Hugging Face's logo

Manirathinam21
/
M-Bert-base-cased-language-detection

Text Classification
Transformers
PyTorch
bert
Inference Endpoints
Model card Files Community
1
Manirathinam21 commited on
Commit
85dc566
•
1 Parent(s): b2691ec

Intial commit

Browse files
Files changed (8) hide show
  1. config.json +80 -0
  2. rng_state.pth +3 -0
  3. scheduler.pt +3 -0
  4. special_tokens_map.json +7 -0
  5. tokenizer.json +0 -0
  6. tokenizer_config.json +13 -0
  7. trainer_state.json +3936 -0
  8. vocab.txt +0 -0
config.json ADDED
@@ -0,0 +1,80 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "bert-base-multilingual-cased",
3
+ "architectures": [
4
+ "BertForSequenceClassification"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.1,
7
+ "classifier_dropout": null,
8
+ "directionality": "bidi",
9
+ "hidden_act": "gelu",
10
+ "hidden_dropout_prob": 0.1,
11
+ "hidden_size": 768,
12
+ "id2label": {
13
+ "0": "Arabic",
14
+ "1": "Chinese",
15
+ "2": "Dutch",
16
+ "3": "English",
17
+ "4": "Estonian",
18
+ "5": "French",
19
+ "6": "Hindi",
20
+ "7": "Indonesian",
21
+ "8": "Japanese",
22
+ "9": "Korean",
23
+ "10": "Latin",
24
+ "11": "Persian",
25
+ "12": "Portugese",
26
+ "13": "Pushto",
27
+ "14": "Romanian",
28
+ "15": "Russian",
29
+ "16": "Spanish",
30
+ "17": "Swedish",
31
+ "18": "Tamil",
32
+ "19": "Thai",
33
+ "20": "Turkish",
34
+ "21": "Urdu"
35
+ },
36
+ "initializer_range": 0.02,
37
+ "intermediate_size": 3072,
38
+ "label2id": {
39
+ "Arabic": 0,
40
+ "Chinese": 1,
41
+ "Dutch": 2,
42
+ "English": 3,
43
+ "Estonian": 4,
44
+ "French": 5,
45
+ "Hindi": 6,
46
+ "Indonesian": 7,
47
+ "Japanese": 8,
48
+ "Korean": 9,
49
+ "Latin": 10,
50
+ "Persian": 11,
51
+ "Portugese": 12,
52
+ "Pushto": 13,
53
+ "Romanian": 14,
54
+ "Russian": 15,
55
+ "Spanish": 16,
56
+ "Swedish": 17,
57
+ "Tamil": 18,
58
+ "Thai": 19,
59
+ "Turkish": 20,
60
+ "Urdu": 21
61
+ },
62
+ "layer_norm_eps": 1e-12,
63
+ "max_position_embeddings": 512,
64
+ "model_type": "bert",
65
+ "num_attention_heads": 12,
66
+ "num_hidden_layers": 12,
67
+ "pad_token_id": 0,
68
+ "pooler_fc_size": 768,
69
+ "pooler_num_attention_heads": 12,
70
+ "pooler_num_fc_layers": 3,
71
+ "pooler_size_per_head": 128,
72
+ "pooler_type": "first_token_transform",
73
+ "position_embedding_type": "absolute",
74
+ "problem_type": "single_label_classification",
75
+ "torch_dtype": "float32",
76
+ "transformers_version": "4.28.0",
77
+ "type_vocab_size": 2,
78
+ "use_cache": true,
79
+ "vocab_size": 119547
80
+ }
rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:737046bb6fc003231864158ed71bb082655d5014e2d3efeacf101ad31c8025c3
3
+ size 14575
scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:13e7b6020e24eba42d999faa1a6eda8c46b428b3cca3741db915220afbe681ba
3
+ size 627
special_tokens_map.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "cls_token": "[CLS]",
3
+ "mask_token": "[MASK]",
4
+ "pad_token": "[PAD]",
5
+ "sep_token": "[SEP]",
6
+ "unk_token": "[UNK]"
7
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "clean_up_tokenization_spaces": true,
3
+ "cls_token": "[CLS]",
4
+ "do_lower_case": false,
5
+ "mask_token": "[MASK]",
6
+ "model_max_length": 512,
7
+ "pad_token": "[PAD]",
8
+ "sep_token": "[SEP]",
9
+ "strip_accents": null,
10
+ "tokenize_chinese_chars": true,
11
+ "tokenizer_class": "BertTokenizer",
12
+ "unk_token": "[UNK]"
13
+ }
trainer_state.json ADDED
@@ -0,0 +1,3936 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 2.9545454545454546,
5
+ "global_step": 6500,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.0,
12
+ "learning_rate": 2.0000000000000003e-06,
13
+ "loss": 3.1215,
14
+ "step": 10
15
+ },
16
+ {
17
+ "epoch": 0.01,
18
+ "learning_rate": 4.000000000000001e-06,
19
+ "loss": 3.0579,
20
+ "step": 20
21
+ },
22
+ {
23
+ "epoch": 0.01,
24
+ "learning_rate": 6e-06,
25
+ "loss": 3.1027,
26
+ "step": 30
27
+ },
28
+ {
29
+ "epoch": 0.02,
30
+ "learning_rate": 8.000000000000001e-06,
31
+ "loss": 3.0934,
32
+ "step": 40
33
+ },
34
+ {
35
+ "epoch": 0.02,
36
+ "learning_rate": 1e-05,
37
+ "loss": 3.0273,
38
+ "step": 50
39
+ },
40
+ {
41
+ "epoch": 0.03,
42
+ "learning_rate": 1.2e-05,
43
+ "loss": 2.9829,
44
+ "step": 60
45
+ },
46
+ {
47
+ "epoch": 0.03,
48
+ "learning_rate": 1.4e-05,
49
+ "loss": 2.8801,
50
+ "step": 70
51
+ },
52
+ {
53
+ "epoch": 0.04,
54
+ "learning_rate": 1.6000000000000003e-05,
55
+ "loss": 2.7154,
56
+ "step": 80
57
+ },
58
+ {
59
+ "epoch": 0.04,
60
+ "learning_rate": 1.8e-05,
61
+ "loss": 2.5808,
62
+ "step": 90
63
+ },
64
+ {
65
+ "epoch": 0.05,
66
+ "learning_rate": 2e-05,
67
+ "loss": 2.2878,
68
+ "step": 100
69
+ },
70
+ {
71
+ "epoch": 0.05,
72
+ "learning_rate": 1.9969230769230773e-05,
73
+ "loss": 1.8999,
74
+ "step": 110
75
+ },
76
+ {
77
+ "epoch": 0.05,
78
+ "learning_rate": 1.993846153846154e-05,
79
+ "loss": 1.5681,
80
+ "step": 120
81
+ },
82
+ {
83
+ "epoch": 0.06,
84
+ "learning_rate": 1.990769230769231e-05,
85
+ "loss": 1.2633,
86
+ "step": 130
87
+ },
88
+ {
89
+ "epoch": 0.06,
90
+ "learning_rate": 1.987692307692308e-05,
91
+ "loss": 1.0283,
92
+ "step": 140
93
+ },
94
+ {
95
+ "epoch": 0.07,
96
+ "learning_rate": 1.9846153846153847e-05,
97
+ "loss": 0.9242,
98
+ "step": 150
99
+ },
100
+ {
101
+ "epoch": 0.07,
102
+ "learning_rate": 1.9815384615384618e-05,
103
+ "loss": 0.6346,
104
+ "step": 160
105
+ },
106
+ {
107
+ "epoch": 0.08,
108
+ "learning_rate": 1.9784615384615386e-05,
109
+ "loss": 0.6004,
110
+ "step": 170
111
+ },
112
+ {
113
+ "epoch": 0.08,
114
+ "learning_rate": 1.9753846153846153e-05,
115
+ "loss": 0.5833,
116
+ "step": 180
117
+ },
118
+ {
119
+ "epoch": 0.09,
120
+ "learning_rate": 1.9723076923076924e-05,
121
+ "loss": 0.3977,
122
+ "step": 190
123
+ },
124
+ {
125
+ "epoch": 0.09,
126
+ "learning_rate": 1.9692307692307696e-05,
127
+ "loss": 0.3478,
128
+ "step": 200
129
+ },
130
+ {
131
+ "epoch": 0.1,
132
+ "learning_rate": 1.9661538461538463e-05,
133
+ "loss": 0.3881,
134
+ "step": 210
135
+ },
136
+ {
137
+ "epoch": 0.1,
138
+ "learning_rate": 1.9630769230769234e-05,
139
+ "loss": 0.2242,
140
+ "step": 220
141
+ },
142
+ {
143
+ "epoch": 0.1,
144
+ "learning_rate": 1.9600000000000002e-05,
145
+ "loss": 0.3317,
146
+ "step": 230
147
+ },
148
+ {
149
+ "epoch": 0.11,
150
+ "learning_rate": 1.9569230769230773e-05,
151
+ "loss": 0.1648,
152
+ "step": 240
153
+ },
154
+ {
155
+ "epoch": 0.11,
156
+ "learning_rate": 1.953846153846154e-05,
157
+ "loss": 0.1313,
158
+ "step": 250
159
+ },
160
+ {
161
+ "epoch": 0.12,
162
+ "learning_rate": 1.950769230769231e-05,
163
+ "loss": 0.238,
164
+ "step": 260
165
+ },
166
+ {
167
+ "epoch": 0.12,
168
+ "learning_rate": 1.9476923076923076e-05,
169
+ "loss": 0.1908,
170
+ "step": 270
171
+ },
172
+ {
173
+ "epoch": 0.13,
174
+ "learning_rate": 1.9446153846153847e-05,
175
+ "loss": 0.085,
176
+ "step": 280
177
+ },
178
+ {
179
+ "epoch": 0.13,
180
+ "learning_rate": 1.9415384615384615e-05,
181
+ "loss": 0.1933,
182
+ "step": 290
183
+ },
184
+ {
185
+ "epoch": 0.14,
186
+ "learning_rate": 1.9384615384615386e-05,
187
+ "loss": 0.304,
188
+ "step": 300
189
+ },
190
+ {
191
+ "epoch": 0.14,
192
+ "learning_rate": 1.9353846153846157e-05,
193
+ "loss": 0.0776,
194
+ "step": 310
195
+ },
196
+ {
197
+ "epoch": 0.15,
198
+ "learning_rate": 1.9323076923076925e-05,
199
+ "loss": 0.2,
200
+ "step": 320
201
+ },
202
+ {
203
+ "epoch": 0.15,
204
+ "learning_rate": 1.9292307692307696e-05,
205
+ "loss": 0.1345,
206
+ "step": 330
207
+ },
208
+ {
209
+ "epoch": 0.15,
210
+ "learning_rate": 1.9261538461538464e-05,
211
+ "loss": 0.1144,
212
+ "step": 340
213
+ },
214
+ {
215
+ "epoch": 0.16,
216
+ "learning_rate": 1.923076923076923e-05,
217
+ "loss": 0.1022,
218
+ "step": 350
219
+ },
220
+ {
221
+ "epoch": 0.16,
222
+ "learning_rate": 1.9200000000000003e-05,
223
+ "loss": 0.0807,
224
+ "step": 360
225
+ },
226
+ {
227
+ "epoch": 0.17,
228
+ "learning_rate": 1.916923076923077e-05,
229
+ "loss": 0.0418,
230
+ "step": 370
231
+ },
232
+ {
233
+ "epoch": 0.17,
234
+ "learning_rate": 1.9138461538461538e-05,
235
+ "loss": 0.1183,
236
+ "step": 380
237
+ },
238
+ {
239
+ "epoch": 0.18,
240
+ "learning_rate": 1.910769230769231e-05,
241
+ "loss": 0.095,
242
+ "step": 390
243
+ },
244
+ {
245
+ "epoch": 0.18,
246
+ "learning_rate": 1.907692307692308e-05,
247
+ "loss": 0.0858,
248
+ "step": 400
249
+ },
250
+ {
251
+ "epoch": 0.19,
252
+ "learning_rate": 1.9046153846153848e-05,
253
+ "loss": 0.1787,
254
+ "step": 410
255
+ },
256
+ {
257
+ "epoch": 0.19,
258
+ "learning_rate": 1.901538461538462e-05,
259
+ "loss": 0.118,
260
+ "step": 420
261
+ },
262
+ {
263
+ "epoch": 0.2,
264
+ "learning_rate": 1.8984615384615387e-05,
265
+ "loss": 0.212,
266
+ "step": 430
267
+ },
268
+ {
269
+ "epoch": 0.2,
270
+ "learning_rate": 1.8953846153846158e-05,
271
+ "loss": 0.0655,
272
+ "step": 440
273
+ },
274
+ {
275
+ "epoch": 0.2,
276
+ "learning_rate": 1.8923076923076925e-05,
277
+ "loss": 0.1352,
278
+ "step": 450
279
+ },
280
+ {
281
+ "epoch": 0.21,
282
+ "learning_rate": 1.8892307692307693e-05,
283
+ "loss": 0.1011,
284
+ "step": 460
285
+ },
286
+ {
287
+ "epoch": 0.21,
288
+ "learning_rate": 1.886153846153846e-05,
289
+ "loss": 0.0258,
290
+ "step": 470
291
+ },
292
+ {
293
+ "epoch": 0.22,
294
+ "learning_rate": 1.8830769230769232e-05,
295
+ "loss": 0.1673,
296
+ "step": 480
297
+ },
298
+ {
299
+ "epoch": 0.22,
300
+ "learning_rate": 1.88e-05,
301
+ "loss": 0.0257,
302
+ "step": 490
303
+ },
304
+ {
305
+ "epoch": 0.23,
306
+ "learning_rate": 1.876923076923077e-05,
307
+ "loss": 0.0242,
308
+ "step": 500
309
+ },
310
+ {
311
+ "epoch": 0.23,
312
+ "learning_rate": 1.8738461538461542e-05,
313
+ "loss": 0.0987,
314
+ "step": 510
315
+ },
316
+ {
317
+ "epoch": 0.24,
318
+ "learning_rate": 1.870769230769231e-05,
319
+ "loss": 0.1029,
320
+ "step": 520
321
+ },
322
+ {
323
+ "epoch": 0.24,
324
+ "learning_rate": 1.867692307692308e-05,
325
+ "loss": 0.0693,
326
+ "step": 530
327
+ },
328
+ {
329
+ "epoch": 0.25,
330
+ "learning_rate": 1.8646153846153848e-05,
331
+ "loss": 0.0703,
332
+ "step": 540
333
+ },
334
+ {
335
+ "epoch": 0.25,
336
+ "learning_rate": 1.8615384615384616e-05,
337
+ "loss": 0.1149,
338
+ "step": 550
339
+ },
340
+ {
341
+ "epoch": 0.25,
342
+ "learning_rate": 1.8584615384615387e-05,
343
+ "loss": 0.0203,
344
+ "step": 560
345
+ },
346
+ {
347
+ "epoch": 0.26,
348
+ "learning_rate": 1.8553846153846155e-05,
349
+ "loss": 0.1444,
350
+ "step": 570
351
+ },
352
+ {
353
+ "epoch": 0.26,
354
+ "learning_rate": 1.8523076923076922e-05,
355
+ "loss": 0.0241,
356
+ "step": 580
357
+ },
358
+ {
359
+ "epoch": 0.27,
360
+ "learning_rate": 1.8492307692307694e-05,
361
+ "loss": 0.0439,
362
+ "step": 590
363
+ },
364
+ {
365
+ "epoch": 0.27,
366
+ "learning_rate": 1.8461538461538465e-05,
367
+ "loss": 0.1108,
368
+ "step": 600
369
+ },
370
+ {
371
+ "epoch": 0.28,
372
+ "learning_rate": 1.8430769230769232e-05,
373
+ "loss": 0.124,
374
+ "step": 610
375
+ },
376
+ {
377
+ "epoch": 0.28,
378
+ "learning_rate": 1.8400000000000003e-05,
379
+ "loss": 0.2086,
380
+ "step": 620
381
+ },
382
+ {
383
+ "epoch": 0.29,
384
+ "learning_rate": 1.836923076923077e-05,
385
+ "loss": 0.0169,
386
+ "step": 630
387
+ },
388
+ {
389
+ "epoch": 0.29,
390
+ "learning_rate": 1.8338461538461542e-05,
391
+ "loss": 0.0163,
392
+ "step": 640
393
+ },
394
+ {
395
+ "epoch": 0.3,
396
+ "learning_rate": 1.830769230769231e-05,
397
+ "loss": 0.0602,
398
+ "step": 650
399
+ },
400
+ {
401
+ "epoch": 0.3,
402
+ "learning_rate": 1.8276923076923078e-05,
403
+ "loss": 0.0148,
404
+ "step": 660
405
+ },
406
+ {
407
+ "epoch": 0.3,
408
+ "learning_rate": 1.8246153846153845e-05,
409
+ "loss": 0.0145,
410
+ "step": 670
411
+ },
412
+ {
413
+ "epoch": 0.31,
414
+ "learning_rate": 1.8215384615384616e-05,
415
+ "loss": 0.0137,
416
+ "step": 680
417
+ },
418
+ {
419
+ "epoch": 0.31,
420
+ "learning_rate": 1.8184615384615384e-05,
421
+ "loss": 0.1912,
422
+ "step": 690
423
+ },
424
+ {
425
+ "epoch": 0.32,
426
+ "learning_rate": 1.8153846153846155e-05,
427
+ "loss": 0.088,
428
+ "step": 700
429
+ },
430
+ {
431
+ "epoch": 0.32,
432
+ "learning_rate": 1.8123076923076926e-05,
433
+ "loss": 0.1182,
434
+ "step": 710
435
+ },
436
+ {
437
+ "epoch": 0.33,
438
+ "learning_rate": 1.8092307692307694e-05,
439
+ "loss": 0.1519,
440
+ "step": 720
441
+ },
442
+ {
443
+ "epoch": 0.33,
444
+ "learning_rate": 1.8061538461538465e-05,
445
+ "loss": 0.0122,
446
+ "step": 730
447
+ },
448
+ {
449
+ "epoch": 0.34,
450
+ "learning_rate": 1.8030769230769233e-05,
451
+ "loss": 0.0829,
452
+ "step": 740
453
+ },
454
+ {
455
+ "epoch": 0.34,
456
+ "learning_rate": 1.8e-05,
457
+ "loss": 0.0192,
458
+ "step": 750
459
+ },
460
+ {
461
+ "epoch": 0.35,
462
+ "learning_rate": 1.796923076923077e-05,
463
+ "loss": 0.0399,
464
+ "step": 760
465
+ },
466
+ {
467
+ "epoch": 0.35,
468
+ "learning_rate": 1.793846153846154e-05,
469
+ "loss": 0.1442,
470
+ "step": 770
471
+ },
472
+ {
473
+ "epoch": 0.35,
474
+ "learning_rate": 1.7907692307692307e-05,
475
+ "loss": 0.2089,
476
+ "step": 780
477
+ },
478
+ {
479
+ "epoch": 0.36,
480
+ "learning_rate": 1.7876923076923078e-05,
481
+ "loss": 0.141,
482
+ "step": 790
483
+ },
484
+ {
485
+ "epoch": 0.36,
486
+ "learning_rate": 1.784615384615385e-05,
487
+ "loss": 0.0447,
488
+ "step": 800
489
+ },
490
+ {
491
+ "epoch": 0.37,
492
+ "learning_rate": 1.7815384615384617e-05,
493
+ "loss": 0.0111,
494
+ "step": 810
495
+ },
496
+ {
497
+ "epoch": 0.37,
498
+ "learning_rate": 1.7784615384615388e-05,
499
+ "loss": 0.262,
500
+ "step": 820
501
+ },
502
+ {
503
+ "epoch": 0.38,
504
+ "learning_rate": 1.7753846153846156e-05,
505
+ "loss": 0.2721,
506
+ "step": 830
507
+ },
508
+ {
509
+ "epoch": 0.38,
510
+ "learning_rate": 1.7723076923076927e-05,
511
+ "loss": 0.0332,
512
+ "step": 840
513
+ },
514
+ {
515
+ "epoch": 0.39,
516
+ "learning_rate": 1.7692307692307694e-05,
517
+ "loss": 0.0683,
518
+ "step": 850
519
+ },
520
+ {
521
+ "epoch": 0.39,
522
+ "learning_rate": 1.7661538461538462e-05,
523
+ "loss": 0.0987,
524
+ "step": 860
525
+ },
526
+ {
527
+ "epoch": 0.4,
528
+ "learning_rate": 1.763076923076923e-05,
529
+ "loss": 0.1833,
530
+ "step": 870
531
+ },
532
+ {
533
+ "epoch": 0.4,
534
+ "learning_rate": 1.76e-05,
535
+ "loss": 0.0347,
536
+ "step": 880
537
+ },
538
+ {
539
+ "epoch": 0.4,
540
+ "learning_rate": 1.7569230769230772e-05,
541
+ "loss": 0.0425,
542
+ "step": 890
543
+ },
544
+ {
545
+ "epoch": 0.41,
546
+ "learning_rate": 1.753846153846154e-05,
547
+ "loss": 0.0615,
548
+ "step": 900
549
+ },
550
+ {
551
+ "epoch": 0.41,
552
+ "learning_rate": 1.750769230769231e-05,
553
+ "loss": 0.1378,
554
+ "step": 910
555
+ },
556
+ {
557
+ "epoch": 0.42,
558
+ "learning_rate": 1.747692307692308e-05,
559
+ "loss": 0.1481,
560
+ "step": 920
561
+ },
562
+ {
563
+ "epoch": 0.42,
564
+ "learning_rate": 1.744615384615385e-05,
565
+ "loss": 0.0604,
566
+ "step": 930
567
+ },
568
+ {
569
+ "epoch": 0.43,
570
+ "learning_rate": 1.7415384615384617e-05,
571
+ "loss": 0.0099,
572
+ "step": 940
573
+ },
574
+ {
575
+ "epoch": 0.43,
576
+ "learning_rate": 1.7384615384615385e-05,
577
+ "loss": 0.0448,
578
+ "step": 950
579
+ },
580
+ {
581
+ "epoch": 0.44,
582
+ "learning_rate": 1.7353846153846156e-05,
583
+ "loss": 0.0093,
584
+ "step": 960
585
+ },
586
+ {
587
+ "epoch": 0.44,
588
+ "learning_rate": 1.7323076923076924e-05,
589
+ "loss": 0.0078,
590
+ "step": 970
591
+ },
592
+ {
593
+ "epoch": 0.45,
594
+ "learning_rate": 1.729230769230769e-05,
595
+ "loss": 0.0731,
596
+ "step": 980
597
+ },
598
+ {
599
+ "epoch": 0.45,
600
+ "learning_rate": 1.7261538461538463e-05,
601
+ "loss": 0.1749,
602
+ "step": 990
603
+ },
604
+ {
605
+ "epoch": 0.45,
606
+ "learning_rate": 1.7230769230769234e-05,
607
+ "loss": 0.0954,
608
+ "step": 1000
609
+ },
610
+ {
611
+ "epoch": 0.46,
612
+ "learning_rate": 1.72e-05,
613
+ "loss": 0.009,
614
+ "step": 1010
615
+ },
616
+ {
617
+ "epoch": 0.46,
618
+ "learning_rate": 1.7169230769230772e-05,
619
+ "loss": 0.0986,
620
+ "step": 1020
621
+ },
622
+ {
623
+ "epoch": 0.47,
624
+ "learning_rate": 1.713846153846154e-05,
625
+ "loss": 0.1027,
626
+ "step": 1030
627
+ },
628
+ {
629
+ "epoch": 0.47,
630
+ "learning_rate": 1.710769230769231e-05,
631
+ "loss": 0.008,
632
+ "step": 1040
633
+ },
634
+ {
635
+ "epoch": 0.48,
636
+ "learning_rate": 1.707692307692308e-05,
637
+ "loss": 0.009,
638
+ "step": 1050
639
+ },
640
+ {
641
+ "epoch": 0.48,
642
+ "learning_rate": 1.7046153846153847e-05,
643
+ "loss": 0.0071,
644
+ "step": 1060
645
+ },
646
+ {
647
+ "epoch": 0.49,
648
+ "learning_rate": 1.7015384615384614e-05,
649
+ "loss": 0.1062,
650
+ "step": 1070
651
+ },
652
+ {
653
+ "epoch": 0.49,
654
+ "learning_rate": 1.6984615384615385e-05,
655
+ "loss": 0.0069,
656
+ "step": 1080
657
+ },
658
+ {
659
+ "epoch": 0.5,
660
+ "learning_rate": 1.6953846153846156e-05,
661
+ "loss": 0.0069,
662
+ "step": 1090
663
+ },
664
+ {
665
+ "epoch": 0.5,
666
+ "learning_rate": 1.6923076923076924e-05,
667
+ "loss": 0.1227,
668
+ "step": 1100
669
+ },
670
+ {
671
+ "epoch": 0.5,
672
+ "learning_rate": 1.6892307692307695e-05,
673
+ "loss": 0.0422,
674
+ "step": 1110
675
+ },
676
+ {
677
+ "epoch": 0.51,
678
+ "learning_rate": 1.6861538461538463e-05,
679
+ "loss": 0.1685,
680
+ "step": 1120
681
+ },
682
+ {
683
+ "epoch": 0.51,
684
+ "learning_rate": 1.6830769230769234e-05,
685
+ "loss": 0.0902,
686
+ "step": 1130
687
+ },
688
+ {
689
+ "epoch": 0.52,
690
+ "learning_rate": 1.6800000000000002e-05,
691
+ "loss": 0.1685,
692
+ "step": 1140
693
+ },
694
+ {
695
+ "epoch": 0.52,
696
+ "learning_rate": 1.676923076923077e-05,
697
+ "loss": 0.0065,
698
+ "step": 1150
699
+ },
700
+ {
701
+ "epoch": 0.53,
702
+ "learning_rate": 1.673846153846154e-05,
703
+ "loss": 0.0173,
704
+ "step": 1160
705
+ },
706
+ {
707
+ "epoch": 0.53,
708
+ "learning_rate": 1.6707692307692308e-05,
709
+ "loss": 0.157,
710
+ "step": 1170
711
+ },
712
+ {
713
+ "epoch": 0.54,
714
+ "learning_rate": 1.6676923076923076e-05,
715
+ "loss": 0.0081,
716
+ "step": 1180
717
+ },
718
+ {
719
+ "epoch": 0.54,
720
+ "learning_rate": 1.6646153846153847e-05,
721
+ "loss": 0.1193,
722
+ "step": 1190
723
+ },
724
+ {
725
+ "epoch": 0.55,
726
+ "learning_rate": 1.6615384615384618e-05,
727
+ "loss": 0.0083,
728
+ "step": 1200
729
+ },
730
+ {
731
+ "epoch": 0.55,
732
+ "learning_rate": 1.6584615384615386e-05,
733
+ "loss": 0.0076,
734
+ "step": 1210
735
+ },
736
+ {
737
+ "epoch": 0.55,
738
+ "learning_rate": 1.6553846153846157e-05,
739
+ "loss": 0.007,
740
+ "step": 1220
741
+ },
742
+ {
743
+ "epoch": 0.56,
744
+ "learning_rate": 1.6523076923076925e-05,
745
+ "loss": 0.0059,
746
+ "step": 1230
747
+ },
748
+ {
749
+ "epoch": 0.56,
750
+ "learning_rate": 1.6492307692307696e-05,
751
+ "loss": 0.0802,
752
+ "step": 1240
753
+ },
754
+ {
755
+ "epoch": 0.57,
756
+ "learning_rate": 1.6461538461538463e-05,
757
+ "loss": 0.1481,
758
+ "step": 1250
759
+ },
760
+ {
761
+ "epoch": 0.57,
762
+ "learning_rate": 1.643076923076923e-05,
763
+ "loss": 0.0699,
764
+ "step": 1260
765
+ },
766
+ {
767
+ "epoch": 0.58,
768
+ "learning_rate": 1.64e-05,
769
+ "loss": 0.007,
770
+ "step": 1270
771
+ },
772
+ {
773
+ "epoch": 0.58,
774
+ "learning_rate": 1.636923076923077e-05,
775
+ "loss": 0.008,
776
+ "step": 1280
777
+ },
778
+ {
779
+ "epoch": 0.59,
780
+ "learning_rate": 1.633846153846154e-05,
781
+ "loss": 0.0059,
782
+ "step": 1290
783
+ },
784
+ {
785
+ "epoch": 0.59,
786
+ "learning_rate": 1.630769230769231e-05,
787
+ "loss": 0.0725,
788
+ "step": 1300
789
+ },
790
+ {
791
+ "epoch": 0.6,
792
+ "learning_rate": 1.627692307692308e-05,
793
+ "loss": 0.2037,
794
+ "step": 1310
795
+ },
796
+ {
797
+ "epoch": 0.6,
798
+ "learning_rate": 1.6246153846153848e-05,
799
+ "loss": 0.1485,
800
+ "step": 1320
801
+ },
802
+ {
803
+ "epoch": 0.6,
804
+ "learning_rate": 1.621538461538462e-05,
805
+ "loss": 0.0085,
806
+ "step": 1330
807
+ },
808
+ {
809
+ "epoch": 0.61,
810
+ "learning_rate": 1.6184615384615386e-05,
811
+ "loss": 0.0591,
812
+ "step": 1340
813
+ },
814
+ {
815
+ "epoch": 0.61,
816
+ "learning_rate": 1.6153846153846154e-05,
817
+ "loss": 0.0824,
818
+ "step": 1350
819
+ },
820
+ {
821
+ "epoch": 0.62,
822
+ "learning_rate": 1.6123076923076925e-05,
823
+ "loss": 0.0402,
824
+ "step": 1360
825
+ },
826
+ {
827
+ "epoch": 0.62,
828
+ "learning_rate": 1.6092307692307693e-05,
829
+ "loss": 0.005,
830
+ "step": 1370
831
+ },
832
+ {
833
+ "epoch": 0.63,
834
+ "learning_rate": 1.606153846153846e-05,
835
+ "loss": 0.0052,
836
+ "step": 1380
837
+ },
838
+ {
839
+ "epoch": 0.63,
840
+ "learning_rate": 1.603076923076923e-05,
841
+ "loss": 0.005,
842
+ "step": 1390
843
+ },
844
+ {
845
+ "epoch": 0.64,
846
+ "learning_rate": 1.6000000000000003e-05,
847
+ "loss": 0.0838,
848
+ "step": 1400
849
+ },
850
+ {
851
+ "epoch": 0.64,
852
+ "learning_rate": 1.596923076923077e-05,
853
+ "loss": 0.0052,
854
+ "step": 1410
855
+ },
856
+ {
857
+ "epoch": 0.65,
858
+ "learning_rate": 1.593846153846154e-05,
859
+ "loss": 0.0244,
860
+ "step": 1420
861
+ },
862
+ {
863
+ "epoch": 0.65,
864
+ "learning_rate": 1.590769230769231e-05,
865
+ "loss": 0.1059,
866
+ "step": 1430
867
+ },
868
+ {
869
+ "epoch": 0.65,
870
+ "learning_rate": 1.587692307692308e-05,
871
+ "loss": 0.1774,
872
+ "step": 1440
873
+ },
874
+ {
875
+ "epoch": 0.66,
876
+ "learning_rate": 1.5846153846153848e-05,
877
+ "loss": 0.1231,
878
+ "step": 1450
879
+ },
880
+ {
881
+ "epoch": 0.66,
882
+ "learning_rate": 1.5815384615384616e-05,
883
+ "loss": 0.0072,
884
+ "step": 1460
885
+ },
886
+ {
887
+ "epoch": 0.67,
888
+ "learning_rate": 1.5784615384615383e-05,
889
+ "loss": 0.0059,
890
+ "step": 1470
891
+ },
892
+ {
893
+ "epoch": 0.67,
894
+ "learning_rate": 1.5753846153846154e-05,
895
+ "loss": 0.0609,
896
+ "step": 1480
897
+ },
898
+ {
899
+ "epoch": 0.68,
900
+ "learning_rate": 1.5723076923076926e-05,
901
+ "loss": 0.2579,
902
+ "step": 1490
903
+ },
904
+ {
905
+ "epoch": 0.68,
906
+ "learning_rate": 1.5692307692307693e-05,
907
+ "loss": 0.0719,
908
+ "step": 1500
909
+ },
910
+ {
911
+ "epoch": 0.69,
912
+ "learning_rate": 1.5661538461538464e-05,
913
+ "loss": 0.0698,
914
+ "step": 1510
915
+ },
916
+ {
917
+ "epoch": 0.69,
918
+ "learning_rate": 1.5630769230769232e-05,
919
+ "loss": 0.0048,
920
+ "step": 1520
921
+ },
922
+ {
923
+ "epoch": 0.7,
924
+ "learning_rate": 1.5600000000000003e-05,
925
+ "loss": 0.0463,
926
+ "step": 1530
927
+ },
928
+ {
929
+ "epoch": 0.7,
930
+ "learning_rate": 1.556923076923077e-05,
931
+ "loss": 0.0206,
932
+ "step": 1540
933
+ },
934
+ {
935
+ "epoch": 0.7,
936
+ "learning_rate": 1.553846153846154e-05,
937
+ "loss": 0.0049,
938
+ "step": 1550
939
+ },
940
+ {
941
+ "epoch": 0.71,
942
+ "learning_rate": 1.550769230769231e-05,
943
+ "loss": 0.0749,
944
+ "step": 1560
945
+ },
946
+ {
947
+ "epoch": 0.71,
948
+ "learning_rate": 1.5476923076923077e-05,
949
+ "loss": 0.0638,
950
+ "step": 1570
951
+ },
952
+ {
953
+ "epoch": 0.72,
954
+ "learning_rate": 1.544615384615385e-05,
955
+ "loss": 0.0781,
956
+ "step": 1580
957
+ },
958
+ {
959
+ "epoch": 0.72,
960
+ "learning_rate": 1.5415384615384616e-05,
961
+ "loss": 0.0125,
962
+ "step": 1590
963
+ },
964
+ {
965
+ "epoch": 0.73,
966
+ "learning_rate": 1.5384615384615387e-05,
967
+ "loss": 0.0055,
968
+ "step": 1600
969
+ },
970
+ {
971
+ "epoch": 0.73,
972
+ "learning_rate": 1.5353846153846155e-05,
973
+ "loss": 0.0053,
974
+ "step": 1610
975
+ },
976
+ {
977
+ "epoch": 0.74,
978
+ "learning_rate": 1.5323076923076926e-05,
979
+ "loss": 0.2656,
980
+ "step": 1620
981
+ },
982
+ {
983
+ "epoch": 0.74,
984
+ "learning_rate": 1.5292307692307694e-05,
985
+ "loss": 0.0034,
986
+ "step": 1630
987
+ },
988
+ {
989
+ "epoch": 0.75,
990
+ "learning_rate": 1.5261538461538465e-05,
991
+ "loss": 0.1076,
992
+ "step": 1640
993
+ },
994
+ {
995
+ "epoch": 0.75,
996
+ "learning_rate": 1.523076923076923e-05,
997
+ "loss": 0.0046,
998
+ "step": 1650
999
+ },
1000
+ {
1001
+ "epoch": 0.75,
1002
+ "learning_rate": 1.5200000000000002e-05,
1003
+ "loss": 0.3507,
1004
+ "step": 1660
1005
+ },
1006
+ {
1007
+ "epoch": 0.76,
1008
+ "learning_rate": 1.516923076923077e-05,
1009
+ "loss": 0.0046,
1010
+ "step": 1670
1011
+ },
1012
+ {
1013
+ "epoch": 0.76,
1014
+ "learning_rate": 1.5138461538461539e-05,
1015
+ "loss": 0.0044,
1016
+ "step": 1680
1017
+ },
1018
+ {
1019
+ "epoch": 0.77,
1020
+ "learning_rate": 1.510769230769231e-05,
1021
+ "loss": 0.0059,
1022
+ "step": 1690
1023
+ },
1024
+ {
1025
+ "epoch": 0.77,
1026
+ "learning_rate": 1.5076923076923078e-05,
1027
+ "loss": 0.0424,
1028
+ "step": 1700
1029
+ },
1030
+ {
1031
+ "epoch": 0.78,
1032
+ "learning_rate": 1.5046153846153849e-05,
1033
+ "loss": 0.0035,
1034
+ "step": 1710
1035
+ },
1036
+ {
1037
+ "epoch": 0.78,
1038
+ "learning_rate": 1.5015384615384617e-05,
1039
+ "loss": 0.1403,
1040
+ "step": 1720
1041
+ },
1042
+ {
1043
+ "epoch": 0.79,
1044
+ "learning_rate": 1.4984615384615386e-05,
1045
+ "loss": 0.1405,
1046
+ "step": 1730
1047
+ },
1048
+ {
1049
+ "epoch": 0.79,
1050
+ "learning_rate": 1.4953846153846154e-05,
1051
+ "loss": 0.0236,
1052
+ "step": 1740
1053
+ },
1054
+ {
1055
+ "epoch": 0.8,
1056
+ "learning_rate": 1.4923076923076925e-05,
1057
+ "loss": 0.1671,
1058
+ "step": 1750
1059
+ },
1060
+ {
1061
+ "epoch": 0.8,
1062
+ "learning_rate": 1.4892307692307692e-05,
1063
+ "loss": 0.0038,
1064
+ "step": 1760
1065
+ },
1066
+ {
1067
+ "epoch": 0.8,
1068
+ "learning_rate": 1.4861538461538464e-05,
1069
+ "loss": 0.0037,
1070
+ "step": 1770
1071
+ },
1072
+ {
1073
+ "epoch": 0.81,
1074
+ "learning_rate": 1.4830769230769233e-05,
1075
+ "loss": 0.0388,
1076
+ "step": 1780
1077
+ },
1078
+ {
1079
+ "epoch": 0.81,
1080
+ "learning_rate": 1.48e-05,
1081
+ "loss": 0.1272,
1082
+ "step": 1790
1083
+ },
1084
+ {
1085
+ "epoch": 0.82,
1086
+ "learning_rate": 1.4769230769230772e-05,
1087
+ "loss": 0.0034,
1088
+ "step": 1800
1089
+ },
1090
+ {
1091
+ "epoch": 0.82,
1092
+ "learning_rate": 1.473846153846154e-05,
1093
+ "loss": 0.045,
1094
+ "step": 1810
1095
+ },
1096
+ {
1097
+ "epoch": 0.83,
1098
+ "learning_rate": 1.4707692307692309e-05,
1099
+ "loss": 0.1702,
1100
+ "step": 1820
1101
+ },
1102
+ {
1103
+ "epoch": 0.83,
1104
+ "learning_rate": 1.4676923076923078e-05,
1105
+ "loss": 0.1103,
1106
+ "step": 1830
1107
+ },
1108
+ {
1109
+ "epoch": 0.84,
1110
+ "learning_rate": 1.4646153846153848e-05,
1111
+ "loss": 0.0579,
1112
+ "step": 1840
1113
+ },
1114
+ {
1115
+ "epoch": 0.84,
1116
+ "learning_rate": 1.4615384615384615e-05,
1117
+ "loss": 0.0034,
1118
+ "step": 1850
1119
+ },
1120
+ {
1121
+ "epoch": 0.85,
1122
+ "learning_rate": 1.4584615384615386e-05,
1123
+ "loss": 0.1787,
1124
+ "step": 1860
1125
+ },
1126
+ {
1127
+ "epoch": 0.85,
1128
+ "learning_rate": 1.4553846153846154e-05,
1129
+ "loss": 0.0039,
1130
+ "step": 1870
1131
+ },
1132
+ {
1133
+ "epoch": 0.85,
1134
+ "learning_rate": 1.4523076923076923e-05,
1135
+ "loss": 0.0044,
1136
+ "step": 1880
1137
+ },
1138
+ {
1139
+ "epoch": 0.86,
1140
+ "learning_rate": 1.4492307692307695e-05,
1141
+ "loss": 0.172,
1142
+ "step": 1890
1143
+ },
1144
+ {
1145
+ "epoch": 0.86,
1146
+ "learning_rate": 1.4461538461538462e-05,
1147
+ "loss": 0.0036,
1148
+ "step": 1900
1149
+ },
1150
+ {
1151
+ "epoch": 0.87,
1152
+ "learning_rate": 1.4430769230769233e-05,
1153
+ "loss": 0.1737,
1154
+ "step": 1910
1155
+ },
1156
+ {
1157
+ "epoch": 0.87,
1158
+ "learning_rate": 1.4400000000000001e-05,
1159
+ "loss": 0.0028,
1160
+ "step": 1920
1161
+ },
1162
+ {
1163
+ "epoch": 0.88,
1164
+ "learning_rate": 1.436923076923077e-05,
1165
+ "loss": 0.2204,
1166
+ "step": 1930
1167
+ },
1168
+ {
1169
+ "epoch": 0.88,
1170
+ "learning_rate": 1.4338461538461538e-05,
1171
+ "loss": 0.0989,
1172
+ "step": 1940
1173
+ },
1174
+ {
1175
+ "epoch": 0.89,
1176
+ "learning_rate": 1.430769230769231e-05,
1177
+ "loss": 0.0718,
1178
+ "step": 1950
1179
+ },
1180
+ {
1181
+ "epoch": 0.89,
1182
+ "learning_rate": 1.4276923076923077e-05,
1183
+ "loss": 0.0097,
1184
+ "step": 1960
1185
+ },
1186
+ {
1187
+ "epoch": 0.9,
1188
+ "learning_rate": 1.4246153846153848e-05,
1189
+ "loss": 0.0874,
1190
+ "step": 1970
1191
+ },
1192
+ {
1193
+ "epoch": 0.9,
1194
+ "learning_rate": 1.4215384615384617e-05,
1195
+ "loss": 0.0045,
1196
+ "step": 1980
1197
+ },
1198
+ {
1199
+ "epoch": 0.9,
1200
+ "learning_rate": 1.4184615384615385e-05,
1201
+ "loss": 0.0039,
1202
+ "step": 1990
1203
+ },
1204
+ {
1205
+ "epoch": 0.91,
1206
+ "learning_rate": 1.4153846153846156e-05,
1207
+ "loss": 0.0046,
1208
+ "step": 2000
1209
+ },
1210
+ {
1211
+ "epoch": 0.91,
1212
+ "learning_rate": 1.4123076923076924e-05,
1213
+ "loss": 0.1695,
1214
+ "step": 2010
1215
+ },
1216
+ {
1217
+ "epoch": 0.92,
1218
+ "learning_rate": 1.4092307692307693e-05,
1219
+ "loss": 0.0039,
1220
+ "step": 2020
1221
+ },
1222
+ {
1223
+ "epoch": 0.92,
1224
+ "learning_rate": 1.4061538461538463e-05,
1225
+ "loss": 0.004,
1226
+ "step": 2030
1227
+ },
1228
+ {
1229
+ "epoch": 0.93,
1230
+ "learning_rate": 1.4030769230769232e-05,
1231
+ "loss": 0.1985,
1232
+ "step": 2040
1233
+ },
1234
+ {
1235
+ "epoch": 0.93,
1236
+ "learning_rate": 1.4e-05,
1237
+ "loss": 0.146,
1238
+ "step": 2050
1239
+ },
1240
+ {
1241
+ "epoch": 0.94,
1242
+ "learning_rate": 1.3969230769230771e-05,
1243
+ "loss": 0.0027,
1244
+ "step": 2060
1245
+ },
1246
+ {
1247
+ "epoch": 0.94,
1248
+ "learning_rate": 1.393846153846154e-05,
1249
+ "loss": 0.1138,
1250
+ "step": 2070
1251
+ },
1252
+ {
1253
+ "epoch": 0.95,
1254
+ "learning_rate": 1.3907692307692308e-05,
1255
+ "loss": 0.0072,
1256
+ "step": 2080
1257
+ },
1258
+ {
1259
+ "epoch": 0.95,
1260
+ "learning_rate": 1.3876923076923079e-05,
1261
+ "loss": 0.0027,
1262
+ "step": 2090
1263
+ },
1264
+ {
1265
+ "epoch": 0.95,
1266
+ "learning_rate": 1.3846153846153847e-05,
1267
+ "loss": 0.003,
1268
+ "step": 2100
1269
+ },
1270
+ {
1271
+ "epoch": 0.96,
1272
+ "learning_rate": 1.3815384615384618e-05,
1273
+ "loss": 0.1025,
1274
+ "step": 2110
1275
+ },
1276
+ {
1277
+ "epoch": 0.96,
1278
+ "learning_rate": 1.3784615384615386e-05,
1279
+ "loss": 0.029,
1280
+ "step": 2120
1281
+ },
1282
+ {
1283
+ "epoch": 0.97,
1284
+ "learning_rate": 1.3753846153846155e-05,
1285
+ "loss": 0.0877,
1286
+ "step": 2130
1287
+ },
1288
+ {
1289
+ "epoch": 0.97,
1290
+ "learning_rate": 1.3723076923076923e-05,
1291
+ "loss": 0.0486,
1292
+ "step": 2140
1293
+ },
1294
+ {
1295
+ "epoch": 0.98,
1296
+ "learning_rate": 1.3692307692307694e-05,
1297
+ "loss": 0.0355,
1298
+ "step": 2150
1299
+ },
1300
+ {
1301
+ "epoch": 0.98,
1302
+ "learning_rate": 1.3661538461538461e-05,
1303
+ "loss": 0.2253,
1304
+ "step": 2160
1305
+ },
1306
+ {
1307
+ "epoch": 0.99,
1308
+ "learning_rate": 1.3630769230769233e-05,
1309
+ "loss": 0.1129,
1310
+ "step": 2170
1311
+ },
1312
+ {
1313
+ "epoch": 0.99,
1314
+ "learning_rate": 1.3600000000000002e-05,
1315
+ "loss": 0.0035,
1316
+ "step": 2180
1317
+ },
1318
+ {
1319
+ "epoch": 1.0,
1320
+ "learning_rate": 1.356923076923077e-05,
1321
+ "loss": 0.0881,
1322
+ "step": 2190
1323
+ },
1324
+ {
1325
+ "epoch": 1.0,
1326
+ "learning_rate": 1.353846153846154e-05,
1327
+ "loss": 0.0028,
1328
+ "step": 2200
1329
+ },
1330
+ {
1331
+ "epoch": 1.0,
1332
+ "eval_Accuracy": 0.9888636363636364,
1333
+ "eval_F1": 0.9890284902407037,
1334
+ "eval_loss": 0.08128699660301208,
1335
+ "eval_runtime": 150.6254,
1336
+ "eval_samples_per_second": 29.212,
1337
+ "eval_steps_per_second": 1.826,
1338
+ "step": 2200
1339
+ },
1340
+ {
1341
+ "epoch": 1.0,
1342
+ "learning_rate": 1.3507692307692308e-05,
1343
+ "loss": 0.0792,
1344
+ "step": 2210
1345
+ },
1346
+ {
1347
+ "epoch": 1.01,
1348
+ "learning_rate": 1.3476923076923078e-05,
1349
+ "loss": 0.1501,
1350
+ "step": 2220
1351
+ },
1352
+ {
1353
+ "epoch": 1.01,
1354
+ "learning_rate": 1.3446153846153847e-05,
1355
+ "loss": 0.1827,
1356
+ "step": 2230
1357
+ },
1358
+ {
1359
+ "epoch": 1.02,
1360
+ "learning_rate": 1.3415384615384617e-05,
1361
+ "loss": 0.0505,
1362
+ "step": 2240
1363
+ },
1364
+ {
1365
+ "epoch": 1.02,
1366
+ "learning_rate": 1.3384615384615384e-05,
1367
+ "loss": 0.0044,
1368
+ "step": 2250
1369
+ },
1370
+ {
1371
+ "epoch": 1.03,
1372
+ "learning_rate": 1.3353846153846155e-05,
1373
+ "loss": 0.113,
1374
+ "step": 2260
1375
+ },
1376
+ {
1377
+ "epoch": 1.03,
1378
+ "learning_rate": 1.3323076923076925e-05,
1379
+ "loss": 0.0928,
1380
+ "step": 2270
1381
+ },
1382
+ {
1383
+ "epoch": 1.04,
1384
+ "learning_rate": 1.3292307692307692e-05,
1385
+ "loss": 0.0496,
1386
+ "step": 2280
1387
+ },
1388
+ {
1389
+ "epoch": 1.04,
1390
+ "learning_rate": 1.3261538461538464e-05,
1391
+ "loss": 0.0048,
1392
+ "step": 2290
1393
+ },
1394
+ {
1395
+ "epoch": 1.05,
1396
+ "learning_rate": 1.3230769230769231e-05,
1397
+ "loss": 0.0026,
1398
+ "step": 2300
1399
+ },
1400
+ {
1401
+ "epoch": 1.05,
1402
+ "learning_rate": 1.3200000000000002e-05,
1403
+ "loss": 0.114,
1404
+ "step": 2310
1405
+ },
1406
+ {
1407
+ "epoch": 1.05,
1408
+ "learning_rate": 1.316923076923077e-05,
1409
+ "loss": 0.0367,
1410
+ "step": 2320
1411
+ },
1412
+ {
1413
+ "epoch": 1.06,
1414
+ "learning_rate": 1.313846153846154e-05,
1415
+ "loss": 0.0756,
1416
+ "step": 2330
1417
+ },
1418
+ {
1419
+ "epoch": 1.06,
1420
+ "learning_rate": 1.3107692307692307e-05,
1421
+ "loss": 0.0752,
1422
+ "step": 2340
1423
+ },
1424
+ {
1425
+ "epoch": 1.07,
1426
+ "learning_rate": 1.3076923076923078e-05,
1427
+ "loss": 0.149,
1428
+ "step": 2350
1429
+ },
1430
+ {
1431
+ "epoch": 1.07,
1432
+ "learning_rate": 1.3046153846153846e-05,
1433
+ "loss": 0.0047,
1434
+ "step": 2360
1435
+ },
1436
+ {
1437
+ "epoch": 1.08,
1438
+ "learning_rate": 1.3015384615384617e-05,
1439
+ "loss": 0.0041,
1440
+ "step": 2370
1441
+ },
1442
+ {
1443
+ "epoch": 1.08,
1444
+ "learning_rate": 1.2984615384615386e-05,
1445
+ "loss": 0.0028,
1446
+ "step": 2380
1447
+ },
1448
+ {
1449
+ "epoch": 1.09,
1450
+ "learning_rate": 1.2953846153846154e-05,
1451
+ "loss": 0.0023,
1452
+ "step": 2390
1453
+ },
1454
+ {
1455
+ "epoch": 1.09,
1456
+ "learning_rate": 1.2923076923076925e-05,
1457
+ "loss": 0.0027,
1458
+ "step": 2400
1459
+ },
1460
+ {
1461
+ "epoch": 1.1,
1462
+ "learning_rate": 1.2892307692307693e-05,
1463
+ "loss": 0.0023,
1464
+ "step": 2410
1465
+ },
1466
+ {
1467
+ "epoch": 1.1,
1468
+ "learning_rate": 1.2861538461538462e-05,
1469
+ "loss": 0.0555,
1470
+ "step": 2420
1471
+ },
1472
+ {
1473
+ "epoch": 1.1,
1474
+ "learning_rate": 1.2830769230769232e-05,
1475
+ "loss": 0.002,
1476
+ "step": 2430
1477
+ },
1478
+ {
1479
+ "epoch": 1.11,
1480
+ "learning_rate": 1.2800000000000001e-05,
1481
+ "loss": 0.0021,
1482
+ "step": 2440
1483
+ },
1484
+ {
1485
+ "epoch": 1.11,
1486
+ "learning_rate": 1.2769230769230769e-05,
1487
+ "loss": 0.0511,
1488
+ "step": 2450
1489
+ },
1490
+ {
1491
+ "epoch": 1.12,
1492
+ "learning_rate": 1.273846153846154e-05,
1493
+ "loss": 0.0022,
1494
+ "step": 2460
1495
+ },
1496
+ {
1497
+ "epoch": 1.12,
1498
+ "learning_rate": 1.270769230769231e-05,
1499
+ "loss": 0.0026,
1500
+ "step": 2470
1501
+ },
1502
+ {
1503
+ "epoch": 1.13,
1504
+ "learning_rate": 1.2676923076923077e-05,
1505
+ "loss": 0.0976,
1506
+ "step": 2480
1507
+ },
1508
+ {
1509
+ "epoch": 1.13,
1510
+ "learning_rate": 1.2646153846153848e-05,
1511
+ "loss": 0.23,
1512
+ "step": 2490
1513
+ },
1514
+ {
1515
+ "epoch": 1.14,
1516
+ "learning_rate": 1.2615384615384616e-05,
1517
+ "loss": 0.1016,
1518
+ "step": 2500
1519
+ },
1520
+ {
1521
+ "epoch": 1.14,
1522
+ "learning_rate": 1.2584615384615387e-05,
1523
+ "loss": 0.2139,
1524
+ "step": 2510
1525
+ },
1526
+ {
1527
+ "epoch": 1.15,
1528
+ "learning_rate": 1.2553846153846155e-05,
1529
+ "loss": 0.2166,
1530
+ "step": 2520
1531
+ },
1532
+ {
1533
+ "epoch": 1.15,
1534
+ "learning_rate": 1.2523076923076924e-05,
1535
+ "loss": 0.0542,
1536
+ "step": 2530
1537
+ },
1538
+ {
1539
+ "epoch": 1.15,
1540
+ "learning_rate": 1.2492307692307692e-05,
1541
+ "loss": 0.0533,
1542
+ "step": 2540
1543
+ },
1544
+ {
1545
+ "epoch": 1.16,
1546
+ "learning_rate": 1.2461538461538463e-05,
1547
+ "loss": 0.057,
1548
+ "step": 2550
1549
+ },
1550
+ {
1551
+ "epoch": 1.16,
1552
+ "learning_rate": 1.243076923076923e-05,
1553
+ "loss": 0.0036,
1554
+ "step": 2560
1555
+ },
1556
+ {
1557
+ "epoch": 1.17,
1558
+ "learning_rate": 1.2400000000000002e-05,
1559
+ "loss": 0.0039,
1560
+ "step": 2570
1561
+ },
1562
+ {
1563
+ "epoch": 1.17,
1564
+ "learning_rate": 1.2369230769230771e-05,
1565
+ "loss": 0.003,
1566
+ "step": 2580
1567
+ },
1568
+ {
1569
+ "epoch": 1.18,
1570
+ "learning_rate": 1.2338461538461539e-05,
1571
+ "loss": 0.0024,
1572
+ "step": 2590
1573
+ },
1574
+ {
1575
+ "epoch": 1.18,
1576
+ "learning_rate": 1.230769230769231e-05,
1577
+ "loss": 0.0022,
1578
+ "step": 2600
1579
+ },
1580
+ {
1581
+ "epoch": 1.19,
1582
+ "learning_rate": 1.2276923076923077e-05,
1583
+ "loss": 0.0967,
1584
+ "step": 2610
1585
+ },
1586
+ {
1587
+ "epoch": 1.19,
1588
+ "learning_rate": 1.2246153846153847e-05,
1589
+ "loss": 0.0695,
1590
+ "step": 2620
1591
+ },
1592
+ {
1593
+ "epoch": 1.2,
1594
+ "learning_rate": 1.2215384615384616e-05,
1595
+ "loss": 0.0028,
1596
+ "step": 2630
1597
+ },
1598
+ {
1599
+ "epoch": 1.2,
1600
+ "learning_rate": 1.2184615384615386e-05,
1601
+ "loss": 0.0233,
1602
+ "step": 2640
1603
+ },
1604
+ {
1605
+ "epoch": 1.2,
1606
+ "learning_rate": 1.2153846153846153e-05,
1607
+ "loss": 0.0908,
1608
+ "step": 2650
1609
+ },
1610
+ {
1611
+ "epoch": 1.21,
1612
+ "learning_rate": 1.2123076923076924e-05,
1613
+ "loss": 0.0058,
1614
+ "step": 2660
1615
+ },
1616
+ {
1617
+ "epoch": 1.21,
1618
+ "learning_rate": 1.2092307692307694e-05,
1619
+ "loss": 0.0031,
1620
+ "step": 2670
1621
+ },
1622
+ {
1623
+ "epoch": 1.22,
1624
+ "learning_rate": 1.2061538461538462e-05,
1625
+ "loss": 0.1291,
1626
+ "step": 2680
1627
+ },
1628
+ {
1629
+ "epoch": 1.22,
1630
+ "learning_rate": 1.2030769230769233e-05,
1631
+ "loss": 0.1178,
1632
+ "step": 2690
1633
+ },
1634
+ {
1635
+ "epoch": 1.23,
1636
+ "learning_rate": 1.2e-05,
1637
+ "loss": 0.0052,
1638
+ "step": 2700
1639
+ },
1640
+ {
1641
+ "epoch": 1.23,
1642
+ "learning_rate": 1.1969230769230771e-05,
1643
+ "loss": 0.0385,
1644
+ "step": 2710
1645
+ },
1646
+ {
1647
+ "epoch": 1.24,
1648
+ "learning_rate": 1.1938461538461539e-05,
1649
+ "loss": 0.1806,
1650
+ "step": 2720
1651
+ },
1652
+ {
1653
+ "epoch": 1.24,
1654
+ "learning_rate": 1.1907692307692308e-05,
1655
+ "loss": 0.0217,
1656
+ "step": 2730
1657
+ },
1658
+ {
1659
+ "epoch": 1.25,
1660
+ "learning_rate": 1.1876923076923076e-05,
1661
+ "loss": 0.0024,
1662
+ "step": 2740
1663
+ },
1664
+ {
1665
+ "epoch": 1.25,
1666
+ "learning_rate": 1.1846153846153847e-05,
1667
+ "loss": 0.0022,
1668
+ "step": 2750
1669
+ },
1670
+ {
1671
+ "epoch": 1.25,
1672
+ "learning_rate": 1.1815384615384617e-05,
1673
+ "loss": 0.0029,
1674
+ "step": 2760
1675
+ },
1676
+ {
1677
+ "epoch": 1.26,
1678
+ "learning_rate": 1.1784615384615386e-05,
1679
+ "loss": 0.1312,
1680
+ "step": 2770
1681
+ },
1682
+ {
1683
+ "epoch": 1.26,
1684
+ "learning_rate": 1.1753846153846155e-05,
1685
+ "loss": 0.0022,
1686
+ "step": 2780
1687
+ },
1688
+ {
1689
+ "epoch": 1.27,
1690
+ "learning_rate": 1.1723076923076923e-05,
1691
+ "loss": 0.0019,
1692
+ "step": 2790
1693
+ },
1694
+ {
1695
+ "epoch": 1.27,
1696
+ "learning_rate": 1.1692307692307694e-05,
1697
+ "loss": 0.197,
1698
+ "step": 2800
1699
+ },
1700
+ {
1701
+ "epoch": 1.28,
1702
+ "learning_rate": 1.1661538461538462e-05,
1703
+ "loss": 0.003,
1704
+ "step": 2810
1705
+ },
1706
+ {
1707
+ "epoch": 1.28,
1708
+ "learning_rate": 1.1630769230769231e-05,
1709
+ "loss": 0.0028,
1710
+ "step": 2820
1711
+ },
1712
+ {
1713
+ "epoch": 1.29,
1714
+ "learning_rate": 1.16e-05,
1715
+ "loss": 0.0024,
1716
+ "step": 2830
1717
+ },
1718
+ {
1719
+ "epoch": 1.29,
1720
+ "learning_rate": 1.156923076923077e-05,
1721
+ "loss": 0.002,
1722
+ "step": 2840
1723
+ },
1724
+ {
1725
+ "epoch": 1.3,
1726
+ "learning_rate": 1.1538461538461538e-05,
1727
+ "loss": 0.0021,
1728
+ "step": 2850
1729
+ },
1730
+ {
1731
+ "epoch": 1.3,
1732
+ "learning_rate": 1.1507692307692309e-05,
1733
+ "loss": 0.0564,
1734
+ "step": 2860
1735
+ },
1736
+ {
1737
+ "epoch": 1.3,
1738
+ "learning_rate": 1.1476923076923078e-05,
1739
+ "loss": 0.002,
1740
+ "step": 2870
1741
+ },
1742
+ {
1743
+ "epoch": 1.31,
1744
+ "learning_rate": 1.1446153846153846e-05,
1745
+ "loss": 0.0243,
1746
+ "step": 2880
1747
+ },
1748
+ {
1749
+ "epoch": 1.31,
1750
+ "learning_rate": 1.1415384615384617e-05,
1751
+ "loss": 0.0363,
1752
+ "step": 2890
1753
+ },
1754
+ {
1755
+ "epoch": 1.32,
1756
+ "learning_rate": 1.1384615384615385e-05,
1757
+ "loss": 0.1064,
1758
+ "step": 2900
1759
+ },
1760
+ {
1761
+ "epoch": 1.32,
1762
+ "learning_rate": 1.1353846153846156e-05,
1763
+ "loss": 0.0019,
1764
+ "step": 2910
1765
+ },
1766
+ {
1767
+ "epoch": 1.33,
1768
+ "learning_rate": 1.1323076923076924e-05,
1769
+ "loss": 0.0049,
1770
+ "step": 2920
1771
+ },
1772
+ {
1773
+ "epoch": 1.33,
1774
+ "learning_rate": 1.1292307692307693e-05,
1775
+ "loss": 0.0814,
1776
+ "step": 2930
1777
+ },
1778
+ {
1779
+ "epoch": 1.34,
1780
+ "learning_rate": 1.126153846153846e-05,
1781
+ "loss": 0.0839,
1782
+ "step": 2940
1783
+ },
1784
+ {
1785
+ "epoch": 1.34,
1786
+ "learning_rate": 1.1230769230769232e-05,
1787
+ "loss": 0.0028,
1788
+ "step": 2950
1789
+ },
1790
+ {
1791
+ "epoch": 1.35,
1792
+ "learning_rate": 1.1200000000000001e-05,
1793
+ "loss": 0.2127,
1794
+ "step": 2960
1795
+ },
1796
+ {
1797
+ "epoch": 1.35,
1798
+ "learning_rate": 1.116923076923077e-05,
1799
+ "loss": 0.0026,
1800
+ "step": 2970
1801
+ },
1802
+ {
1803
+ "epoch": 1.35,
1804
+ "learning_rate": 1.113846153846154e-05,
1805
+ "loss": 0.0368,
1806
+ "step": 2980
1807
+ },
1808
+ {
1809
+ "epoch": 1.36,
1810
+ "learning_rate": 1.1107692307692308e-05,
1811
+ "loss": 0.0468,
1812
+ "step": 2990
1813
+ },
1814
+ {
1815
+ "epoch": 1.36,
1816
+ "learning_rate": 1.1076923076923079e-05,
1817
+ "loss": 0.0019,
1818
+ "step": 3000
1819
+ },
1820
+ {
1821
+ "epoch": 1.37,
1822
+ "learning_rate": 1.1046153846153846e-05,
1823
+ "loss": 0.0017,
1824
+ "step": 3010
1825
+ },
1826
+ {
1827
+ "epoch": 1.37,
1828
+ "learning_rate": 1.1015384615384616e-05,
1829
+ "loss": 0.1788,
1830
+ "step": 3020
1831
+ },
1832
+ {
1833
+ "epoch": 1.38,
1834
+ "learning_rate": 1.0984615384615385e-05,
1835
+ "loss": 0.0705,
1836
+ "step": 3030
1837
+ },
1838
+ {
1839
+ "epoch": 1.38,
1840
+ "learning_rate": 1.0953846153846155e-05,
1841
+ "loss": 0.0524,
1842
+ "step": 3040
1843
+ },
1844
+ {
1845
+ "epoch": 1.39,
1846
+ "learning_rate": 1.0923076923076922e-05,
1847
+ "loss": 0.0035,
1848
+ "step": 3050
1849
+ },
1850
+ {
1851
+ "epoch": 1.39,
1852
+ "learning_rate": 1.0892307692307693e-05,
1853
+ "loss": 0.002,
1854
+ "step": 3060
1855
+ },
1856
+ {
1857
+ "epoch": 1.4,
1858
+ "learning_rate": 1.0861538461538463e-05,
1859
+ "loss": 0.0015,
1860
+ "step": 3070
1861
+ },
1862
+ {
1863
+ "epoch": 1.4,
1864
+ "learning_rate": 1.083076923076923e-05,
1865
+ "loss": 0.0017,
1866
+ "step": 3080
1867
+ },
1868
+ {
1869
+ "epoch": 1.4,
1870
+ "learning_rate": 1.0800000000000002e-05,
1871
+ "loss": 0.0015,
1872
+ "step": 3090
1873
+ },
1874
+ {
1875
+ "epoch": 1.41,
1876
+ "learning_rate": 1.076923076923077e-05,
1877
+ "loss": 0.0017,
1878
+ "step": 3100
1879
+ },
1880
+ {
1881
+ "epoch": 1.41,
1882
+ "learning_rate": 1.073846153846154e-05,
1883
+ "loss": 0.1966,
1884
+ "step": 3110
1885
+ },
1886
+ {
1887
+ "epoch": 1.42,
1888
+ "learning_rate": 1.0707692307692308e-05,
1889
+ "loss": 0.0829,
1890
+ "step": 3120
1891
+ },
1892
+ {
1893
+ "epoch": 1.42,
1894
+ "learning_rate": 1.0676923076923078e-05,
1895
+ "loss": 0.0827,
1896
+ "step": 3130
1897
+ },
1898
+ {
1899
+ "epoch": 1.43,
1900
+ "learning_rate": 1.0646153846153845e-05,
1901
+ "loss": 0.0049,
1902
+ "step": 3140
1903
+ },
1904
+ {
1905
+ "epoch": 1.43,
1906
+ "learning_rate": 1.0615384615384616e-05,
1907
+ "loss": 0.0194,
1908
+ "step": 3150
1909
+ },
1910
+ {
1911
+ "epoch": 1.44,
1912
+ "learning_rate": 1.0584615384615386e-05,
1913
+ "loss": 0.0026,
1914
+ "step": 3160
1915
+ },
1916
+ {
1917
+ "epoch": 1.44,
1918
+ "learning_rate": 1.0553846153846155e-05,
1919
+ "loss": 0.0679,
1920
+ "step": 3170
1921
+ },
1922
+ {
1923
+ "epoch": 1.45,
1924
+ "learning_rate": 1.0523076923076924e-05,
1925
+ "loss": 0.115,
1926
+ "step": 3180
1927
+ },
1928
+ {
1929
+ "epoch": 1.45,
1930
+ "learning_rate": 1.0492307692307692e-05,
1931
+ "loss": 0.0629,
1932
+ "step": 3190
1933
+ },
1934
+ {
1935
+ "epoch": 1.45,
1936
+ "learning_rate": 1.0461538461538463e-05,
1937
+ "loss": 0.0267,
1938
+ "step": 3200
1939
+ },
1940
+ {
1941
+ "epoch": 1.46,
1942
+ "learning_rate": 1.0430769230769231e-05,
1943
+ "loss": 0.1529,
1944
+ "step": 3210
1945
+ },
1946
+ {
1947
+ "epoch": 1.46,
1948
+ "learning_rate": 1.04e-05,
1949
+ "loss": 0.002,
1950
+ "step": 3220
1951
+ },
1952
+ {
1953
+ "epoch": 1.47,
1954
+ "learning_rate": 1.036923076923077e-05,
1955
+ "loss": 0.0018,
1956
+ "step": 3230
1957
+ },
1958
+ {
1959
+ "epoch": 1.47,
1960
+ "learning_rate": 1.033846153846154e-05,
1961
+ "loss": 0.002,
1962
+ "step": 3240
1963
+ },
1964
+ {
1965
+ "epoch": 1.48,
1966
+ "learning_rate": 1.0307692307692307e-05,
1967
+ "loss": 0.0348,
1968
+ "step": 3250
1969
+ },
1970
+ {
1971
+ "epoch": 1.48,
1972
+ "learning_rate": 1.0276923076923078e-05,
1973
+ "loss": 0.0019,
1974
+ "step": 3260
1975
+ },
1976
+ {
1977
+ "epoch": 1.49,
1978
+ "learning_rate": 1.0246153846153847e-05,
1979
+ "loss": 0.0017,
1980
+ "step": 3270
1981
+ },
1982
+ {
1983
+ "epoch": 1.49,
1984
+ "learning_rate": 1.0215384615384615e-05,
1985
+ "loss": 0.0688,
1986
+ "step": 3280
1987
+ },
1988
+ {
1989
+ "epoch": 1.5,
1990
+ "learning_rate": 1.0184615384615386e-05,
1991
+ "loss": 0.2175,
1992
+ "step": 3290
1993
+ },
1994
+ {
1995
+ "epoch": 1.5,
1996
+ "learning_rate": 1.0153846153846154e-05,
1997
+ "loss": 0.0031,
1998
+ "step": 3300
1999
+ },
2000
+ {
2001
+ "epoch": 1.5,
2002
+ "learning_rate": 1.0123076923076925e-05,
2003
+ "loss": 0.0017,
2004
+ "step": 3310
2005
+ },
2006
+ {
2007
+ "epoch": 1.51,
2008
+ "learning_rate": 1.0092307692307693e-05,
2009
+ "loss": 0.0024,
2010
+ "step": 3320
2011
+ },
2012
+ {
2013
+ "epoch": 1.51,
2014
+ "learning_rate": 1.0061538461538462e-05,
2015
+ "loss": 0.065,
2016
+ "step": 3330
2017
+ },
2018
+ {
2019
+ "epoch": 1.52,
2020
+ "learning_rate": 1.0030769230769231e-05,
2021
+ "loss": 0.0027,
2022
+ "step": 3340
2023
+ },
2024
+ {
2025
+ "epoch": 1.52,
2026
+ "learning_rate": 1e-05,
2027
+ "loss": 0.0389,
2028
+ "step": 3350
2029
+ },
2030
+ {
2031
+ "epoch": 1.53,
2032
+ "learning_rate": 9.96923076923077e-06,
2033
+ "loss": 0.1488,
2034
+ "step": 3360
2035
+ },
2036
+ {
2037
+ "epoch": 1.53,
2038
+ "learning_rate": 9.93846153846154e-06,
2039
+ "loss": 0.1107,
2040
+ "step": 3370
2041
+ },
2042
+ {
2043
+ "epoch": 1.54,
2044
+ "learning_rate": 9.907692307692309e-06,
2045
+ "loss": 0.2026,
2046
+ "step": 3380
2047
+ },
2048
+ {
2049
+ "epoch": 1.54,
2050
+ "learning_rate": 9.876923076923077e-06,
2051
+ "loss": 0.0154,
2052
+ "step": 3390
2053
+ },
2054
+ {
2055
+ "epoch": 1.55,
2056
+ "learning_rate": 9.846153846153848e-06,
2057
+ "loss": 0.0568,
2058
+ "step": 3400
2059
+ },
2060
+ {
2061
+ "epoch": 1.55,
2062
+ "learning_rate": 9.815384615384617e-06,
2063
+ "loss": 0.0015,
2064
+ "step": 3410
2065
+ },
2066
+ {
2067
+ "epoch": 1.55,
2068
+ "learning_rate": 9.784615384615387e-06,
2069
+ "loss": 0.1328,
2070
+ "step": 3420
2071
+ },
2072
+ {
2073
+ "epoch": 1.56,
2074
+ "learning_rate": 9.753846153846154e-06,
2075
+ "loss": 0.073,
2076
+ "step": 3430
2077
+ },
2078
+ {
2079
+ "epoch": 1.56,
2080
+ "learning_rate": 9.723076923076924e-06,
2081
+ "loss": 0.0713,
2082
+ "step": 3440
2083
+ },
2084
+ {
2085
+ "epoch": 1.57,
2086
+ "learning_rate": 9.692307692307693e-06,
2087
+ "loss": 0.0484,
2088
+ "step": 3450
2089
+ },
2090
+ {
2091
+ "epoch": 1.57,
2092
+ "learning_rate": 9.661538461538462e-06,
2093
+ "loss": 0.0703,
2094
+ "step": 3460
2095
+ },
2096
+ {
2097
+ "epoch": 1.58,
2098
+ "learning_rate": 9.630769230769232e-06,
2099
+ "loss": 0.1891,
2100
+ "step": 3470
2101
+ },
2102
+ {
2103
+ "epoch": 1.58,
2104
+ "learning_rate": 9.600000000000001e-06,
2105
+ "loss": 0.1082,
2106
+ "step": 3480
2107
+ },
2108
+ {
2109
+ "epoch": 1.59,
2110
+ "learning_rate": 9.569230769230769e-06,
2111
+ "loss": 0.0017,
2112
+ "step": 3490
2113
+ },
2114
+ {
2115
+ "epoch": 1.59,
2116
+ "learning_rate": 9.53846153846154e-06,
2117
+ "loss": 0.0096,
2118
+ "step": 3500
2119
+ },
2120
+ {
2121
+ "epoch": 1.6,
2122
+ "learning_rate": 9.50769230769231e-06,
2123
+ "loss": 0.022,
2124
+ "step": 3510
2125
+ },
2126
+ {
2127
+ "epoch": 1.6,
2128
+ "learning_rate": 9.476923076923079e-06,
2129
+ "loss": 0.0016,
2130
+ "step": 3520
2131
+ },
2132
+ {
2133
+ "epoch": 1.6,
2134
+ "learning_rate": 9.446153846153847e-06,
2135
+ "loss": 0.0013,
2136
+ "step": 3530
2137
+ },
2138
+ {
2139
+ "epoch": 1.61,
2140
+ "learning_rate": 9.415384615384616e-06,
2141
+ "loss": 0.0017,
2142
+ "step": 3540
2143
+ },
2144
+ {
2145
+ "epoch": 1.61,
2146
+ "learning_rate": 9.384615384615385e-06,
2147
+ "loss": 0.0016,
2148
+ "step": 3550
2149
+ },
2150
+ {
2151
+ "epoch": 1.62,
2152
+ "learning_rate": 9.353846153846155e-06,
2153
+ "loss": 0.0012,
2154
+ "step": 3560
2155
+ },
2156
+ {
2157
+ "epoch": 1.62,
2158
+ "learning_rate": 9.323076923076924e-06,
2159
+ "loss": 0.1051,
2160
+ "step": 3570
2161
+ },
2162
+ {
2163
+ "epoch": 1.63,
2164
+ "learning_rate": 9.292307692307694e-06,
2165
+ "loss": 0.0012,
2166
+ "step": 3580
2167
+ },
2168
+ {
2169
+ "epoch": 1.63,
2170
+ "learning_rate": 9.261538461538461e-06,
2171
+ "loss": 0.0013,
2172
+ "step": 3590
2173
+ },
2174
+ {
2175
+ "epoch": 1.64,
2176
+ "learning_rate": 9.230769230769232e-06,
2177
+ "loss": 0.0446,
2178
+ "step": 3600
2179
+ },
2180
+ {
2181
+ "epoch": 1.64,
2182
+ "learning_rate": 9.200000000000002e-06,
2183
+ "loss": 0.0728,
2184
+ "step": 3610
2185
+ },
2186
+ {
2187
+ "epoch": 1.65,
2188
+ "learning_rate": 9.169230769230771e-06,
2189
+ "loss": 0.0021,
2190
+ "step": 3620
2191
+ },
2192
+ {
2193
+ "epoch": 1.65,
2194
+ "learning_rate": 9.138461538461539e-06,
2195
+ "loss": 0.002,
2196
+ "step": 3630
2197
+ },
2198
+ {
2199
+ "epoch": 1.65,
2200
+ "learning_rate": 9.107692307692308e-06,
2201
+ "loss": 0.0564,
2202
+ "step": 3640
2203
+ },
2204
+ {
2205
+ "epoch": 1.66,
2206
+ "learning_rate": 9.076923076923078e-06,
2207
+ "loss": 0.0688,
2208
+ "step": 3650
2209
+ },
2210
+ {
2211
+ "epoch": 1.66,
2212
+ "learning_rate": 9.046153846153847e-06,
2213
+ "loss": 0.1502,
2214
+ "step": 3660
2215
+ },
2216
+ {
2217
+ "epoch": 1.67,
2218
+ "learning_rate": 9.015384615384616e-06,
2219
+ "loss": 0.0015,
2220
+ "step": 3670
2221
+ },
2222
+ {
2223
+ "epoch": 1.67,
2224
+ "learning_rate": 8.984615384615386e-06,
2225
+ "loss": 0.0627,
2226
+ "step": 3680
2227
+ },
2228
+ {
2229
+ "epoch": 1.68,
2230
+ "learning_rate": 8.953846153846153e-06,
2231
+ "loss": 0.002,
2232
+ "step": 3690
2233
+ },
2234
+ {
2235
+ "epoch": 1.68,
2236
+ "learning_rate": 8.923076923076925e-06,
2237
+ "loss": 0.0399,
2238
+ "step": 3700
2239
+ },
2240
+ {
2241
+ "epoch": 1.69,
2242
+ "learning_rate": 8.892307692307694e-06,
2243
+ "loss": 0.0023,
2244
+ "step": 3710
2245
+ },
2246
+ {
2247
+ "epoch": 1.69,
2248
+ "learning_rate": 8.861538461538463e-06,
2249
+ "loss": 0.0719,
2250
+ "step": 3720
2251
+ },
2252
+ {
2253
+ "epoch": 1.7,
2254
+ "learning_rate": 8.830769230769231e-06,
2255
+ "loss": 0.0019,
2256
+ "step": 3730
2257
+ },
2258
+ {
2259
+ "epoch": 1.7,
2260
+ "learning_rate": 8.8e-06,
2261
+ "loss": 0.0579,
2262
+ "step": 3740
2263
+ },
2264
+ {
2265
+ "epoch": 1.7,
2266
+ "learning_rate": 8.76923076923077e-06,
2267
+ "loss": 0.0016,
2268
+ "step": 3750
2269
+ },
2270
+ {
2271
+ "epoch": 1.71,
2272
+ "learning_rate": 8.73846153846154e-06,
2273
+ "loss": 0.0521,
2274
+ "step": 3760
2275
+ },
2276
+ {
2277
+ "epoch": 1.71,
2278
+ "learning_rate": 8.707692307692309e-06,
2279
+ "loss": 0.1428,
2280
+ "step": 3770
2281
+ },
2282
+ {
2283
+ "epoch": 1.72,
2284
+ "learning_rate": 8.676923076923078e-06,
2285
+ "loss": 0.0464,
2286
+ "step": 3780
2287
+ },
2288
+ {
2289
+ "epoch": 1.72,
2290
+ "learning_rate": 8.646153846153846e-06,
2291
+ "loss": 0.0394,
2292
+ "step": 3790
2293
+ },
2294
+ {
2295
+ "epoch": 1.73,
2296
+ "learning_rate": 8.615384615384617e-06,
2297
+ "loss": 0.0908,
2298
+ "step": 3800
2299
+ },
2300
+ {
2301
+ "epoch": 1.73,
2302
+ "learning_rate": 8.584615384615386e-06,
2303
+ "loss": 0.0027,
2304
+ "step": 3810
2305
+ },
2306
+ {
2307
+ "epoch": 1.74,
2308
+ "learning_rate": 8.553846153846156e-06,
2309
+ "loss": 0.0717,
2310
+ "step": 3820
2311
+ },
2312
+ {
2313
+ "epoch": 1.74,
2314
+ "learning_rate": 8.523076923076923e-06,
2315
+ "loss": 0.178,
2316
+ "step": 3830
2317
+ },
2318
+ {
2319
+ "epoch": 1.75,
2320
+ "learning_rate": 8.492307692307693e-06,
2321
+ "loss": 0.1158,
2322
+ "step": 3840
2323
+ },
2324
+ {
2325
+ "epoch": 1.75,
2326
+ "learning_rate": 8.461538461538462e-06,
2327
+ "loss": 0.0016,
2328
+ "step": 3850
2329
+ },
2330
+ {
2331
+ "epoch": 1.75,
2332
+ "learning_rate": 8.430769230769231e-06,
2333
+ "loss": 0.0504,
2334
+ "step": 3860
2335
+ },
2336
+ {
2337
+ "epoch": 1.76,
2338
+ "learning_rate": 8.400000000000001e-06,
2339
+ "loss": 0.1092,
2340
+ "step": 3870
2341
+ },
2342
+ {
2343
+ "epoch": 1.76,
2344
+ "learning_rate": 8.36923076923077e-06,
2345
+ "loss": 0.114,
2346
+ "step": 3880
2347
+ },
2348
+ {
2349
+ "epoch": 1.77,
2350
+ "learning_rate": 8.338461538461538e-06,
2351
+ "loss": 0.002,
2352
+ "step": 3890
2353
+ },
2354
+ {
2355
+ "epoch": 1.77,
2356
+ "learning_rate": 8.307692307692309e-06,
2357
+ "loss": 0.0029,
2358
+ "step": 3900
2359
+ },
2360
+ {
2361
+ "epoch": 1.78,
2362
+ "learning_rate": 8.276923076923078e-06,
2363
+ "loss": 0.002,
2364
+ "step": 3910
2365
+ },
2366
+ {
2367
+ "epoch": 1.78,
2368
+ "learning_rate": 8.246153846153848e-06,
2369
+ "loss": 0.1239,
2370
+ "step": 3920
2371
+ },
2372
+ {
2373
+ "epoch": 1.79,
2374
+ "learning_rate": 8.215384615384616e-06,
2375
+ "loss": 0.0016,
2376
+ "step": 3930
2377
+ },
2378
+ {
2379
+ "epoch": 1.79,
2380
+ "learning_rate": 8.184615384615385e-06,
2381
+ "loss": 0.0265,
2382
+ "step": 3940
2383
+ },
2384
+ {
2385
+ "epoch": 1.8,
2386
+ "learning_rate": 8.153846153846154e-06,
2387
+ "loss": 0.0836,
2388
+ "step": 3950
2389
+ },
2390
+ {
2391
+ "epoch": 1.8,
2392
+ "learning_rate": 8.123076923076924e-06,
2393
+ "loss": 0.0017,
2394
+ "step": 3960
2395
+ },
2396
+ {
2397
+ "epoch": 1.8,
2398
+ "learning_rate": 8.092307692307693e-06,
2399
+ "loss": 0.0014,
2400
+ "step": 3970
2401
+ },
2402
+ {
2403
+ "epoch": 1.81,
2404
+ "learning_rate": 8.061538461538463e-06,
2405
+ "loss": 0.1186,
2406
+ "step": 3980
2407
+ },
2408
+ {
2409
+ "epoch": 1.81,
2410
+ "learning_rate": 8.03076923076923e-06,
2411
+ "loss": 0.0791,
2412
+ "step": 3990
2413
+ },
2414
+ {
2415
+ "epoch": 1.82,
2416
+ "learning_rate": 8.000000000000001e-06,
2417
+ "loss": 0.0021,
2418
+ "step": 4000
2419
+ },
2420
+ {
2421
+ "epoch": 1.82,
2422
+ "learning_rate": 7.96923076923077e-06,
2423
+ "loss": 0.2008,
2424
+ "step": 4010
2425
+ },
2426
+ {
2427
+ "epoch": 1.83,
2428
+ "learning_rate": 7.93846153846154e-06,
2429
+ "loss": 0.1339,
2430
+ "step": 4020
2431
+ },
2432
+ {
2433
+ "epoch": 1.83,
2434
+ "learning_rate": 7.907692307692308e-06,
2435
+ "loss": 0.0873,
2436
+ "step": 4030
2437
+ },
2438
+ {
2439
+ "epoch": 1.84,
2440
+ "learning_rate": 7.876923076923077e-06,
2441
+ "loss": 0.0076,
2442
+ "step": 4040
2443
+ },
2444
+ {
2445
+ "epoch": 1.84,
2446
+ "learning_rate": 7.846153846153847e-06,
2447
+ "loss": 0.0695,
2448
+ "step": 4050
2449
+ },
2450
+ {
2451
+ "epoch": 1.85,
2452
+ "learning_rate": 7.815384615384616e-06,
2453
+ "loss": 0.0023,
2454
+ "step": 4060
2455
+ },
2456
+ {
2457
+ "epoch": 1.85,
2458
+ "learning_rate": 7.784615384615385e-06,
2459
+ "loss": 0.1105,
2460
+ "step": 4070
2461
+ },
2462
+ {
2463
+ "epoch": 1.85,
2464
+ "learning_rate": 7.753846153846155e-06,
2465
+ "loss": 0.0645,
2466
+ "step": 4080
2467
+ },
2468
+ {
2469
+ "epoch": 1.86,
2470
+ "learning_rate": 7.723076923076924e-06,
2471
+ "loss": 0.0019,
2472
+ "step": 4090
2473
+ },
2474
+ {
2475
+ "epoch": 1.86,
2476
+ "learning_rate": 7.692307692307694e-06,
2477
+ "loss": 0.0652,
2478
+ "step": 4100
2479
+ },
2480
+ {
2481
+ "epoch": 1.87,
2482
+ "learning_rate": 7.661538461538463e-06,
2483
+ "loss": 0.1311,
2484
+ "step": 4110
2485
+ },
2486
+ {
2487
+ "epoch": 1.87,
2488
+ "learning_rate": 7.630769230769232e-06,
2489
+ "loss": 0.0037,
2490
+ "step": 4120
2491
+ },
2492
+ {
2493
+ "epoch": 1.88,
2494
+ "learning_rate": 7.600000000000001e-06,
2495
+ "loss": 0.0161,
2496
+ "step": 4130
2497
+ },
2498
+ {
2499
+ "epoch": 1.88,
2500
+ "learning_rate": 7.5692307692307695e-06,
2501
+ "loss": 0.0014,
2502
+ "step": 4140
2503
+ },
2504
+ {
2505
+ "epoch": 1.89,
2506
+ "learning_rate": 7.538461538461539e-06,
2507
+ "loss": 0.0014,
2508
+ "step": 4150
2509
+ },
2510
+ {
2511
+ "epoch": 1.89,
2512
+ "learning_rate": 7.507692307692308e-06,
2513
+ "loss": 0.1131,
2514
+ "step": 4160
2515
+ },
2516
+ {
2517
+ "epoch": 1.9,
2518
+ "learning_rate": 7.476923076923077e-06,
2519
+ "loss": 0.0515,
2520
+ "step": 4170
2521
+ },
2522
+ {
2523
+ "epoch": 1.9,
2524
+ "learning_rate": 7.446153846153846e-06,
2525
+ "loss": 0.0013,
2526
+ "step": 4180
2527
+ },
2528
+ {
2529
+ "epoch": 1.9,
2530
+ "learning_rate": 7.4153846153846164e-06,
2531
+ "loss": 0.01,
2532
+ "step": 4190
2533
+ },
2534
+ {
2535
+ "epoch": 1.91,
2536
+ "learning_rate": 7.384615384615386e-06,
2537
+ "loss": 0.0011,
2538
+ "step": 4200
2539
+ },
2540
+ {
2541
+ "epoch": 1.91,
2542
+ "learning_rate": 7.353846153846154e-06,
2543
+ "loss": 0.001,
2544
+ "step": 4210
2545
+ },
2546
+ {
2547
+ "epoch": 1.92,
2548
+ "learning_rate": 7.323076923076924e-06,
2549
+ "loss": 0.1202,
2550
+ "step": 4220
2551
+ },
2552
+ {
2553
+ "epoch": 1.92,
2554
+ "learning_rate": 7.292307692307693e-06,
2555
+ "loss": 0.001,
2556
+ "step": 4230
2557
+ },
2558
+ {
2559
+ "epoch": 1.93,
2560
+ "learning_rate": 7.261538461538462e-06,
2561
+ "loss": 0.0686,
2562
+ "step": 4240
2563
+ },
2564
+ {
2565
+ "epoch": 1.93,
2566
+ "learning_rate": 7.230769230769231e-06,
2567
+ "loss": 0.0012,
2568
+ "step": 4250
2569
+ },
2570
+ {
2571
+ "epoch": 1.94,
2572
+ "learning_rate": 7.2000000000000005e-06,
2573
+ "loss": 0.0012,
2574
+ "step": 4260
2575
+ },
2576
+ {
2577
+ "epoch": 1.94,
2578
+ "learning_rate": 7.169230769230769e-06,
2579
+ "loss": 0.0012,
2580
+ "step": 4270
2581
+ },
2582
+ {
2583
+ "epoch": 1.95,
2584
+ "learning_rate": 7.1384615384615385e-06,
2585
+ "loss": 0.1015,
2586
+ "step": 4280
2587
+ },
2588
+ {
2589
+ "epoch": 1.95,
2590
+ "learning_rate": 7.107692307692309e-06,
2591
+ "loss": 0.1873,
2592
+ "step": 4290
2593
+ },
2594
+ {
2595
+ "epoch": 1.95,
2596
+ "learning_rate": 7.076923076923078e-06,
2597
+ "loss": 0.0013,
2598
+ "step": 4300
2599
+ },
2600
+ {
2601
+ "epoch": 1.96,
2602
+ "learning_rate": 7.046153846153847e-06,
2603
+ "loss": 0.0024,
2604
+ "step": 4310
2605
+ },
2606
+ {
2607
+ "epoch": 1.96,
2608
+ "learning_rate": 7.015384615384616e-06,
2609
+ "loss": 0.0013,
2610
+ "step": 4320
2611
+ },
2612
+ {
2613
+ "epoch": 1.97,
2614
+ "learning_rate": 6.9846153846153854e-06,
2615
+ "loss": 0.0289,
2616
+ "step": 4330
2617
+ },
2618
+ {
2619
+ "epoch": 1.97,
2620
+ "learning_rate": 6.953846153846154e-06,
2621
+ "loss": 0.0013,
2622
+ "step": 4340
2623
+ },
2624
+ {
2625
+ "epoch": 1.98,
2626
+ "learning_rate": 6.923076923076923e-06,
2627
+ "loss": 0.0804,
2628
+ "step": 4350
2629
+ },
2630
+ {
2631
+ "epoch": 1.98,
2632
+ "learning_rate": 6.892307692307693e-06,
2633
+ "loss": 0.0897,
2634
+ "step": 4360
2635
+ },
2636
+ {
2637
+ "epoch": 1.99,
2638
+ "learning_rate": 6.861538461538461e-06,
2639
+ "loss": 0.0316,
2640
+ "step": 4370
2641
+ },
2642
+ {
2643
+ "epoch": 1.99,
2644
+ "learning_rate": 6.830769230769231e-06,
2645
+ "loss": 0.0015,
2646
+ "step": 4380
2647
+ },
2648
+ {
2649
+ "epoch": 2.0,
2650
+ "learning_rate": 6.800000000000001e-06,
2651
+ "loss": 0.0921,
2652
+ "step": 4390
2653
+ },
2654
+ {
2655
+ "epoch": 2.0,
2656
+ "learning_rate": 6.76923076923077e-06,
2657
+ "loss": 0.0011,
2658
+ "step": 4400
2659
+ },
2660
+ {
2661
+ "epoch": 2.0,
2662
+ "eval_Accuracy": 0.9897727272727272,
2663
+ "eval_F1": 0.9898706979395142,
2664
+ "eval_loss": 0.06447935849428177,
2665
+ "eval_runtime": 150.7492,
2666
+ "eval_samples_per_second": 29.188,
2667
+ "eval_steps_per_second": 1.824,
2668
+ "step": 4400
2669
+ },
2670
+ {
2671
+ "epoch": 2.0,
2672
+ "learning_rate": 6.738461538461539e-06,
2673
+ "loss": 0.001,
2674
+ "step": 4410
2675
+ },
2676
+ {
2677
+ "epoch": 2.01,
2678
+ "learning_rate": 6.707692307692308e-06,
2679
+ "loss": 0.0389,
2680
+ "step": 4420
2681
+ },
2682
+ {
2683
+ "epoch": 2.01,
2684
+ "learning_rate": 6.676923076923078e-06,
2685
+ "loss": 0.1424,
2686
+ "step": 4430
2687
+ },
2688
+ {
2689
+ "epoch": 2.02,
2690
+ "learning_rate": 6.646153846153846e-06,
2691
+ "loss": 0.0012,
2692
+ "step": 4440
2693
+ },
2694
+ {
2695
+ "epoch": 2.02,
2696
+ "learning_rate": 6.615384615384616e-06,
2697
+ "loss": 0.0662,
2698
+ "step": 4450
2699
+ },
2700
+ {
2701
+ "epoch": 2.03,
2702
+ "learning_rate": 6.584615384615385e-06,
2703
+ "loss": 0.0536,
2704
+ "step": 4460
2705
+ },
2706
+ {
2707
+ "epoch": 2.03,
2708
+ "learning_rate": 6.553846153846154e-06,
2709
+ "loss": 0.0683,
2710
+ "step": 4470
2711
+ },
2712
+ {
2713
+ "epoch": 2.04,
2714
+ "learning_rate": 6.523076923076923e-06,
2715
+ "loss": 0.0012,
2716
+ "step": 4480
2717
+ },
2718
+ {
2719
+ "epoch": 2.04,
2720
+ "learning_rate": 6.492307692307693e-06,
2721
+ "loss": 0.0011,
2722
+ "step": 4490
2723
+ },
2724
+ {
2725
+ "epoch": 2.05,
2726
+ "learning_rate": 6.461538461538463e-06,
2727
+ "loss": 0.0014,
2728
+ "step": 4500
2729
+ },
2730
+ {
2731
+ "epoch": 2.05,
2732
+ "learning_rate": 6.430769230769231e-06,
2733
+ "loss": 0.0317,
2734
+ "step": 4510
2735
+ },
2736
+ {
2737
+ "epoch": 2.05,
2738
+ "learning_rate": 6.4000000000000006e-06,
2739
+ "loss": 0.0974,
2740
+ "step": 4520
2741
+ },
2742
+ {
2743
+ "epoch": 2.06,
2744
+ "learning_rate": 6.36923076923077e-06,
2745
+ "loss": 0.0994,
2746
+ "step": 4530
2747
+ },
2748
+ {
2749
+ "epoch": 2.06,
2750
+ "learning_rate": 6.3384615384615385e-06,
2751
+ "loss": 0.0027,
2752
+ "step": 4540
2753
+ },
2754
+ {
2755
+ "epoch": 2.07,
2756
+ "learning_rate": 6.307692307692308e-06,
2757
+ "loss": 0.0508,
2758
+ "step": 4550
2759
+ },
2760
+ {
2761
+ "epoch": 2.07,
2762
+ "learning_rate": 6.276923076923077e-06,
2763
+ "loss": 0.0017,
2764
+ "step": 4560
2765
+ },
2766
+ {
2767
+ "epoch": 2.08,
2768
+ "learning_rate": 6.246153846153846e-06,
2769
+ "loss": 0.0453,
2770
+ "step": 4570
2771
+ },
2772
+ {
2773
+ "epoch": 2.08,
2774
+ "learning_rate": 6.215384615384615e-06,
2775
+ "loss": 0.0022,
2776
+ "step": 4580
2777
+ },
2778
+ {
2779
+ "epoch": 2.09,
2780
+ "learning_rate": 6.1846153846153855e-06,
2781
+ "loss": 0.0014,
2782
+ "step": 4590
2783
+ },
2784
+ {
2785
+ "epoch": 2.09,
2786
+ "learning_rate": 6.153846153846155e-06,
2787
+ "loss": 0.0015,
2788
+ "step": 4600
2789
+ },
2790
+ {
2791
+ "epoch": 2.1,
2792
+ "learning_rate": 6.123076923076923e-06,
2793
+ "loss": 0.0015,
2794
+ "step": 4610
2795
+ },
2796
+ {
2797
+ "epoch": 2.1,
2798
+ "learning_rate": 6.092307692307693e-06,
2799
+ "loss": 0.0015,
2800
+ "step": 4620
2801
+ },
2802
+ {
2803
+ "epoch": 2.1,
2804
+ "learning_rate": 6.061538461538462e-06,
2805
+ "loss": 0.065,
2806
+ "step": 4630
2807
+ },
2808
+ {
2809
+ "epoch": 2.11,
2810
+ "learning_rate": 6.030769230769231e-06,
2811
+ "loss": 0.0013,
2812
+ "step": 4640
2813
+ },
2814
+ {
2815
+ "epoch": 2.11,
2816
+ "learning_rate": 6e-06,
2817
+ "loss": 0.0018,
2818
+ "step": 4650
2819
+ },
2820
+ {
2821
+ "epoch": 2.12,
2822
+ "learning_rate": 5.9692307692307695e-06,
2823
+ "loss": 0.0021,
2824
+ "step": 4660
2825
+ },
2826
+ {
2827
+ "epoch": 2.12,
2828
+ "learning_rate": 5.938461538461538e-06,
2829
+ "loss": 0.0708,
2830
+ "step": 4670
2831
+ },
2832
+ {
2833
+ "epoch": 2.13,
2834
+ "learning_rate": 5.907692307692308e-06,
2835
+ "loss": 0.0318,
2836
+ "step": 4680
2837
+ },
2838
+ {
2839
+ "epoch": 2.13,
2840
+ "learning_rate": 5.876923076923078e-06,
2841
+ "loss": 0.0107,
2842
+ "step": 4690
2843
+ },
2844
+ {
2845
+ "epoch": 2.14,
2846
+ "learning_rate": 5.846153846153847e-06,
2847
+ "loss": 0.0012,
2848
+ "step": 4700
2849
+ },
2850
+ {
2851
+ "epoch": 2.14,
2852
+ "learning_rate": 5.815384615384616e-06,
2853
+ "loss": 0.0635,
2854
+ "step": 4710
2855
+ },
2856
+ {
2857
+ "epoch": 2.15,
2858
+ "learning_rate": 5.784615384615385e-06,
2859
+ "loss": 0.0056,
2860
+ "step": 4720
2861
+ },
2862
+ {
2863
+ "epoch": 2.15,
2864
+ "learning_rate": 5.7538461538461545e-06,
2865
+ "loss": 0.0009,
2866
+ "step": 4730
2867
+ },
2868
+ {
2869
+ "epoch": 2.15,
2870
+ "learning_rate": 5.723076923076923e-06,
2871
+ "loss": 0.0404,
2872
+ "step": 4740
2873
+ },
2874
+ {
2875
+ "epoch": 2.16,
2876
+ "learning_rate": 5.692307692307692e-06,
2877
+ "loss": 0.0012,
2878
+ "step": 4750
2879
+ },
2880
+ {
2881
+ "epoch": 2.16,
2882
+ "learning_rate": 5.661538461538462e-06,
2883
+ "loss": 0.1147,
2884
+ "step": 4760
2885
+ },
2886
+ {
2887
+ "epoch": 2.17,
2888
+ "learning_rate": 5.63076923076923e-06,
2889
+ "loss": 0.0591,
2890
+ "step": 4770
2891
+ },
2892
+ {
2893
+ "epoch": 2.17,
2894
+ "learning_rate": 5.600000000000001e-06,
2895
+ "loss": 0.0515,
2896
+ "step": 4780
2897
+ },
2898
+ {
2899
+ "epoch": 2.18,
2900
+ "learning_rate": 5.56923076923077e-06,
2901
+ "loss": 0.0009,
2902
+ "step": 4790
2903
+ },
2904
+ {
2905
+ "epoch": 2.18,
2906
+ "learning_rate": 5.538461538461539e-06,
2907
+ "loss": 0.001,
2908
+ "step": 4800
2909
+ },
2910
+ {
2911
+ "epoch": 2.19,
2912
+ "learning_rate": 5.507692307692308e-06,
2913
+ "loss": 0.0009,
2914
+ "step": 4810
2915
+ },
2916
+ {
2917
+ "epoch": 2.19,
2918
+ "learning_rate": 5.476923076923077e-06,
2919
+ "loss": 0.0013,
2920
+ "step": 4820
2921
+ },
2922
+ {
2923
+ "epoch": 2.2,
2924
+ "learning_rate": 5.446153846153847e-06,
2925
+ "loss": 0.0009,
2926
+ "step": 4830
2927
+ },
2928
+ {
2929
+ "epoch": 2.2,
2930
+ "learning_rate": 5.415384615384615e-06,
2931
+ "loss": 0.0332,
2932
+ "step": 4840
2933
+ },
2934
+ {
2935
+ "epoch": 2.2,
2936
+ "learning_rate": 5.384615384615385e-06,
2937
+ "loss": 0.0977,
2938
+ "step": 4850
2939
+ },
2940
+ {
2941
+ "epoch": 2.21,
2942
+ "learning_rate": 5.353846153846154e-06,
2943
+ "loss": 0.0376,
2944
+ "step": 4860
2945
+ },
2946
+ {
2947
+ "epoch": 2.21,
2948
+ "learning_rate": 5.323076923076923e-06,
2949
+ "loss": 0.0396,
2950
+ "step": 4870
2951
+ },
2952
+ {
2953
+ "epoch": 2.22,
2954
+ "learning_rate": 5.292307692307693e-06,
2955
+ "loss": 0.0016,
2956
+ "step": 4880
2957
+ },
2958
+ {
2959
+ "epoch": 2.22,
2960
+ "learning_rate": 5.261538461538462e-06,
2961
+ "loss": 0.0212,
2962
+ "step": 4890
2963
+ },
2964
+ {
2965
+ "epoch": 2.23,
2966
+ "learning_rate": 5.230769230769232e-06,
2967
+ "loss": 0.0588,
2968
+ "step": 4900
2969
+ },
2970
+ {
2971
+ "epoch": 2.23,
2972
+ "learning_rate": 5.2e-06,
2973
+ "loss": 0.0011,
2974
+ "step": 4910
2975
+ },
2976
+ {
2977
+ "epoch": 2.24,
2978
+ "learning_rate": 5.16923076923077e-06,
2979
+ "loss": 0.0009,
2980
+ "step": 4920
2981
+ },
2982
+ {
2983
+ "epoch": 2.24,
2984
+ "learning_rate": 5.138461538461539e-06,
2985
+ "loss": 0.1418,
2986
+ "step": 4930
2987
+ },
2988
+ {
2989
+ "epoch": 2.25,
2990
+ "learning_rate": 5.1076923076923075e-06,
2991
+ "loss": 0.0538,
2992
+ "step": 4940
2993
+ },
2994
+ {
2995
+ "epoch": 2.25,
2996
+ "learning_rate": 5.076923076923077e-06,
2997
+ "loss": 0.0155,
2998
+ "step": 4950
2999
+ },
3000
+ {
3001
+ "epoch": 2.25,
3002
+ "learning_rate": 5.046153846153846e-06,
3003
+ "loss": 0.1014,
3004
+ "step": 4960
3005
+ },
3006
+ {
3007
+ "epoch": 2.26,
3008
+ "learning_rate": 5.015384615384616e-06,
3009
+ "loss": 0.0013,
3010
+ "step": 4970
3011
+ },
3012
+ {
3013
+ "epoch": 2.26,
3014
+ "learning_rate": 4.984615384615385e-06,
3015
+ "loss": 0.0126,
3016
+ "step": 4980
3017
+ },
3018
+ {
3019
+ "epoch": 2.27,
3020
+ "learning_rate": 4.9538461538461545e-06,
3021
+ "loss": 0.0569,
3022
+ "step": 4990
3023
+ },
3024
+ {
3025
+ "epoch": 2.27,
3026
+ "learning_rate": 4.923076923076924e-06,
3027
+ "loss": 0.0241,
3028
+ "step": 5000
3029
+ },
3030
+ {
3031
+ "epoch": 2.28,
3032
+ "learning_rate": 4.892307692307693e-06,
3033
+ "loss": 0.001,
3034
+ "step": 5010
3035
+ },
3036
+ {
3037
+ "epoch": 2.28,
3038
+ "learning_rate": 4.861538461538462e-06,
3039
+ "loss": 0.0029,
3040
+ "step": 5020
3041
+ },
3042
+ {
3043
+ "epoch": 2.29,
3044
+ "learning_rate": 4.830769230769231e-06,
3045
+ "loss": 0.0042,
3046
+ "step": 5030
3047
+ },
3048
+ {
3049
+ "epoch": 2.29,
3050
+ "learning_rate": 4.800000000000001e-06,
3051
+ "loss": 0.04,
3052
+ "step": 5040
3053
+ },
3054
+ {
3055
+ "epoch": 2.3,
3056
+ "learning_rate": 4.76923076923077e-06,
3057
+ "loss": 0.0008,
3058
+ "step": 5050
3059
+ },
3060
+ {
3061
+ "epoch": 2.3,
3062
+ "learning_rate": 4.738461538461539e-06,
3063
+ "loss": 0.0014,
3064
+ "step": 5060
3065
+ },
3066
+ {
3067
+ "epoch": 2.3,
3068
+ "learning_rate": 4.707692307692308e-06,
3069
+ "loss": 0.0959,
3070
+ "step": 5070
3071
+ },
3072
+ {
3073
+ "epoch": 2.31,
3074
+ "learning_rate": 4.676923076923077e-06,
3075
+ "loss": 0.0009,
3076
+ "step": 5080
3077
+ },
3078
+ {
3079
+ "epoch": 2.31,
3080
+ "learning_rate": 4.646153846153847e-06,
3081
+ "loss": 0.0137,
3082
+ "step": 5090
3083
+ },
3084
+ {
3085
+ "epoch": 2.32,
3086
+ "learning_rate": 4.615384615384616e-06,
3087
+ "loss": 0.1642,
3088
+ "step": 5100
3089
+ },
3090
+ {
3091
+ "epoch": 2.32,
3092
+ "learning_rate": 4.5846153846153855e-06,
3093
+ "loss": 0.0026,
3094
+ "step": 5110
3095
+ },
3096
+ {
3097
+ "epoch": 2.33,
3098
+ "learning_rate": 4.553846153846154e-06,
3099
+ "loss": 0.1011,
3100
+ "step": 5120
3101
+ },
3102
+ {
3103
+ "epoch": 2.33,
3104
+ "learning_rate": 4.5230769230769235e-06,
3105
+ "loss": 0.0456,
3106
+ "step": 5130
3107
+ },
3108
+ {
3109
+ "epoch": 2.34,
3110
+ "learning_rate": 4.492307692307693e-06,
3111
+ "loss": 0.0484,
3112
+ "step": 5140
3113
+ },
3114
+ {
3115
+ "epoch": 2.34,
3116
+ "learning_rate": 4.461538461538462e-06,
3117
+ "loss": 0.0272,
3118
+ "step": 5150
3119
+ },
3120
+ {
3121
+ "epoch": 2.35,
3122
+ "learning_rate": 4.430769230769232e-06,
3123
+ "loss": 0.0009,
3124
+ "step": 5160
3125
+ },
3126
+ {
3127
+ "epoch": 2.35,
3128
+ "learning_rate": 4.4e-06,
3129
+ "loss": 0.1856,
3130
+ "step": 5170
3131
+ },
3132
+ {
3133
+ "epoch": 2.35,
3134
+ "learning_rate": 4.36923076923077e-06,
3135
+ "loss": 0.0075,
3136
+ "step": 5180
3137
+ },
3138
+ {
3139
+ "epoch": 2.36,
3140
+ "learning_rate": 4.338461538461539e-06,
3141
+ "loss": 0.0019,
3142
+ "step": 5190
3143
+ },
3144
+ {
3145
+ "epoch": 2.36,
3146
+ "learning_rate": 4.307692307692308e-06,
3147
+ "loss": 0.0012,
3148
+ "step": 5200
3149
+ },
3150
+ {
3151
+ "epoch": 2.37,
3152
+ "learning_rate": 4.276923076923078e-06,
3153
+ "loss": 0.001,
3154
+ "step": 5210
3155
+ },
3156
+ {
3157
+ "epoch": 2.37,
3158
+ "learning_rate": 4.246153846153846e-06,
3159
+ "loss": 0.0975,
3160
+ "step": 5220
3161
+ },
3162
+ {
3163
+ "epoch": 2.38,
3164
+ "learning_rate": 4.215384615384616e-06,
3165
+ "loss": 0.0011,
3166
+ "step": 5230
3167
+ },
3168
+ {
3169
+ "epoch": 2.38,
3170
+ "learning_rate": 4.184615384615385e-06,
3171
+ "loss": 0.0011,
3172
+ "step": 5240
3173
+ },
3174
+ {
3175
+ "epoch": 2.39,
3176
+ "learning_rate": 4.1538461538461545e-06,
3177
+ "loss": 0.0012,
3178
+ "step": 5250
3179
+ },
3180
+ {
3181
+ "epoch": 2.39,
3182
+ "learning_rate": 4.123076923076924e-06,
3183
+ "loss": 0.1065,
3184
+ "step": 5260
3185
+ },
3186
+ {
3187
+ "epoch": 2.4,
3188
+ "learning_rate": 4.0923076923076925e-06,
3189
+ "loss": 0.0019,
3190
+ "step": 5270
3191
+ },
3192
+ {
3193
+ "epoch": 2.4,
3194
+ "learning_rate": 4.061538461538462e-06,
3195
+ "loss": 0.0016,
3196
+ "step": 5280
3197
+ },
3198
+ {
3199
+ "epoch": 2.4,
3200
+ "learning_rate": 4.030769230769231e-06,
3201
+ "loss": 0.0017,
3202
+ "step": 5290
3203
+ },
3204
+ {
3205
+ "epoch": 2.41,
3206
+ "learning_rate": 4.000000000000001e-06,
3207
+ "loss": 0.0372,
3208
+ "step": 5300
3209
+ },
3210
+ {
3211
+ "epoch": 2.41,
3212
+ "learning_rate": 3.96923076923077e-06,
3213
+ "loss": 0.088,
3214
+ "step": 5310
3215
+ },
3216
+ {
3217
+ "epoch": 2.42,
3218
+ "learning_rate": 3.938461538461539e-06,
3219
+ "loss": 0.1187,
3220
+ "step": 5320
3221
+ },
3222
+ {
3223
+ "epoch": 2.42,
3224
+ "learning_rate": 3.907692307692308e-06,
3225
+ "loss": 0.001,
3226
+ "step": 5330
3227
+ },
3228
+ {
3229
+ "epoch": 2.43,
3230
+ "learning_rate": 3.876923076923077e-06,
3231
+ "loss": 0.0011,
3232
+ "step": 5340
3233
+ },
3234
+ {
3235
+ "epoch": 2.43,
3236
+ "learning_rate": 3.846153846153847e-06,
3237
+ "loss": 0.0494,
3238
+ "step": 5350
3239
+ },
3240
+ {
3241
+ "epoch": 2.44,
3242
+ "learning_rate": 3.815384615384616e-06,
3243
+ "loss": 0.001,
3244
+ "step": 5360
3245
+ },
3246
+ {
3247
+ "epoch": 2.44,
3248
+ "learning_rate": 3.7846153846153847e-06,
3249
+ "loss": 0.0008,
3250
+ "step": 5370
3251
+ },
3252
+ {
3253
+ "epoch": 2.45,
3254
+ "learning_rate": 3.753846153846154e-06,
3255
+ "loss": 0.085,
3256
+ "step": 5380
3257
+ },
3258
+ {
3259
+ "epoch": 2.45,
3260
+ "learning_rate": 3.723076923076923e-06,
3261
+ "loss": 0.001,
3262
+ "step": 5390
3263
+ },
3264
+ {
3265
+ "epoch": 2.45,
3266
+ "learning_rate": 3.692307692307693e-06,
3267
+ "loss": 0.1068,
3268
+ "step": 5400
3269
+ },
3270
+ {
3271
+ "epoch": 2.46,
3272
+ "learning_rate": 3.661538461538462e-06,
3273
+ "loss": 0.086,
3274
+ "step": 5410
3275
+ },
3276
+ {
3277
+ "epoch": 2.46,
3278
+ "learning_rate": 3.630769230769231e-06,
3279
+ "loss": 0.0531,
3280
+ "step": 5420
3281
+ },
3282
+ {
3283
+ "epoch": 2.47,
3284
+ "learning_rate": 3.6000000000000003e-06,
3285
+ "loss": 0.051,
3286
+ "step": 5430
3287
+ },
3288
+ {
3289
+ "epoch": 2.47,
3290
+ "learning_rate": 3.5692307692307692e-06,
3291
+ "loss": 0.0009,
3292
+ "step": 5440
3293
+ },
3294
+ {
3295
+ "epoch": 2.48,
3296
+ "learning_rate": 3.538461538461539e-06,
3297
+ "loss": 0.001,
3298
+ "step": 5450
3299
+ },
3300
+ {
3301
+ "epoch": 2.48,
3302
+ "learning_rate": 3.507692307692308e-06,
3303
+ "loss": 0.0015,
3304
+ "step": 5460
3305
+ },
3306
+ {
3307
+ "epoch": 2.49,
3308
+ "learning_rate": 3.476923076923077e-06,
3309
+ "loss": 0.0118,
3310
+ "step": 5470
3311
+ },
3312
+ {
3313
+ "epoch": 2.49,
3314
+ "learning_rate": 3.4461538461538464e-06,
3315
+ "loss": 0.0124,
3316
+ "step": 5480
3317
+ },
3318
+ {
3319
+ "epoch": 2.5,
3320
+ "learning_rate": 3.4153846153846154e-06,
3321
+ "loss": 0.0032,
3322
+ "step": 5490
3323
+ },
3324
+ {
3325
+ "epoch": 2.5,
3326
+ "learning_rate": 3.384615384615385e-06,
3327
+ "loss": 0.0013,
3328
+ "step": 5500
3329
+ },
3330
+ {
3331
+ "epoch": 2.5,
3332
+ "learning_rate": 3.353846153846154e-06,
3333
+ "loss": 0.1963,
3334
+ "step": 5510
3335
+ },
3336
+ {
3337
+ "epoch": 2.51,
3338
+ "learning_rate": 3.323076923076923e-06,
3339
+ "loss": 0.053,
3340
+ "step": 5520
3341
+ },
3342
+ {
3343
+ "epoch": 2.51,
3344
+ "learning_rate": 3.2923076923076925e-06,
3345
+ "loss": 0.001,
3346
+ "step": 5530
3347
+ },
3348
+ {
3349
+ "epoch": 2.52,
3350
+ "learning_rate": 3.2615384615384615e-06,
3351
+ "loss": 0.0281,
3352
+ "step": 5540
3353
+ },
3354
+ {
3355
+ "epoch": 2.52,
3356
+ "learning_rate": 3.2307692307692313e-06,
3357
+ "loss": 0.0941,
3358
+ "step": 5550
3359
+ },
3360
+ {
3361
+ "epoch": 2.53,
3362
+ "learning_rate": 3.2000000000000003e-06,
3363
+ "loss": 0.001,
3364
+ "step": 5560
3365
+ },
3366
+ {
3367
+ "epoch": 2.53,
3368
+ "learning_rate": 3.1692307692307693e-06,
3369
+ "loss": 0.001,
3370
+ "step": 5570
3371
+ },
3372
+ {
3373
+ "epoch": 2.54,
3374
+ "learning_rate": 3.1384615384615386e-06,
3375
+ "loss": 0.001,
3376
+ "step": 5580
3377
+ },
3378
+ {
3379
+ "epoch": 2.54,
3380
+ "learning_rate": 3.1076923076923076e-06,
3381
+ "loss": 0.0008,
3382
+ "step": 5590
3383
+ },
3384
+ {
3385
+ "epoch": 2.55,
3386
+ "learning_rate": 3.0769230769230774e-06,
3387
+ "loss": 0.0318,
3388
+ "step": 5600
3389
+ },
3390
+ {
3391
+ "epoch": 2.55,
3392
+ "learning_rate": 3.0461538461538464e-06,
3393
+ "loss": 0.0009,
3394
+ "step": 5610
3395
+ },
3396
+ {
3397
+ "epoch": 2.55,
3398
+ "learning_rate": 3.0153846153846154e-06,
3399
+ "loss": 0.0156,
3400
+ "step": 5620
3401
+ },
3402
+ {
3403
+ "epoch": 2.56,
3404
+ "learning_rate": 2.9846153846153848e-06,
3405
+ "loss": 0.1011,
3406
+ "step": 5630
3407
+ },
3408
+ {
3409
+ "epoch": 2.56,
3410
+ "learning_rate": 2.953846153846154e-06,
3411
+ "loss": 0.0008,
3412
+ "step": 5640
3413
+ },
3414
+ {
3415
+ "epoch": 2.57,
3416
+ "learning_rate": 2.9230769230769236e-06,
3417
+ "loss": 0.0009,
3418
+ "step": 5650
3419
+ },
3420
+ {
3421
+ "epoch": 2.57,
3422
+ "learning_rate": 2.8923076923076925e-06,
3423
+ "loss": 0.0008,
3424
+ "step": 5660
3425
+ },
3426
+ {
3427
+ "epoch": 2.58,
3428
+ "learning_rate": 2.8615384615384615e-06,
3429
+ "loss": 0.0008,
3430
+ "step": 5670
3431
+ },
3432
+ {
3433
+ "epoch": 2.58,
3434
+ "learning_rate": 2.830769230769231e-06,
3435
+ "loss": 0.0266,
3436
+ "step": 5680
3437
+ },
3438
+ {
3439
+ "epoch": 2.59,
3440
+ "learning_rate": 2.8000000000000003e-06,
3441
+ "loss": 0.0008,
3442
+ "step": 5690
3443
+ },
3444
+ {
3445
+ "epoch": 2.59,
3446
+ "learning_rate": 2.7692307692307697e-06,
3447
+ "loss": 0.1243,
3448
+ "step": 5700
3449
+ },
3450
+ {
3451
+ "epoch": 2.6,
3452
+ "learning_rate": 2.7384615384615387e-06,
3453
+ "loss": 0.0009,
3454
+ "step": 5710
3455
+ },
3456
+ {
3457
+ "epoch": 2.6,
3458
+ "learning_rate": 2.7076923076923076e-06,
3459
+ "loss": 0.1052,
3460
+ "step": 5720
3461
+ },
3462
+ {
3463
+ "epoch": 2.6,
3464
+ "learning_rate": 2.676923076923077e-06,
3465
+ "loss": 0.1307,
3466
+ "step": 5730
3467
+ },
3468
+ {
3469
+ "epoch": 2.61,
3470
+ "learning_rate": 2.6461538461538464e-06,
3471
+ "loss": 0.1628,
3472
+ "step": 5740
3473
+ },
3474
+ {
3475
+ "epoch": 2.61,
3476
+ "learning_rate": 2.615384615384616e-06,
3477
+ "loss": 0.0011,
3478
+ "step": 5750
3479
+ },
3480
+ {
3481
+ "epoch": 2.62,
3482
+ "learning_rate": 2.584615384615385e-06,
3483
+ "loss": 0.001,
3484
+ "step": 5760
3485
+ },
3486
+ {
3487
+ "epoch": 2.62,
3488
+ "learning_rate": 2.5538461538461538e-06,
3489
+ "loss": 0.013,
3490
+ "step": 5770
3491
+ },
3492
+ {
3493
+ "epoch": 2.63,
3494
+ "learning_rate": 2.523076923076923e-06,
3495
+ "loss": 0.1045,
3496
+ "step": 5780
3497
+ },
3498
+ {
3499
+ "epoch": 2.63,
3500
+ "learning_rate": 2.4923076923076926e-06,
3501
+ "loss": 0.003,
3502
+ "step": 5790
3503
+ },
3504
+ {
3505
+ "epoch": 2.64,
3506
+ "learning_rate": 2.461538461538462e-06,
3507
+ "loss": 0.0009,
3508
+ "step": 5800
3509
+ },
3510
+ {
3511
+ "epoch": 2.64,
3512
+ "learning_rate": 2.430769230769231e-06,
3513
+ "loss": 0.0008,
3514
+ "step": 5810
3515
+ },
3516
+ {
3517
+ "epoch": 2.65,
3518
+ "learning_rate": 2.4000000000000003e-06,
3519
+ "loss": 0.0085,
3520
+ "step": 5820
3521
+ },
3522
+ {
3523
+ "epoch": 2.65,
3524
+ "learning_rate": 2.3692307692307697e-06,
3525
+ "loss": 0.0008,
3526
+ "step": 5830
3527
+ },
3528
+ {
3529
+ "epoch": 2.65,
3530
+ "learning_rate": 2.3384615384615387e-06,
3531
+ "loss": 0.0321,
3532
+ "step": 5840
3533
+ },
3534
+ {
3535
+ "epoch": 2.66,
3536
+ "learning_rate": 2.307692307692308e-06,
3537
+ "loss": 0.0397,
3538
+ "step": 5850
3539
+ },
3540
+ {
3541
+ "epoch": 2.66,
3542
+ "learning_rate": 2.276923076923077e-06,
3543
+ "loss": 0.002,
3544
+ "step": 5860
3545
+ },
3546
+ {
3547
+ "epoch": 2.67,
3548
+ "learning_rate": 2.2461538461538464e-06,
3549
+ "loss": 0.0009,
3550
+ "step": 5870
3551
+ },
3552
+ {
3553
+ "epoch": 2.67,
3554
+ "learning_rate": 2.215384615384616e-06,
3555
+ "loss": 0.0025,
3556
+ "step": 5880
3557
+ },
3558
+ {
3559
+ "epoch": 2.68,
3560
+ "learning_rate": 2.184615384615385e-06,
3561
+ "loss": 0.1094,
3562
+ "step": 5890
3563
+ },
3564
+ {
3565
+ "epoch": 2.68,
3566
+ "learning_rate": 2.153846153846154e-06,
3567
+ "loss": 0.0009,
3568
+ "step": 5900
3569
+ },
3570
+ {
3571
+ "epoch": 2.69,
3572
+ "learning_rate": 2.123076923076923e-06,
3573
+ "loss": 0.0009,
3574
+ "step": 5910
3575
+ },
3576
+ {
3577
+ "epoch": 2.69,
3578
+ "learning_rate": 2.0923076923076926e-06,
3579
+ "loss": 0.0699,
3580
+ "step": 5920
3581
+ },
3582
+ {
3583
+ "epoch": 2.7,
3584
+ "learning_rate": 2.061538461538462e-06,
3585
+ "loss": 0.0011,
3586
+ "step": 5930
3587
+ },
3588
+ {
3589
+ "epoch": 2.7,
3590
+ "learning_rate": 2.030769230769231e-06,
3591
+ "loss": 0.0288,
3592
+ "step": 5940
3593
+ },
3594
+ {
3595
+ "epoch": 2.7,
3596
+ "learning_rate": 2.0000000000000003e-06,
3597
+ "loss": 0.0018,
3598
+ "step": 5950
3599
+ },
3600
+ {
3601
+ "epoch": 2.71,
3602
+ "learning_rate": 1.9692307692307693e-06,
3603
+ "loss": 0.0479,
3604
+ "step": 5960
3605
+ },
3606
+ {
3607
+ "epoch": 2.71,
3608
+ "learning_rate": 1.9384615384615387e-06,
3609
+ "loss": 0.0015,
3610
+ "step": 5970
3611
+ },
3612
+ {
3613
+ "epoch": 2.72,
3614
+ "learning_rate": 1.907692307692308e-06,
3615
+ "loss": 0.0009,
3616
+ "step": 5980
3617
+ },
3618
+ {
3619
+ "epoch": 2.72,
3620
+ "learning_rate": 1.876923076923077e-06,
3621
+ "loss": 0.0009,
3622
+ "step": 5990
3623
+ },
3624
+ {
3625
+ "epoch": 2.73,
3626
+ "learning_rate": 1.8461538461538465e-06,
3627
+ "loss": 0.0009,
3628
+ "step": 6000
3629
+ },
3630
+ {
3631
+ "epoch": 2.73,
3632
+ "learning_rate": 1.8153846153846154e-06,
3633
+ "loss": 0.0603,
3634
+ "step": 6010
3635
+ },
3636
+ {
3637
+ "epoch": 2.74,
3638
+ "learning_rate": 1.7846153846153846e-06,
3639
+ "loss": 0.013,
3640
+ "step": 6020
3641
+ },
3642
+ {
3643
+ "epoch": 2.74,
3644
+ "learning_rate": 1.753846153846154e-06,
3645
+ "loss": 0.0023,
3646
+ "step": 6030
3647
+ },
3648
+ {
3649
+ "epoch": 2.75,
3650
+ "learning_rate": 1.7230769230769232e-06,
3651
+ "loss": 0.0009,
3652
+ "step": 6040
3653
+ },
3654
+ {
3655
+ "epoch": 2.75,
3656
+ "learning_rate": 1.6923076923076926e-06,
3657
+ "loss": 0.0007,
3658
+ "step": 6050
3659
+ },
3660
+ {
3661
+ "epoch": 2.75,
3662
+ "learning_rate": 1.6615384615384616e-06,
3663
+ "loss": 0.0305,
3664
+ "step": 6060
3665
+ },
3666
+ {
3667
+ "epoch": 2.76,
3668
+ "learning_rate": 1.6307692307692307e-06,
3669
+ "loss": 0.0008,
3670
+ "step": 6070
3671
+ },
3672
+ {
3673
+ "epoch": 2.76,
3674
+ "learning_rate": 1.6000000000000001e-06,
3675
+ "loss": 0.0544,
3676
+ "step": 6080
3677
+ },
3678
+ {
3679
+ "epoch": 2.77,
3680
+ "learning_rate": 1.5692307692307693e-06,
3681
+ "loss": 0.0691,
3682
+ "step": 6090
3683
+ },
3684
+ {
3685
+ "epoch": 2.77,
3686
+ "learning_rate": 1.5384615384615387e-06,
3687
+ "loss": 0.0217,
3688
+ "step": 6100
3689
+ },
3690
+ {
3691
+ "epoch": 2.78,
3692
+ "learning_rate": 1.5076923076923077e-06,
3693
+ "loss": 0.0579,
3694
+ "step": 6110
3695
+ },
3696
+ {
3697
+ "epoch": 2.78,
3698
+ "learning_rate": 1.476923076923077e-06,
3699
+ "loss": 0.0013,
3700
+ "step": 6120
3701
+ },
3702
+ {
3703
+ "epoch": 2.79,
3704
+ "learning_rate": 1.4461538461538463e-06,
3705
+ "loss": 0.0008,
3706
+ "step": 6130
3707
+ },
3708
+ {
3709
+ "epoch": 2.79,
3710
+ "learning_rate": 1.4153846153846155e-06,
3711
+ "loss": 0.0007,
3712
+ "step": 6140
3713
+ },
3714
+ {
3715
+ "epoch": 2.8,
3716
+ "learning_rate": 1.3846153846153848e-06,
3717
+ "loss": 0.0126,
3718
+ "step": 6150
3719
+ },
3720
+ {
3721
+ "epoch": 2.8,
3722
+ "learning_rate": 1.3538461538461538e-06,
3723
+ "loss": 0.0534,
3724
+ "step": 6160
3725
+ },
3726
+ {
3727
+ "epoch": 2.8,
3728
+ "learning_rate": 1.3230769230769232e-06,
3729
+ "loss": 0.1074,
3730
+ "step": 6170
3731
+ },
3732
+ {
3733
+ "epoch": 2.81,
3734
+ "learning_rate": 1.2923076923076924e-06,
3735
+ "loss": 0.001,
3736
+ "step": 6180
3737
+ },
3738
+ {
3739
+ "epoch": 2.81,
3740
+ "learning_rate": 1.2615384615384616e-06,
3741
+ "loss": 0.001,
3742
+ "step": 6190
3743
+ },
3744
+ {
3745
+ "epoch": 2.82,
3746
+ "learning_rate": 1.230769230769231e-06,
3747
+ "loss": 0.0106,
3748
+ "step": 6200
3749
+ },
3750
+ {
3751
+ "epoch": 2.82,
3752
+ "learning_rate": 1.2000000000000002e-06,
3753
+ "loss": 0.0008,
3754
+ "step": 6210
3755
+ },
3756
+ {
3757
+ "epoch": 2.83,
3758
+ "learning_rate": 1.1692307692307693e-06,
3759
+ "loss": 0.0681,
3760
+ "step": 6220
3761
+ },
3762
+ {
3763
+ "epoch": 2.83,
3764
+ "learning_rate": 1.1384615384615385e-06,
3765
+ "loss": 0.0008,
3766
+ "step": 6230
3767
+ },
3768
+ {
3769
+ "epoch": 2.84,
3770
+ "learning_rate": 1.107692307692308e-06,
3771
+ "loss": 0.0008,
3772
+ "step": 6240
3773
+ },
3774
+ {
3775
+ "epoch": 2.84,
3776
+ "learning_rate": 1.076923076923077e-06,
3777
+ "loss": 0.001,
3778
+ "step": 6250
3779
+ },
3780
+ {
3781
+ "epoch": 2.85,
3782
+ "learning_rate": 1.0461538461538463e-06,
3783
+ "loss": 0.0008,
3784
+ "step": 6260
3785
+ },
3786
+ {
3787
+ "epoch": 2.85,
3788
+ "learning_rate": 1.0153846153846155e-06,
3789
+ "loss": 0.0197,
3790
+ "step": 6270
3791
+ },
3792
+ {
3793
+ "epoch": 2.85,
3794
+ "learning_rate": 9.846153846153847e-07,
3795
+ "loss": 0.0016,
3796
+ "step": 6280
3797
+ },
3798
+ {
3799
+ "epoch": 2.86,
3800
+ "learning_rate": 9.53846153846154e-07,
3801
+ "loss": 0.0817,
3802
+ "step": 6290
3803
+ },
3804
+ {
3805
+ "epoch": 2.86,
3806
+ "learning_rate": 9.230769230769232e-07,
3807
+ "loss": 0.0281,
3808
+ "step": 6300
3809
+ },
3810
+ {
3811
+ "epoch": 2.87,
3812
+ "learning_rate": 8.923076923076923e-07,
3813
+ "loss": 0.0009,
3814
+ "step": 6310
3815
+ },
3816
+ {
3817
+ "epoch": 2.87,
3818
+ "learning_rate": 8.615384615384616e-07,
3819
+ "loss": 0.0158,
3820
+ "step": 6320
3821
+ },
3822
+ {
3823
+ "epoch": 2.88,
3824
+ "learning_rate": 8.307692307692308e-07,
3825
+ "loss": 0.0296,
3826
+ "step": 6330
3827
+ },
3828
+ {
3829
+ "epoch": 2.88,
3830
+ "learning_rate": 8.000000000000001e-07,
3831
+ "loss": 0.0334,
3832
+ "step": 6340
3833
+ },
3834
+ {
3835
+ "epoch": 2.89,
3836
+ "learning_rate": 7.692307692307694e-07,
3837
+ "loss": 0.0612,
3838
+ "step": 6350
3839
+ },
3840
+ {
3841
+ "epoch": 2.89,
3842
+ "learning_rate": 7.384615384615385e-07,
3843
+ "loss": 0.102,
3844
+ "step": 6360
3845
+ },
3846
+ {
3847
+ "epoch": 2.9,
3848
+ "learning_rate": 7.076923076923077e-07,
3849
+ "loss": 0.0007,
3850
+ "step": 6370
3851
+ },
3852
+ {
3853
+ "epoch": 2.9,
3854
+ "learning_rate": 6.769230769230769e-07,
3855
+ "loss": 0.097,
3856
+ "step": 6380
3857
+ },
3858
+ {
3859
+ "epoch": 2.9,
3860
+ "learning_rate": 6.461538461538462e-07,
3861
+ "loss": 0.1572,
3862
+ "step": 6390
3863
+ },
3864
+ {
3865
+ "epoch": 2.91,
3866
+ "learning_rate": 6.153846153846155e-07,
3867
+ "loss": 0.0007,
3868
+ "step": 6400
3869
+ },
3870
+ {
3871
+ "epoch": 2.91,
3872
+ "learning_rate": 5.846153846153847e-07,
3873
+ "loss": 0.0551,
3874
+ "step": 6410
3875
+ },
3876
+ {
3877
+ "epoch": 2.92,
3878
+ "learning_rate": 5.53846153846154e-07,
3879
+ "loss": 0.0014,
3880
+ "step": 6420
3881
+ },
3882
+ {
3883
+ "epoch": 2.92,
3884
+ "learning_rate": 5.230769230769231e-07,
3885
+ "loss": 0.001,
3886
+ "step": 6430
3887
+ },
3888
+ {
3889
+ "epoch": 2.93,
3890
+ "learning_rate": 4.923076923076923e-07,
3891
+ "loss": 0.0007,
3892
+ "step": 6440
3893
+ },
3894
+ {
3895
+ "epoch": 2.93,
3896
+ "learning_rate": 4.615384615384616e-07,
3897
+ "loss": 0.0009,
3898
+ "step": 6450
3899
+ },
3900
+ {
3901
+ "epoch": 2.94,
3902
+ "learning_rate": 4.307692307692308e-07,
3903
+ "loss": 0.0082,
3904
+ "step": 6460
3905
+ },
3906
+ {
3907
+ "epoch": 2.94,
3908
+ "learning_rate": 4.0000000000000003e-07,
3909
+ "loss": 0.0008,
3910
+ "step": 6470
3911
+ },
3912
+ {
3913
+ "epoch": 2.95,
3914
+ "learning_rate": 3.6923076923076927e-07,
3915
+ "loss": 0.0009,
3916
+ "step": 6480
3917
+ },
3918
+ {
3919
+ "epoch": 2.95,
3920
+ "learning_rate": 3.3846153846153845e-07,
3921
+ "loss": 0.0008,
3922
+ "step": 6490
3923
+ },
3924
+ {
3925
+ "epoch": 2.95,
3926
+ "learning_rate": 3.0769230769230774e-07,
3927
+ "loss": 0.0008,
3928
+ "step": 6500
3929
+ }
3930
+ ],
3931
+ "max_steps": 6600,
3932
+ "num_train_epochs": 3,
3933
+ "total_flos": 1.368423174144e+16,
3934
+ "trial_name": null,
3935
+ "trial_params": null
3936
+ }
vocab.txt ADDED
The diff for this file is too large to render. See raw diff