Training in progress, step 20000, checkpoint
Browse files- checkpoint-20000/config.json +30 -0
- checkpoint-20000/generation_config.json +6 -0
- checkpoint-20000/model.safetensors +3 -0
- checkpoint-20000/optimizer.pt +3 -0
- checkpoint-20000/rng_state.pth +3 -0
- checkpoint-20000/scheduler.pt +3 -0
- checkpoint-20000/special_tokens_map.json +24 -0
- checkpoint-20000/tokenizer.json +96 -0
- checkpoint-20000/tokenizer_config.json +38 -0
- checkpoint-20000/trainer_state.json +1433 -0
- checkpoint-20000/training_args.bin +3 -0
checkpoint-20000/config.json
ADDED
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "EleutherAI/pythia-70m",
|
3 |
+
"architectures": [
|
4 |
+
"GPTNeoXForCausalLM"
|
5 |
+
],
|
6 |
+
"attention_bias": true,
|
7 |
+
"attention_dropout": 0.0,
|
8 |
+
"bos_token_id": 0,
|
9 |
+
"classifier_dropout": 0.1,
|
10 |
+
"eos_token_id": 0,
|
11 |
+
"hidden_act": "gelu",
|
12 |
+
"hidden_dropout": 0.0,
|
13 |
+
"hidden_size": 512,
|
14 |
+
"initializer_range": 0.02,
|
15 |
+
"intermediate_size": 2048,
|
16 |
+
"layer_norm_eps": 1e-05,
|
17 |
+
"max_position_embeddings": 2048,
|
18 |
+
"model_type": "gpt_neox",
|
19 |
+
"num_attention_heads": 8,
|
20 |
+
"num_hidden_layers": 6,
|
21 |
+
"rope_scaling": null,
|
22 |
+
"rotary_emb_base": 10000,
|
23 |
+
"rotary_pct": 0.25,
|
24 |
+
"tie_word_embeddings": false,
|
25 |
+
"torch_dtype": "float32",
|
26 |
+
"transformers_version": "4.42.3",
|
27 |
+
"use_cache": true,
|
28 |
+
"use_parallel_residual": true,
|
29 |
+
"vocab_size": 8
|
30 |
+
}
|
checkpoint-20000/generation_config.json
ADDED
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_from_model_config": true,
|
3 |
+
"bos_token_id": 0,
|
4 |
+
"eos_token_id": 0,
|
5 |
+
"transformers_version": "4.42.3"
|
6 |
+
}
|
checkpoint-20000/model.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:6a744392d80db7cd4293dbca4fec0475fa1b73efbdc0b88c9fb7b3b12009d90f
|
3 |
+
size 75702552
|
checkpoint-20000/optimizer.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:6981cbe51605329dd1bae02454c5ae8522322b1cd749f3293864b50fab2b4524
|
3 |
+
size 151452026
|
checkpoint-20000/rng_state.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:835f869ea325fd6edf27b48b589309fb66641cb92b45f2fc13d1bb6e8814106c
|
3 |
+
size 14244
|
checkpoint-20000/scheduler.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9d0925a4706094b70adf03dc767ec5ca4bdf4659b432e6f32e334a22542a84f5
|
3 |
+
size 1064
|
checkpoint-20000/special_tokens_map.json
ADDED
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"bos_token": {
|
3 |
+
"content": "<|beginoftext|>",
|
4 |
+
"lstrip": false,
|
5 |
+
"normalized": true,
|
6 |
+
"rstrip": false,
|
7 |
+
"single_word": false
|
8 |
+
},
|
9 |
+
"eos_token": {
|
10 |
+
"content": "<|endoftext|>",
|
11 |
+
"lstrip": false,
|
12 |
+
"normalized": false,
|
13 |
+
"rstrip": false,
|
14 |
+
"single_word": false
|
15 |
+
},
|
16 |
+
"pad_token": "<|endoftext|>",
|
17 |
+
"unk_token": {
|
18 |
+
"content": "<|unknown|>",
|
19 |
+
"lstrip": false,
|
20 |
+
"normalized": false,
|
21 |
+
"rstrip": false,
|
22 |
+
"single_word": false
|
23 |
+
}
|
24 |
+
}
|
checkpoint-20000/tokenizer.json
ADDED
@@ -0,0 +1,96 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"version": "1.0",
|
3 |
+
"truncation": null,
|
4 |
+
"padding": null,
|
5 |
+
"added_tokens": [
|
6 |
+
{
|
7 |
+
"id": 0,
|
8 |
+
"content": "<|beginoftext|>",
|
9 |
+
"single_word": false,
|
10 |
+
"lstrip": false,
|
11 |
+
"rstrip": false,
|
12 |
+
"normalized": true,
|
13 |
+
"special": true
|
14 |
+
},
|
15 |
+
{
|
16 |
+
"id": 1,
|
17 |
+
"content": "<|endoftext|>",
|
18 |
+
"single_word": false,
|
19 |
+
"lstrip": false,
|
20 |
+
"rstrip": false,
|
21 |
+
"normalized": false,
|
22 |
+
"special": true
|
23 |
+
},
|
24 |
+
{
|
25 |
+
"id": 2,
|
26 |
+
"content": "<|unknown|>",
|
27 |
+
"single_word": false,
|
28 |
+
"lstrip": false,
|
29 |
+
"rstrip": false,
|
30 |
+
"normalized": false,
|
31 |
+
"special": true
|
32 |
+
}
|
33 |
+
],
|
34 |
+
"normalizer": {
|
35 |
+
"type": "NFC"
|
36 |
+
},
|
37 |
+
"pre_tokenizer": {
|
38 |
+
"type": "ByteLevel",
|
39 |
+
"add_prefix_space": false,
|
40 |
+
"trim_offsets": true,
|
41 |
+
"use_regex": true
|
42 |
+
},
|
43 |
+
"post_processor": {
|
44 |
+
"type": "TemplateProcessing",
|
45 |
+
"single": [
|
46 |
+
{
|
47 |
+
"Sequence": {
|
48 |
+
"id": "A",
|
49 |
+
"type_id": 0
|
50 |
+
}
|
51 |
+
}
|
52 |
+
],
|
53 |
+
"pair": [
|
54 |
+
{
|
55 |
+
"Sequence": {
|
56 |
+
"id": "A",
|
57 |
+
"type_id": 0
|
58 |
+
}
|
59 |
+
},
|
60 |
+
{
|
61 |
+
"Sequence": {
|
62 |
+
"id": "B",
|
63 |
+
"type_id": 1
|
64 |
+
}
|
65 |
+
}
|
66 |
+
],
|
67 |
+
"special_tokens": {}
|
68 |
+
},
|
69 |
+
"decoder": {
|
70 |
+
"type": "ByteLevel",
|
71 |
+
"add_prefix_space": false,
|
72 |
+
"trim_offsets": true,
|
73 |
+
"use_regex": true
|
74 |
+
},
|
75 |
+
"model": {
|
76 |
+
"type": "BPE",
|
77 |
+
"dropout": null,
|
78 |
+
"unk_token": "<|unknown|>",
|
79 |
+
"continuing_subword_prefix": null,
|
80 |
+
"end_of_word_suffix": null,
|
81 |
+
"fuse_unk": false,
|
82 |
+
"byte_fallback": false,
|
83 |
+
"ignore_merges": false,
|
84 |
+
"vocab": {
|
85 |
+
"<|beginoftext|>": 0,
|
86 |
+
"<|padding|>": 1,
|
87 |
+
"<|unknown|>": 2,
|
88 |
+
"a": 3,
|
89 |
+
"c": 4,
|
90 |
+
"g": 5,
|
91 |
+
"t": 6,
|
92 |
+
"n": 7
|
93 |
+
},
|
94 |
+
"merges": []
|
95 |
+
}
|
96 |
+
}
|
checkpoint-20000/tokenizer_config.json
ADDED
@@ -0,0 +1,38 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"add_bos_token": false,
|
3 |
+
"add_eos_token": false,
|
4 |
+
"add_prefix_space": false,
|
5 |
+
"added_tokens_decoder": {
|
6 |
+
"0": {
|
7 |
+
"content": "<|beginoftext|>",
|
8 |
+
"lstrip": false,
|
9 |
+
"normalized": true,
|
10 |
+
"rstrip": false,
|
11 |
+
"single_word": false,
|
12 |
+
"special": true
|
13 |
+
},
|
14 |
+
"1": {
|
15 |
+
"content": "<|endoftext|>",
|
16 |
+
"lstrip": false,
|
17 |
+
"normalized": false,
|
18 |
+
"rstrip": false,
|
19 |
+
"single_word": false,
|
20 |
+
"special": true
|
21 |
+
},
|
22 |
+
"2": {
|
23 |
+
"content": "<|unknown|>",
|
24 |
+
"lstrip": false,
|
25 |
+
"normalized": false,
|
26 |
+
"rstrip": false,
|
27 |
+
"single_word": false,
|
28 |
+
"special": true
|
29 |
+
}
|
30 |
+
},
|
31 |
+
"bos_token": "<|beginoftext|>",
|
32 |
+
"clean_up_tokenization_spaces": true,
|
33 |
+
"eos_token": "<|endoftext|>",
|
34 |
+
"model_max_length": 1000000000000000019884624838656,
|
35 |
+
"pad_token": "<|endoftext|>",
|
36 |
+
"tokenizer_class": "GPTNeoXTokenizer",
|
37 |
+
"unk_token": "<|unknown|>"
|
38 |
+
}
|
checkpoint-20000/trainer_state.json
ADDED
@@ -0,0 +1,1433 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 0.5863039399624765,
|
5 |
+
"eval_steps": 500,
|
6 |
+
"global_step": 20000,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.0029315196998123826,
|
13 |
+
"grad_norm": 0.7248865962028503,
|
14 |
+
"learning_rate": 4.9868042930033434e-05,
|
15 |
+
"loss": 1.3593,
|
16 |
+
"step": 100
|
17 |
+
},
|
18 |
+
{
|
19 |
+
"epoch": 0.005863039399624765,
|
20 |
+
"grad_norm": 0.3698488771915436,
|
21 |
+
"learning_rate": 4.972142396340391e-05,
|
22 |
+
"loss": 1.3285,
|
23 |
+
"step": 200
|
24 |
+
},
|
25 |
+
{
|
26 |
+
"epoch": 0.008794559099437148,
|
27 |
+
"grad_norm": 0.24848727881908417,
|
28 |
+
"learning_rate": 4.957480499677439e-05,
|
29 |
+
"loss": 1.324,
|
30 |
+
"step": 300
|
31 |
+
},
|
32 |
+
{
|
33 |
+
"epoch": 0.01172607879924953,
|
34 |
+
"grad_norm": 0.2648352086544037,
|
35 |
+
"learning_rate": 4.942818603014486e-05,
|
36 |
+
"loss": 1.3145,
|
37 |
+
"step": 400
|
38 |
+
},
|
39 |
+
{
|
40 |
+
"epoch": 0.014657598499061914,
|
41 |
+
"grad_norm": 0.3995356559753418,
|
42 |
+
"learning_rate": 4.928156706351534e-05,
|
43 |
+
"loss": 1.3141,
|
44 |
+
"step": 500
|
45 |
+
},
|
46 |
+
{
|
47 |
+
"epoch": 0.017589118198874296,
|
48 |
+
"grad_norm": 0.32299748063087463,
|
49 |
+
"learning_rate": 4.913494809688581e-05,
|
50 |
+
"loss": 1.3087,
|
51 |
+
"step": 600
|
52 |
+
},
|
53 |
+
{
|
54 |
+
"epoch": 0.02052063789868668,
|
55 |
+
"grad_norm": 0.33237338066101074,
|
56 |
+
"learning_rate": 4.898832913025629e-05,
|
57 |
+
"loss": 1.3071,
|
58 |
+
"step": 700
|
59 |
+
},
|
60 |
+
{
|
61 |
+
"epoch": 0.02345215759849906,
|
62 |
+
"grad_norm": 0.2716487646102905,
|
63 |
+
"learning_rate": 4.8841710163626766e-05,
|
64 |
+
"loss": 1.3048,
|
65 |
+
"step": 800
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 0.026383677298311446,
|
69 |
+
"grad_norm": 0.8654017448425293,
|
70 |
+
"learning_rate": 4.869509119699725e-05,
|
71 |
+
"loss": 1.3087,
|
72 |
+
"step": 900
|
73 |
+
},
|
74 |
+
{
|
75 |
+
"epoch": 0.02931519699812383,
|
76 |
+
"grad_norm": 0.27945375442504883,
|
77 |
+
"learning_rate": 4.854847223036773e-05,
|
78 |
+
"loss": 1.3015,
|
79 |
+
"step": 1000
|
80 |
+
},
|
81 |
+
{
|
82 |
+
"epoch": 0.03224671669793621,
|
83 |
+
"grad_norm": 0.27838173508644104,
|
84 |
+
"learning_rate": 4.84018532637382e-05,
|
85 |
+
"loss": 1.2998,
|
86 |
+
"step": 1100
|
87 |
+
},
|
88 |
+
{
|
89 |
+
"epoch": 0.03517823639774859,
|
90 |
+
"grad_norm": 0.24157749116420746,
|
91 |
+
"learning_rate": 4.8255234297108676e-05,
|
92 |
+
"loss": 1.3016,
|
93 |
+
"step": 1200
|
94 |
+
},
|
95 |
+
{
|
96 |
+
"epoch": 0.038109756097560975,
|
97 |
+
"grad_norm": 0.45793822407722473,
|
98 |
+
"learning_rate": 4.810861533047915e-05,
|
99 |
+
"loss": 1.2983,
|
100 |
+
"step": 1300
|
101 |
+
},
|
102 |
+
{
|
103 |
+
"epoch": 0.04104127579737336,
|
104 |
+
"grad_norm": 0.1312064826488495,
|
105 |
+
"learning_rate": 4.796199636384963e-05,
|
106 |
+
"loss": 1.2978,
|
107 |
+
"step": 1400
|
108 |
+
},
|
109 |
+
{
|
110 |
+
"epoch": 0.04397279549718574,
|
111 |
+
"grad_norm": 0.33062976598739624,
|
112 |
+
"learning_rate": 4.7815377397220105e-05,
|
113 |
+
"loss": 1.2972,
|
114 |
+
"step": 1500
|
115 |
+
},
|
116 |
+
{
|
117 |
+
"epoch": 0.04690431519699812,
|
118 |
+
"grad_norm": 0.29819390177726746,
|
119 |
+
"learning_rate": 4.766875843059058e-05,
|
120 |
+
"loss": 1.2945,
|
121 |
+
"step": 1600
|
122 |
+
},
|
123 |
+
{
|
124 |
+
"epoch": 0.0498358348968105,
|
125 |
+
"grad_norm": 0.3396131694316864,
|
126 |
+
"learning_rate": 4.752213946396106e-05,
|
127 |
+
"loss": 1.2921,
|
128 |
+
"step": 1700
|
129 |
+
},
|
130 |
+
{
|
131 |
+
"epoch": 0.05276735459662289,
|
132 |
+
"grad_norm": 0.3231181502342224,
|
133 |
+
"learning_rate": 4.737552049733154e-05,
|
134 |
+
"loss": 1.2932,
|
135 |
+
"step": 1800
|
136 |
+
},
|
137 |
+
{
|
138 |
+
"epoch": 0.055698874296435275,
|
139 |
+
"grad_norm": 0.4015660285949707,
|
140 |
+
"learning_rate": 4.7228901530702015e-05,
|
141 |
+
"loss": 1.2899,
|
142 |
+
"step": 1900
|
143 |
+
},
|
144 |
+
{
|
145 |
+
"epoch": 0.05863039399624766,
|
146 |
+
"grad_norm": 0.436213880777359,
|
147 |
+
"learning_rate": 4.708228256407249e-05,
|
148 |
+
"loss": 1.2906,
|
149 |
+
"step": 2000
|
150 |
+
},
|
151 |
+
{
|
152 |
+
"epoch": 0.06156191369606004,
|
153 |
+
"grad_norm": 0.3451833426952362,
|
154 |
+
"learning_rate": 4.693566359744297e-05,
|
155 |
+
"loss": 1.2884,
|
156 |
+
"step": 2100
|
157 |
+
},
|
158 |
+
{
|
159 |
+
"epoch": 0.06449343339587242,
|
160 |
+
"grad_norm": 0.41890543699264526,
|
161 |
+
"learning_rate": 4.6789044630813445e-05,
|
162 |
+
"loss": 1.2892,
|
163 |
+
"step": 2200
|
164 |
+
},
|
165 |
+
{
|
166 |
+
"epoch": 0.0674249530956848,
|
167 |
+
"grad_norm": 0.3117181062698364,
|
168 |
+
"learning_rate": 4.664242566418392e-05,
|
169 |
+
"loss": 1.2866,
|
170 |
+
"step": 2300
|
171 |
+
},
|
172 |
+
{
|
173 |
+
"epoch": 0.07035647279549719,
|
174 |
+
"grad_norm": 0.2703840136528015,
|
175 |
+
"learning_rate": 4.64958066975544e-05,
|
176 |
+
"loss": 1.2869,
|
177 |
+
"step": 2400
|
178 |
+
},
|
179 |
+
{
|
180 |
+
"epoch": 0.07328799249530957,
|
181 |
+
"grad_norm": 0.31400740146636963,
|
182 |
+
"learning_rate": 4.6349187730924874e-05,
|
183 |
+
"loss": 1.2844,
|
184 |
+
"step": 2500
|
185 |
+
},
|
186 |
+
{
|
187 |
+
"epoch": 0.07621951219512195,
|
188 |
+
"grad_norm": 0.36265355348587036,
|
189 |
+
"learning_rate": 4.620256876429535e-05,
|
190 |
+
"loss": 1.2847,
|
191 |
+
"step": 2600
|
192 |
+
},
|
193 |
+
{
|
194 |
+
"epoch": 0.07915103189493433,
|
195 |
+
"grad_norm": 0.32774218916893005,
|
196 |
+
"learning_rate": 4.605594979766583e-05,
|
197 |
+
"loss": 1.2889,
|
198 |
+
"step": 2700
|
199 |
+
},
|
200 |
+
{
|
201 |
+
"epoch": 0.08208255159474671,
|
202 |
+
"grad_norm": 0.26954421401023865,
|
203 |
+
"learning_rate": 4.590933083103631e-05,
|
204 |
+
"loss": 1.2823,
|
205 |
+
"step": 2800
|
206 |
+
},
|
207 |
+
{
|
208 |
+
"epoch": 0.0850140712945591,
|
209 |
+
"grad_norm": 0.4042912423610687,
|
210 |
+
"learning_rate": 4.5762711864406784e-05,
|
211 |
+
"loss": 1.2869,
|
212 |
+
"step": 2900
|
213 |
+
},
|
214 |
+
{
|
215 |
+
"epoch": 0.08794559099437148,
|
216 |
+
"grad_norm": 0.28043991327285767,
|
217 |
+
"learning_rate": 4.561609289777726e-05,
|
218 |
+
"loss": 1.2851,
|
219 |
+
"step": 3000
|
220 |
+
},
|
221 |
+
{
|
222 |
+
"epoch": 0.09087711069418386,
|
223 |
+
"grad_norm": 0.35115665197372437,
|
224 |
+
"learning_rate": 4.546947393114774e-05,
|
225 |
+
"loss": 1.2816,
|
226 |
+
"step": 3100
|
227 |
+
},
|
228 |
+
{
|
229 |
+
"epoch": 0.09380863039399624,
|
230 |
+
"grad_norm": 0.3325822651386261,
|
231 |
+
"learning_rate": 4.532285496451821e-05,
|
232 |
+
"loss": 1.2821,
|
233 |
+
"step": 3200
|
234 |
+
},
|
235 |
+
{
|
236 |
+
"epoch": 0.09674015009380862,
|
237 |
+
"grad_norm": 0.14176355302333832,
|
238 |
+
"learning_rate": 4.517623599788869e-05,
|
239 |
+
"loss": 1.2808,
|
240 |
+
"step": 3300
|
241 |
+
},
|
242 |
+
{
|
243 |
+
"epoch": 0.099671669793621,
|
244 |
+
"grad_norm": 0.3418448567390442,
|
245 |
+
"learning_rate": 4.502961703125916e-05,
|
246 |
+
"loss": 1.2808,
|
247 |
+
"step": 3400
|
248 |
+
},
|
249 |
+
{
|
250 |
+
"epoch": 0.1026031894934334,
|
251 |
+
"grad_norm": 0.3579727113246918,
|
252 |
+
"learning_rate": 4.488299806462964e-05,
|
253 |
+
"loss": 1.2768,
|
254 |
+
"step": 3500
|
255 |
+
},
|
256 |
+
{
|
257 |
+
"epoch": 0.10553470919324578,
|
258 |
+
"grad_norm": 0.2470945566892624,
|
259 |
+
"learning_rate": 4.473637909800012e-05,
|
260 |
+
"loss": 1.2812,
|
261 |
+
"step": 3600
|
262 |
+
},
|
263 |
+
{
|
264 |
+
"epoch": 0.10846622889305817,
|
265 |
+
"grad_norm": 0.32134631276130676,
|
266 |
+
"learning_rate": 4.45897601313706e-05,
|
267 |
+
"loss": 1.2797,
|
268 |
+
"step": 3700
|
269 |
+
},
|
270 |
+
{
|
271 |
+
"epoch": 0.11139774859287055,
|
272 |
+
"grad_norm": 0.2982274889945984,
|
273 |
+
"learning_rate": 4.444314116474108e-05,
|
274 |
+
"loss": 1.2762,
|
275 |
+
"step": 3800
|
276 |
+
},
|
277 |
+
{
|
278 |
+
"epoch": 0.11432926829268293,
|
279 |
+
"grad_norm": 0.2519618570804596,
|
280 |
+
"learning_rate": 4.429652219811155e-05,
|
281 |
+
"loss": 1.2767,
|
282 |
+
"step": 3900
|
283 |
+
},
|
284 |
+
{
|
285 |
+
"epoch": 0.11726078799249531,
|
286 |
+
"grad_norm": 0.30841368436813354,
|
287 |
+
"learning_rate": 4.4149903231482026e-05,
|
288 |
+
"loss": 1.2765,
|
289 |
+
"step": 4000
|
290 |
+
},
|
291 |
+
{
|
292 |
+
"epoch": 0.1201923076923077,
|
293 |
+
"grad_norm": 0.3413415551185608,
|
294 |
+
"learning_rate": 4.40032842648525e-05,
|
295 |
+
"loss": 1.274,
|
296 |
+
"step": 4100
|
297 |
+
},
|
298 |
+
{
|
299 |
+
"epoch": 0.12312382739212008,
|
300 |
+
"grad_norm": 0.3912579119205475,
|
301 |
+
"learning_rate": 4.385666529822298e-05,
|
302 |
+
"loss": 1.2694,
|
303 |
+
"step": 4200
|
304 |
+
},
|
305 |
+
{
|
306 |
+
"epoch": 0.12605534709193245,
|
307 |
+
"grad_norm": 0.39514127373695374,
|
308 |
+
"learning_rate": 4.3710046331593455e-05,
|
309 |
+
"loss": 1.2742,
|
310 |
+
"step": 4300
|
311 |
+
},
|
312 |
+
{
|
313 |
+
"epoch": 0.12898686679174484,
|
314 |
+
"grad_norm": 0.5046316385269165,
|
315 |
+
"learning_rate": 4.356342736496393e-05,
|
316 |
+
"loss": 1.2723,
|
317 |
+
"step": 4400
|
318 |
+
},
|
319 |
+
{
|
320 |
+
"epoch": 0.1319183864915572,
|
321 |
+
"grad_norm": 0.40433749556541443,
|
322 |
+
"learning_rate": 4.341680839833441e-05,
|
323 |
+
"loss": 1.2714,
|
324 |
+
"step": 4500
|
325 |
+
},
|
326 |
+
{
|
327 |
+
"epoch": 0.1348499061913696,
|
328 |
+
"grad_norm": 0.41701552271842957,
|
329 |
+
"learning_rate": 4.327018943170489e-05,
|
330 |
+
"loss": 1.2686,
|
331 |
+
"step": 4600
|
332 |
+
},
|
333 |
+
{
|
334 |
+
"epoch": 0.137781425891182,
|
335 |
+
"grad_norm": 0.6152161359786987,
|
336 |
+
"learning_rate": 4.3123570465075365e-05,
|
337 |
+
"loss": 1.2642,
|
338 |
+
"step": 4700
|
339 |
+
},
|
340 |
+
{
|
341 |
+
"epoch": 0.14071294559099437,
|
342 |
+
"grad_norm": 0.6579223275184631,
|
343 |
+
"learning_rate": 4.297695149844584e-05,
|
344 |
+
"loss": 1.2648,
|
345 |
+
"step": 4800
|
346 |
+
},
|
347 |
+
{
|
348 |
+
"epoch": 0.14364446529080677,
|
349 |
+
"grad_norm": 0.47039562463760376,
|
350 |
+
"learning_rate": 4.283033253181632e-05,
|
351 |
+
"loss": 1.2623,
|
352 |
+
"step": 4900
|
353 |
+
},
|
354 |
+
{
|
355 |
+
"epoch": 0.14657598499061913,
|
356 |
+
"grad_norm": 0.5053458213806152,
|
357 |
+
"learning_rate": 4.2683713565186795e-05,
|
358 |
+
"loss": 1.2581,
|
359 |
+
"step": 5000
|
360 |
+
},
|
361 |
+
{
|
362 |
+
"epoch": 0.14950750469043153,
|
363 |
+
"grad_norm": 0.6096323728561401,
|
364 |
+
"learning_rate": 4.253709459855727e-05,
|
365 |
+
"loss": 1.2526,
|
366 |
+
"step": 5100
|
367 |
+
},
|
368 |
+
{
|
369 |
+
"epoch": 0.1524390243902439,
|
370 |
+
"grad_norm": 0.6972830891609192,
|
371 |
+
"learning_rate": 4.239047563192775e-05,
|
372 |
+
"loss": 1.2541,
|
373 |
+
"step": 5200
|
374 |
+
},
|
375 |
+
{
|
376 |
+
"epoch": 0.1553705440900563,
|
377 |
+
"grad_norm": 0.27344340085983276,
|
378 |
+
"learning_rate": 4.2243856665298224e-05,
|
379 |
+
"loss": 1.2496,
|
380 |
+
"step": 5300
|
381 |
+
},
|
382 |
+
{
|
383 |
+
"epoch": 0.15830206378986866,
|
384 |
+
"grad_norm": 0.6050379276275635,
|
385 |
+
"learning_rate": 4.20972376986687e-05,
|
386 |
+
"loss": 1.243,
|
387 |
+
"step": 5400
|
388 |
+
},
|
389 |
+
{
|
390 |
+
"epoch": 0.16123358348968106,
|
391 |
+
"grad_norm": 0.4654715061187744,
|
392 |
+
"learning_rate": 4.195061873203918e-05,
|
393 |
+
"loss": 1.2398,
|
394 |
+
"step": 5500
|
395 |
+
},
|
396 |
+
{
|
397 |
+
"epoch": 0.16416510318949343,
|
398 |
+
"grad_norm": 0.8257409334182739,
|
399 |
+
"learning_rate": 4.180399976540966e-05,
|
400 |
+
"loss": 1.2377,
|
401 |
+
"step": 5600
|
402 |
+
},
|
403 |
+
{
|
404 |
+
"epoch": 0.16709662288930582,
|
405 |
+
"grad_norm": 0.74887615442276,
|
406 |
+
"learning_rate": 4.1657380798780134e-05,
|
407 |
+
"loss": 1.236,
|
408 |
+
"step": 5700
|
409 |
+
},
|
410 |
+
{
|
411 |
+
"epoch": 0.1700281425891182,
|
412 |
+
"grad_norm": 0.7653385996818542,
|
413 |
+
"learning_rate": 4.151076183215061e-05,
|
414 |
+
"loss": 1.2276,
|
415 |
+
"step": 5800
|
416 |
+
},
|
417 |
+
{
|
418 |
+
"epoch": 0.1729596622889306,
|
419 |
+
"grad_norm": 0.7907219529151917,
|
420 |
+
"learning_rate": 4.136414286552109e-05,
|
421 |
+
"loss": 1.221,
|
422 |
+
"step": 5900
|
423 |
+
},
|
424 |
+
{
|
425 |
+
"epoch": 0.17589118198874296,
|
426 |
+
"grad_norm": 0.6932777762413025,
|
427 |
+
"learning_rate": 4.121752389889156e-05,
|
428 |
+
"loss": 1.2205,
|
429 |
+
"step": 6000
|
430 |
+
},
|
431 |
+
{
|
432 |
+
"epoch": 0.17882270168855535,
|
433 |
+
"grad_norm": 0.7687219381332397,
|
434 |
+
"learning_rate": 4.107090493226204e-05,
|
435 |
+
"loss": 1.2145,
|
436 |
+
"step": 6100
|
437 |
+
},
|
438 |
+
{
|
439 |
+
"epoch": 0.18175422138836772,
|
440 |
+
"grad_norm": 0.850978434085846,
|
441 |
+
"learning_rate": 4.092428596563251e-05,
|
442 |
+
"loss": 1.2066,
|
443 |
+
"step": 6200
|
444 |
+
},
|
445 |
+
{
|
446 |
+
"epoch": 0.18468574108818012,
|
447 |
+
"grad_norm": 0.8447002172470093,
|
448 |
+
"learning_rate": 4.077766699900299e-05,
|
449 |
+
"loss": 1.2091,
|
450 |
+
"step": 6300
|
451 |
+
},
|
452 |
+
{
|
453 |
+
"epoch": 0.18761726078799248,
|
454 |
+
"grad_norm": 1.415692687034607,
|
455 |
+
"learning_rate": 4.063104803237347e-05,
|
456 |
+
"loss": 1.1981,
|
457 |
+
"step": 6400
|
458 |
+
},
|
459 |
+
{
|
460 |
+
"epoch": 0.19054878048780488,
|
461 |
+
"grad_norm": 0.9456603527069092,
|
462 |
+
"learning_rate": 4.048442906574395e-05,
|
463 |
+
"loss": 1.2041,
|
464 |
+
"step": 6500
|
465 |
+
},
|
466 |
+
{
|
467 |
+
"epoch": 0.19348030018761725,
|
468 |
+
"grad_norm": 1.8772284984588623,
|
469 |
+
"learning_rate": 4.033781009911443e-05,
|
470 |
+
"loss": 1.199,
|
471 |
+
"step": 6600
|
472 |
+
},
|
473 |
+
{
|
474 |
+
"epoch": 0.19641181988742965,
|
475 |
+
"grad_norm": 0.8293085694313049,
|
476 |
+
"learning_rate": 4.01911911324849e-05,
|
477 |
+
"loss": 1.1886,
|
478 |
+
"step": 6700
|
479 |
+
},
|
480 |
+
{
|
481 |
+
"epoch": 0.199343339587242,
|
482 |
+
"grad_norm": 1.4174566268920898,
|
483 |
+
"learning_rate": 4.0044572165855376e-05,
|
484 |
+
"loss": 1.1851,
|
485 |
+
"step": 6800
|
486 |
+
},
|
487 |
+
{
|
488 |
+
"epoch": 0.2022748592870544,
|
489 |
+
"grad_norm": 1.6964879035949707,
|
490 |
+
"learning_rate": 3.989795319922585e-05,
|
491 |
+
"loss": 1.1851,
|
492 |
+
"step": 6900
|
493 |
+
},
|
494 |
+
{
|
495 |
+
"epoch": 0.2052063789868668,
|
496 |
+
"grad_norm": 2.025838613510132,
|
497 |
+
"learning_rate": 3.975133423259633e-05,
|
498 |
+
"loss": 1.1778,
|
499 |
+
"step": 7000
|
500 |
+
},
|
501 |
+
{
|
502 |
+
"epoch": 0.20813789868667917,
|
503 |
+
"grad_norm": 1.919710397720337,
|
504 |
+
"learning_rate": 3.9604715265966805e-05,
|
505 |
+
"loss": 1.1717,
|
506 |
+
"step": 7100
|
507 |
+
},
|
508 |
+
{
|
509 |
+
"epoch": 0.21106941838649157,
|
510 |
+
"grad_norm": 1.467106580734253,
|
511 |
+
"learning_rate": 3.945809629933728e-05,
|
512 |
+
"loss": 1.1593,
|
513 |
+
"step": 7200
|
514 |
+
},
|
515 |
+
{
|
516 |
+
"epoch": 0.21400093808630394,
|
517 |
+
"grad_norm": 2.1268837451934814,
|
518 |
+
"learning_rate": 3.931147733270776e-05,
|
519 |
+
"loss": 1.162,
|
520 |
+
"step": 7300
|
521 |
+
},
|
522 |
+
{
|
523 |
+
"epoch": 0.21693245778611633,
|
524 |
+
"grad_norm": 1.1182403564453125,
|
525 |
+
"learning_rate": 3.916485836607824e-05,
|
526 |
+
"loss": 1.1487,
|
527 |
+
"step": 7400
|
528 |
+
},
|
529 |
+
{
|
530 |
+
"epoch": 0.2198639774859287,
|
531 |
+
"grad_norm": 1.3415223360061646,
|
532 |
+
"learning_rate": 3.9018239399448715e-05,
|
533 |
+
"loss": 1.1471,
|
534 |
+
"step": 7500
|
535 |
+
},
|
536 |
+
{
|
537 |
+
"epoch": 0.2227954971857411,
|
538 |
+
"grad_norm": 1.8642910718917847,
|
539 |
+
"learning_rate": 3.887162043281919e-05,
|
540 |
+
"loss": 1.1555,
|
541 |
+
"step": 7600
|
542 |
+
},
|
543 |
+
{
|
544 |
+
"epoch": 0.22572701688555347,
|
545 |
+
"grad_norm": 1.912855625152588,
|
546 |
+
"learning_rate": 3.872500146618967e-05,
|
547 |
+
"loss": 1.1328,
|
548 |
+
"step": 7700
|
549 |
+
},
|
550 |
+
{
|
551 |
+
"epoch": 0.22865853658536586,
|
552 |
+
"grad_norm": 2.582273244857788,
|
553 |
+
"learning_rate": 3.8578382499560145e-05,
|
554 |
+
"loss": 1.1394,
|
555 |
+
"step": 7800
|
556 |
+
},
|
557 |
+
{
|
558 |
+
"epoch": 0.23159005628517823,
|
559 |
+
"grad_norm": 1.6247990131378174,
|
560 |
+
"learning_rate": 3.843176353293062e-05,
|
561 |
+
"loss": 1.1413,
|
562 |
+
"step": 7900
|
563 |
+
},
|
564 |
+
{
|
565 |
+
"epoch": 0.23452157598499063,
|
566 |
+
"grad_norm": 2.7654881477355957,
|
567 |
+
"learning_rate": 3.82851445663011e-05,
|
568 |
+
"loss": 1.1219,
|
569 |
+
"step": 8000
|
570 |
+
},
|
571 |
+
{
|
572 |
+
"epoch": 0.237453095684803,
|
573 |
+
"grad_norm": 1.3285764455795288,
|
574 |
+
"learning_rate": 3.8138525599671574e-05,
|
575 |
+
"loss": 1.1267,
|
576 |
+
"step": 8100
|
577 |
+
},
|
578 |
+
{
|
579 |
+
"epoch": 0.2403846153846154,
|
580 |
+
"grad_norm": 3.288980484008789,
|
581 |
+
"learning_rate": 3.7991906633042055e-05,
|
582 |
+
"loss": 1.1151,
|
583 |
+
"step": 8200
|
584 |
+
},
|
585 |
+
{
|
586 |
+
"epoch": 0.24331613508442776,
|
587 |
+
"grad_norm": 1.586205244064331,
|
588 |
+
"learning_rate": 3.784528766641253e-05,
|
589 |
+
"loss": 1.1202,
|
590 |
+
"step": 8300
|
591 |
+
},
|
592 |
+
{
|
593 |
+
"epoch": 0.24624765478424016,
|
594 |
+
"grad_norm": 2.017575979232788,
|
595 |
+
"learning_rate": 3.769866869978301e-05,
|
596 |
+
"loss": 1.12,
|
597 |
+
"step": 8400
|
598 |
+
},
|
599 |
+
{
|
600 |
+
"epoch": 0.24917917448405252,
|
601 |
+
"grad_norm": 3.5337297916412354,
|
602 |
+
"learning_rate": 3.7552049733153484e-05,
|
603 |
+
"loss": 1.109,
|
604 |
+
"step": 8500
|
605 |
+
},
|
606 |
+
{
|
607 |
+
"epoch": 0.2521106941838649,
|
608 |
+
"grad_norm": 2.3067338466644287,
|
609 |
+
"learning_rate": 3.740543076652396e-05,
|
610 |
+
"loss": 1.1157,
|
611 |
+
"step": 8600
|
612 |
+
},
|
613 |
+
{
|
614 |
+
"epoch": 0.2550422138836773,
|
615 |
+
"grad_norm": 1.8356772661209106,
|
616 |
+
"learning_rate": 3.725881179989444e-05,
|
617 |
+
"loss": 1.1013,
|
618 |
+
"step": 8700
|
619 |
+
},
|
620 |
+
{
|
621 |
+
"epoch": 0.2579737335834897,
|
622 |
+
"grad_norm": 1.1588494777679443,
|
623 |
+
"learning_rate": 3.711219283326491e-05,
|
624 |
+
"loss": 1.0838,
|
625 |
+
"step": 8800
|
626 |
+
},
|
627 |
+
{
|
628 |
+
"epoch": 0.26090525328330205,
|
629 |
+
"grad_norm": 2.9327263832092285,
|
630 |
+
"learning_rate": 3.696557386663539e-05,
|
631 |
+
"loss": 1.0998,
|
632 |
+
"step": 8900
|
633 |
+
},
|
634 |
+
{
|
635 |
+
"epoch": 0.2638367729831144,
|
636 |
+
"grad_norm": 1.1504569053649902,
|
637 |
+
"learning_rate": 3.681895490000586e-05,
|
638 |
+
"loss": 1.1067,
|
639 |
+
"step": 9000
|
640 |
+
},
|
641 |
+
{
|
642 |
+
"epoch": 0.26676829268292684,
|
643 |
+
"grad_norm": 1.546099066734314,
|
644 |
+
"learning_rate": 3.667233593337634e-05,
|
645 |
+
"loss": 1.0985,
|
646 |
+
"step": 9100
|
647 |
+
},
|
648 |
+
{
|
649 |
+
"epoch": 0.2696998123827392,
|
650 |
+
"grad_norm": 2.004364252090454,
|
651 |
+
"learning_rate": 3.652571696674682e-05,
|
652 |
+
"loss": 1.0906,
|
653 |
+
"step": 9200
|
654 |
+
},
|
655 |
+
{
|
656 |
+
"epoch": 0.2726313320825516,
|
657 |
+
"grad_norm": 2.3435940742492676,
|
658 |
+
"learning_rate": 3.63790980001173e-05,
|
659 |
+
"loss": 1.0815,
|
660 |
+
"step": 9300
|
661 |
+
},
|
662 |
+
{
|
663 |
+
"epoch": 0.275562851782364,
|
664 |
+
"grad_norm": 1.5270298719406128,
|
665 |
+
"learning_rate": 3.623247903348778e-05,
|
666 |
+
"loss": 1.08,
|
667 |
+
"step": 9400
|
668 |
+
},
|
669 |
+
{
|
670 |
+
"epoch": 0.2784943714821764,
|
671 |
+
"grad_norm": 4.2500386238098145,
|
672 |
+
"learning_rate": 3.608586006685825e-05,
|
673 |
+
"loss": 1.0857,
|
674 |
+
"step": 9500
|
675 |
+
},
|
676 |
+
{
|
677 |
+
"epoch": 0.28142589118198874,
|
678 |
+
"grad_norm": 1.3212778568267822,
|
679 |
+
"learning_rate": 3.5939241100228726e-05,
|
680 |
+
"loss": 1.0717,
|
681 |
+
"step": 9600
|
682 |
+
},
|
683 |
+
{
|
684 |
+
"epoch": 0.2843574108818011,
|
685 |
+
"grad_norm": 3.6155266761779785,
|
686 |
+
"learning_rate": 3.57926221335992e-05,
|
687 |
+
"loss": 1.0798,
|
688 |
+
"step": 9700
|
689 |
+
},
|
690 |
+
{
|
691 |
+
"epoch": 0.28728893058161353,
|
692 |
+
"grad_norm": 3.934211492538452,
|
693 |
+
"learning_rate": 3.564600316696968e-05,
|
694 |
+
"loss": 1.0777,
|
695 |
+
"step": 9800
|
696 |
+
},
|
697 |
+
{
|
698 |
+
"epoch": 0.2902204502814259,
|
699 |
+
"grad_norm": 1.3743681907653809,
|
700 |
+
"learning_rate": 3.5499384200340155e-05,
|
701 |
+
"loss": 1.0776,
|
702 |
+
"step": 9900
|
703 |
+
},
|
704 |
+
{
|
705 |
+
"epoch": 0.29315196998123827,
|
706 |
+
"grad_norm": 1.2762603759765625,
|
707 |
+
"learning_rate": 3.5352765233710636e-05,
|
708 |
+
"loss": 1.0653,
|
709 |
+
"step": 10000
|
710 |
+
},
|
711 |
+
{
|
712 |
+
"epoch": 0.29608348968105064,
|
713 |
+
"grad_norm": 1.7702666521072388,
|
714 |
+
"learning_rate": 3.520614626708112e-05,
|
715 |
+
"loss": 1.0623,
|
716 |
+
"step": 10100
|
717 |
+
},
|
718 |
+
{
|
719 |
+
"epoch": 0.29901500938086306,
|
720 |
+
"grad_norm": 1.9537125825881958,
|
721 |
+
"learning_rate": 3.505952730045159e-05,
|
722 |
+
"loss": 1.077,
|
723 |
+
"step": 10200
|
724 |
+
},
|
725 |
+
{
|
726 |
+
"epoch": 0.30194652908067543,
|
727 |
+
"grad_norm": 1.6890079975128174,
|
728 |
+
"learning_rate": 3.4912908333822065e-05,
|
729 |
+
"loss": 1.0689,
|
730 |
+
"step": 10300
|
731 |
+
},
|
732 |
+
{
|
733 |
+
"epoch": 0.3048780487804878,
|
734 |
+
"grad_norm": 2.9840946197509766,
|
735 |
+
"learning_rate": 3.476628936719254e-05,
|
736 |
+
"loss": 1.0683,
|
737 |
+
"step": 10400
|
738 |
+
},
|
739 |
+
{
|
740 |
+
"epoch": 0.30780956848030017,
|
741 |
+
"grad_norm": 0.620637834072113,
|
742 |
+
"learning_rate": 3.461967040056302e-05,
|
743 |
+
"loss": 1.0599,
|
744 |
+
"step": 10500
|
745 |
+
},
|
746 |
+
{
|
747 |
+
"epoch": 0.3107410881801126,
|
748 |
+
"grad_norm": 2.518418073654175,
|
749 |
+
"learning_rate": 3.4473051433933494e-05,
|
750 |
+
"loss": 1.0511,
|
751 |
+
"step": 10600
|
752 |
+
},
|
753 |
+
{
|
754 |
+
"epoch": 0.31367260787992496,
|
755 |
+
"grad_norm": 4.154684543609619,
|
756 |
+
"learning_rate": 3.432643246730397e-05,
|
757 |
+
"loss": 1.0639,
|
758 |
+
"step": 10700
|
759 |
+
},
|
760 |
+
{
|
761 |
+
"epoch": 0.3166041275797373,
|
762 |
+
"grad_norm": 2.329303741455078,
|
763 |
+
"learning_rate": 3.417981350067445e-05,
|
764 |
+
"loss": 1.0606,
|
765 |
+
"step": 10800
|
766 |
+
},
|
767 |
+
{
|
768 |
+
"epoch": 0.3195356472795497,
|
769 |
+
"grad_norm": 1.5261120796203613,
|
770 |
+
"learning_rate": 3.4033194534044924e-05,
|
771 |
+
"loss": 1.0651,
|
772 |
+
"step": 10900
|
773 |
+
},
|
774 |
+
{
|
775 |
+
"epoch": 0.3224671669793621,
|
776 |
+
"grad_norm": 1.4638831615447998,
|
777 |
+
"learning_rate": 3.3886575567415405e-05,
|
778 |
+
"loss": 1.0508,
|
779 |
+
"step": 11000
|
780 |
+
},
|
781 |
+
{
|
782 |
+
"epoch": 0.3253986866791745,
|
783 |
+
"grad_norm": 1.0534781217575073,
|
784 |
+
"learning_rate": 3.373995660078588e-05,
|
785 |
+
"loss": 1.0593,
|
786 |
+
"step": 11100
|
787 |
+
},
|
788 |
+
{
|
789 |
+
"epoch": 0.32833020637898686,
|
790 |
+
"grad_norm": 1.870732307434082,
|
791 |
+
"learning_rate": 3.359333763415636e-05,
|
792 |
+
"loss": 1.0464,
|
793 |
+
"step": 11200
|
794 |
+
},
|
795 |
+
{
|
796 |
+
"epoch": 0.3312617260787992,
|
797 |
+
"grad_norm": 1.8039565086364746,
|
798 |
+
"learning_rate": 3.3446718667526834e-05,
|
799 |
+
"loss": 1.0496,
|
800 |
+
"step": 11300
|
801 |
+
},
|
802 |
+
{
|
803 |
+
"epoch": 0.33419324577861165,
|
804 |
+
"grad_norm": 1.551080584526062,
|
805 |
+
"learning_rate": 3.330009970089731e-05,
|
806 |
+
"loss": 1.03,
|
807 |
+
"step": 11400
|
808 |
+
},
|
809 |
+
{
|
810 |
+
"epoch": 0.337124765478424,
|
811 |
+
"grad_norm": 1.5708621740341187,
|
812 |
+
"learning_rate": 3.315348073426779e-05,
|
813 |
+
"loss": 1.0538,
|
814 |
+
"step": 11500
|
815 |
+
},
|
816 |
+
{
|
817 |
+
"epoch": 0.3400562851782364,
|
818 |
+
"grad_norm": 2.1092281341552734,
|
819 |
+
"learning_rate": 3.300686176763826e-05,
|
820 |
+
"loss": 1.0436,
|
821 |
+
"step": 11600
|
822 |
+
},
|
823 |
+
{
|
824 |
+
"epoch": 0.3429878048780488,
|
825 |
+
"grad_norm": 1.5008190870285034,
|
826 |
+
"learning_rate": 3.286024280100874e-05,
|
827 |
+
"loss": 1.0379,
|
828 |
+
"step": 11700
|
829 |
+
},
|
830 |
+
{
|
831 |
+
"epoch": 0.3459193245778612,
|
832 |
+
"grad_norm": 0.9910880923271179,
|
833 |
+
"learning_rate": 3.271362383437922e-05,
|
834 |
+
"loss": 1.0252,
|
835 |
+
"step": 11800
|
836 |
+
},
|
837 |
+
{
|
838 |
+
"epoch": 0.34885084427767354,
|
839 |
+
"grad_norm": 0.7434718012809753,
|
840 |
+
"learning_rate": 3.256700486774969e-05,
|
841 |
+
"loss": 1.0419,
|
842 |
+
"step": 11900
|
843 |
+
},
|
844 |
+
{
|
845 |
+
"epoch": 0.3517823639774859,
|
846 |
+
"grad_norm": 5.62740421295166,
|
847 |
+
"learning_rate": 3.242038590112017e-05,
|
848 |
+
"loss": 1.0448,
|
849 |
+
"step": 12000
|
850 |
+
},
|
851 |
+
{
|
852 |
+
"epoch": 0.35471388367729834,
|
853 |
+
"grad_norm": 0.7728565335273743,
|
854 |
+
"learning_rate": 3.227376693449065e-05,
|
855 |
+
"loss": 1.0381,
|
856 |
+
"step": 12100
|
857 |
+
},
|
858 |
+
{
|
859 |
+
"epoch": 0.3576454033771107,
|
860 |
+
"grad_norm": 0.6013042330741882,
|
861 |
+
"learning_rate": 3.212714796786113e-05,
|
862 |
+
"loss": 1.0316,
|
863 |
+
"step": 12200
|
864 |
+
},
|
865 |
+
{
|
866 |
+
"epoch": 0.3605769230769231,
|
867 |
+
"grad_norm": 1.558634638786316,
|
868 |
+
"learning_rate": 3.19805290012316e-05,
|
869 |
+
"loss": 1.0423,
|
870 |
+
"step": 12300
|
871 |
+
},
|
872 |
+
{
|
873 |
+
"epoch": 0.36350844277673544,
|
874 |
+
"grad_norm": 2.510350465774536,
|
875 |
+
"learning_rate": 3.1833910034602076e-05,
|
876 |
+
"loss": 1.0261,
|
877 |
+
"step": 12400
|
878 |
+
},
|
879 |
+
{
|
880 |
+
"epoch": 0.36643996247654786,
|
881 |
+
"grad_norm": 1.5596210956573486,
|
882 |
+
"learning_rate": 3.168729106797255e-05,
|
883 |
+
"loss": 1.0289,
|
884 |
+
"step": 12500
|
885 |
+
},
|
886 |
+
{
|
887 |
+
"epoch": 0.36937148217636023,
|
888 |
+
"grad_norm": 1.2188498973846436,
|
889 |
+
"learning_rate": 3.154067210134303e-05,
|
890 |
+
"loss": 1.0307,
|
891 |
+
"step": 12600
|
892 |
+
},
|
893 |
+
{
|
894 |
+
"epoch": 0.3723030018761726,
|
895 |
+
"grad_norm": 1.0569968223571777,
|
896 |
+
"learning_rate": 3.1394053134713505e-05,
|
897 |
+
"loss": 1.0114,
|
898 |
+
"step": 12700
|
899 |
+
},
|
900 |
+
{
|
901 |
+
"epoch": 0.37523452157598497,
|
902 |
+
"grad_norm": 1.9288491010665894,
|
903 |
+
"learning_rate": 3.1247434168083986e-05,
|
904 |
+
"loss": 1.0382,
|
905 |
+
"step": 12800
|
906 |
+
},
|
907 |
+
{
|
908 |
+
"epoch": 0.3781660412757974,
|
909 |
+
"grad_norm": 2.289350748062134,
|
910 |
+
"learning_rate": 3.110081520145447e-05,
|
911 |
+
"loss": 1.0367,
|
912 |
+
"step": 12900
|
913 |
+
},
|
914 |
+
{
|
915 |
+
"epoch": 0.38109756097560976,
|
916 |
+
"grad_norm": 1.0306146144866943,
|
917 |
+
"learning_rate": 3.095419623482494e-05,
|
918 |
+
"loss": 1.0417,
|
919 |
+
"step": 13000
|
920 |
+
},
|
921 |
+
{
|
922 |
+
"epoch": 0.38402908067542213,
|
923 |
+
"grad_norm": 1.7335386276245117,
|
924 |
+
"learning_rate": 3.0807577268195415e-05,
|
925 |
+
"loss": 1.0433,
|
926 |
+
"step": 13100
|
927 |
+
},
|
928 |
+
{
|
929 |
+
"epoch": 0.3869606003752345,
|
930 |
+
"grad_norm": 1.5590994358062744,
|
931 |
+
"learning_rate": 3.066095830156589e-05,
|
932 |
+
"loss": 1.0313,
|
933 |
+
"step": 13200
|
934 |
+
},
|
935 |
+
{
|
936 |
+
"epoch": 0.3898921200750469,
|
937 |
+
"grad_norm": 0.944471001625061,
|
938 |
+
"learning_rate": 3.051433933493637e-05,
|
939 |
+
"loss": 1.0338,
|
940 |
+
"step": 13300
|
941 |
+
},
|
942 |
+
{
|
943 |
+
"epoch": 0.3928236397748593,
|
944 |
+
"grad_norm": 0.9511914849281311,
|
945 |
+
"learning_rate": 3.0367720368306844e-05,
|
946 |
+
"loss": 1.0313,
|
947 |
+
"step": 13400
|
948 |
+
},
|
949 |
+
{
|
950 |
+
"epoch": 0.39575515947467166,
|
951 |
+
"grad_norm": 0.8080090880393982,
|
952 |
+
"learning_rate": 3.0221101401677322e-05,
|
953 |
+
"loss": 1.0286,
|
954 |
+
"step": 13500
|
955 |
+
},
|
956 |
+
{
|
957 |
+
"epoch": 0.398686679174484,
|
958 |
+
"grad_norm": 1.507084846496582,
|
959 |
+
"learning_rate": 3.0074482435047803e-05,
|
960 |
+
"loss": 1.0267,
|
961 |
+
"step": 13600
|
962 |
+
},
|
963 |
+
{
|
964 |
+
"epoch": 0.40161819887429645,
|
965 |
+
"grad_norm": 0.6863479614257812,
|
966 |
+
"learning_rate": 2.9927863468418277e-05,
|
967 |
+
"loss": 1.0271,
|
968 |
+
"step": 13700
|
969 |
+
},
|
970 |
+
{
|
971 |
+
"epoch": 0.4045497185741088,
|
972 |
+
"grad_norm": 2.595979690551758,
|
973 |
+
"learning_rate": 2.978124450178875e-05,
|
974 |
+
"loss": 1.0119,
|
975 |
+
"step": 13800
|
976 |
+
},
|
977 |
+
{
|
978 |
+
"epoch": 0.4074812382739212,
|
979 |
+
"grad_norm": 0.9219207763671875,
|
980 |
+
"learning_rate": 2.963462553515923e-05,
|
981 |
+
"loss": 1.015,
|
982 |
+
"step": 13900
|
983 |
+
},
|
984 |
+
{
|
985 |
+
"epoch": 0.4104127579737336,
|
986 |
+
"grad_norm": 3.082329750061035,
|
987 |
+
"learning_rate": 2.948800656852971e-05,
|
988 |
+
"loss": 1.02,
|
989 |
+
"step": 14000
|
990 |
+
},
|
991 |
+
{
|
992 |
+
"epoch": 0.413344277673546,
|
993 |
+
"grad_norm": 2.0359930992126465,
|
994 |
+
"learning_rate": 2.9341387601900184e-05,
|
995 |
+
"loss": 1.0227,
|
996 |
+
"step": 14100
|
997 |
+
},
|
998 |
+
{
|
999 |
+
"epoch": 0.41627579737335835,
|
1000 |
+
"grad_norm": 1.199521541595459,
|
1001 |
+
"learning_rate": 2.9194768635270658e-05,
|
1002 |
+
"loss": 1.0234,
|
1003 |
+
"step": 14200
|
1004 |
+
},
|
1005 |
+
{
|
1006 |
+
"epoch": 0.4192073170731707,
|
1007 |
+
"grad_norm": 4.709115982055664,
|
1008 |
+
"learning_rate": 2.904814966864114e-05,
|
1009 |
+
"loss": 1.0262,
|
1010 |
+
"step": 14300
|
1011 |
+
},
|
1012 |
+
{
|
1013 |
+
"epoch": 0.42213883677298314,
|
1014 |
+
"grad_norm": 2.031787872314453,
|
1015 |
+
"learning_rate": 2.8901530702011613e-05,
|
1016 |
+
"loss": 1.0129,
|
1017 |
+
"step": 14400
|
1018 |
+
},
|
1019 |
+
{
|
1020 |
+
"epoch": 0.4250703564727955,
|
1021 |
+
"grad_norm": 1.0663975477218628,
|
1022 |
+
"learning_rate": 2.875491173538209e-05,
|
1023 |
+
"loss": 1.0231,
|
1024 |
+
"step": 14500
|
1025 |
+
},
|
1026 |
+
{
|
1027 |
+
"epoch": 0.4280018761726079,
|
1028 |
+
"grad_norm": 1.1190619468688965,
|
1029 |
+
"learning_rate": 2.8608292768752564e-05,
|
1030 |
+
"loss": 1.0203,
|
1031 |
+
"step": 14600
|
1032 |
+
},
|
1033 |
+
{
|
1034 |
+
"epoch": 0.43093339587242024,
|
1035 |
+
"grad_norm": 1.1025962829589844,
|
1036 |
+
"learning_rate": 2.8461673802123045e-05,
|
1037 |
+
"loss": 1.0098,
|
1038 |
+
"step": 14700
|
1039 |
+
},
|
1040 |
+
{
|
1041 |
+
"epoch": 0.43386491557223267,
|
1042 |
+
"grad_norm": 2.02659273147583,
|
1043 |
+
"learning_rate": 2.831505483549352e-05,
|
1044 |
+
"loss": 1.0146,
|
1045 |
+
"step": 14800
|
1046 |
+
},
|
1047 |
+
{
|
1048 |
+
"epoch": 0.43679643527204504,
|
1049 |
+
"grad_norm": 1.0219608545303345,
|
1050 |
+
"learning_rate": 2.8168435868863997e-05,
|
1051 |
+
"loss": 1.0233,
|
1052 |
+
"step": 14900
|
1053 |
+
},
|
1054 |
+
{
|
1055 |
+
"epoch": 0.4397279549718574,
|
1056 |
+
"grad_norm": 1.0353147983551025,
|
1057 |
+
"learning_rate": 2.8021816902234478e-05,
|
1058 |
+
"loss": 1.0094,
|
1059 |
+
"step": 15000
|
1060 |
+
},
|
1061 |
+
{
|
1062 |
+
"epoch": 0.4426594746716698,
|
1063 |
+
"grad_norm": 1.9201918840408325,
|
1064 |
+
"learning_rate": 2.7875197935604952e-05,
|
1065 |
+
"loss": 1.0042,
|
1066 |
+
"step": 15100
|
1067 |
+
},
|
1068 |
+
{
|
1069 |
+
"epoch": 0.4455909943714822,
|
1070 |
+
"grad_norm": 0.9451847672462463,
|
1071 |
+
"learning_rate": 2.7728578968975426e-05,
|
1072 |
+
"loss": 1.0179,
|
1073 |
+
"step": 15200
|
1074 |
+
},
|
1075 |
+
{
|
1076 |
+
"epoch": 0.44852251407129456,
|
1077 |
+
"grad_norm": 2.44089937210083,
|
1078 |
+
"learning_rate": 2.7581960002345904e-05,
|
1079 |
+
"loss": 1.0182,
|
1080 |
+
"step": 15300
|
1081 |
+
},
|
1082 |
+
{
|
1083 |
+
"epoch": 0.45145403377110693,
|
1084 |
+
"grad_norm": 1.49729585647583,
|
1085 |
+
"learning_rate": 2.7435341035716384e-05,
|
1086 |
+
"loss": 1.019,
|
1087 |
+
"step": 15400
|
1088 |
+
},
|
1089 |
+
{
|
1090 |
+
"epoch": 0.4543855534709193,
|
1091 |
+
"grad_norm": 3.6466991901397705,
|
1092 |
+
"learning_rate": 2.728872206908686e-05,
|
1093 |
+
"loss": 1.0255,
|
1094 |
+
"step": 15500
|
1095 |
+
},
|
1096 |
+
{
|
1097 |
+
"epoch": 0.4573170731707317,
|
1098 |
+
"grad_norm": 1.428072452545166,
|
1099 |
+
"learning_rate": 2.7142103102457333e-05,
|
1100 |
+
"loss": 1.0137,
|
1101 |
+
"step": 15600
|
1102 |
+
},
|
1103 |
+
{
|
1104 |
+
"epoch": 0.4602485928705441,
|
1105 |
+
"grad_norm": 1.0874691009521484,
|
1106 |
+
"learning_rate": 2.6995484135827814e-05,
|
1107 |
+
"loss": 0.9944,
|
1108 |
+
"step": 15700
|
1109 |
+
},
|
1110 |
+
{
|
1111 |
+
"epoch": 0.46318011257035646,
|
1112 |
+
"grad_norm": 0.8218618631362915,
|
1113 |
+
"learning_rate": 2.684886516919829e-05,
|
1114 |
+
"loss": 0.9995,
|
1115 |
+
"step": 15800
|
1116 |
+
},
|
1117 |
+
{
|
1118 |
+
"epoch": 0.46611163227016883,
|
1119 |
+
"grad_norm": 1.0309679508209229,
|
1120 |
+
"learning_rate": 2.6702246202568765e-05,
|
1121 |
+
"loss": 1.0211,
|
1122 |
+
"step": 15900
|
1123 |
+
},
|
1124 |
+
{
|
1125 |
+
"epoch": 0.46904315196998125,
|
1126 |
+
"grad_norm": 1.7576292753219604,
|
1127 |
+
"learning_rate": 2.655562723593924e-05,
|
1128 |
+
"loss": 1.016,
|
1129 |
+
"step": 16000
|
1130 |
+
},
|
1131 |
+
{
|
1132 |
+
"epoch": 0.4719746716697936,
|
1133 |
+
"grad_norm": 1.233428716659546,
|
1134 |
+
"learning_rate": 2.640900826930972e-05,
|
1135 |
+
"loss": 1.0005,
|
1136 |
+
"step": 16100
|
1137 |
+
},
|
1138 |
+
{
|
1139 |
+
"epoch": 0.474906191369606,
|
1140 |
+
"grad_norm": 1.013920545578003,
|
1141 |
+
"learning_rate": 2.6262389302680194e-05,
|
1142 |
+
"loss": 0.9947,
|
1143 |
+
"step": 16200
|
1144 |
+
},
|
1145 |
+
{
|
1146 |
+
"epoch": 0.4778377110694184,
|
1147 |
+
"grad_norm": 0.98600834608078,
|
1148 |
+
"learning_rate": 2.6115770336050672e-05,
|
1149 |
+
"loss": 1.002,
|
1150 |
+
"step": 16300
|
1151 |
+
},
|
1152 |
+
{
|
1153 |
+
"epoch": 0.4807692307692308,
|
1154 |
+
"grad_norm": 0.656637966632843,
|
1155 |
+
"learning_rate": 2.5969151369421153e-05,
|
1156 |
+
"loss": 1.009,
|
1157 |
+
"step": 16400
|
1158 |
+
},
|
1159 |
+
{
|
1160 |
+
"epoch": 0.48370075046904315,
|
1161 |
+
"grad_norm": 1.2900489568710327,
|
1162 |
+
"learning_rate": 2.5822532402791627e-05,
|
1163 |
+
"loss": 1.0082,
|
1164 |
+
"step": 16500
|
1165 |
+
},
|
1166 |
+
{
|
1167 |
+
"epoch": 0.4866322701688555,
|
1168 |
+
"grad_norm": 1.068027377128601,
|
1169 |
+
"learning_rate": 2.56759134361621e-05,
|
1170 |
+
"loss": 1.0088,
|
1171 |
+
"step": 16600
|
1172 |
+
},
|
1173 |
+
{
|
1174 |
+
"epoch": 0.48956378986866794,
|
1175 |
+
"grad_norm": 0.6388201117515564,
|
1176 |
+
"learning_rate": 2.552929446953258e-05,
|
1177 |
+
"loss": 0.9925,
|
1178 |
+
"step": 16700
|
1179 |
+
},
|
1180 |
+
{
|
1181 |
+
"epoch": 0.4924953095684803,
|
1182 |
+
"grad_norm": 0.9185531139373779,
|
1183 |
+
"learning_rate": 2.538267550290306e-05,
|
1184 |
+
"loss": 1.019,
|
1185 |
+
"step": 16800
|
1186 |
+
},
|
1187 |
+
{
|
1188 |
+
"epoch": 0.4954268292682927,
|
1189 |
+
"grad_norm": 1.1571044921875,
|
1190 |
+
"learning_rate": 2.5236056536273534e-05,
|
1191 |
+
"loss": 1.0118,
|
1192 |
+
"step": 16900
|
1193 |
+
},
|
1194 |
+
{
|
1195 |
+
"epoch": 0.49835834896810505,
|
1196 |
+
"grad_norm": 1.3585349321365356,
|
1197 |
+
"learning_rate": 2.5089437569644008e-05,
|
1198 |
+
"loss": 1.0092,
|
1199 |
+
"step": 17000
|
1200 |
+
},
|
1201 |
+
{
|
1202 |
+
"epoch": 0.5012898686679175,
|
1203 |
+
"grad_norm": 1.0178191661834717,
|
1204 |
+
"learning_rate": 2.4942818603014485e-05,
|
1205 |
+
"loss": 0.9954,
|
1206 |
+
"step": 17100
|
1207 |
+
},
|
1208 |
+
{
|
1209 |
+
"epoch": 0.5042213883677298,
|
1210 |
+
"grad_norm": 1.0959707498550415,
|
1211 |
+
"learning_rate": 2.4796199636384966e-05,
|
1212 |
+
"loss": 1.0117,
|
1213 |
+
"step": 17200
|
1214 |
+
},
|
1215 |
+
{
|
1216 |
+
"epoch": 0.5071529080675422,
|
1217 |
+
"grad_norm": 1.6161094903945923,
|
1218 |
+
"learning_rate": 2.464958066975544e-05,
|
1219 |
+
"loss": 0.9874,
|
1220 |
+
"step": 17300
|
1221 |
+
},
|
1222 |
+
{
|
1223 |
+
"epoch": 0.5100844277673546,
|
1224 |
+
"grad_norm": 1.3352782726287842,
|
1225 |
+
"learning_rate": 2.4502961703125918e-05,
|
1226 |
+
"loss": 1.0053,
|
1227 |
+
"step": 17400
|
1228 |
+
},
|
1229 |
+
{
|
1230 |
+
"epoch": 0.5130159474671669,
|
1231 |
+
"grad_norm": 0.9680103659629822,
|
1232 |
+
"learning_rate": 2.4356342736496392e-05,
|
1233 |
+
"loss": 0.9874,
|
1234 |
+
"step": 17500
|
1235 |
+
},
|
1236 |
+
{
|
1237 |
+
"epoch": 0.5159474671669794,
|
1238 |
+
"grad_norm": 1.894599437713623,
|
1239 |
+
"learning_rate": 2.420972376986687e-05,
|
1240 |
+
"loss": 0.9793,
|
1241 |
+
"step": 17600
|
1242 |
+
},
|
1243 |
+
{
|
1244 |
+
"epoch": 0.5188789868667918,
|
1245 |
+
"grad_norm": 2.3807919025421143,
|
1246 |
+
"learning_rate": 2.406310480323735e-05,
|
1247 |
+
"loss": 1.001,
|
1248 |
+
"step": 17700
|
1249 |
+
},
|
1250 |
+
{
|
1251 |
+
"epoch": 0.5218105065666041,
|
1252 |
+
"grad_norm": 1.3010458946228027,
|
1253 |
+
"learning_rate": 2.3916485836607824e-05,
|
1254 |
+
"loss": 0.9976,
|
1255 |
+
"step": 17800
|
1256 |
+
},
|
1257 |
+
{
|
1258 |
+
"epoch": 0.5247420262664165,
|
1259 |
+
"grad_norm": 0.8739995360374451,
|
1260 |
+
"learning_rate": 2.3769866869978302e-05,
|
1261 |
+
"loss": 0.9832,
|
1262 |
+
"step": 17900
|
1263 |
+
},
|
1264 |
+
{
|
1265 |
+
"epoch": 0.5276735459662288,
|
1266 |
+
"grad_norm": 1.0226731300354004,
|
1267 |
+
"learning_rate": 2.3623247903348776e-05,
|
1268 |
+
"loss": 0.9889,
|
1269 |
+
"step": 18000
|
1270 |
+
},
|
1271 |
+
{
|
1272 |
+
"epoch": 0.5306050656660413,
|
1273 |
+
"grad_norm": 0.9787300229072571,
|
1274 |
+
"learning_rate": 2.3476628936719257e-05,
|
1275 |
+
"loss": 1.0001,
|
1276 |
+
"step": 18100
|
1277 |
+
},
|
1278 |
+
{
|
1279 |
+
"epoch": 0.5335365853658537,
|
1280 |
+
"grad_norm": 0.8458353877067566,
|
1281 |
+
"learning_rate": 2.333000997008973e-05,
|
1282 |
+
"loss": 0.9963,
|
1283 |
+
"step": 18200
|
1284 |
+
},
|
1285 |
+
{
|
1286 |
+
"epoch": 0.536468105065666,
|
1287 |
+
"grad_norm": 2.051384210586548,
|
1288 |
+
"learning_rate": 2.318339100346021e-05,
|
1289 |
+
"loss": 0.9974,
|
1290 |
+
"step": 18300
|
1291 |
+
},
|
1292 |
+
{
|
1293 |
+
"epoch": 0.5393996247654784,
|
1294 |
+
"grad_norm": 1.1882303953170776,
|
1295 |
+
"learning_rate": 2.3036772036830686e-05,
|
1296 |
+
"loss": 0.9999,
|
1297 |
+
"step": 18400
|
1298 |
+
},
|
1299 |
+
{
|
1300 |
+
"epoch": 0.5423311444652908,
|
1301 |
+
"grad_norm": 1.2132831811904907,
|
1302 |
+
"learning_rate": 2.289015307020116e-05,
|
1303 |
+
"loss": 0.9826,
|
1304 |
+
"step": 18500
|
1305 |
+
},
|
1306 |
+
{
|
1307 |
+
"epoch": 0.5452626641651032,
|
1308 |
+
"grad_norm": 0.7028972506523132,
|
1309 |
+
"learning_rate": 2.274353410357164e-05,
|
1310 |
+
"loss": 0.9785,
|
1311 |
+
"step": 18600
|
1312 |
+
},
|
1313 |
+
{
|
1314 |
+
"epoch": 0.5481941838649156,
|
1315 |
+
"grad_norm": 0.5096247792243958,
|
1316 |
+
"learning_rate": 2.2596915136942115e-05,
|
1317 |
+
"loss": 0.9985,
|
1318 |
+
"step": 18700
|
1319 |
+
},
|
1320 |
+
{
|
1321 |
+
"epoch": 0.551125703564728,
|
1322 |
+
"grad_norm": 0.7240370512008667,
|
1323 |
+
"learning_rate": 2.2450296170312593e-05,
|
1324 |
+
"loss": 0.9938,
|
1325 |
+
"step": 18800
|
1326 |
+
},
|
1327 |
+
{
|
1328 |
+
"epoch": 0.5540572232645403,
|
1329 |
+
"grad_norm": 1.274845004081726,
|
1330 |
+
"learning_rate": 2.2303677203683067e-05,
|
1331 |
+
"loss": 0.9979,
|
1332 |
+
"step": 18900
|
1333 |
+
},
|
1334 |
+
{
|
1335 |
+
"epoch": 0.5569887429643527,
|
1336 |
+
"grad_norm": 1.0941996574401855,
|
1337 |
+
"learning_rate": 2.2157058237053548e-05,
|
1338 |
+
"loss": 1.0095,
|
1339 |
+
"step": 19000
|
1340 |
+
},
|
1341 |
+
{
|
1342 |
+
"epoch": 0.5599202626641651,
|
1343 |
+
"grad_norm": 0.6764891147613525,
|
1344 |
+
"learning_rate": 2.2010439270424025e-05,
|
1345 |
+
"loss": 0.9839,
|
1346 |
+
"step": 19100
|
1347 |
+
},
|
1348 |
+
{
|
1349 |
+
"epoch": 0.5628517823639775,
|
1350 |
+
"grad_norm": 1.7382996082305908,
|
1351 |
+
"learning_rate": 2.18638203037945e-05,
|
1352 |
+
"loss": 0.9846,
|
1353 |
+
"step": 19200
|
1354 |
+
},
|
1355 |
+
{
|
1356 |
+
"epoch": 0.5657833020637899,
|
1357 |
+
"grad_norm": 0.9387032389640808,
|
1358 |
+
"learning_rate": 2.1717201337164977e-05,
|
1359 |
+
"loss": 0.9895,
|
1360 |
+
"step": 19300
|
1361 |
+
},
|
1362 |
+
{
|
1363 |
+
"epoch": 0.5687148217636022,
|
1364 |
+
"grad_norm": 1.1649117469787598,
|
1365 |
+
"learning_rate": 2.157058237053545e-05,
|
1366 |
+
"loss": 0.9927,
|
1367 |
+
"step": 19400
|
1368 |
+
},
|
1369 |
+
{
|
1370 |
+
"epoch": 0.5716463414634146,
|
1371 |
+
"grad_norm": 0.8027826547622681,
|
1372 |
+
"learning_rate": 2.1423963403905932e-05,
|
1373 |
+
"loss": 0.9856,
|
1374 |
+
"step": 19500
|
1375 |
+
},
|
1376 |
+
{
|
1377 |
+
"epoch": 0.5745778611632271,
|
1378 |
+
"grad_norm": 1.4004380702972412,
|
1379 |
+
"learning_rate": 2.1277344437276406e-05,
|
1380 |
+
"loss": 0.9868,
|
1381 |
+
"step": 19600
|
1382 |
+
},
|
1383 |
+
{
|
1384 |
+
"epoch": 0.5775093808630394,
|
1385 |
+
"grad_norm": 0.5555911064147949,
|
1386 |
+
"learning_rate": 2.1130725470646884e-05,
|
1387 |
+
"loss": 0.9996,
|
1388 |
+
"step": 19700
|
1389 |
+
},
|
1390 |
+
{
|
1391 |
+
"epoch": 0.5804409005628518,
|
1392 |
+
"grad_norm": 1.0119162797927856,
|
1393 |
+
"learning_rate": 2.098410650401736e-05,
|
1394 |
+
"loss": 0.9783,
|
1395 |
+
"step": 19800
|
1396 |
+
},
|
1397 |
+
{
|
1398 |
+
"epoch": 0.5833724202626641,
|
1399 |
+
"grad_norm": 1.3984887599945068,
|
1400 |
+
"learning_rate": 2.083748753738784e-05,
|
1401 |
+
"loss": 0.9743,
|
1402 |
+
"step": 19900
|
1403 |
+
},
|
1404 |
+
{
|
1405 |
+
"epoch": 0.5863039399624765,
|
1406 |
+
"grad_norm": 2.691563844680786,
|
1407 |
+
"learning_rate": 2.0690868570758316e-05,
|
1408 |
+
"loss": 1.0043,
|
1409 |
+
"step": 20000
|
1410 |
+
}
|
1411 |
+
],
|
1412 |
+
"logging_steps": 100,
|
1413 |
+
"max_steps": 34112,
|
1414 |
+
"num_input_tokens_seen": 0,
|
1415 |
+
"num_train_epochs": 1,
|
1416 |
+
"save_steps": 500,
|
1417 |
+
"stateful_callbacks": {
|
1418 |
+
"TrainerControl": {
|
1419 |
+
"args": {
|
1420 |
+
"should_epoch_stop": false,
|
1421 |
+
"should_evaluate": false,
|
1422 |
+
"should_log": false,
|
1423 |
+
"should_save": true,
|
1424 |
+
"should_training_stop": false
|
1425 |
+
},
|
1426 |
+
"attributes": {}
|
1427 |
+
}
|
1428 |
+
},
|
1429 |
+
"total_flos": 1.6738695512064e+17,
|
1430 |
+
"train_batch_size": 36,
|
1431 |
+
"trial_name": null,
|
1432 |
+
"trial_params": null
|
1433 |
+
}
|
checkpoint-20000/training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9591953a84ca47c12016b5607aa6af8b03d94b1b1c3e735d4e7c3645542c9a3c
|
3 |
+
size 5176
|