
|
| 2012 |
+
"step": 33400
|
| 2013 |
+
},
|
| 2014 |
+
{
|
| 2015 |
+
"epoch": 10.7,
|
| 2016 |
+
"learning_rate": 3.256254657723837e-05,
|
| 2017 |
+
"loss": 0.1881,
|
| 2018 |
+
"step": 33500
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 10.73,
|
| 2022 |
+
"learning_rate": 3.252705915752866e-05,
|
| 2023 |
+
"loss": 0.1792,
|
| 2024 |
+
"step": 33600
|
| 2025 |
+
},
|
| 2026 |
+
{
|
| 2027 |
+
"epoch": 10.76,
|
| 2028 |
+
"learning_rate": 3.2491571737818945e-05,
|
| 2029 |
+
"loss": 0.1832,
|
| 2030 |
+
"step": 33700
|
| 2031 |
+
},
|
| 2032 |
+
{
|
| 2033 |
+
"epoch": 10.8,
|
| 2034 |
+
"learning_rate": 3.245608431810923e-05,
|
| 2035 |
+
"loss": 0.1889,
|
| 2036 |
+
"step": 33800
|
| 2037 |
+
},
|
| 2038 |
+
{
|
| 2039 |
+
"epoch": 10.83,
|
| 2040 |
+
"learning_rate": 3.242059689839952e-05,
|
| 2041 |
+
"loss": 0.1824,
|
| 2042 |
+
"step": 33900
|
| 2043 |
+
},
|
| 2044 |
+
{
|
| 2045 |
+
"epoch": 10.86,
|
| 2046 |
+
"learning_rate": 3.238510947868981e-05,
|
| 2047 |
+
"loss": 0.1746,
|
| 2048 |
+
"step": 34000
|
| 2049 |
+
},
|
| 2050 |
+
{
|
| 2051 |
+
"epoch": 10.89,
|
| 2052 |
+
"learning_rate": 3.2349622058980096e-05,
|
| 2053 |
+
"loss": 0.1735,
|
| 2054 |
+
"step": 34100
|
| 2055 |
+
},
|
| 2056 |
+
{
|
| 2057 |
+
"epoch": 10.92,
|
| 2058 |
+
"learning_rate": 3.2314134639270384e-05,
|
| 2059 |
+
"loss": 0.1822,
|
| 2060 |
+
"step": 34200
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 10.95,
|
| 2064 |
+
"learning_rate": 3.227864721956067e-05,
|
| 2065 |
+
"loss": 0.1768,
|
| 2066 |
+
"step": 34300
|
| 2067 |
+
},
|
| 2068 |
+
{
|
| 2069 |
+
"epoch": 10.99,
|
| 2070 |
+
"learning_rate": 3.224315979985095e-05,
|
| 2071 |
+
"loss": 0.1812,
|
| 2072 |
+
"step": 34400
|
| 2073 |
+
},
|
| 2074 |
+
{
|
| 2075 |
+
"epoch": 11.02,
|
| 2076 |
+
"learning_rate": 3.220767238014124e-05,
|
| 2077 |
+
"loss": 0.1645,
|
| 2078 |
+
"step": 34500
|
| 2079 |
+
},
|
| 2080 |
+
{
|
| 2081 |
+
"epoch": 11.05,
|
| 2082 |
+
"learning_rate": 3.217218496043153e-05,
|
| 2083 |
+
"loss": 0.1483,
|
| 2084 |
+
"step": 34600
|
| 2085 |
+
},
|
| 2086 |
+
{
|
| 2087 |
+
"epoch": 11.08,
|
| 2088 |
+
"learning_rate": 3.213669754072182e-05,
|
| 2089 |
+
"loss": 0.1615,
|
| 2090 |
+
"step": 34700
|
| 2091 |
+
},
|
| 2092 |
+
{
|
| 2093 |
+
"epoch": 11.11,
|
| 2094 |
+
"learning_rate": 3.2101210121012105e-05,
|
| 2095 |
+
"loss": 0.1637,
|
| 2096 |
+
"step": 34800
|
| 2097 |
+
},
|
| 2098 |
+
{
|
| 2099 |
+
"epoch": 11.15,
|
| 2100 |
+
"learning_rate": 3.206572270130239e-05,
|
| 2101 |
+
"loss": 0.1497,
|
| 2102 |
+
"step": 34900
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 11.18,
|
| 2106 |
+
"learning_rate": 3.2030235281592674e-05,
|
| 2107 |
+
"loss": 0.1524,
|
| 2108 |
+
"step": 35000
|
| 2109 |
+
},
|
| 2110 |
+
{
|
| 2111 |
+
"epoch": 11.21,
|
| 2112 |
+
"learning_rate": 3.199474786188296e-05,
|
| 2113 |
+
"loss": 0.1545,
|
| 2114 |
+
"step": 35100
|
| 2115 |
+
},
|
| 2116 |
+
{
|
| 2117 |
+
"epoch": 11.24,
|
| 2118 |
+
"learning_rate": 3.195926044217325e-05,
|
| 2119 |
+
"loss": 0.1526,
|
| 2120 |
+
"step": 35200
|
| 2121 |
+
},
|
| 2122 |
+
{
|
| 2123 |
+
"epoch": 11.27,
|
| 2124 |
+
"learning_rate": 3.192377302246354e-05,
|
| 2125 |
+
"loss": 0.1499,
|
| 2126 |
+
"step": 35300
|
| 2127 |
+
},
|
| 2128 |
+
{
|
| 2129 |
+
"epoch": 11.31,
|
| 2130 |
+
"learning_rate": 3.1888285602753826e-05,
|
| 2131 |
+
"loss": 0.1553,
|
| 2132 |
+
"step": 35400
|
| 2133 |
+
},
|
| 2134 |
+
{
|
| 2135 |
+
"epoch": 11.34,
|
| 2136 |
+
"learning_rate": 3.1852798183044114e-05,
|
| 2137 |
+
"loss": 0.1583,
|
| 2138 |
+
"step": 35500
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 11.37,
|
| 2142 |
+
"learning_rate": 3.18173107633344e-05,
|
| 2143 |
+
"loss": 0.156,
|
| 2144 |
+
"step": 35600
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 11.4,
|
| 2148 |
+
"learning_rate": 3.178182334362469e-05,
|
| 2149 |
+
"loss": 0.1582,
|
| 2150 |
+
"step": 35700
|
| 2151 |
+
},
|
| 2152 |
+
{
|
| 2153 |
+
"epoch": 11.43,
|
| 2154 |
+
"learning_rate": 3.174633592391498e-05,
|
| 2155 |
+
"loss": 0.155,
|
| 2156 |
+
"step": 35800
|
| 2157 |
+
},
|
| 2158 |
+
{
|
| 2159 |
+
"epoch": 11.47,
|
| 2160 |
+
"learning_rate": 3.1710848504205265e-05,
|
| 2161 |
+
"loss": 0.1546,
|
| 2162 |
+
"step": 35900
|
| 2163 |
+
},
|
| 2164 |
+
{
|
| 2165 |
+
"epoch": 11.5,
|
| 2166 |
+
"learning_rate": 3.167536108449555e-05,
|
| 2167 |
+
"loss": 0.157,
|
| 2168 |
+
"step": 36000
|
| 2169 |
+
},
|
| 2170 |
+
{
|
| 2171 |
+
"epoch": 11.53,
|
| 2172 |
+
"learning_rate": 3.1639873664785834e-05,
|
| 2173 |
+
"loss": 0.1644,
|
| 2174 |
+
"step": 36100
|
| 2175 |
+
},
|
| 2176 |
+
{
|
| 2177 |
+
"epoch": 11.56,
|
| 2178 |
+
"learning_rate": 3.160438624507612e-05,
|
| 2179 |
+
"loss": 0.1627,
|
| 2180 |
+
"step": 36200
|
| 2181 |
+
},
|
| 2182 |
+
{
|
| 2183 |
+
"epoch": 11.59,
|
| 2184 |
+
"learning_rate": 3.156889882536641e-05,
|
| 2185 |
+
"loss": 0.165,
|
| 2186 |
+
"step": 36300
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 11.63,
|
| 2190 |
+
"learning_rate": 3.15334114056567e-05,
|
| 2191 |
+
"loss": 0.1602,
|
| 2192 |
+
"step": 36400
|
| 2193 |
+
},
|
| 2194 |
+
{
|
| 2195 |
+
"epoch": 11.66,
|
| 2196 |
+
"learning_rate": 3.1497923985946986e-05,
|
| 2197 |
+
"loss": 0.1694,
|
| 2198 |
+
"step": 36500
|
| 2199 |
+
},
|
| 2200 |
+
{
|
| 2201 |
+
"epoch": 11.69,
|
| 2202 |
+
"learning_rate": 3.1462436566237274e-05,
|
| 2203 |
+
"loss": 0.1604,
|
| 2204 |
+
"step": 36600
|
| 2205 |
+
},
|
| 2206 |
+
{
|
| 2207 |
+
"epoch": 11.72,
|
| 2208 |
+
"learning_rate": 3.1426949146527555e-05,
|
| 2209 |
+
"loss": 0.1725,
|
| 2210 |
+
"step": 36700
|
| 2211 |
+
},
|
| 2212 |
+
{
|
| 2213 |
+
"epoch": 11.75,
|
| 2214 |
+
"learning_rate": 3.139146172681784e-05,
|
| 2215 |
+
"loss": 0.1577,
|
| 2216 |
+
"step": 36800
|
| 2217 |
+
},
|
| 2218 |
+
{
|
| 2219 |
+
"epoch": 11.79,
|
| 2220 |
+
"learning_rate": 3.135597430710813e-05,
|
| 2221 |
+
"loss": 0.1741,
|
| 2222 |
+
"step": 36900
|
| 2223 |
+
},
|
| 2224 |
+
{
|
| 2225 |
+
"epoch": 11.82,
|
| 2226 |
+
"learning_rate": 3.132048688739842e-05,
|
| 2227 |
+
"loss": 0.1547,
|
| 2228 |
+
"step": 37000
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 11.85,
|
| 2232 |
+
"learning_rate": 3.1284999467688707e-05,
|
| 2233 |
+
"loss": 0.1621,
|
| 2234 |
+
"step": 37100
|
| 2235 |
+
},
|
| 2236 |
+
{
|
| 2237 |
+
"epoch": 11.88,
|
| 2238 |
+
"learning_rate": 3.1249512047978994e-05,
|
| 2239 |
+
"loss": 0.1663,
|
| 2240 |
+
"step": 37200
|
| 2241 |
+
},
|
| 2242 |
+
{
|
| 2243 |
+
"epoch": 11.91,
|
| 2244 |
+
"learning_rate": 3.121437950246638e-05,
|
| 2245 |
+
"loss": 0.1594,
|
| 2246 |
+
"step": 37300
|
| 2247 |
+
},
|
| 2248 |
+
{
|
| 2249 |
+
"epoch": 11.95,
|
| 2250 |
+
"learning_rate": 3.117889208275667e-05,
|
| 2251 |
+
"loss": 0.174,
|
| 2252 |
+
"step": 37400
|
| 2253 |
+
},
|
| 2254 |
+
{
|
| 2255 |
+
"epoch": 11.98,
|
| 2256 |
+
"learning_rate": 3.114340466304696e-05,
|
| 2257 |
+
"loss": 0.1755,
|
| 2258 |
+
"step": 37500
|
| 2259 |
+
},
|
| 2260 |
+
{
|
| 2261 |
+
"epoch": 12.01,
|
| 2262 |
+
"learning_rate": 3.110791724333724e-05,
|
| 2263 |
+
"loss": 0.1627,
|
| 2264 |
+
"step": 37600
|
| 2265 |
+
},
|
| 2266 |
+
{
|
| 2267 |
+
"epoch": 12.04,
|
| 2268 |
+
"learning_rate": 3.1072429823627526e-05,
|
| 2269 |
+
"loss": 0.1445,
|
| 2270 |
+
"step": 37700
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 12.07,
|
| 2274 |
+
"learning_rate": 3.1036942403917814e-05,
|
| 2275 |
+
"loss": 0.1399,
|
| 2276 |
+
"step": 37800
|
| 2277 |
+
},
|
| 2278 |
+
{
|
| 2279 |
+
"epoch": 12.1,
|
| 2280 |
+
"learning_rate": 3.10014549842081e-05,
|
| 2281 |
+
"loss": 0.1475,
|
| 2282 |
+
"step": 37900
|
| 2283 |
+
},
|
| 2284 |
+
{
|
| 2285 |
+
"epoch": 12.14,
|
| 2286 |
+
"learning_rate": 3.096596756449839e-05,
|
| 2287 |
+
"loss": 0.1373,
|
| 2288 |
+
"step": 38000
|
| 2289 |
+
},
|
| 2290 |
+
{
|
| 2291 |
+
"epoch": 12.17,
|
| 2292 |
+
"learning_rate": 3.093048014478868e-05,
|
| 2293 |
+
"loss": 0.1379,
|
| 2294 |
+
"step": 38100
|
| 2295 |
+
},
|
| 2296 |
+
{
|
| 2297 |
+
"epoch": 12.2,
|
| 2298 |
+
"learning_rate": 3.089499272507896e-05,
|
| 2299 |
+
"loss": 0.1389,
|
| 2300 |
+
"step": 38200
|
| 2301 |
+
},
|
| 2302 |
+
{
|
| 2303 |
+
"epoch": 12.23,
|
| 2304 |
+
"learning_rate": 3.085950530536925e-05,
|
| 2305 |
+
"loss": 0.1394,
|
| 2306 |
+
"step": 38300
|
| 2307 |
+
},
|
| 2308 |
+
{
|
| 2309 |
+
"epoch": 12.26,
|
| 2310 |
+
"learning_rate": 3.0824017885659535e-05,
|
| 2311 |
+
"loss": 0.1434,
|
| 2312 |
+
"step": 38400
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 12.3,
|
| 2316 |
+
"learning_rate": 3.078853046594982e-05,
|
| 2317 |
+
"loss": 0.1467,
|
| 2318 |
+
"step": 38500
|
| 2319 |
+
},
|
| 2320 |
+
{
|
| 2321 |
+
"epoch": 12.33,
|
| 2322 |
+
"learning_rate": 3.075339792043721e-05,
|
| 2323 |
+
"loss": 0.1272,
|
| 2324 |
+
"step": 38600
|
| 2325 |
+
},
|
| 2326 |
+
{
|
| 2327 |
+
"epoch": 12.36,
|
| 2328 |
+
"learning_rate": 3.07179105007275e-05,
|
| 2329 |
+
"loss": 0.1363,
|
| 2330 |
+
"step": 38700
|
| 2331 |
+
},
|
| 2332 |
+
{
|
| 2333 |
+
"epoch": 12.39,
|
| 2334 |
+
"learning_rate": 3.0682423081017786e-05,
|
| 2335 |
+
"loss": 0.1467,
|
| 2336 |
+
"step": 38800
|
| 2337 |
+
},
|
| 2338 |
+
{
|
| 2339 |
+
"epoch": 12.42,
|
| 2340 |
+
"learning_rate": 3.0646935661308074e-05,
|
| 2341 |
+
"loss": 0.1389,
|
| 2342 |
+
"step": 38900
|
| 2343 |
+
},
|
| 2344 |
+
{
|
| 2345 |
+
"epoch": 12.46,
|
| 2346 |
+
"learning_rate": 3.0611448241598355e-05,
|
| 2347 |
+
"loss": 0.1447,
|
| 2348 |
+
"step": 39000
|
| 2349 |
+
},
|
| 2350 |
+
{
|
| 2351 |
+
"epoch": 12.49,
|
| 2352 |
+
"learning_rate": 3.057596082188864e-05,
|
| 2353 |
+
"loss": 0.1524,
|
| 2354 |
+
"step": 39100
|
| 2355 |
+
},
|
| 2356 |
+
{
|
| 2357 |
+
"epoch": 12.52,
|
| 2358 |
+
"learning_rate": 3.054047340217893e-05,
|
| 2359 |
+
"loss": 0.1434,
|
| 2360 |
+
"step": 39200
|
| 2361 |
+
},
|
| 2362 |
+
{
|
| 2363 |
+
"epoch": 12.55,
|
| 2364 |
+
"learning_rate": 3.050498598246922e-05,
|
| 2365 |
+
"loss": 0.142,
|
| 2366 |
+
"step": 39300
|
| 2367 |
+
},
|
| 2368 |
+
{
|
| 2369 |
+
"epoch": 12.58,
|
| 2370 |
+
"learning_rate": 3.0469498562759506e-05,
|
| 2371 |
+
"loss": 0.1495,
|
| 2372 |
+
"step": 39400
|
| 2373 |
+
},
|
| 2374 |
+
{
|
| 2375 |
+
"epoch": 12.62,
|
| 2376 |
+
"learning_rate": 3.0434011143049794e-05,
|
| 2377 |
+
"loss": 0.1451,
|
| 2378 |
+
"step": 39500
|
| 2379 |
+
},
|
| 2380 |
+
{
|
| 2381 |
+
"epoch": 12.65,
|
| 2382 |
+
"learning_rate": 3.0398523723340075e-05,
|
| 2383 |
+
"loss": 0.1464,
|
| 2384 |
+
"step": 39600
|
| 2385 |
+
},
|
| 2386 |
+
{
|
| 2387 |
+
"epoch": 12.68,
|
| 2388 |
+
"learning_rate": 3.0363036303630363e-05,
|
| 2389 |
+
"loss": 0.1419,
|
| 2390 |
+
"step": 39700
|
| 2391 |
+
},
|
| 2392 |
+
{
|
| 2393 |
+
"epoch": 12.71,
|
| 2394 |
+
"learning_rate": 3.032754888392065e-05,
|
| 2395 |
+
"loss": 0.1451,
|
| 2396 |
+
"step": 39800
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 12.74,
|
| 2400 |
+
"learning_rate": 3.029206146421094e-05,
|
| 2401 |
+
"loss": 0.151,
|
| 2402 |
+
"step": 39900
|
| 2403 |
+
},
|
| 2404 |
+
{
|
| 2405 |
+
"epoch": 12.78,
|
| 2406 |
+
"learning_rate": 3.0256574044501227e-05,
|
| 2407 |
+
"loss": 0.1442,
|
| 2408 |
+
"step": 40000
|
| 2409 |
+
},
|
| 2410 |
+
{
|
| 2411 |
+
"epoch": 12.81,
|
| 2412 |
+
"learning_rate": 3.0221086624791515e-05,
|
| 2413 |
+
"loss": 0.1553,
|
| 2414 |
+
"step": 40100
|
| 2415 |
+
},
|
| 2416 |
+
{
|
| 2417 |
+
"epoch": 12.84,
|
| 2418 |
+
"learning_rate": 3.01855992050818e-05,
|
| 2419 |
+
"loss": 0.1481,
|
| 2420 |
+
"step": 40200
|
| 2421 |
+
},
|
| 2422 |
+
{
|
| 2423 |
+
"epoch": 12.87,
|
| 2424 |
+
"learning_rate": 3.0150111785372087e-05,
|
| 2425 |
+
"loss": 0.1442,
|
| 2426 |
+
"step": 40300
|
| 2427 |
+
},
|
| 2428 |
+
{
|
| 2429 |
+
"epoch": 12.9,
|
| 2430 |
+
"learning_rate": 3.0114624365662375e-05,
|
| 2431 |
+
"loss": 0.1504,
|
| 2432 |
+
"step": 40400
|
| 2433 |
+
},
|
| 2434 |
+
{
|
| 2435 |
+
"epoch": 12.94,
|
| 2436 |
+
"learning_rate": 3.0079136945952663e-05,
|
| 2437 |
+
"loss": 0.1476,
|
| 2438 |
+
"step": 40500
|
| 2439 |
+
},
|
| 2440 |
+
{
|
| 2441 |
+
"epoch": 12.97,
|
| 2442 |
+
"learning_rate": 3.004364952624295e-05,
|
| 2443 |
+
"loss": 0.145,
|
| 2444 |
+
"step": 40600
|
| 2445 |
+
},
|
| 2446 |
+
{
|
| 2447 |
+
"epoch": 13.0,
|
| 2448 |
+
"learning_rate": 3.000816210653324e-05,
|
| 2449 |
+
"loss": 0.1549,
|
| 2450 |
+
"step": 40700
|
| 2451 |
+
},
|
| 2452 |
+
{
|
| 2453 |
+
"epoch": 13.03,
|
| 2454 |
+
"learning_rate": 2.997267468682352e-05,
|
| 2455 |
+
"loss": 0.1253,
|
| 2456 |
+
"step": 40800
|
| 2457 |
+
},
|
| 2458 |
+
{
|
| 2459 |
+
"epoch": 13.06,
|
| 2460 |
+
"learning_rate": 2.9937187267113808e-05,
|
| 2461 |
+
"loss": 0.1259,
|
| 2462 |
+
"step": 40900
|
| 2463 |
+
},
|
| 2464 |
+
{
|
| 2465 |
+
"epoch": 13.09,
|
| 2466 |
+
"learning_rate": 2.9901699847404096e-05,
|
| 2467 |
+
"loss": 0.1334,
|
| 2468 |
+
"step": 41000
|
| 2469 |
+
},
|
| 2470 |
+
{
|
| 2471 |
+
"epoch": 13.13,
|
| 2472 |
+
"learning_rate": 2.9866212427694384e-05,
|
| 2473 |
+
"loss": 0.1184,
|
| 2474 |
+
"step": 41100
|
| 2475 |
+
},
|
| 2476 |
+
{
|
| 2477 |
+
"epoch": 13.16,
|
| 2478 |
+
"learning_rate": 2.9830725007984672e-05,
|
| 2479 |
+
"loss": 0.1323,
|
| 2480 |
+
"step": 41200
|
| 2481 |
+
},
|
| 2482 |
+
{
|
| 2483 |
+
"epoch": 13.19,
|
| 2484 |
+
"learning_rate": 2.979523758827496e-05,
|
| 2485 |
+
"loss": 0.1294,
|
| 2486 |
+
"step": 41300
|
| 2487 |
+
},
|
| 2488 |
+
{
|
| 2489 |
+
"epoch": 13.22,
|
| 2490 |
+
"learning_rate": 2.9759750168565244e-05,
|
| 2491 |
+
"loss": 0.1283,
|
| 2492 |
+
"step": 41400
|
| 2493 |
+
},
|
| 2494 |
+
{
|
| 2495 |
+
"epoch": 13.25,
|
| 2496 |
+
"learning_rate": 2.9724262748855532e-05,
|
| 2497 |
+
"loss": 0.1313,
|
| 2498 |
+
"step": 41500
|
| 2499 |
+
},
|
| 2500 |
+
{
|
| 2501 |
+
"epoch": 13.29,
|
| 2502 |
+
"learning_rate": 2.968877532914582e-05,
|
| 2503 |
+
"loss": 0.1336,
|
| 2504 |
+
"step": 41600
|
| 2505 |
+
},
|
| 2506 |
+
{
|
| 2507 |
+
"epoch": 13.32,
|
| 2508 |
+
"learning_rate": 2.9653287909436108e-05,
|
| 2509 |
+
"loss": 0.1329,
|
| 2510 |
+
"step": 41700
|
| 2511 |
+
},
|
| 2512 |
+
{
|
| 2513 |
+
"epoch": 13.35,
|
| 2514 |
+
"learning_rate": 2.9617800489726396e-05,
|
| 2515 |
+
"loss": 0.1336,
|
| 2516 |
+
"step": 41800
|
| 2517 |
+
},
|
| 2518 |
+
{
|
| 2519 |
+
"epoch": 13.38,
|
| 2520 |
+
"learning_rate": 2.958231307001668e-05,
|
| 2521 |
+
"loss": 0.1305,
|
| 2522 |
+
"step": 41900
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 13.41,
|
| 2526 |
+
"learning_rate": 2.954682565030697e-05,
|
| 2527 |
+
"loss": 0.1308,
|
| 2528 |
+
"step": 42000
|
| 2529 |
+
},
|
| 2530 |
+
{
|
| 2531 |
+
"epoch": 13.45,
|
| 2532 |
+
"learning_rate": 2.9511338230597256e-05,
|
| 2533 |
+
"loss": 0.1381,
|
| 2534 |
+
"step": 42100
|
| 2535 |
+
},
|
| 2536 |
+
{
|
| 2537 |
+
"epoch": 13.48,
|
| 2538 |
+
"learning_rate": 2.9475850810887544e-05,
|
| 2539 |
+
"loss": 0.1379,
|
| 2540 |
+
"step": 42200
|
| 2541 |
+
},
|
| 2542 |
+
{
|
| 2543 |
+
"epoch": 13.51,
|
| 2544 |
+
"learning_rate": 2.9440363391177832e-05,
|
| 2545 |
+
"loss": 0.1359,
|
| 2546 |
+
"step": 42300
|
| 2547 |
+
},
|
| 2548 |
+
{
|
| 2549 |
+
"epoch": 13.54,
|
| 2550 |
+
"learning_rate": 2.940487597146812e-05,
|
| 2551 |
+
"loss": 0.1273,
|
| 2552 |
+
"step": 42400
|
| 2553 |
+
},
|
| 2554 |
+
{
|
| 2555 |
+
"epoch": 13.57,
|
| 2556 |
+
"learning_rate": 2.93693885517584e-05,
|
| 2557 |
+
"loss": 0.1331,
|
| 2558 |
+
"step": 42500
|
| 2559 |
+
},
|
| 2560 |
+
{
|
| 2561 |
+
"epoch": 13.61,
|
| 2562 |
+
"learning_rate": 2.933390113204869e-05,
|
| 2563 |
+
"loss": 0.1315,
|
| 2564 |
+
"step": 42600
|
| 2565 |
+
},
|
| 2566 |
+
{
|
| 2567 |
+
"epoch": 13.64,
|
| 2568 |
+
"learning_rate": 2.9298413712338977e-05,
|
| 2569 |
+
"loss": 0.1335,
|
| 2570 |
+
"step": 42700
|
| 2571 |
+
},
|
| 2572 |
+
{
|
| 2573 |
+
"epoch": 13.67,
|
| 2574 |
+
"learning_rate": 2.9262926292629265e-05,
|
| 2575 |
+
"loss": 0.1344,
|
| 2576 |
+
"step": 42800
|
| 2577 |
+
},
|
| 2578 |
+
{
|
| 2579 |
+
"epoch": 13.7,
|
| 2580 |
+
"learning_rate": 2.9227438872919553e-05,
|
| 2581 |
+
"loss": 0.1411,
|
| 2582 |
+
"step": 42900
|
| 2583 |
+
},
|
| 2584 |
+
{
|
| 2585 |
+
"epoch": 13.73,
|
| 2586 |
+
"learning_rate": 2.919195145320984e-05,
|
| 2587 |
+
"loss": 0.132,
|
| 2588 |
+
"step": 43000
|
| 2589 |
+
},
|
| 2590 |
+
{
|
| 2591 |
+
"epoch": 13.77,
|
| 2592 |
+
"learning_rate": 2.9156464033500125e-05,
|
| 2593 |
+
"loss": 0.1338,
|
| 2594 |
+
"step": 43100
|
| 2595 |
+
},
|
| 2596 |
+
{
|
| 2597 |
+
"epoch": 13.8,
|
| 2598 |
+
"learning_rate": 2.9120976613790413e-05,
|
| 2599 |
+
"loss": 0.125,
|
| 2600 |
+
"step": 43200
|
| 2601 |
+
},
|
| 2602 |
+
{
|
| 2603 |
+
"epoch": 13.83,
|
| 2604 |
+
"learning_rate": 2.90858440682778e-05,
|
| 2605 |
+
"loss": 0.1389,
|
| 2606 |
+
"step": 43300
|
| 2607 |
+
},
|
| 2608 |
+
{
|
| 2609 |
+
"epoch": 13.86,
|
| 2610 |
+
"learning_rate": 2.9050356648568085e-05,
|
| 2611 |
+
"loss": 0.1386,
|
| 2612 |
+
"step": 43400
|
| 2613 |
+
},
|
| 2614 |
+
{
|
| 2615 |
+
"epoch": 13.89,
|
| 2616 |
+
"learning_rate": 2.9014869228858373e-05,
|
| 2617 |
+
"loss": 0.13,
|
| 2618 |
+
"step": 43500
|
| 2619 |
+
},
|
| 2620 |
+
{
|
| 2621 |
+
"epoch": 13.93,
|
| 2622 |
+
"learning_rate": 2.897938180914866e-05,
|
| 2623 |
+
"loss": 0.1351,
|
| 2624 |
+
"step": 43600
|
| 2625 |
+
},
|
| 2626 |
+
{
|
| 2627 |
+
"epoch": 13.96,
|
| 2628 |
+
"learning_rate": 2.894389438943895e-05,
|
| 2629 |
+
"loss": 0.1341,
|
| 2630 |
+
"step": 43700
|
| 2631 |
+
},
|
| 2632 |
+
{
|
| 2633 |
+
"epoch": 13.99,
|
| 2634 |
+
"learning_rate": 2.8908406969729236e-05,
|
| 2635 |
+
"loss": 0.1399,
|
| 2636 |
+
"step": 43800
|
| 2637 |
+
},
|
| 2638 |
+
{
|
| 2639 |
+
"epoch": 14.02,
|
| 2640 |
+
"learning_rate": 2.887291955001952e-05,
|
| 2641 |
+
"loss": 0.1186,
|
| 2642 |
+
"step": 43900
|
| 2643 |
+
},
|
| 2644 |
+
{
|
| 2645 |
+
"epoch": 14.05,
|
| 2646 |
+
"learning_rate": 2.8837432130309805e-05,
|
| 2647 |
+
"loss": 0.1124,
|
| 2648 |
+
"step": 44000
|
| 2649 |
+
},
|
| 2650 |
+
{
|
| 2651 |
+
"epoch": 14.08,
|
| 2652 |
+
"learning_rate": 2.8801944710600093e-05,
|
| 2653 |
+
"loss": 0.1156,
|
| 2654 |
+
"step": 44100
|
| 2655 |
+
},
|
| 2656 |
+
{
|
| 2657 |
+
"epoch": 14.12,
|
| 2658 |
+
"learning_rate": 2.876645729089038e-05,
|
| 2659 |
+
"loss": 0.1161,
|
| 2660 |
+
"step": 44200
|
| 2661 |
+
},
|
| 2662 |
+
{
|
| 2663 |
+
"epoch": 14.15,
|
| 2664 |
+
"learning_rate": 2.873096987118067e-05,
|
| 2665 |
+
"loss": 0.113,
|
| 2666 |
+
"step": 44300
|
| 2667 |
+
},
|
| 2668 |
+
{
|
| 2669 |
+
"epoch": 14.18,
|
| 2670 |
+
"learning_rate": 2.8695482451470957e-05,
|
| 2671 |
+
"loss": 0.1219,
|
| 2672 |
+
"step": 44400
|
| 2673 |
+
},
|
| 2674 |
+
{
|
| 2675 |
+
"epoch": 14.21,
|
| 2676 |
+
"learning_rate": 2.8659995031761245e-05,
|
| 2677 |
+
"loss": 0.1228,
|
| 2678 |
+
"step": 44500
|
| 2679 |
+
},
|
| 2680 |
+
{
|
| 2681 |
+
"epoch": 14.24,
|
| 2682 |
+
"learning_rate": 2.862450761205153e-05,
|
| 2683 |
+
"loss": 0.1208,
|
| 2684 |
+
"step": 44600
|
| 2685 |
+
},
|
| 2686 |
+
{
|
| 2687 |
+
"epoch": 14.28,
|
| 2688 |
+
"learning_rate": 2.8589020192341817e-05,
|
| 2689 |
+
"loss": 0.1213,
|
| 2690 |
+
"step": 44700
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 14.31,
|
| 2694 |
+
"learning_rate": 2.8553532772632105e-05,
|
| 2695 |
+
"loss": 0.1147,
|
| 2696 |
+
"step": 44800
|
| 2697 |
+
},
|
| 2698 |
+
{
|
| 2699 |
+
"epoch": 14.34,
|
| 2700 |
+
"learning_rate": 2.8518045352922393e-05,
|
| 2701 |
+
"loss": 0.1198,
|
| 2702 |
+
"step": 44900
|
| 2703 |
+
},
|
| 2704 |
+
{
|
| 2705 |
+
"epoch": 14.37,
|
| 2706 |
+
"learning_rate": 2.848255793321268e-05,
|
| 2707 |
+
"loss": 0.1144,
|
| 2708 |
+
"step": 45000
|
| 2709 |
+
},
|
| 2710 |
+
{
|
| 2711 |
+
"epoch": 14.4,
|
| 2712 |
+
"learning_rate": 2.8447070513502962e-05,
|
| 2713 |
+
"loss": 0.1172,
|
| 2714 |
+
"step": 45100
|
| 2715 |
+
},
|
| 2716 |
+
{
|
| 2717 |
+
"epoch": 14.44,
|
| 2718 |
+
"learning_rate": 2.841158309379325e-05,
|
| 2719 |
+
"loss": 0.1158,
|
| 2720 |
+
"step": 45200
|
| 2721 |
+
},
|
| 2722 |
+
{
|
| 2723 |
+
"epoch": 14.47,
|
| 2724 |
+
"learning_rate": 2.8376095674083538e-05,
|
| 2725 |
+
"loss": 0.1188,
|
| 2726 |
+
"step": 45300
|
| 2727 |
+
},
|
| 2728 |
+
{
|
| 2729 |
+
"epoch": 14.5,
|
| 2730 |
+
"learning_rate": 2.8340608254373826e-05,
|
| 2731 |
+
"loss": 0.1249,
|
| 2732 |
+
"step": 45400
|
| 2733 |
+
},
|
| 2734 |
+
{
|
| 2735 |
+
"epoch": 14.53,
|
| 2736 |
+
"learning_rate": 2.8305120834664114e-05,
|
| 2737 |
+
"loss": 0.1208,
|
| 2738 |
+
"step": 45500
|
| 2739 |
+
},
|
| 2740 |
+
{
|
| 2741 |
+
"epoch": 14.56,
|
| 2742 |
+
"learning_rate": 2.8269633414954402e-05,
|
| 2743 |
+
"loss": 0.125,
|
| 2744 |
+
"step": 45600
|
| 2745 |
+
},
|
| 2746 |
+
{
|
| 2747 |
+
"epoch": 14.6,
|
| 2748 |
+
"learning_rate": 2.8234145995244686e-05,
|
| 2749 |
+
"loss": 0.1153,
|
| 2750 |
+
"step": 45700
|
| 2751 |
+
},
|
| 2752 |
+
{
|
| 2753 |
+
"epoch": 14.63,
|
| 2754 |
+
"learning_rate": 2.8199013449732073e-05,
|
| 2755 |
+
"loss": 0.1262,
|
| 2756 |
+
"step": 45800
|
| 2757 |
+
},
|
| 2758 |
+
{
|
| 2759 |
+
"epoch": 14.66,
|
| 2760 |
+
"learning_rate": 2.816352603002236e-05,
|
| 2761 |
+
"loss": 0.1242,
|
| 2762 |
+
"step": 45900
|
| 2763 |
+
},
|
| 2764 |
+
{
|
| 2765 |
+
"epoch": 14.69,
|
| 2766 |
+
"learning_rate": 2.8128038610312646e-05,
|
| 2767 |
+
"loss": 0.1286,
|
| 2768 |
+
"step": 46000
|
| 2769 |
+
},
|
| 2770 |
+
{
|
| 2771 |
+
"epoch": 14.72,
|
| 2772 |
+
"learning_rate": 2.8092551190602934e-05,
|
| 2773 |
+
"loss": 0.1238,
|
| 2774 |
+
"step": 46100
|
| 2775 |
+
},
|
| 2776 |
+
{
|
| 2777 |
+
"epoch": 14.76,
|
| 2778 |
+
"learning_rate": 2.805706377089322e-05,
|
| 2779 |
+
"loss": 0.1303,
|
| 2780 |
+
"step": 46200
|
| 2781 |
+
},
|
| 2782 |
+
{
|
| 2783 |
+
"epoch": 14.79,
|
| 2784 |
+
"learning_rate": 2.802157635118351e-05,
|
| 2785 |
+
"loss": 0.1249,
|
| 2786 |
+
"step": 46300
|
| 2787 |
+
},
|
| 2788 |
+
{
|
| 2789 |
+
"epoch": 14.82,
|
| 2790 |
+
"learning_rate": 2.7986088931473797e-05,
|
| 2791 |
+
"loss": 0.1301,
|
| 2792 |
+
"step": 46400
|
| 2793 |
+
},
|
| 2794 |
+
{
|
| 2795 |
+
"epoch": 14.85,
|
| 2796 |
+
"learning_rate": 2.7950601511764085e-05,
|
| 2797 |
+
"loss": 0.1276,
|
| 2798 |
+
"step": 46500
|
| 2799 |
+
},
|
| 2800 |
+
{
|
| 2801 |
+
"epoch": 14.88,
|
| 2802 |
+
"learning_rate": 2.7915114092054366e-05,
|
| 2803 |
+
"loss": 0.1271,
|
| 2804 |
+
"step": 46600
|
| 2805 |
+
},
|
| 2806 |
+
{
|
| 2807 |
+
"epoch": 14.92,
|
| 2808 |
+
"learning_rate": 2.7879626672344654e-05,
|
| 2809 |
+
"loss": 0.1275,
|
| 2810 |
+
"step": 46700
|
| 2811 |
+
},
|
| 2812 |
+
{
|
| 2813 |
+
"epoch": 14.95,
|
| 2814 |
+
"learning_rate": 2.7844139252634942e-05,
|
| 2815 |
+
"loss": 0.1239,
|
| 2816 |
+
"step": 46800
|
| 2817 |
+
},
|
| 2818 |
+
{
|
| 2819 |
+
"epoch": 14.98,
|
| 2820 |
+
"learning_rate": 2.780865183292523e-05,
|
| 2821 |
+
"loss": 0.1177,
|
| 2822 |
+
"step": 46900
|
| 2823 |
+
},
|
| 2824 |
+
{
|
| 2825 |
+
"epoch": 15.01,
|
| 2826 |
+
"learning_rate": 2.7773164413215518e-05,
|
| 2827 |
+
"loss": 0.1172,
|
| 2828 |
+
"step": 47000
|
| 2829 |
+
},
|
| 2830 |
+
{
|
| 2831 |
+
"epoch": 15.04,
|
| 2832 |
+
"learning_rate": 2.7737676993505806e-05,
|
| 2833 |
+
"loss": 0.105,
|
| 2834 |
+
"step": 47100
|
| 2835 |
+
},
|
| 2836 |
+
{
|
| 2837 |
+
"epoch": 15.08,
|
| 2838 |
+
"learning_rate": 2.770254444799319e-05,
|
| 2839 |
+
"loss": 0.1079,
|
| 2840 |
+
"step": 47200
|
| 2841 |
+
},
|
| 2842 |
+
{
|
| 2843 |
+
"epoch": 15.11,
|
| 2844 |
+
"learning_rate": 2.7667057028283478e-05,
|
| 2845 |
+
"loss": 0.1038,
|
| 2846 |
+
"step": 47300
|
| 2847 |
+
},
|
| 2848 |
+
{
|
| 2849 |
+
"epoch": 15.14,
|
| 2850 |
+
"learning_rate": 2.7631569608573765e-05,
|
| 2851 |
+
"loss": 0.1056,
|
| 2852 |
+
"step": 47400
|
| 2853 |
+
},
|
| 2854 |
+
{
|
| 2855 |
+
"epoch": 15.17,
|
| 2856 |
+
"learning_rate": 2.759608218886405e-05,
|
| 2857 |
+
"loss": 0.1102,
|
| 2858 |
+
"step": 47500
|
| 2859 |
+
},
|
| 2860 |
+
{
|
| 2861 |
+
"epoch": 15.2,
|
| 2862 |
+
"learning_rate": 2.7560594769154338e-05,
|
| 2863 |
+
"loss": 0.1105,
|
| 2864 |
+
"step": 47600
|
| 2865 |
+
},
|
| 2866 |
+
{
|
| 2867 |
+
"epoch": 15.23,
|
| 2868 |
+
"learning_rate": 2.7525107349444626e-05,
|
| 2869 |
+
"loss": 0.1076,
|
| 2870 |
+
"step": 47700
|
| 2871 |
+
},
|
| 2872 |
+
{
|
| 2873 |
+
"epoch": 15.27,
|
| 2874 |
+
"learning_rate": 2.748961992973491e-05,
|
| 2875 |
+
"loss": 0.1072,
|
| 2876 |
+
"step": 47800
|
| 2877 |
+
},
|
| 2878 |
+
{
|
| 2879 |
+
"epoch": 15.3,
|
| 2880 |
+
"learning_rate": 2.7454132510025198e-05,
|
| 2881 |
+
"loss": 0.109,
|
| 2882 |
+
"step": 47900
|
| 2883 |
+
},
|
| 2884 |
+
{
|
| 2885 |
+
"epoch": 15.33,
|
| 2886 |
+
"learning_rate": 2.7418645090315483e-05,
|
| 2887 |
+
"loss": 0.1042,
|
| 2888 |
+
"step": 48000
|
| 2889 |
+
},
|
| 2890 |
+
{
|
| 2891 |
+
"epoch": 15.36,
|
| 2892 |
+
"learning_rate": 2.738315767060577e-05,
|
| 2893 |
+
"loss": 0.1077,
|
| 2894 |
+
"step": 48100
|
| 2895 |
+
},
|
| 2896 |
+
{
|
| 2897 |
+
"epoch": 15.39,
|
| 2898 |
+
"learning_rate": 2.734767025089606e-05,
|
| 2899 |
+
"loss": 0.1127,
|
| 2900 |
+
"step": 48200
|
| 2901 |
+
},
|
| 2902 |
+
{
|
| 2903 |
+
"epoch": 15.43,
|
| 2904 |
+
"learning_rate": 2.7312182831186346e-05,
|
| 2905 |
+
"loss": 0.1048,
|
| 2906 |
+
"step": 48300
|
| 2907 |
+
},
|
| 2908 |
+
{
|
| 2909 |
+
"epoch": 15.46,
|
| 2910 |
+
"learning_rate": 2.7276695411476634e-05,
|
| 2911 |
+
"loss": 0.1131,
|
| 2912 |
+
"step": 48400
|
| 2913 |
+
},
|
| 2914 |
+
{
|
| 2915 |
+
"epoch": 15.49,
|
| 2916 |
+
"learning_rate": 2.7241207991766922e-05,
|
| 2917 |
+
"loss": 0.1096,
|
| 2918 |
+
"step": 48500
|
| 2919 |
+
},
|
| 2920 |
+
{
|
| 2921 |
+
"epoch": 15.52,
|
| 2922 |
+
"learning_rate": 2.7205720572057207e-05,
|
| 2923 |
+
"loss": 0.1134,
|
| 2924 |
+
"step": 48600
|
| 2925 |
+
},
|
| 2926 |
+
{
|
| 2927 |
+
"epoch": 15.55,
|
| 2928 |
+
"learning_rate": 2.7170233152347495e-05,
|
| 2929 |
+
"loss": 0.1117,
|
| 2930 |
+
"step": 48700
|
| 2931 |
+
},
|
| 2932 |
+
{
|
| 2933 |
+
"epoch": 15.59,
|
| 2934 |
+
"learning_rate": 2.7134745732637783e-05,
|
| 2935 |
+
"loss": 0.111,
|
| 2936 |
+
"step": 48800
|
| 2937 |
+
},
|
| 2938 |
+
{
|
| 2939 |
+
"epoch": 15.62,
|
| 2940 |
+
"learning_rate": 2.709925831292807e-05,
|
| 2941 |
+
"loss": 0.1171,
|
| 2942 |
+
"step": 48900
|
| 2943 |
+
},
|
| 2944 |
+
{
|
| 2945 |
+
"epoch": 15.65,
|
| 2946 |
+
"learning_rate": 2.706377089321836e-05,
|
| 2947 |
+
"loss": 0.1266,
|
| 2948 |
+
"step": 49000
|
| 2949 |
+
},
|
| 2950 |
+
{
|
| 2951 |
+
"epoch": 15.68,
|
| 2952 |
+
"learning_rate": 2.7028283473508646e-05,
|
| 2953 |
+
"loss": 0.114,
|
| 2954 |
+
"step": 49100
|
| 2955 |
+
},
|
| 2956 |
+
{
|
| 2957 |
+
"epoch": 15.71,
|
| 2958 |
+
"learning_rate": 2.6992796053798928e-05,
|
| 2959 |
+
"loss": 0.1158,
|
| 2960 |
+
"step": 49200
|
| 2961 |
+
},
|
| 2962 |
+
{
|
| 2963 |
+
"epoch": 15.75,
|
| 2964 |
+
"learning_rate": 2.6957663508286315e-05,
|
| 2965 |
+
"loss": 0.1097,
|
| 2966 |
+
"step": 49300
|
| 2967 |
+
},
|
| 2968 |
+
{
|
| 2969 |
+
"epoch": 15.78,
|
| 2970 |
+
"learning_rate": 2.6922176088576602e-05,
|
| 2971 |
+
"loss": 0.118,
|
| 2972 |
+
"step": 49400
|
| 2973 |
+
},
|
| 2974 |
+
{
|
| 2975 |
+
"epoch": 15.81,
|
| 2976 |
+
"learning_rate": 2.6886688668866887e-05,
|
| 2977 |
+
"loss": 0.1153,
|
| 2978 |
+
"step": 49500
|
| 2979 |
+
},
|
| 2980 |
+
{
|
| 2981 |
+
"epoch": 15.84,
|
| 2982 |
+
"learning_rate": 2.6851201249157175e-05,
|
| 2983 |
+
"loss": 0.111,
|
| 2984 |
+
"step": 49600
|
| 2985 |
+
},
|
| 2986 |
+
{
|
| 2987 |
+
"epoch": 15.87,
|
| 2988 |
+
"learning_rate": 2.6815713829447463e-05,
|
| 2989 |
+
"loss": 0.1213,
|
| 2990 |
+
"step": 49700
|
| 2991 |
+
},
|
| 2992 |
+
{
|
| 2993 |
+
"epoch": 15.91,
|
| 2994 |
+
"learning_rate": 2.678022640973775e-05,
|
| 2995 |
+
"loss": 0.1105,
|
| 2996 |
+
"step": 49800
|
| 2997 |
+
},
|
| 2998 |
+
{
|
| 2999 |
+
"epoch": 15.94,
|
| 3000 |
+
"learning_rate": 2.674473899002804e-05,
|
| 3001 |
+
"loss": 0.1128,
|
| 3002 |
+
"step": 49900
|
| 3003 |
+
},
|
| 3004 |
+
{
|
| 3005 |
+
"epoch": 15.97,
|
| 3006 |
+
"learning_rate": 2.6709251570318327e-05,
|
| 3007 |
+
"loss": 0.1198,
|
| 3008 |
+
"step": 50000
|
| 3009 |
+
}
|
| 3010 |
+
],
|
| 3011 |
+
"max_steps": 125240,
|
| 3012 |
+
"num_train_epochs": 40,
|
| 3013 |
+
"total_flos": 0.0,
|
| 3014 |
+
"trial_name": null,
|
| 3015 |
+
"trial_params": null
|
| 3016 |
+
}
|
checkpoint-55000/config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "Davlan/afro-xlmr-base",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"XLMRobertaModel"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 768,
|
| 13 |
+
"id2label": {
|
| 14 |
+
"0": "LABEL_0"
|
| 15 |
+
},
|
| 16 |
+
"initializer_range": 0.02,
|
| 17 |
+
"intermediate_size": 3072,
|
| 18 |
+
"label2id": {
|
| 19 |
+
"LABEL_0": 0
|
| 20 |
+
},
|
| 21 |
+
"layer_norm_eps": 1e-05,
|
| 22 |
+
"max_position_embeddings": 514,
|
| 23 |
+
"model_type": "xlm-roberta",
|
| 24 |
+
"num_attention_heads": 12,
|
| 25 |
+
"num_hidden_layers": 12,
|
| 26 |
+
"output_past": true,
|
| 27 |
+
"pad_token_id": 1,
|
| 28 |
+
"position_embedding_type": "absolute",
|
| 29 |
+
"torch_dtype": "float32",
|
| 30 |
+
"transformers_version": "4.29.0",
|
| 31 |
+
"type_vocab_size": 1,
|
| 32 |
+
"use_cache": true,
|
| 33 |
+
"vocab_size": 250002
|
| 34 |
+
}
|
checkpoint-55000/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4b8e2fc9b9e0c1791910882c54f04df5aa6b71684b4495b4e2d4e343687ad5f5
|
| 3 |
+
size 1112246250
|