Training in progress, step 13500, checkpoint
Browse files- checkpoint-13500/config.json +30 -0
- checkpoint-13500/generation_config.json +6 -0
- checkpoint-13500/model.safetensors +3 -0
- checkpoint-13500/optimizer.pt +3 -0
- checkpoint-13500/rng_state.pth +3 -0
- checkpoint-13500/scheduler.pt +3 -0
- checkpoint-13500/special_tokens_map.json +24 -0
- checkpoint-13500/tokenizer.json +96 -0
- checkpoint-13500/tokenizer_config.json +38 -0
- checkpoint-13500/trainer_state.json +978 -0
- checkpoint-13500/training_args.bin +3 -0
checkpoint-13500/config.json
ADDED
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "EleutherAI/pythia-160m",
|
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": 768,
|
14 |
+
"initializer_range": 0.02,
|
15 |
+
"intermediate_size": 3072,
|
16 |
+
"layer_norm_eps": 1e-05,
|
17 |
+
"max_position_embeddings": 2048,
|
18 |
+
"model_type": "gpt_neox",
|
19 |
+
"num_attention_heads": 12,
|
20 |
+
"num_hidden_layers": 12,
|
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-13500/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-13500/model.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:28ee27c0961c7c60f95b7783624efac8bb68a284b10891aee513af1f0ca9fe25
|
3 |
+
size 340289960
|
checkpoint-13500/optimizer.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:620e2fbe5eecdfb83d21db11f994d268d201b161bf7946703e667bd973a403df
|
3 |
+
size 680670138
|
checkpoint-13500/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-13500/scheduler.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e79ee010ec30fbb68836ca4e58171bd109eca8e966f24a9b0f3de97a5a245396
|
3 |
+
size 1064
|
checkpoint-13500/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-13500/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-13500/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-13500/trainer_state.json
ADDED
@@ -0,0 +1,978 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 0.13191967557531636,
|
5 |
+
"eval_steps": 500,
|
6 |
+
"global_step": 13500,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.0009771827820393806,
|
13 |
+
"grad_norm": 0.5417118072509766,
|
14 |
+
"learning_rate": 4.995602247740044e-05,
|
15 |
+
"loss": 1.378,
|
16 |
+
"step": 100
|
17 |
+
},
|
18 |
+
{
|
19 |
+
"epoch": 0.001954365564078761,
|
20 |
+
"grad_norm": 0.6493918895721436,
|
21 |
+
"learning_rate": 4.990715856340093e-05,
|
22 |
+
"loss": 1.3304,
|
23 |
+
"step": 200
|
24 |
+
},
|
25 |
+
{
|
26 |
+
"epoch": 0.0029315483461181415,
|
27 |
+
"grad_norm": 0.9062462449073792,
|
28 |
+
"learning_rate": 4.9858294649401425e-05,
|
29 |
+
"loss": 1.3284,
|
30 |
+
"step": 300
|
31 |
+
},
|
32 |
+
{
|
33 |
+
"epoch": 0.003908731128157522,
|
34 |
+
"grad_norm": 0.750052273273468,
|
35 |
+
"learning_rate": 4.9809430735401906e-05,
|
36 |
+
"loss": 1.3166,
|
37 |
+
"step": 400
|
38 |
+
},
|
39 |
+
{
|
40 |
+
"epoch": 0.004885913910196903,
|
41 |
+
"grad_norm": 0.6602022051811218,
|
42 |
+
"learning_rate": 4.97605668214024e-05,
|
43 |
+
"loss": 1.3166,
|
44 |
+
"step": 500
|
45 |
+
},
|
46 |
+
{
|
47 |
+
"epoch": 0.005863096692236283,
|
48 |
+
"grad_norm": 0.4193927049636841,
|
49 |
+
"learning_rate": 4.971170290740288e-05,
|
50 |
+
"loss": 1.3098,
|
51 |
+
"step": 600
|
52 |
+
},
|
53 |
+
{
|
54 |
+
"epoch": 0.006840279474275663,
|
55 |
+
"grad_norm": 0.6095415949821472,
|
56 |
+
"learning_rate": 4.966283899340338e-05,
|
57 |
+
"loss": 1.3103,
|
58 |
+
"step": 700
|
59 |
+
},
|
60 |
+
{
|
61 |
+
"epoch": 0.007817462256315045,
|
62 |
+
"grad_norm": 0.9943467378616333,
|
63 |
+
"learning_rate": 4.9613975079403865e-05,
|
64 |
+
"loss": 1.3096,
|
65 |
+
"step": 800
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 0.008794645038354424,
|
69 |
+
"grad_norm": 1.2263585329055786,
|
70 |
+
"learning_rate": 4.9565111165404346e-05,
|
71 |
+
"loss": 1.3067,
|
72 |
+
"step": 900
|
73 |
+
},
|
74 |
+
{
|
75 |
+
"epoch": 0.009771827820393805,
|
76 |
+
"grad_norm": 0.7198677659034729,
|
77 |
+
"learning_rate": 4.951624725140484e-05,
|
78 |
+
"loss": 1.3041,
|
79 |
+
"step": 1000
|
80 |
+
},
|
81 |
+
{
|
82 |
+
"epoch": 0.010749010602433185,
|
83 |
+
"grad_norm": 0.7370775938034058,
|
84 |
+
"learning_rate": 4.946738333740533e-05,
|
85 |
+
"loss": 1.302,
|
86 |
+
"step": 1100
|
87 |
+
},
|
88 |
+
{
|
89 |
+
"epoch": 0.011726193384472566,
|
90 |
+
"grad_norm": 0.5109437704086304,
|
91 |
+
"learning_rate": 4.941851942340582e-05,
|
92 |
+
"loss": 1.3089,
|
93 |
+
"step": 1200
|
94 |
+
},
|
95 |
+
{
|
96 |
+
"epoch": 0.012703376166511945,
|
97 |
+
"grad_norm": 0.1879555583000183,
|
98 |
+
"learning_rate": 4.9369655509406305e-05,
|
99 |
+
"loss": 1.3043,
|
100 |
+
"step": 1300
|
101 |
+
},
|
102 |
+
{
|
103 |
+
"epoch": 0.013680558948551327,
|
104 |
+
"grad_norm": 0.951046884059906,
|
105 |
+
"learning_rate": 4.932079159540679e-05,
|
106 |
+
"loss": 1.3098,
|
107 |
+
"step": 1400
|
108 |
+
},
|
109 |
+
{
|
110 |
+
"epoch": 0.014657741730590706,
|
111 |
+
"grad_norm": 0.2478829026222229,
|
112 |
+
"learning_rate": 4.927192768140728e-05,
|
113 |
+
"loss": 1.3026,
|
114 |
+
"step": 1500
|
115 |
+
},
|
116 |
+
{
|
117 |
+
"epoch": 0.01563492451263009,
|
118 |
+
"grad_norm": 0.5585843324661255,
|
119 |
+
"learning_rate": 4.9223063767407776e-05,
|
120 |
+
"loss": 1.3014,
|
121 |
+
"step": 1600
|
122 |
+
},
|
123 |
+
{
|
124 |
+
"epoch": 0.016612107294669467,
|
125 |
+
"grad_norm": 0.48532453179359436,
|
126 |
+
"learning_rate": 4.917419985340826e-05,
|
127 |
+
"loss": 1.2981,
|
128 |
+
"step": 1700
|
129 |
+
},
|
130 |
+
{
|
131 |
+
"epoch": 0.017589290076708848,
|
132 |
+
"grad_norm": 0.4233573079109192,
|
133 |
+
"learning_rate": 4.912533593940875e-05,
|
134 |
+
"loss": 1.2992,
|
135 |
+
"step": 1800
|
136 |
+
},
|
137 |
+
{
|
138 |
+
"epoch": 0.01856647285874823,
|
139 |
+
"grad_norm": 0.3272475600242615,
|
140 |
+
"learning_rate": 4.9076472025409234e-05,
|
141 |
+
"loss": 1.292,
|
142 |
+
"step": 1900
|
143 |
+
},
|
144 |
+
{
|
145 |
+
"epoch": 0.01954365564078761,
|
146 |
+
"grad_norm": 0.5299385786056519,
|
147 |
+
"learning_rate": 4.902760811140973e-05,
|
148 |
+
"loss": 1.2963,
|
149 |
+
"step": 2000
|
150 |
+
},
|
151 |
+
{
|
152 |
+
"epoch": 0.02052083842282699,
|
153 |
+
"grad_norm": 0.1614024043083191,
|
154 |
+
"learning_rate": 4.8978744197410216e-05,
|
155 |
+
"loss": 1.2945,
|
156 |
+
"step": 2100
|
157 |
+
},
|
158 |
+
{
|
159 |
+
"epoch": 0.02149802120486637,
|
160 |
+
"grad_norm": 0.6039963960647583,
|
161 |
+
"learning_rate": 4.8929880283410705e-05,
|
162 |
+
"loss": 1.2913,
|
163 |
+
"step": 2200
|
164 |
+
},
|
165 |
+
{
|
166 |
+
"epoch": 0.02247520398690575,
|
167 |
+
"grad_norm": 0.5772804021835327,
|
168 |
+
"learning_rate": 4.888101636941119e-05,
|
169 |
+
"loss": 1.2895,
|
170 |
+
"step": 2300
|
171 |
+
},
|
172 |
+
{
|
173 |
+
"epoch": 0.023452386768945132,
|
174 |
+
"grad_norm": 0.7489622235298157,
|
175 |
+
"learning_rate": 4.883215245541168e-05,
|
176 |
+
"loss": 1.2847,
|
177 |
+
"step": 2400
|
178 |
+
},
|
179 |
+
{
|
180 |
+
"epoch": 0.024429569550984513,
|
181 |
+
"grad_norm": 0.30208253860473633,
|
182 |
+
"learning_rate": 4.878328854141217e-05,
|
183 |
+
"loss": 1.2924,
|
184 |
+
"step": 2500
|
185 |
+
},
|
186 |
+
{
|
187 |
+
"epoch": 0.02540675233302389,
|
188 |
+
"grad_norm": 0.36944472789764404,
|
189 |
+
"learning_rate": 4.873442462741266e-05,
|
190 |
+
"loss": 1.2916,
|
191 |
+
"step": 2600
|
192 |
+
},
|
193 |
+
{
|
194 |
+
"epoch": 0.026383935115063272,
|
195 |
+
"grad_norm": 0.3268676698207855,
|
196 |
+
"learning_rate": 4.8685560713413145e-05,
|
197 |
+
"loss": 1.2893,
|
198 |
+
"step": 2700
|
199 |
+
},
|
200 |
+
{
|
201 |
+
"epoch": 0.027361117897102653,
|
202 |
+
"grad_norm": 0.2795974910259247,
|
203 |
+
"learning_rate": 4.863669679941363e-05,
|
204 |
+
"loss": 1.282,
|
205 |
+
"step": 2800
|
206 |
+
},
|
207 |
+
{
|
208 |
+
"epoch": 0.028338300679142035,
|
209 |
+
"grad_norm": 0.36298853158950806,
|
210 |
+
"learning_rate": 4.858783288541413e-05,
|
211 |
+
"loss": 1.2832,
|
212 |
+
"step": 2900
|
213 |
+
},
|
214 |
+
{
|
215 |
+
"epoch": 0.029315483461181412,
|
216 |
+
"grad_norm": 0.5242423415184021,
|
217 |
+
"learning_rate": 4.853896897141461e-05,
|
218 |
+
"loss": 1.2819,
|
219 |
+
"step": 3000
|
220 |
+
},
|
221 |
+
{
|
222 |
+
"epoch": 0.030292666243220794,
|
223 |
+
"grad_norm": 0.25340864062309265,
|
224 |
+
"learning_rate": 4.8490105057415104e-05,
|
225 |
+
"loss": 1.2809,
|
226 |
+
"step": 3100
|
227 |
+
},
|
228 |
+
{
|
229 |
+
"epoch": 0.03126984902526018,
|
230 |
+
"grad_norm": 0.7241976261138916,
|
231 |
+
"learning_rate": 4.844124114341559e-05,
|
232 |
+
"loss": 1.2802,
|
233 |
+
"step": 3200
|
234 |
+
},
|
235 |
+
{
|
236 |
+
"epoch": 0.032247031807299556,
|
237 |
+
"grad_norm": 0.5154001712799072,
|
238 |
+
"learning_rate": 4.839237722941608e-05,
|
239 |
+
"loss": 1.2748,
|
240 |
+
"step": 3300
|
241 |
+
},
|
242 |
+
{
|
243 |
+
"epoch": 0.033224214589338934,
|
244 |
+
"grad_norm": 0.5323473811149597,
|
245 |
+
"learning_rate": 4.834351331541657e-05,
|
246 |
+
"loss": 1.284,
|
247 |
+
"step": 3400
|
248 |
+
},
|
249 |
+
{
|
250 |
+
"epoch": 0.03420139737137832,
|
251 |
+
"grad_norm": 0.3947168290615082,
|
252 |
+
"learning_rate": 4.8294649401417056e-05,
|
253 |
+
"loss": 1.276,
|
254 |
+
"step": 3500
|
255 |
+
},
|
256 |
+
{
|
257 |
+
"epoch": 0.035178580153417696,
|
258 |
+
"grad_norm": 0.4776057302951813,
|
259 |
+
"learning_rate": 4.8245785487417544e-05,
|
260 |
+
"loss": 1.2783,
|
261 |
+
"step": 3600
|
262 |
+
},
|
263 |
+
{
|
264 |
+
"epoch": 0.036155762935457074,
|
265 |
+
"grad_norm": 0.4884164035320282,
|
266 |
+
"learning_rate": 4.819692157341804e-05,
|
267 |
+
"loss": 1.2745,
|
268 |
+
"step": 3700
|
269 |
+
},
|
270 |
+
{
|
271 |
+
"epoch": 0.03713294571749646,
|
272 |
+
"grad_norm": 0.5210428833961487,
|
273 |
+
"learning_rate": 4.814805765941852e-05,
|
274 |
+
"loss": 1.2707,
|
275 |
+
"step": 3800
|
276 |
+
},
|
277 |
+
{
|
278 |
+
"epoch": 0.038110128499535836,
|
279 |
+
"grad_norm": 0.46214359998703003,
|
280 |
+
"learning_rate": 4.809919374541901e-05,
|
281 |
+
"loss": 1.2727,
|
282 |
+
"step": 3900
|
283 |
+
},
|
284 |
+
{
|
285 |
+
"epoch": 0.03908731128157522,
|
286 |
+
"grad_norm": 0.2656782865524292,
|
287 |
+
"learning_rate": 4.8050329831419496e-05,
|
288 |
+
"loss": 1.2694,
|
289 |
+
"step": 4000
|
290 |
+
},
|
291 |
+
{
|
292 |
+
"epoch": 0.0400644940636146,
|
293 |
+
"grad_norm": 0.4923059940338135,
|
294 |
+
"learning_rate": 4.8001465917419985e-05,
|
295 |
+
"loss": 1.2665,
|
296 |
+
"step": 4100
|
297 |
+
},
|
298 |
+
{
|
299 |
+
"epoch": 0.04104167684565398,
|
300 |
+
"grad_norm": 0.92928147315979,
|
301 |
+
"learning_rate": 4.795260200342048e-05,
|
302 |
+
"loss": 1.2627,
|
303 |
+
"step": 4200
|
304 |
+
},
|
305 |
+
{
|
306 |
+
"epoch": 0.04201885962769336,
|
307 |
+
"grad_norm": 1.0651229619979858,
|
308 |
+
"learning_rate": 4.790373808942096e-05,
|
309 |
+
"loss": 1.2623,
|
310 |
+
"step": 4300
|
311 |
+
},
|
312 |
+
{
|
313 |
+
"epoch": 0.04299604240973274,
|
314 |
+
"grad_norm": 0.9612557888031006,
|
315 |
+
"learning_rate": 4.7854874175421456e-05,
|
316 |
+
"loss": 1.2482,
|
317 |
+
"step": 4400
|
318 |
+
},
|
319 |
+
{
|
320 |
+
"epoch": 0.043973225191772124,
|
321 |
+
"grad_norm": 1.0120874643325806,
|
322 |
+
"learning_rate": 4.7806010261421944e-05,
|
323 |
+
"loss": 1.2589,
|
324 |
+
"step": 4500
|
325 |
+
},
|
326 |
+
{
|
327 |
+
"epoch": 0.0449504079738115,
|
328 |
+
"grad_norm": 0.6250020861625671,
|
329 |
+
"learning_rate": 4.775714634742243e-05,
|
330 |
+
"loss": 1.2499,
|
331 |
+
"step": 4600
|
332 |
+
},
|
333 |
+
{
|
334 |
+
"epoch": 0.04592759075585088,
|
335 |
+
"grad_norm": 0.2850038707256317,
|
336 |
+
"learning_rate": 4.770828243342292e-05,
|
337 |
+
"loss": 1.2446,
|
338 |
+
"step": 4700
|
339 |
+
},
|
340 |
+
{
|
341 |
+
"epoch": 0.046904773537890264,
|
342 |
+
"grad_norm": 1.2032625675201416,
|
343 |
+
"learning_rate": 4.765941851942341e-05,
|
344 |
+
"loss": 1.2238,
|
345 |
+
"step": 4800
|
346 |
+
},
|
347 |
+
{
|
348 |
+
"epoch": 0.04788195631992964,
|
349 |
+
"grad_norm": 0.42024949193000793,
|
350 |
+
"learning_rate": 4.7610554605423896e-05,
|
351 |
+
"loss": 1.2255,
|
352 |
+
"step": 4900
|
353 |
+
},
|
354 |
+
{
|
355 |
+
"epoch": 0.048859139101969026,
|
356 |
+
"grad_norm": 0.7451406121253967,
|
357 |
+
"learning_rate": 4.756169069142439e-05,
|
358 |
+
"loss": 1.2071,
|
359 |
+
"step": 5000
|
360 |
+
},
|
361 |
+
{
|
362 |
+
"epoch": 0.049836321884008404,
|
363 |
+
"grad_norm": 0.8735096454620361,
|
364 |
+
"learning_rate": 4.751282677742487e-05,
|
365 |
+
"loss": 1.2126,
|
366 |
+
"step": 5100
|
367 |
+
},
|
368 |
+
{
|
369 |
+
"epoch": 0.05081350466604778,
|
370 |
+
"grad_norm": 0.73675137758255,
|
371 |
+
"learning_rate": 4.746396286342537e-05,
|
372 |
+
"loss": 1.2036,
|
373 |
+
"step": 5200
|
374 |
+
},
|
375 |
+
{
|
376 |
+
"epoch": 0.051790687448087167,
|
377 |
+
"grad_norm": 0.6540606617927551,
|
378 |
+
"learning_rate": 4.741509894942585e-05,
|
379 |
+
"loss": 1.1825,
|
380 |
+
"step": 5300
|
381 |
+
},
|
382 |
+
{
|
383 |
+
"epoch": 0.052767870230126544,
|
384 |
+
"grad_norm": 0.825066864490509,
|
385 |
+
"learning_rate": 4.7366235035426336e-05,
|
386 |
+
"loss": 1.1655,
|
387 |
+
"step": 5400
|
388 |
+
},
|
389 |
+
{
|
390 |
+
"epoch": 0.05374505301216593,
|
391 |
+
"grad_norm": 1.6421219110488892,
|
392 |
+
"learning_rate": 4.731737112142683e-05,
|
393 |
+
"loss": 1.1716,
|
394 |
+
"step": 5500
|
395 |
+
},
|
396 |
+
{
|
397 |
+
"epoch": 0.05472223579420531,
|
398 |
+
"grad_norm": 1.0644057989120483,
|
399 |
+
"learning_rate": 4.726850720742731e-05,
|
400 |
+
"loss": 1.1384,
|
401 |
+
"step": 5600
|
402 |
+
},
|
403 |
+
{
|
404 |
+
"epoch": 0.055699418576244684,
|
405 |
+
"grad_norm": 1.1611616611480713,
|
406 |
+
"learning_rate": 4.721964329342781e-05,
|
407 |
+
"loss": 1.1499,
|
408 |
+
"step": 5700
|
409 |
+
},
|
410 |
+
{
|
411 |
+
"epoch": 0.05667660135828407,
|
412 |
+
"grad_norm": 2.0900723934173584,
|
413 |
+
"learning_rate": 4.7170779379428295e-05,
|
414 |
+
"loss": 1.1323,
|
415 |
+
"step": 5800
|
416 |
+
},
|
417 |
+
{
|
418 |
+
"epoch": 0.05765378414032345,
|
419 |
+
"grad_norm": 1.0580404996871948,
|
420 |
+
"learning_rate": 4.712191546542878e-05,
|
421 |
+
"loss": 1.112,
|
422 |
+
"step": 5900
|
423 |
+
},
|
424 |
+
{
|
425 |
+
"epoch": 0.058630966922362825,
|
426 |
+
"grad_norm": 0.6299407482147217,
|
427 |
+
"learning_rate": 4.707305155142927e-05,
|
428 |
+
"loss": 1.104,
|
429 |
+
"step": 6000
|
430 |
+
},
|
431 |
+
{
|
432 |
+
"epoch": 0.05960814970440221,
|
433 |
+
"grad_norm": 0.6816271543502808,
|
434 |
+
"learning_rate": 4.702418763742976e-05,
|
435 |
+
"loss": 1.1128,
|
436 |
+
"step": 6100
|
437 |
+
},
|
438 |
+
{
|
439 |
+
"epoch": 0.06058533248644159,
|
440 |
+
"grad_norm": 0.654796302318573,
|
441 |
+
"learning_rate": 4.697532372343025e-05,
|
442 |
+
"loss": 1.0942,
|
443 |
+
"step": 6200
|
444 |
+
},
|
445 |
+
{
|
446 |
+
"epoch": 0.06156251526848097,
|
447 |
+
"grad_norm": 1.0433884859085083,
|
448 |
+
"learning_rate": 4.692645980943074e-05,
|
449 |
+
"loss": 1.0862,
|
450 |
+
"step": 6300
|
451 |
+
},
|
452 |
+
{
|
453 |
+
"epoch": 0.06253969805052036,
|
454 |
+
"grad_norm": 0.6256537437438965,
|
455 |
+
"learning_rate": 4.6877595895431224e-05,
|
456 |
+
"loss": 1.081,
|
457 |
+
"step": 6400
|
458 |
+
},
|
459 |
+
{
|
460 |
+
"epoch": 0.06351688083255973,
|
461 |
+
"grad_norm": 0.8173975348472595,
|
462 |
+
"learning_rate": 4.682873198143172e-05,
|
463 |
+
"loss": 1.0767,
|
464 |
+
"step": 6500
|
465 |
+
},
|
466 |
+
{
|
467 |
+
"epoch": 0.06449406361459911,
|
468 |
+
"grad_norm": 0.7856473922729492,
|
469 |
+
"learning_rate": 4.6779868067432206e-05,
|
470 |
+
"loss": 1.0767,
|
471 |
+
"step": 6600
|
472 |
+
},
|
473 |
+
{
|
474 |
+
"epoch": 0.0654712463966385,
|
475 |
+
"grad_norm": 0.6337741017341614,
|
476 |
+
"learning_rate": 4.6731004153432695e-05,
|
477 |
+
"loss": 1.0829,
|
478 |
+
"step": 6700
|
479 |
+
},
|
480 |
+
{
|
481 |
+
"epoch": 0.06644842917867787,
|
482 |
+
"grad_norm": 0.5813809037208557,
|
483 |
+
"learning_rate": 4.668214023943318e-05,
|
484 |
+
"loss": 1.0571,
|
485 |
+
"step": 6800
|
486 |
+
},
|
487 |
+
{
|
488 |
+
"epoch": 0.06742561196071725,
|
489 |
+
"grad_norm": 0.4155445992946625,
|
490 |
+
"learning_rate": 4.6633276325433664e-05,
|
491 |
+
"loss": 1.0707,
|
492 |
+
"step": 6900
|
493 |
+
},
|
494 |
+
{
|
495 |
+
"epoch": 0.06840279474275664,
|
496 |
+
"grad_norm": 0.6730567812919617,
|
497 |
+
"learning_rate": 4.658441241143416e-05,
|
498 |
+
"loss": 1.0477,
|
499 |
+
"step": 7000
|
500 |
+
},
|
501 |
+
{
|
502 |
+
"epoch": 0.06937997752479601,
|
503 |
+
"grad_norm": 0.8348300457000732,
|
504 |
+
"learning_rate": 4.653554849743465e-05,
|
505 |
+
"loss": 1.0644,
|
506 |
+
"step": 7100
|
507 |
+
},
|
508 |
+
{
|
509 |
+
"epoch": 0.07035716030683539,
|
510 |
+
"grad_norm": 2.2414326667785645,
|
511 |
+
"learning_rate": 4.6486684583435135e-05,
|
512 |
+
"loss": 1.0577,
|
513 |
+
"step": 7200
|
514 |
+
},
|
515 |
+
{
|
516 |
+
"epoch": 0.07133434308887478,
|
517 |
+
"grad_norm": 1.6573911905288696,
|
518 |
+
"learning_rate": 4.643782066943562e-05,
|
519 |
+
"loss": 1.0836,
|
520 |
+
"step": 7300
|
521 |
+
},
|
522 |
+
{
|
523 |
+
"epoch": 0.07231152587091415,
|
524 |
+
"grad_norm": 0.5690039396286011,
|
525 |
+
"learning_rate": 4.638895675543611e-05,
|
526 |
+
"loss": 1.0541,
|
527 |
+
"step": 7400
|
528 |
+
},
|
529 |
+
{
|
530 |
+
"epoch": 0.07328870865295353,
|
531 |
+
"grad_norm": 0.527215301990509,
|
532 |
+
"learning_rate": 4.63400928414366e-05,
|
533 |
+
"loss": 1.0164,
|
534 |
+
"step": 7500
|
535 |
+
},
|
536 |
+
{
|
537 |
+
"epoch": 0.07426589143499292,
|
538 |
+
"grad_norm": 0.7997362613677979,
|
539 |
+
"learning_rate": 4.6291228927437094e-05,
|
540 |
+
"loss": 1.0447,
|
541 |
+
"step": 7600
|
542 |
+
},
|
543 |
+
{
|
544 |
+
"epoch": 0.0752430742170323,
|
545 |
+
"grad_norm": 2.257143259048462,
|
546 |
+
"learning_rate": 4.6242365013437575e-05,
|
547 |
+
"loss": 1.0365,
|
548 |
+
"step": 7700
|
549 |
+
},
|
550 |
+
{
|
551 |
+
"epoch": 0.07622025699907167,
|
552 |
+
"grad_norm": 0.9132490158081055,
|
553 |
+
"learning_rate": 4.619350109943807e-05,
|
554 |
+
"loss": 1.0498,
|
555 |
+
"step": 7800
|
556 |
+
},
|
557 |
+
{
|
558 |
+
"epoch": 0.07719743978111106,
|
559 |
+
"grad_norm": 0.5229859948158264,
|
560 |
+
"learning_rate": 4.614463718543856e-05,
|
561 |
+
"loss": 1.0342,
|
562 |
+
"step": 7900
|
563 |
+
},
|
564 |
+
{
|
565 |
+
"epoch": 0.07817462256315044,
|
566 |
+
"grad_norm": 0.6948792338371277,
|
567 |
+
"learning_rate": 4.6095773271439046e-05,
|
568 |
+
"loss": 1.0325,
|
569 |
+
"step": 8000
|
570 |
+
},
|
571 |
+
{
|
572 |
+
"epoch": 0.07915180534518981,
|
573 |
+
"grad_norm": 0.8526360988616943,
|
574 |
+
"learning_rate": 4.6046909357439534e-05,
|
575 |
+
"loss": 1.0183,
|
576 |
+
"step": 8100
|
577 |
+
},
|
578 |
+
{
|
579 |
+
"epoch": 0.0801289881272292,
|
580 |
+
"grad_norm": 1.1457374095916748,
|
581 |
+
"learning_rate": 4.599804544344002e-05,
|
582 |
+
"loss": 1.0243,
|
583 |
+
"step": 8200
|
584 |
+
},
|
585 |
+
{
|
586 |
+
"epoch": 0.08110617090926858,
|
587 |
+
"grad_norm": 0.9335997700691223,
|
588 |
+
"learning_rate": 4.594918152944051e-05,
|
589 |
+
"loss": 1.046,
|
590 |
+
"step": 8300
|
591 |
+
},
|
592 |
+
{
|
593 |
+
"epoch": 0.08208335369130795,
|
594 |
+
"grad_norm": 0.8367229700088501,
|
595 |
+
"learning_rate": 4.5900317615441e-05,
|
596 |
+
"loss": 1.0176,
|
597 |
+
"step": 8400
|
598 |
+
},
|
599 |
+
{
|
600 |
+
"epoch": 0.08306053647334734,
|
601 |
+
"grad_norm": 3.7648801803588867,
|
602 |
+
"learning_rate": 4.5851453701441486e-05,
|
603 |
+
"loss": 1.0047,
|
604 |
+
"step": 8500
|
605 |
+
},
|
606 |
+
{
|
607 |
+
"epoch": 0.08403771925538672,
|
608 |
+
"grad_norm": 0.5877612829208374,
|
609 |
+
"learning_rate": 4.5802589787441975e-05,
|
610 |
+
"loss": 1.0346,
|
611 |
+
"step": 8600
|
612 |
+
},
|
613 |
+
{
|
614 |
+
"epoch": 0.08501490203742611,
|
615 |
+
"grad_norm": 0.5145990252494812,
|
616 |
+
"learning_rate": 4.575372587344246e-05,
|
617 |
+
"loss": 1.0268,
|
618 |
+
"step": 8700
|
619 |
+
},
|
620 |
+
{
|
621 |
+
"epoch": 0.08599208481946548,
|
622 |
+
"grad_norm": 0.9310688376426697,
|
623 |
+
"learning_rate": 4.570486195944295e-05,
|
624 |
+
"loss": 1.0109,
|
625 |
+
"step": 8800
|
626 |
+
},
|
627 |
+
{
|
628 |
+
"epoch": 0.08696926760150486,
|
629 |
+
"grad_norm": 0.5182886719703674,
|
630 |
+
"learning_rate": 4.5655998045443445e-05,
|
631 |
+
"loss": 1.0117,
|
632 |
+
"step": 8900
|
633 |
+
},
|
634 |
+
{
|
635 |
+
"epoch": 0.08794645038354425,
|
636 |
+
"grad_norm": 0.4319695234298706,
|
637 |
+
"learning_rate": 4.560713413144393e-05,
|
638 |
+
"loss": 1.0053,
|
639 |
+
"step": 9000
|
640 |
+
},
|
641 |
+
{
|
642 |
+
"epoch": 0.08892363316558362,
|
643 |
+
"grad_norm": 4.307732582092285,
|
644 |
+
"learning_rate": 4.555827021744442e-05,
|
645 |
+
"loss": 1.0151,
|
646 |
+
"step": 9100
|
647 |
+
},
|
648 |
+
{
|
649 |
+
"epoch": 0.089900815947623,
|
650 |
+
"grad_norm": 0.46516236662864685,
|
651 |
+
"learning_rate": 4.550940630344491e-05,
|
652 |
+
"loss": 0.9945,
|
653 |
+
"step": 9200
|
654 |
+
},
|
655 |
+
{
|
656 |
+
"epoch": 0.09087799872966239,
|
657 |
+
"grad_norm": 1.2372952699661255,
|
658 |
+
"learning_rate": 4.54605423894454e-05,
|
659 |
+
"loss": 0.9865,
|
660 |
+
"step": 9300
|
661 |
+
},
|
662 |
+
{
|
663 |
+
"epoch": 0.09185518151170176,
|
664 |
+
"grad_norm": 0.7494595646858215,
|
665 |
+
"learning_rate": 4.5411678475445886e-05,
|
666 |
+
"loss": 0.9824,
|
667 |
+
"step": 9400
|
668 |
+
},
|
669 |
+
{
|
670 |
+
"epoch": 0.09283236429374114,
|
671 |
+
"grad_norm": 0.5540333390235901,
|
672 |
+
"learning_rate": 4.5362814561446374e-05,
|
673 |
+
"loss": 1.0132,
|
674 |
+
"step": 9500
|
675 |
+
},
|
676 |
+
{
|
677 |
+
"epoch": 0.09380954707578053,
|
678 |
+
"grad_norm": 0.48533427715301514,
|
679 |
+
"learning_rate": 4.531395064744686e-05,
|
680 |
+
"loss": 1.0173,
|
681 |
+
"step": 9600
|
682 |
+
},
|
683 |
+
{
|
684 |
+
"epoch": 0.0947867298578199,
|
685 |
+
"grad_norm": 0.4972572922706604,
|
686 |
+
"learning_rate": 4.526508673344736e-05,
|
687 |
+
"loss": 1.0078,
|
688 |
+
"step": 9700
|
689 |
+
},
|
690 |
+
{
|
691 |
+
"epoch": 0.09576391263985928,
|
692 |
+
"grad_norm": 0.6748878955841064,
|
693 |
+
"learning_rate": 4.521622281944784e-05,
|
694 |
+
"loss": 1.0172,
|
695 |
+
"step": 9800
|
696 |
+
},
|
697 |
+
{
|
698 |
+
"epoch": 0.09674109542189867,
|
699 |
+
"grad_norm": 0.5261876583099365,
|
700 |
+
"learning_rate": 4.5167358905448326e-05,
|
701 |
+
"loss": 1.0189,
|
702 |
+
"step": 9900
|
703 |
+
},
|
704 |
+
{
|
705 |
+
"epoch": 0.09771827820393805,
|
706 |
+
"grad_norm": 0.4164600670337677,
|
707 |
+
"learning_rate": 4.5118494991448814e-05,
|
708 |
+
"loss": 0.9978,
|
709 |
+
"step": 10000
|
710 |
+
},
|
711 |
+
{
|
712 |
+
"epoch": 0.09869546098597742,
|
713 |
+
"grad_norm": 0.40417763590812683,
|
714 |
+
"learning_rate": 4.50696310774493e-05,
|
715 |
+
"loss": 1.0103,
|
716 |
+
"step": 10100
|
717 |
+
},
|
718 |
+
{
|
719 |
+
"epoch": 0.09967264376801681,
|
720 |
+
"grad_norm": 0.8591890931129456,
|
721 |
+
"learning_rate": 4.50207671634498e-05,
|
722 |
+
"loss": 1.0065,
|
723 |
+
"step": 10200
|
724 |
+
},
|
725 |
+
{
|
726 |
+
"epoch": 0.10064982655005619,
|
727 |
+
"grad_norm": 0.5676371455192566,
|
728 |
+
"learning_rate": 4.497190324945028e-05,
|
729 |
+
"loss": 1.0089,
|
730 |
+
"step": 10300
|
731 |
+
},
|
732 |
+
{
|
733 |
+
"epoch": 0.10162700933209556,
|
734 |
+
"grad_norm": 0.616646945476532,
|
735 |
+
"learning_rate": 4.492303933545077e-05,
|
736 |
+
"loss": 0.9897,
|
737 |
+
"step": 10400
|
738 |
+
},
|
739 |
+
{
|
740 |
+
"epoch": 0.10260419211413495,
|
741 |
+
"grad_norm": 0.37536484003067017,
|
742 |
+
"learning_rate": 4.487417542145126e-05,
|
743 |
+
"loss": 0.9989,
|
744 |
+
"step": 10500
|
745 |
+
},
|
746 |
+
{
|
747 |
+
"epoch": 0.10358137489617433,
|
748 |
+
"grad_norm": 0.6801789402961731,
|
749 |
+
"learning_rate": 4.482531150745175e-05,
|
750 |
+
"loss": 0.9923,
|
751 |
+
"step": 10600
|
752 |
+
},
|
753 |
+
{
|
754 |
+
"epoch": 0.1045585576782137,
|
755 |
+
"grad_norm": 0.5848776698112488,
|
756 |
+
"learning_rate": 4.477644759345224e-05,
|
757 |
+
"loss": 0.9919,
|
758 |
+
"step": 10700
|
759 |
+
},
|
760 |
+
{
|
761 |
+
"epoch": 0.10553574046025309,
|
762 |
+
"grad_norm": 0.7715157866477966,
|
763 |
+
"learning_rate": 4.4727583679452725e-05,
|
764 |
+
"loss": 0.9814,
|
765 |
+
"step": 10800
|
766 |
+
},
|
767 |
+
{
|
768 |
+
"epoch": 0.10651292324229247,
|
769 |
+
"grad_norm": 0.8080986142158508,
|
770 |
+
"learning_rate": 4.4678719765453214e-05,
|
771 |
+
"loss": 0.9935,
|
772 |
+
"step": 10900
|
773 |
+
},
|
774 |
+
{
|
775 |
+
"epoch": 0.10749010602433186,
|
776 |
+
"grad_norm": 0.4375016391277313,
|
777 |
+
"learning_rate": 4.462985585145371e-05,
|
778 |
+
"loss": 0.988,
|
779 |
+
"step": 11000
|
780 |
+
},
|
781 |
+
{
|
782 |
+
"epoch": 0.10846728880637123,
|
783 |
+
"grad_norm": 0.8055805563926697,
|
784 |
+
"learning_rate": 4.458099193745419e-05,
|
785 |
+
"loss": 0.9861,
|
786 |
+
"step": 11100
|
787 |
+
},
|
788 |
+
{
|
789 |
+
"epoch": 0.10944447158841061,
|
790 |
+
"grad_norm": 1.1914618015289307,
|
791 |
+
"learning_rate": 4.4532128023454685e-05,
|
792 |
+
"loss": 0.9622,
|
793 |
+
"step": 11200
|
794 |
+
},
|
795 |
+
{
|
796 |
+
"epoch": 0.11042165437045,
|
797 |
+
"grad_norm": 0.4247540533542633,
|
798 |
+
"learning_rate": 4.448326410945517e-05,
|
799 |
+
"loss": 0.9602,
|
800 |
+
"step": 11300
|
801 |
+
},
|
802 |
+
{
|
803 |
+
"epoch": 0.11139883715248937,
|
804 |
+
"grad_norm": 0.5454650521278381,
|
805 |
+
"learning_rate": 4.4434400195455654e-05,
|
806 |
+
"loss": 0.9696,
|
807 |
+
"step": 11400
|
808 |
+
},
|
809 |
+
{
|
810 |
+
"epoch": 0.11237601993452875,
|
811 |
+
"grad_norm": 0.5259748697280884,
|
812 |
+
"learning_rate": 4.438553628145615e-05,
|
813 |
+
"loss": 1.0021,
|
814 |
+
"step": 11500
|
815 |
+
},
|
816 |
+
{
|
817 |
+
"epoch": 0.11335320271656814,
|
818 |
+
"grad_norm": 0.5165246725082397,
|
819 |
+
"learning_rate": 4.433667236745663e-05,
|
820 |
+
"loss": 0.982,
|
821 |
+
"step": 11600
|
822 |
+
},
|
823 |
+
{
|
824 |
+
"epoch": 0.11433038549860751,
|
825 |
+
"grad_norm": 0.6768147945404053,
|
826 |
+
"learning_rate": 4.4287808453457125e-05,
|
827 |
+
"loss": 0.9398,
|
828 |
+
"step": 11700
|
829 |
+
},
|
830 |
+
{
|
831 |
+
"epoch": 0.1153075682806469,
|
832 |
+
"grad_norm": 1.0245041847229004,
|
833 |
+
"learning_rate": 4.423894453945761e-05,
|
834 |
+
"loss": 0.9934,
|
835 |
+
"step": 11800
|
836 |
+
},
|
837 |
+
{
|
838 |
+
"epoch": 0.11628475106268628,
|
839 |
+
"grad_norm": 0.6241583228111267,
|
840 |
+
"learning_rate": 4.41900806254581e-05,
|
841 |
+
"loss": 0.9697,
|
842 |
+
"step": 11900
|
843 |
+
},
|
844 |
+
{
|
845 |
+
"epoch": 0.11726193384472565,
|
846 |
+
"grad_norm": 0.4234873652458191,
|
847 |
+
"learning_rate": 4.414121671145859e-05,
|
848 |
+
"loss": 0.9723,
|
849 |
+
"step": 12000
|
850 |
+
},
|
851 |
+
{
|
852 |
+
"epoch": 0.11823911662676503,
|
853 |
+
"grad_norm": 0.3932545781135559,
|
854 |
+
"learning_rate": 4.409235279745908e-05,
|
855 |
+
"loss": 0.9826,
|
856 |
+
"step": 12100
|
857 |
+
},
|
858 |
+
{
|
859 |
+
"epoch": 0.11921629940880442,
|
860 |
+
"grad_norm": 1.5067880153656006,
|
861 |
+
"learning_rate": 4.4043488883459565e-05,
|
862 |
+
"loss": 0.9581,
|
863 |
+
"step": 12200
|
864 |
+
},
|
865 |
+
{
|
866 |
+
"epoch": 0.1201934821908438,
|
867 |
+
"grad_norm": 0.41707366704940796,
|
868 |
+
"learning_rate": 4.399462496946006e-05,
|
869 |
+
"loss": 0.9666,
|
870 |
+
"step": 12300
|
871 |
+
},
|
872 |
+
{
|
873 |
+
"epoch": 0.12117066497288317,
|
874 |
+
"grad_norm": 1.1278653144836426,
|
875 |
+
"learning_rate": 4.394576105546054e-05,
|
876 |
+
"loss": 0.9553,
|
877 |
+
"step": 12400
|
878 |
+
},
|
879 |
+
{
|
880 |
+
"epoch": 0.12214784775492256,
|
881 |
+
"grad_norm": 0.350543737411499,
|
882 |
+
"learning_rate": 4.3896897141461036e-05,
|
883 |
+
"loss": 0.9422,
|
884 |
+
"step": 12500
|
885 |
+
},
|
886 |
+
{
|
887 |
+
"epoch": 0.12312503053696194,
|
888 |
+
"grad_norm": 0.3775838315486908,
|
889 |
+
"learning_rate": 4.3848033227461524e-05,
|
890 |
+
"loss": 0.9626,
|
891 |
+
"step": 12600
|
892 |
+
},
|
893 |
+
{
|
894 |
+
"epoch": 0.12410221331900131,
|
895 |
+
"grad_norm": 0.8341017365455627,
|
896 |
+
"learning_rate": 4.379916931346201e-05,
|
897 |
+
"loss": 0.9289,
|
898 |
+
"step": 12700
|
899 |
+
},
|
900 |
+
{
|
901 |
+
"epoch": 0.1250793961010407,
|
902 |
+
"grad_norm": 0.805614173412323,
|
903 |
+
"learning_rate": 4.37503053994625e-05,
|
904 |
+
"loss": 0.9474,
|
905 |
+
"step": 12800
|
906 |
+
},
|
907 |
+
{
|
908 |
+
"epoch": 0.12605657888308008,
|
909 |
+
"grad_norm": 0.8439397215843201,
|
910 |
+
"learning_rate": 4.370144148546299e-05,
|
911 |
+
"loss": 0.9661,
|
912 |
+
"step": 12900
|
913 |
+
},
|
914 |
+
{
|
915 |
+
"epoch": 0.12703376166511945,
|
916 |
+
"grad_norm": 1.1272892951965332,
|
917 |
+
"learning_rate": 4.3652577571463476e-05,
|
918 |
+
"loss": 0.9514,
|
919 |
+
"step": 13000
|
920 |
+
},
|
921 |
+
{
|
922 |
+
"epoch": 0.12801094444715885,
|
923 |
+
"grad_norm": 0.6426375508308411,
|
924 |
+
"learning_rate": 4.3603713657463965e-05,
|
925 |
+
"loss": 0.9448,
|
926 |
+
"step": 13100
|
927 |
+
},
|
928 |
+
{
|
929 |
+
"epoch": 0.12898812722919822,
|
930 |
+
"grad_norm": 1.3205431699752808,
|
931 |
+
"learning_rate": 4.355484974346445e-05,
|
932 |
+
"loss": 0.9511,
|
933 |
+
"step": 13200
|
934 |
+
},
|
935 |
+
{
|
936 |
+
"epoch": 0.1299653100112376,
|
937 |
+
"grad_norm": 0.3671954870223999,
|
938 |
+
"learning_rate": 4.350598582946494e-05,
|
939 |
+
"loss": 0.9506,
|
940 |
+
"step": 13300
|
941 |
+
},
|
942 |
+
{
|
943 |
+
"epoch": 0.130942492793277,
|
944 |
+
"grad_norm": 0.7566332817077637,
|
945 |
+
"learning_rate": 4.345712191546543e-05,
|
946 |
+
"loss": 0.9363,
|
947 |
+
"step": 13400
|
948 |
+
},
|
949 |
+
{
|
950 |
+
"epoch": 0.13191967557531636,
|
951 |
+
"grad_norm": 0.8800159692764282,
|
952 |
+
"learning_rate": 4.340825800146592e-05,
|
953 |
+
"loss": 0.9388,
|
954 |
+
"step": 13500
|
955 |
+
}
|
956 |
+
],
|
957 |
+
"logging_steps": 100,
|
958 |
+
"max_steps": 102335,
|
959 |
+
"num_input_tokens_seen": 0,
|
960 |
+
"num_train_epochs": 1,
|
961 |
+
"save_steps": 500,
|
962 |
+
"stateful_callbacks": {
|
963 |
+
"TrainerControl": {
|
964 |
+
"args": {
|
965 |
+
"should_epoch_stop": false,
|
966 |
+
"should_evaluate": false,
|
967 |
+
"should_log": false,
|
968 |
+
"should_save": true,
|
969 |
+
"should_training_stop": false
|
970 |
+
},
|
971 |
+
"attributes": {}
|
972 |
+
}
|
973 |
+
},
|
974 |
+
"total_flos": 1.6932946279661568e+17,
|
975 |
+
"train_batch_size": 12,
|
976 |
+
"trial_name": null,
|
977 |
+
"trial_params": null
|
978 |
+
}
|
checkpoint-13500/training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:278b8ae4ca6b721ea0f2d20061c63ec30cd0ca06fdf7af3d153eaddb11a3fd6b
|
3 |
+
size 5176
|