Hugging Face's logo

dheerajpai
/
indicner-oriya-finetuned

Token Classification
Transformers
PyTorch
albert
Inference Endpoints
Model card Files Community
dheerajpai commited on
Commit
744ad18
·
1 Parent(s): 46873f2

Upload 12 files

Browse files
Files changed (13) hide show
  1. .gitattributes +1 -0
  2. added_tokens.json +7 -0
  3. config.json +51 -0
  4. optimizer.pt +3 -0
  5. pytorch_model.bin +3 -0
  6. rng_state.pth +3 -0
  7. scheduler.pt +3 -0
  8. special_tokens_map.json +9 -0
  9. spiece.model +3 -0
  10. tokenizer.json +3 -0
  11. tokenizer_config.json +59 -0
  12. trainer_state.json +4321 -0
  13. training_args.bin +3 -0
.gitattributes CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ tokenizer.json filter=lfs diff=lfs merge=lfs -text
added_tokens.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "<pad>": 0,
3
+ "<unk>": 1,
4
+ "[CLS]": 2,
5
+ "[MASK]": 4,
6
+ "[SEP]": 3
7
+ }
config.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "ai4bharat/indic-bert",
3
+ "architectures": [
4
+ "AlbertForTokenClassification"
5
+ ],
6
+ "attention_probs_dropout_prob": 0,
7
+ "bos_token_id": 2,
8
+ "classifier_dropout_prob": 0.1,
9
+ "down_scale_factor": 1,
10
+ "embedding_size": 128,
11
+ "eos_token_id": 3,
12
+ "gap_size": 0,
13
+ "hidden_act": "gelu",
14
+ "hidden_dropout_prob": 0,
15
+ "hidden_size": 768,
16
+ "id2label": {
17
+ "0": "LABEL_0",
18
+ "1": "LABEL_1",
19
+ "2": "LABEL_2",
20
+ "3": "LABEL_3",
21
+ "4": "LABEL_4",
22
+ "5": "LABEL_5",
23
+ "6": "LABEL_6"
24
+ },
25
+ "initializer_range": 0.02,
26
+ "inner_group_num": 1,
27
+ "intermediate_size": 3072,
28
+ "label2id": {
29
+ "LABEL_0": 0,
30
+ "LABEL_1": 1,
31
+ "LABEL_2": 2,
32
+ "LABEL_3": 3,
33
+ "LABEL_4": 4,
34
+ "LABEL_5": 5,
35
+ "LABEL_6": 6
36
+ },
37
+ "layer_norm_eps": 1e-12,
38
+ "max_position_embeddings": 512,
39
+ "model_type": "albert",
40
+ "net_structure_type": 0,
41
+ "num_attention_heads": 12,
42
+ "num_hidden_groups": 1,
43
+ "num_hidden_layers": 12,
44
+ "num_memory_blocks": 0,
45
+ "pad_token_id": 0,
46
+ "position_embedding_type": "absolute",
47
+ "torch_dtype": "float32",
48
+ "transformers_version": "4.34.0",
49
+ "type_vocab_size": 2,
50
+ "vocab_size": 200000
51
+ }
optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:68d7b9b051d981b462bddb775a9749de871604ff7a3c21c8cbd1a17ff1d1bd60
3
+ size 262888339
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3a46be1f22f79f1e5308761c894dc12e15c90812b172b60de58cf277efe2a48d
3
+ size 131442617
rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6010e7ddbf5a5aae8144d7425c7b15d7a52079ebc57deb1571ebb5421cd0498b
3
+ size 14244
scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:866b60f6c69cb6eec815d849808de1d627f8164d87d2f29c7a55271cf6e150e6
3
+ size 1064
special_tokens_map.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "[CLS]",
3
+ "cls_token": "[CLS]",
4
+ "eos_token": "[SEP]",
5
+ "mask_token": "[MASK]",
6
+ "pad_token": "<pad>",
7
+ "sep_token": "[SEP]",
8
+ "unk_token": "<unk>"
9
+ }
spiece.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3a1173c2b6e144a02c001e289a05b5dbefddf247c50d4dcf42633158b2968fcb
3
+ size 5646064
tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1380ca9cde3e5552469ae7f5c6b897faa4e96bfd39bd8a64ac5b4d05c5add12f
3
+ size 15285634
tokenizer_config.json ADDED
@@ -0,0 +1,59 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<pad>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<unk>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "[CLS]",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "[SEP]",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "4": {
36
+ "content": "[MASK]",
37
+ "lstrip": true,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ }
43
+ },
44
+ "additional_special_tokens": [],
45
+ "bos_token": "[CLS]",
46
+ "clean_up_tokenization_spaces": true,
47
+ "cls_token": "[CLS]",
48
+ "do_lower_case": true,
49
+ "eos_token": "[SEP]",
50
+ "keep_accents": false,
51
+ "mask_token": "[MASK]",
52
+ "model_max_length": 1000000000000000019884624838656,
53
+ "pad_token": "<pad>",
54
+ "remove_space": true,
55
+ "sep_token": "[SEP]",
56
+ "sp_model_kwargs": {},
57
+ "tokenizer_class": "AlbertTokenizer",
58
+ "unk_token": "<unk>"
59
+ }
trainer_state.json ADDED
@@ -0,0 +1,4321 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 0.9992697108389406,
5
+ "eval_steps": 500,
6
+ "global_step": 358500,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.0,
13
+ "learning_rate": 4.993031591974624e-05,
14
+ "loss": 0.6149,
15
+ "step": 500
16
+ },
17
+ {
18
+ "epoch": 0.0,
19
+ "learning_rate": 4.986063183949248e-05,
20
+ "loss": 0.4578,
21
+ "step": 1000
22
+ },
23
+ {
24
+ "epoch": 0.0,
25
+ "learning_rate": 4.9790947759238714e-05,
26
+ "loss": 0.4321,
27
+ "step": 1500
28
+ },
29
+ {
30
+ "epoch": 0.01,
31
+ "learning_rate": 4.972126367898496e-05,
32
+ "loss": 0.4103,
33
+ "step": 2000
34
+ },
35
+ {
36
+ "epoch": 0.01,
37
+ "learning_rate": 4.9651579598731195e-05,
38
+ "loss": 0.4003,
39
+ "step": 2500
40
+ },
41
+ {
42
+ "epoch": 0.01,
43
+ "learning_rate": 4.958189551847743e-05,
44
+ "loss": 0.3793,
45
+ "step": 3000
46
+ },
47
+ {
48
+ "epoch": 0.01,
49
+ "learning_rate": 4.951221143822367e-05,
50
+ "loss": 0.3678,
51
+ "step": 3500
52
+ },
53
+ {
54
+ "epoch": 0.01,
55
+ "learning_rate": 4.944252735796991e-05,
56
+ "loss": 0.386,
57
+ "step": 4000
58
+ },
59
+ {
60
+ "epoch": 0.01,
61
+ "learning_rate": 4.9372843277716144e-05,
62
+ "loss": 0.3704,
63
+ "step": 4500
64
+ },
65
+ {
66
+ "epoch": 0.01,
67
+ "learning_rate": 4.930315919746239e-05,
68
+ "loss": 0.3623,
69
+ "step": 5000
70
+ },
71
+ {
72
+ "epoch": 0.02,
73
+ "learning_rate": 4.9233475117208625e-05,
74
+ "loss": 0.3508,
75
+ "step": 5500
76
+ },
77
+ {
78
+ "epoch": 0.02,
79
+ "learning_rate": 4.916379103695486e-05,
80
+ "loss": 0.3505,
81
+ "step": 6000
82
+ },
83
+ {
84
+ "epoch": 0.02,
85
+ "learning_rate": 4.90941069567011e-05,
86
+ "loss": 0.3406,
87
+ "step": 6500
88
+ },
89
+ {
90
+ "epoch": 0.02,
91
+ "learning_rate": 4.902442287644734e-05,
92
+ "loss": 0.337,
93
+ "step": 7000
94
+ },
95
+ {
96
+ "epoch": 0.02,
97
+ "learning_rate": 4.895473879619358e-05,
98
+ "loss": 0.342,
99
+ "step": 7500
100
+ },
101
+ {
102
+ "epoch": 0.02,
103
+ "learning_rate": 4.888505471593982e-05,
104
+ "loss": 0.3322,
105
+ "step": 8000
106
+ },
107
+ {
108
+ "epoch": 0.02,
109
+ "learning_rate": 4.881537063568606e-05,
110
+ "loss": 0.3352,
111
+ "step": 8500
112
+ },
113
+ {
114
+ "epoch": 0.03,
115
+ "learning_rate": 4.87456865554323e-05,
116
+ "loss": 0.3293,
117
+ "step": 9000
118
+ },
119
+ {
120
+ "epoch": 0.03,
121
+ "learning_rate": 4.8676002475178536e-05,
122
+ "loss": 0.3249,
123
+ "step": 9500
124
+ },
125
+ {
126
+ "epoch": 0.03,
127
+ "learning_rate": 4.860631839492477e-05,
128
+ "loss": 0.3265,
129
+ "step": 10000
130
+ },
131
+ {
132
+ "epoch": 0.03,
133
+ "learning_rate": 4.853663431467101e-05,
134
+ "loss": 0.3125,
135
+ "step": 10500
136
+ },
137
+ {
138
+ "epoch": 0.03,
139
+ "learning_rate": 4.846695023441725e-05,
140
+ "loss": 0.3207,
141
+ "step": 11000
142
+ },
143
+ {
144
+ "epoch": 0.03,
145
+ "learning_rate": 4.8397266154163485e-05,
146
+ "loss": 0.3271,
147
+ "step": 11500
148
+ },
149
+ {
150
+ "epoch": 0.03,
151
+ "learning_rate": 4.832758207390973e-05,
152
+ "loss": 0.3187,
153
+ "step": 12000
154
+ },
155
+ {
156
+ "epoch": 0.03,
157
+ "learning_rate": 4.8257897993655966e-05,
158
+ "loss": 0.3163,
159
+ "step": 12500
160
+ },
161
+ {
162
+ "epoch": 0.04,
163
+ "learning_rate": 4.81882139134022e-05,
164
+ "loss": 0.3197,
165
+ "step": 13000
166
+ },
167
+ {
168
+ "epoch": 0.04,
169
+ "learning_rate": 4.811852983314844e-05,
170
+ "loss": 0.3132,
171
+ "step": 13500
172
+ },
173
+ {
174
+ "epoch": 0.04,
175
+ "learning_rate": 4.804884575289468e-05,
176
+ "loss": 0.306,
177
+ "step": 14000
178
+ },
179
+ {
180
+ "epoch": 0.04,
181
+ "learning_rate": 4.7979161672640915e-05,
182
+ "loss": 0.3053,
183
+ "step": 14500
184
+ },
185
+ {
186
+ "epoch": 0.04,
187
+ "learning_rate": 4.790947759238716e-05,
188
+ "loss": 0.3011,
189
+ "step": 15000
190
+ },
191
+ {
192
+ "epoch": 0.04,
193
+ "learning_rate": 4.7839793512133396e-05,
194
+ "loss": 0.3033,
195
+ "step": 15500
196
+ },
197
+ {
198
+ "epoch": 0.04,
199
+ "learning_rate": 4.777010943187963e-05,
200
+ "loss": 0.3103,
201
+ "step": 16000
202
+ },
203
+ {
204
+ "epoch": 0.05,
205
+ "learning_rate": 4.770042535162587e-05,
206
+ "loss": 0.3059,
207
+ "step": 16500
208
+ },
209
+ {
210
+ "epoch": 0.05,
211
+ "learning_rate": 4.763074127137211e-05,
212
+ "loss": 0.298,
213
+ "step": 17000
214
+ },
215
+ {
216
+ "epoch": 0.05,
217
+ "learning_rate": 4.7561057191118345e-05,
218
+ "loss": 0.2967,
219
+ "step": 17500
220
+ },
221
+ {
222
+ "epoch": 0.05,
223
+ "learning_rate": 4.749137311086458e-05,
224
+ "loss": 0.2844,
225
+ "step": 18000
226
+ },
227
+ {
228
+ "epoch": 0.05,
229
+ "learning_rate": 4.7421689030610826e-05,
230
+ "loss": 0.3001,
231
+ "step": 18500
232
+ },
233
+ {
234
+ "epoch": 0.05,
235
+ "learning_rate": 4.735200495035706e-05,
236
+ "loss": 0.3088,
237
+ "step": 19000
238
+ },
239
+ {
240
+ "epoch": 0.05,
241
+ "learning_rate": 4.72823208701033e-05,
242
+ "loss": 0.3079,
243
+ "step": 19500
244
+ },
245
+ {
246
+ "epoch": 0.06,
247
+ "learning_rate": 4.721263678984954e-05,
248
+ "loss": 0.2908,
249
+ "step": 20000
250
+ },
251
+ {
252
+ "epoch": 0.06,
253
+ "learning_rate": 4.714295270959578e-05,
254
+ "loss": 0.2879,
255
+ "step": 20500
256
+ },
257
+ {
258
+ "epoch": 0.06,
259
+ "learning_rate": 4.707326862934202e-05,
260
+ "loss": 0.2891,
261
+ "step": 21000
262
+ },
263
+ {
264
+ "epoch": 0.06,
265
+ "learning_rate": 4.7003584549088256e-05,
266
+ "loss": 0.3005,
267
+ "step": 21500
268
+ },
269
+ {
270
+ "epoch": 0.06,
271
+ "learning_rate": 4.69339004688345e-05,
272
+ "loss": 0.3008,
273
+ "step": 22000
274
+ },
275
+ {
276
+ "epoch": 0.06,
277
+ "learning_rate": 4.686421638858074e-05,
278
+ "loss": 0.2914,
279
+ "step": 22500
280
+ },
281
+ {
282
+ "epoch": 0.06,
283
+ "learning_rate": 4.6794532308326974e-05,
284
+ "loss": 0.2864,
285
+ "step": 23000
286
+ },
287
+ {
288
+ "epoch": 0.07,
289
+ "learning_rate": 4.672484822807321e-05,
290
+ "loss": 0.2885,
291
+ "step": 23500
292
+ },
293
+ {
294
+ "epoch": 0.07,
295
+ "learning_rate": 4.665516414781945e-05,
296
+ "loss": 0.289,
297
+ "step": 24000
298
+ },
299
+ {
300
+ "epoch": 0.07,
301
+ "learning_rate": 4.6585480067565685e-05,
302
+ "loss": 0.283,
303
+ "step": 24500
304
+ },
305
+ {
306
+ "epoch": 0.07,
307
+ "learning_rate": 4.651579598731193e-05,
308
+ "loss": 0.2917,
309
+ "step": 25000
310
+ },
311
+ {
312
+ "epoch": 0.07,
313
+ "learning_rate": 4.6446111907058167e-05,
314
+ "loss": 0.2895,
315
+ "step": 25500
316
+ },
317
+ {
318
+ "epoch": 0.07,
319
+ "learning_rate": 4.6376427826804404e-05,
320
+ "loss": 0.276,
321
+ "step": 26000
322
+ },
323
+ {
324
+ "epoch": 0.07,
325
+ "learning_rate": 4.630674374655064e-05,
326
+ "loss": 0.2847,
327
+ "step": 26500
328
+ },
329
+ {
330
+ "epoch": 0.08,
331
+ "learning_rate": 4.623705966629688e-05,
332
+ "loss": 0.2774,
333
+ "step": 27000
334
+ },
335
+ {
336
+ "epoch": 0.08,
337
+ "learning_rate": 4.6167375586043115e-05,
338
+ "loss": 0.283,
339
+ "step": 27500
340
+ },
341
+ {
342
+ "epoch": 0.08,
343
+ "learning_rate": 4.609769150578935e-05,
344
+ "loss": 0.2859,
345
+ "step": 28000
346
+ },
347
+ {
348
+ "epoch": 0.08,
349
+ "learning_rate": 4.6028007425535596e-05,
350
+ "loss": 0.2793,
351
+ "step": 28500
352
+ },
353
+ {
354
+ "epoch": 0.08,
355
+ "learning_rate": 4.5958323345281834e-05,
356
+ "loss": 0.2769,
357
+ "step": 29000
358
+ },
359
+ {
360
+ "epoch": 0.08,
361
+ "learning_rate": 4.588863926502807e-05,
362
+ "loss": 0.2763,
363
+ "step": 29500
364
+ },
365
+ {
366
+ "epoch": 0.08,
367
+ "learning_rate": 4.581895518477431e-05,
368
+ "loss": 0.2886,
369
+ "step": 30000
370
+ },
371
+ {
372
+ "epoch": 0.09,
373
+ "learning_rate": 4.5749271104520545e-05,
374
+ "loss": 0.2721,
375
+ "step": 30500
376
+ },
377
+ {
378
+ "epoch": 0.09,
379
+ "learning_rate": 4.567958702426678e-05,
380
+ "loss": 0.2805,
381
+ "step": 31000
382
+ },
383
+ {
384
+ "epoch": 0.09,
385
+ "learning_rate": 4.560990294401302e-05,
386
+ "loss": 0.2779,
387
+ "step": 31500
388
+ },
389
+ {
390
+ "epoch": 0.09,
391
+ "learning_rate": 4.5540218863759264e-05,
392
+ "loss": 0.2833,
393
+ "step": 32000
394
+ },
395
+ {
396
+ "epoch": 0.09,
397
+ "learning_rate": 4.54705347835055e-05,
398
+ "loss": 0.2881,
399
+ "step": 32500
400
+ },
401
+ {
402
+ "epoch": 0.09,
403
+ "learning_rate": 4.540085070325174e-05,
404
+ "loss": 0.2802,
405
+ "step": 33000
406
+ },
407
+ {
408
+ "epoch": 0.09,
409
+ "learning_rate": 4.5331166622997975e-05,
410
+ "loss": 0.2683,
411
+ "step": 33500
412
+ },
413
+ {
414
+ "epoch": 0.09,
415
+ "learning_rate": 4.526148254274422e-05,
416
+ "loss": 0.2732,
417
+ "step": 34000
418
+ },
419
+ {
420
+ "epoch": 0.1,
421
+ "learning_rate": 4.5191798462490456e-05,
422
+ "loss": 0.2773,
423
+ "step": 34500
424
+ },
425
+ {
426
+ "epoch": 0.1,
427
+ "learning_rate": 4.5122114382236693e-05,
428
+ "loss": 0.2791,
429
+ "step": 35000
430
+ },
431
+ {
432
+ "epoch": 0.1,
433
+ "learning_rate": 4.505243030198294e-05,
434
+ "loss": 0.2842,
435
+ "step": 35500
436
+ },
437
+ {
438
+ "epoch": 0.1,
439
+ "learning_rate": 4.4982746221729175e-05,
440
+ "loss": 0.278,
441
+ "step": 36000
442
+ },
443
+ {
444
+ "epoch": 0.1,
445
+ "learning_rate": 4.491306214147541e-05,
446
+ "loss": 0.2757,
447
+ "step": 36500
448
+ },
449
+ {
450
+ "epoch": 0.1,
451
+ "learning_rate": 4.484337806122165e-05,
452
+ "loss": 0.2741,
453
+ "step": 37000
454
+ },
455
+ {
456
+ "epoch": 0.1,
457
+ "learning_rate": 4.4773693980967886e-05,
458
+ "loss": 0.2708,
459
+ "step": 37500
460
+ },
461
+ {
462
+ "epoch": 0.11,
463
+ "learning_rate": 4.470400990071412e-05,
464
+ "loss": 0.2725,
465
+ "step": 38000
466
+ },
467
+ {
468
+ "epoch": 0.11,
469
+ "learning_rate": 4.463432582046037e-05,
470
+ "loss": 0.2735,
471
+ "step": 38500
472
+ },
473
+ {
474
+ "epoch": 0.11,
475
+ "learning_rate": 4.4564641740206604e-05,
476
+ "loss": 0.2735,
477
+ "step": 39000
478
+ },
479
+ {
480
+ "epoch": 0.11,
481
+ "learning_rate": 4.449495765995284e-05,
482
+ "loss": 0.2704,
483
+ "step": 39500
484
+ },
485
+ {
486
+ "epoch": 0.11,
487
+ "learning_rate": 4.442527357969908e-05,
488
+ "loss": 0.2757,
489
+ "step": 40000
490
+ },
491
+ {
492
+ "epoch": 0.11,
493
+ "learning_rate": 4.4355589499445316e-05,
494
+ "loss": 0.2851,
495
+ "step": 40500
496
+ },
497
+ {
498
+ "epoch": 0.11,
499
+ "learning_rate": 4.428590541919155e-05,
500
+ "loss": 0.2681,
501
+ "step": 41000
502
+ },
503
+ {
504
+ "epoch": 0.12,
505
+ "learning_rate": 4.421622133893779e-05,
506
+ "loss": 0.27,
507
+ "step": 41500
508
+ },
509
+ {
510
+ "epoch": 0.12,
511
+ "learning_rate": 4.4146537258684034e-05,
512
+ "loss": 0.2701,
513
+ "step": 42000
514
+ },
515
+ {
516
+ "epoch": 0.12,
517
+ "learning_rate": 4.407685317843027e-05,
518
+ "loss": 0.2648,
519
+ "step": 42500
520
+ },
521
+ {
522
+ "epoch": 0.12,
523
+ "learning_rate": 4.400716909817651e-05,
524
+ "loss": 0.2698,
525
+ "step": 43000
526
+ },
527
+ {
528
+ "epoch": 0.12,
529
+ "learning_rate": 4.3937485017922746e-05,
530
+ "loss": 0.2693,
531
+ "step": 43500
532
+ },
533
+ {
534
+ "epoch": 0.12,
535
+ "learning_rate": 4.386780093766898e-05,
536
+ "loss": 0.2665,
537
+ "step": 44000
538
+ },
539
+ {
540
+ "epoch": 0.12,
541
+ "learning_rate": 4.379811685741522e-05,
542
+ "loss": 0.2688,
543
+ "step": 44500
544
+ },
545
+ {
546
+ "epoch": 0.13,
547
+ "learning_rate": 4.3728432777161464e-05,
548
+ "loss": 0.2689,
549
+ "step": 45000
550
+ },
551
+ {
552
+ "epoch": 0.13,
553
+ "learning_rate": 4.36587486969077e-05,
554
+ "loss": 0.273,
555
+ "step": 45500
556
+ },
557
+ {
558
+ "epoch": 0.13,
559
+ "learning_rate": 4.358906461665394e-05,
560
+ "loss": 0.2717,
561
+ "step": 46000
562
+ },
563
+ {
564
+ "epoch": 0.13,
565
+ "learning_rate": 4.3519380536400176e-05,
566
+ "loss": 0.2691,
567
+ "step": 46500
568
+ },
569
+ {
570
+ "epoch": 0.13,
571
+ "learning_rate": 4.344969645614641e-05,
572
+ "loss": 0.2596,
573
+ "step": 47000
574
+ },
575
+ {
576
+ "epoch": 0.13,
577
+ "learning_rate": 4.338001237589266e-05,
578
+ "loss": 0.2639,
579
+ "step": 47500
580
+ },
581
+ {
582
+ "epoch": 0.13,
583
+ "learning_rate": 4.3310328295638894e-05,
584
+ "loss": 0.2724,
585
+ "step": 48000
586
+ },
587
+ {
588
+ "epoch": 0.14,
589
+ "learning_rate": 4.324064421538513e-05,
590
+ "loss": 0.2572,
591
+ "step": 48500
592
+ },
593
+ {
594
+ "epoch": 0.14,
595
+ "learning_rate": 4.3170960135131375e-05,
596
+ "loss": 0.2657,
597
+ "step": 49000
598
+ },
599
+ {
600
+ "epoch": 0.14,
601
+ "learning_rate": 4.310127605487761e-05,
602
+ "loss": 0.2681,
603
+ "step": 49500
604
+ },
605
+ {
606
+ "epoch": 0.14,
607
+ "learning_rate": 4.303159197462385e-05,
608
+ "loss": 0.2654,
609
+ "step": 50000
610
+ },
611
+ {
612
+ "epoch": 0.14,
613
+ "learning_rate": 4.296190789437009e-05,
614
+ "loss": 0.2655,
615
+ "step": 50500
616
+ },
617
+ {
618
+ "epoch": 0.14,
619
+ "learning_rate": 4.2892223814116324e-05,
620
+ "loss": 0.2658,
621
+ "step": 51000
622
+ },
623
+ {
624
+ "epoch": 0.14,
625
+ "learning_rate": 4.282253973386256e-05,
626
+ "loss": 0.2582,
627
+ "step": 51500
628
+ },
629
+ {
630
+ "epoch": 0.14,
631
+ "learning_rate": 4.2752855653608805e-05,
632
+ "loss": 0.2664,
633
+ "step": 52000
634
+ },
635
+ {
636
+ "epoch": 0.15,
637
+ "learning_rate": 4.268317157335504e-05,
638
+ "loss": 0.2508,
639
+ "step": 52500
640
+ },
641
+ {
642
+ "epoch": 0.15,
643
+ "learning_rate": 4.261348749310128e-05,
644
+ "loss": 0.2569,
645
+ "step": 53000
646
+ },
647
+ {
648
+ "epoch": 0.15,
649
+ "learning_rate": 4.254380341284752e-05,
650
+ "loss": 0.2592,
651
+ "step": 53500
652
+ },
653
+ {
654
+ "epoch": 0.15,
655
+ "learning_rate": 4.2474119332593754e-05,
656
+ "loss": 0.2645,
657
+ "step": 54000
658
+ },
659
+ {
660
+ "epoch": 0.15,
661
+ "learning_rate": 4.240443525233999e-05,
662
+ "loss": 0.2597,
663
+ "step": 54500
664
+ },
665
+ {
666
+ "epoch": 0.15,
667
+ "learning_rate": 4.2334751172086235e-05,
668
+ "loss": 0.2545,
669
+ "step": 55000
670
+ },
671
+ {
672
+ "epoch": 0.15,
673
+ "learning_rate": 4.226506709183247e-05,
674
+ "loss": 0.2652,
675
+ "step": 55500
676
+ },
677
+ {
678
+ "epoch": 0.16,
679
+ "learning_rate": 4.219538301157871e-05,
680
+ "loss": 0.2677,
681
+ "step": 56000
682
+ },
683
+ {
684
+ "epoch": 0.16,
685
+ "learning_rate": 4.2125698931324947e-05,
686
+ "loss": 0.2671,
687
+ "step": 56500
688
+ },
689
+ {
690
+ "epoch": 0.16,
691
+ "learning_rate": 4.2056014851071184e-05,
692
+ "loss": 0.2601,
693
+ "step": 57000
694
+ },
695
+ {
696
+ "epoch": 0.16,
697
+ "learning_rate": 4.198633077081742e-05,
698
+ "loss": 0.2728,
699
+ "step": 57500
700
+ },
701
+ {
702
+ "epoch": 0.16,
703
+ "learning_rate": 4.191664669056366e-05,
704
+ "loss": 0.2628,
705
+ "step": 58000
706
+ },
707
+ {
708
+ "epoch": 0.16,
709
+ "learning_rate": 4.18469626103099e-05,
710
+ "loss": 0.2628,
711
+ "step": 58500
712
+ },
713
+ {
714
+ "epoch": 0.16,
715
+ "learning_rate": 4.177727853005614e-05,
716
+ "loss": 0.2508,
717
+ "step": 59000
718
+ },
719
+ {
720
+ "epoch": 0.17,
721
+ "learning_rate": 4.1707594449802376e-05,
722
+ "loss": 0.2661,
723
+ "step": 59500
724
+ },
725
+ {
726
+ "epoch": 0.17,
727
+ "learning_rate": 4.1637910369548614e-05,
728
+ "loss": 0.2575,
729
+ "step": 60000
730
+ },
731
+ {
732
+ "epoch": 0.17,
733
+ "learning_rate": 4.156822628929485e-05,
734
+ "loss": 0.2537,
735
+ "step": 60500
736
+ },
737
+ {
738
+ "epoch": 0.17,
739
+ "learning_rate": 4.149854220904109e-05,
740
+ "loss": 0.2495,
741
+ "step": 61000
742
+ },
743
+ {
744
+ "epoch": 0.17,
745
+ "learning_rate": 4.142885812878733e-05,
746
+ "loss": 0.2564,
747
+ "step": 61500
748
+ },
749
+ {
750
+ "epoch": 0.17,
751
+ "learning_rate": 4.135917404853357e-05,
752
+ "loss": 0.2515,
753
+ "step": 62000
754
+ },
755
+ {
756
+ "epoch": 0.17,
757
+ "learning_rate": 4.128948996827981e-05,
758
+ "loss": 0.2582,
759
+ "step": 62500
760
+ },
761
+ {
762
+ "epoch": 0.18,
763
+ "learning_rate": 4.121980588802605e-05,
764
+ "loss": 0.266,
765
+ "step": 63000
766
+ },
767
+ {
768
+ "epoch": 0.18,
769
+ "learning_rate": 4.115012180777229e-05,
770
+ "loss": 0.2492,
771
+ "step": 63500
772
+ },
773
+ {
774
+ "epoch": 0.18,
775
+ "learning_rate": 4.1080437727518525e-05,
776
+ "loss": 0.2677,
777
+ "step": 64000
778
+ },
779
+ {
780
+ "epoch": 0.18,
781
+ "learning_rate": 4.101075364726476e-05,
782
+ "loss": 0.257,
783
+ "step": 64500
784
+ },
785
+ {
786
+ "epoch": 0.18,
787
+ "learning_rate": 4.0941069567011006e-05,
788
+ "loss": 0.2586,
789
+ "step": 65000
790
+ },
791
+ {
792
+ "epoch": 0.18,
793
+ "learning_rate": 4.087138548675724e-05,
794
+ "loss": 0.2571,
795
+ "step": 65500
796
+ },
797
+ {
798
+ "epoch": 0.18,
799
+ "learning_rate": 4.080170140650348e-05,
800
+ "loss": 0.2549,
801
+ "step": 66000
802
+ },
803
+ {
804
+ "epoch": 0.19,
805
+ "learning_rate": 4.073201732624972e-05,
806
+ "loss": 0.258,
807
+ "step": 66500
808
+ },
809
+ {
810
+ "epoch": 0.19,
811
+ "learning_rate": 4.0662333245995955e-05,
812
+ "loss": 0.2594,
813
+ "step": 67000
814
+ },
815
+ {
816
+ "epoch": 0.19,
817
+ "learning_rate": 4.059264916574219e-05,
818
+ "loss": 0.2564,
819
+ "step": 67500
820
+ },
821
+ {
822
+ "epoch": 0.19,
823
+ "learning_rate": 4.052296508548843e-05,
824
+ "loss": 0.2646,
825
+ "step": 68000
826
+ },
827
+ {
828
+ "epoch": 0.19,
829
+ "learning_rate": 4.045328100523467e-05,
830
+ "loss": 0.2577,
831
+ "step": 68500
832
+ },
833
+ {
834
+ "epoch": 0.19,
835
+ "learning_rate": 4.038359692498091e-05,
836
+ "loss": 0.2567,
837
+ "step": 69000
838
+ },
839
+ {
840
+ "epoch": 0.19,
841
+ "learning_rate": 4.031391284472715e-05,
842
+ "loss": 0.2518,
843
+ "step": 69500
844
+ },
845
+ {
846
+ "epoch": 0.2,
847
+ "learning_rate": 4.0244228764473384e-05,
848
+ "loss": 0.2541,
849
+ "step": 70000
850
+ },
851
+ {
852
+ "epoch": 0.2,
853
+ "learning_rate": 4.017454468421962e-05,
854
+ "loss": 0.2504,
855
+ "step": 70500
856
+ },
857
+ {
858
+ "epoch": 0.2,
859
+ "learning_rate": 4.010486060396586e-05,
860
+ "loss": 0.2595,
861
+ "step": 71000
862
+ },
863
+ {
864
+ "epoch": 0.2,
865
+ "learning_rate": 4.00351765237121e-05,
866
+ "loss": 0.2618,
867
+ "step": 71500
868
+ },
869
+ {
870
+ "epoch": 0.2,
871
+ "learning_rate": 3.996549244345834e-05,
872
+ "loss": 0.2604,
873
+ "step": 72000
874
+ },
875
+ {
876
+ "epoch": 0.2,
877
+ "learning_rate": 3.989580836320458e-05,
878
+ "loss": 0.2525,
879
+ "step": 72500
880
+ },
881
+ {
882
+ "epoch": 0.2,
883
+ "learning_rate": 3.9826124282950814e-05,
884
+ "loss": 0.2447,
885
+ "step": 73000
886
+ },
887
+ {
888
+ "epoch": 0.2,
889
+ "learning_rate": 3.975644020269705e-05,
890
+ "loss": 0.2535,
891
+ "step": 73500
892
+ },
893
+ {
894
+ "epoch": 0.21,
895
+ "learning_rate": 3.968675612244329e-05,
896
+ "loss": 0.256,
897
+ "step": 74000
898
+ },
899
+ {
900
+ "epoch": 0.21,
901
+ "learning_rate": 3.9617072042189526e-05,
902
+ "loss": 0.2475,
903
+ "step": 74500
904
+ },
905
+ {
906
+ "epoch": 0.21,
907
+ "learning_rate": 3.954738796193577e-05,
908
+ "loss": 0.2591,
909
+ "step": 75000
910
+ },
911
+ {
912
+ "epoch": 0.21,
913
+ "learning_rate": 3.947770388168201e-05,
914
+ "loss": 0.2464,
915
+ "step": 75500
916
+ },
917
+ {
918
+ "epoch": 0.21,
919
+ "learning_rate": 3.9408019801428244e-05,
920
+ "loss": 0.2535,
921
+ "step": 76000
922
+ },
923
+ {
924
+ "epoch": 0.21,
925
+ "learning_rate": 3.933833572117449e-05,
926
+ "loss": 0.2514,
927
+ "step": 76500
928
+ },
929
+ {
930
+ "epoch": 0.21,
931
+ "learning_rate": 3.9268651640920725e-05,
932
+ "loss": 0.2484,
933
+ "step": 77000
934
+ },
935
+ {
936
+ "epoch": 0.22,
937
+ "learning_rate": 3.919896756066696e-05,
938
+ "loss": 0.2483,
939
+ "step": 77500
940
+ },
941
+ {
942
+ "epoch": 0.22,
943
+ "learning_rate": 3.91292834804132e-05,
944
+ "loss": 0.2559,
945
+ "step": 78000
946
+ },
947
+ {
948
+ "epoch": 0.22,
949
+ "learning_rate": 3.9059599400159444e-05,
950
+ "loss": 0.2526,
951
+ "step": 78500
952
+ },
953
+ {
954
+ "epoch": 0.22,
955
+ "learning_rate": 3.898991531990568e-05,
956
+ "loss": 0.2537,
957
+ "step": 79000
958
+ },
959
+ {
960
+ "epoch": 0.22,
961
+ "learning_rate": 3.892023123965192e-05,
962
+ "loss": 0.2586,
963
+ "step": 79500
964
+ },
965
+ {
966
+ "epoch": 0.22,
967
+ "learning_rate": 3.8850547159398155e-05,
968
+ "loss": 0.2407,
969
+ "step": 80000
970
+ },
971
+ {
972
+ "epoch": 0.22,
973
+ "learning_rate": 3.878086307914439e-05,
974
+ "loss": 0.2518,
975
+ "step": 80500
976
+ },
977
+ {
978
+ "epoch": 0.23,
979
+ "learning_rate": 3.871117899889063e-05,
980
+ "loss": 0.254,
981
+ "step": 81000
982
+ },
983
+ {
984
+ "epoch": 0.23,
985
+ "learning_rate": 3.8641494918636874e-05,
986
+ "loss": 0.2517,
987
+ "step": 81500
988
+ },
989
+ {
990
+ "epoch": 0.23,
991
+ "learning_rate": 3.857181083838311e-05,
992
+ "loss": 0.248,
993
+ "step": 82000
994
+ },
995
+ {
996
+ "epoch": 0.23,
997
+ "learning_rate": 3.850212675812935e-05,
998
+ "loss": 0.2483,
999
+ "step": 82500
1000
+ },
1001
+ {
1002
+ "epoch": 0.23,
1003
+ "learning_rate": 3.8432442677875585e-05,
1004
+ "loss": 0.2465,
1005
+ "step": 83000
1006
+ },
1007
+ {
1008
+ "epoch": 0.23,
1009
+ "learning_rate": 3.836275859762182e-05,
1010
+ "loss": 0.243,
1011
+ "step": 83500
1012
+ },
1013
+ {
1014
+ "epoch": 0.23,
1015
+ "learning_rate": 3.829307451736806e-05,
1016
+ "loss": 0.2518,
1017
+ "step": 84000
1018
+ },
1019
+ {
1020
+ "epoch": 0.24,
1021
+ "learning_rate": 3.82233904371143e-05,
1022
+ "loss": 0.2548,
1023
+ "step": 84500
1024
+ },
1025
+ {
1026
+ "epoch": 0.24,
1027
+ "learning_rate": 3.815370635686054e-05,
1028
+ "loss": 0.2492,
1029
+ "step": 85000
1030
+ },
1031
+ {
1032
+ "epoch": 0.24,
1033
+ "learning_rate": 3.808402227660678e-05,
1034
+ "loss": 0.2444,
1035
+ "step": 85500
1036
+ },
1037
+ {
1038
+ "epoch": 0.24,
1039
+ "learning_rate": 3.8014338196353015e-05,
1040
+ "loss": 0.2436,
1041
+ "step": 86000
1042
+ },
1043
+ {
1044
+ "epoch": 0.24,
1045
+ "learning_rate": 3.794465411609925e-05,
1046
+ "loss": 0.2494,
1047
+ "step": 86500
1048
+ },
1049
+ {
1050
+ "epoch": 0.24,
1051
+ "learning_rate": 3.787497003584549e-05,
1052
+ "loss": 0.2595,
1053
+ "step": 87000
1054
+ },
1055
+ {
1056
+ "epoch": 0.24,
1057
+ "learning_rate": 3.7805285955591727e-05,
1058
+ "loss": 0.2507,
1059
+ "step": 87500
1060
+ },
1061
+ {
1062
+ "epoch": 0.25,
1063
+ "learning_rate": 3.7735601875337964e-05,
1064
+ "loss": 0.2388,
1065
+ "step": 88000
1066
+ },
1067
+ {
1068
+ "epoch": 0.25,
1069
+ "learning_rate": 3.766591779508421e-05,
1070
+ "loss": 0.2455,
1071
+ "step": 88500
1072
+ },
1073
+ {
1074
+ "epoch": 0.25,
1075
+ "learning_rate": 3.7596233714830445e-05,
1076
+ "loss": 0.2364,
1077
+ "step": 89000
1078
+ },
1079
+ {
1080
+ "epoch": 0.25,
1081
+ "learning_rate": 3.752654963457668e-05,
1082
+ "loss": 0.2433,
1083
+ "step": 89500
1084
+ },
1085
+ {
1086
+ "epoch": 0.25,
1087
+ "learning_rate": 3.7456865554322926e-05,
1088
+ "loss": 0.2519,
1089
+ "step": 90000
1090
+ },
1091
+ {
1092
+ "epoch": 0.25,
1093
+ "learning_rate": 3.738718147406916e-05,
1094
+ "loss": 0.2495,
1095
+ "step": 90500
1096
+ },
1097
+ {
1098
+ "epoch": 0.25,
1099
+ "learning_rate": 3.73174973938154e-05,
1100
+ "loss": 0.2405,
1101
+ "step": 91000
1102
+ },
1103
+ {
1104
+ "epoch": 0.26,
1105
+ "learning_rate": 3.7247813313561644e-05,
1106
+ "loss": 0.2431,
1107
+ "step": 91500
1108
+ },
1109
+ {
1110
+ "epoch": 0.26,
1111
+ "learning_rate": 3.717812923330788e-05,
1112
+ "loss": 0.2474,
1113
+ "step": 92000
1114
+ },
1115
+ {
1116
+ "epoch": 0.26,
1117
+ "learning_rate": 3.710844515305412e-05,
1118
+ "loss": 0.2468,
1119
+ "step": 92500
1120
+ },
1121
+ {
1122
+ "epoch": 0.26,
1123
+ "learning_rate": 3.7038761072800356e-05,
1124
+ "loss": 0.2464,
1125
+ "step": 93000
1126
+ },
1127
+ {
1128
+ "epoch": 0.26,
1129
+ "learning_rate": 3.696907699254659e-05,
1130
+ "loss": 0.2464,
1131
+ "step": 93500
1132
+ },
1133
+ {
1134
+ "epoch": 0.26,
1135
+ "learning_rate": 3.689939291229283e-05,
1136
+ "loss": 0.2424,
1137
+ "step": 94000
1138
+ },
1139
+ {
1140
+ "epoch": 0.26,
1141
+ "learning_rate": 3.682970883203907e-05,
1142
+ "loss": 0.2379,
1143
+ "step": 94500
1144
+ },
1145
+ {
1146
+ "epoch": 0.26,
1147
+ "learning_rate": 3.676002475178531e-05,
1148
+ "loss": 0.2442,
1149
+ "step": 95000
1150
+ },
1151
+ {
1152
+ "epoch": 0.27,
1153
+ "learning_rate": 3.669034067153155e-05,
1154
+ "loss": 0.246,
1155
+ "step": 95500
1156
+ },
1157
+ {
1158
+ "epoch": 0.27,
1159
+ "learning_rate": 3.6620656591277786e-05,
1160
+ "loss": 0.2403,
1161
+ "step": 96000
1162
+ },
1163
+ {
1164
+ "epoch": 0.27,
1165
+ "learning_rate": 3.655097251102402e-05,
1166
+ "loss": 0.2429,
1167
+ "step": 96500
1168
+ },
1169
+ {
1170
+ "epoch": 0.27,
1171
+ "learning_rate": 3.648128843077026e-05,
1172
+ "loss": 0.2482,
1173
+ "step": 97000
1174
+ },
1175
+ {
1176
+ "epoch": 0.27,
1177
+ "learning_rate": 3.64116043505165e-05,
1178
+ "loss": 0.2432,
1179
+ "step": 97500
1180
+ },
1181
+ {
1182
+ "epoch": 0.27,
1183
+ "learning_rate": 3.6341920270262735e-05,
1184
+ "loss": 0.2374,
1185
+ "step": 98000
1186
+ },
1187
+ {
1188
+ "epoch": 0.27,
1189
+ "learning_rate": 3.627223619000898e-05,
1190
+ "loss": 0.2539,
1191
+ "step": 98500
1192
+ },
1193
+ {
1194
+ "epoch": 0.28,
1195
+ "learning_rate": 3.6202552109755216e-05,
1196
+ "loss": 0.2496,
1197
+ "step": 99000
1198
+ },
1199
+ {
1200
+ "epoch": 0.28,
1201
+ "learning_rate": 3.613286802950145e-05,
1202
+ "loss": 0.2447,
1203
+ "step": 99500
1204
+ },
1205
+ {
1206
+ "epoch": 0.28,
1207
+ "learning_rate": 3.606318394924769e-05,
1208
+ "loss": 0.2396,
1209
+ "step": 100000
1210
+ },
1211
+ {
1212
+ "epoch": 0.28,
1213
+ "learning_rate": 3.599349986899393e-05,
1214
+ "loss": 0.2472,
1215
+ "step": 100500
1216
+ },
1217
+ {
1218
+ "epoch": 0.28,
1219
+ "learning_rate": 3.5923815788740164e-05,
1220
+ "loss": 0.2431,
1221
+ "step": 101000
1222
+ },
1223
+ {
1224
+ "epoch": 0.28,
1225
+ "learning_rate": 3.585413170848641e-05,
1226
+ "loss": 0.2414,
1227
+ "step": 101500
1228
+ },
1229
+ {
1230
+ "epoch": 0.28,
1231
+ "learning_rate": 3.5784447628232646e-05,
1232
+ "loss": 0.2398,
1233
+ "step": 102000
1234
+ },
1235
+ {
1236
+ "epoch": 0.29,
1237
+ "learning_rate": 3.571476354797888e-05,
1238
+ "loss": 0.2415,
1239
+ "step": 102500
1240
+ },
1241
+ {
1242
+ "epoch": 0.29,
1243
+ "learning_rate": 3.564507946772512e-05,
1244
+ "loss": 0.2428,
1245
+ "step": 103000
1246
+ },
1247
+ {
1248
+ "epoch": 0.29,
1249
+ "learning_rate": 3.5575395387471364e-05,
1250
+ "loss": 0.2341,
1251
+ "step": 103500
1252
+ },
1253
+ {
1254
+ "epoch": 0.29,
1255
+ "learning_rate": 3.55057113072176e-05,
1256
+ "loss": 0.2461,
1257
+ "step": 104000
1258
+ },
1259
+ {
1260
+ "epoch": 0.29,
1261
+ "learning_rate": 3.543602722696384e-05,
1262
+ "loss": 0.2481,
1263
+ "step": 104500
1264
+ },
1265
+ {
1266
+ "epoch": 0.29,
1267
+ "learning_rate": 3.536634314671008e-05,
1268
+ "loss": 0.2409,
1269
+ "step": 105000
1270
+ },
1271
+ {
1272
+ "epoch": 0.29,
1273
+ "learning_rate": 3.529665906645632e-05,
1274
+ "loss": 0.2388,
1275
+ "step": 105500
1276
+ },
1277
+ {
1278
+ "epoch": 0.3,
1279
+ "learning_rate": 3.522697498620256e-05,
1280
+ "loss": 0.2427,
1281
+ "step": 106000
1282
+ },
1283
+ {
1284
+ "epoch": 0.3,
1285
+ "learning_rate": 3.5157290905948794e-05,
1286
+ "loss": 0.238,
1287
+ "step": 106500
1288
+ },
1289
+ {
1290
+ "epoch": 0.3,
1291
+ "learning_rate": 3.508760682569503e-05,
1292
+ "loss": 0.2453,
1293
+ "step": 107000
1294
+ },
1295
+ {
1296
+ "epoch": 0.3,
1297
+ "learning_rate": 3.501792274544127e-05,
1298
+ "loss": 0.245,
1299
+ "step": 107500
1300
+ },
1301
+ {
1302
+ "epoch": 0.3,
1303
+ "learning_rate": 3.4948238665187505e-05,
1304
+ "loss": 0.2395,
1305
+ "step": 108000
1306
+ },
1307
+ {
1308
+ "epoch": 0.3,
1309
+ "learning_rate": 3.487855458493375e-05,
1310
+ "loss": 0.2369,
1311
+ "step": 108500
1312
+ },
1313
+ {
1314
+ "epoch": 0.3,
1315
+ "learning_rate": 3.4808870504679986e-05,
1316
+ "loss": 0.2369,
1317
+ "step": 109000
1318
+ },
1319
+ {
1320
+ "epoch": 0.31,
1321
+ "learning_rate": 3.4739186424426224e-05,
1322
+ "loss": 0.2476,
1323
+ "step": 109500
1324
+ },
1325
+ {
1326
+ "epoch": 0.31,
1327
+ "learning_rate": 3.466950234417246e-05,
1328
+ "loss": 0.238,
1329
+ "step": 110000
1330
+ },
1331
+ {
1332
+ "epoch": 0.31,
1333
+ "learning_rate": 3.45998182639187e-05,
1334
+ "loss": 0.2417,
1335
+ "step": 110500
1336
+ },
1337
+ {
1338
+ "epoch": 0.31,
1339
+ "learning_rate": 3.4530134183664935e-05,
1340
+ "loss": 0.2433,
1341
+ "step": 111000
1342
+ },
1343
+ {
1344
+ "epoch": 0.31,
1345
+ "learning_rate": 3.446045010341118e-05,
1346
+ "loss": 0.2431,
1347
+ "step": 111500
1348
+ },
1349
+ {
1350
+ "epoch": 0.31,
1351
+ "learning_rate": 3.4390766023157416e-05,
1352
+ "loss": 0.2434,
1353
+ "step": 112000
1354
+ },
1355
+ {
1356
+ "epoch": 0.31,
1357
+ "learning_rate": 3.4321081942903654e-05,
1358
+ "loss": 0.2459,
1359
+ "step": 112500
1360
+ },
1361
+ {
1362
+ "epoch": 0.31,
1363
+ "learning_rate": 3.425139786264989e-05,
1364
+ "loss": 0.2464,
1365
+ "step": 113000
1366
+ },
1367
+ {
1368
+ "epoch": 0.32,
1369
+ "learning_rate": 3.418171378239613e-05,
1370
+ "loss": 0.2357,
1371
+ "step": 113500
1372
+ },
1373
+ {
1374
+ "epoch": 0.32,
1375
+ "learning_rate": 3.4112029702142365e-05,
1376
+ "loss": 0.2394,
1377
+ "step": 114000
1378
+ },
1379
+ {
1380
+ "epoch": 0.32,
1381
+ "learning_rate": 3.40423456218886e-05,
1382
+ "loss": 0.2404,
1383
+ "step": 114500
1384
+ },
1385
+ {
1386
+ "epoch": 0.32,
1387
+ "learning_rate": 3.3972661541634846e-05,
1388
+ "loss": 0.2333,
1389
+ "step": 115000
1390
+ },
1391
+ {
1392
+ "epoch": 0.32,
1393
+ "learning_rate": 3.3902977461381083e-05,
1394
+ "loss": 0.2323,
1395
+ "step": 115500
1396
+ },
1397
+ {
1398
+ "epoch": 0.32,
1399
+ "learning_rate": 3.383329338112732e-05,
1400
+ "loss": 0.2562,
1401
+ "step": 116000
1402
+ },
1403
+ {
1404
+ "epoch": 0.32,
1405
+ "learning_rate": 3.376360930087356e-05,
1406
+ "loss": 0.238,
1407
+ "step": 116500
1408
+ },
1409
+ {
1410
+ "epoch": 0.33,
1411
+ "learning_rate": 3.36939252206198e-05,
1412
+ "loss": 0.2409,
1413
+ "step": 117000
1414
+ },
1415
+ {
1416
+ "epoch": 0.33,
1417
+ "learning_rate": 3.362424114036604e-05,
1418
+ "loss": 0.2435,
1419
+ "step": 117500
1420
+ },
1421
+ {
1422
+ "epoch": 0.33,
1423
+ "learning_rate": 3.3554557060112276e-05,
1424
+ "loss": 0.2377,
1425
+ "step": 118000
1426
+ },
1427
+ {
1428
+ "epoch": 0.33,
1429
+ "learning_rate": 3.348487297985852e-05,
1430
+ "loss": 0.2287,
1431
+ "step": 118500
1432
+ },
1433
+ {
1434
+ "epoch": 0.33,
1435
+ "learning_rate": 3.341518889960476e-05,
1436
+ "loss": 0.2427,
1437
+ "step": 119000
1438
+ },
1439
+ {
1440
+ "epoch": 0.33,
1441
+ "learning_rate": 3.3345504819350994e-05,
1442
+ "loss": 0.2438,
1443
+ "step": 119500
1444
+ },
1445
+ {
1446
+ "epoch": 0.33,
1447
+ "learning_rate": 3.327582073909723e-05,
1448
+ "loss": 0.2447,
1449
+ "step": 120000
1450
+ },
1451
+ {
1452
+ "epoch": 0.34,
1453
+ "learning_rate": 3.320613665884347e-05,
1454
+ "loss": 0.2375,
1455
+ "step": 120500
1456
+ },
1457
+ {
1458
+ "epoch": 0.34,
1459
+ "learning_rate": 3.3136452578589706e-05,
1460
+ "loss": 0.2399,
1461
+ "step": 121000
1462
+ },
1463
+ {
1464
+ "epoch": 0.34,
1465
+ "learning_rate": 3.306676849833595e-05,
1466
+ "loss": 0.2416,
1467
+ "step": 121500
1468
+ },
1469
+ {
1470
+ "epoch": 0.34,
1471
+ "learning_rate": 3.299708441808219e-05,
1472
+ "loss": 0.2405,
1473
+ "step": 122000
1474
+ },
1475
+ {
1476
+ "epoch": 0.34,
1477
+ "learning_rate": 3.2927400337828424e-05,
1478
+ "loss": 0.2322,
1479
+ "step": 122500
1480
+ },
1481
+ {
1482
+ "epoch": 0.34,
1483
+ "learning_rate": 3.285771625757466e-05,
1484
+ "loss": 0.2337,
1485
+ "step": 123000
1486
+ },
1487
+ {
1488
+ "epoch": 0.34,
1489
+ "learning_rate": 3.27880321773209e-05,
1490
+ "loss": 0.2391,
1491
+ "step": 123500
1492
+ },
1493
+ {
1494
+ "epoch": 0.35,
1495
+ "learning_rate": 3.2718348097067136e-05,
1496
+ "loss": 0.239,
1497
+ "step": 124000
1498
+ },
1499
+ {
1500
+ "epoch": 0.35,
1501
+ "learning_rate": 3.264866401681337e-05,
1502
+ "loss": 0.2363,
1503
+ "step": 124500
1504
+ },
1505
+ {
1506
+ "epoch": 0.35,
1507
+ "learning_rate": 3.257897993655962e-05,
1508
+ "loss": 0.239,
1509
+ "step": 125000
1510
+ },
1511
+ {
1512
+ "epoch": 0.35,
1513
+ "learning_rate": 3.2509295856305854e-05,
1514
+ "loss": 0.243,
1515
+ "step": 125500
1516
+ },
1517
+ {
1518
+ "epoch": 0.35,
1519
+ "learning_rate": 3.243961177605209e-05,
1520
+ "loss": 0.2338,
1521
+ "step": 126000
1522
+ },
1523
+ {
1524
+ "epoch": 0.35,
1525
+ "learning_rate": 3.236992769579833e-05,
1526
+ "loss": 0.2386,
1527
+ "step": 126500
1528
+ },
1529
+ {
1530
+ "epoch": 0.35,
1531
+ "learning_rate": 3.2300243615544566e-05,
1532
+ "loss": 0.2339,
1533
+ "step": 127000
1534
+ },
1535
+ {
1536
+ "epoch": 0.36,
1537
+ "learning_rate": 3.22305595352908e-05,
1538
+ "loss": 0.2372,
1539
+ "step": 127500
1540
+ },
1541
+ {
1542
+ "epoch": 0.36,
1543
+ "learning_rate": 3.216087545503705e-05,
1544
+ "loss": 0.2368,
1545
+ "step": 128000
1546
+ },
1547
+ {
1548
+ "epoch": 0.36,
1549
+ "learning_rate": 3.2091191374783284e-05,
1550
+ "loss": 0.2344,
1551
+ "step": 128500
1552
+ },
1553
+ {
1554
+ "epoch": 0.36,
1555
+ "learning_rate": 3.202150729452952e-05,
1556
+ "loss": 0.2402,
1557
+ "step": 129000
1558
+ },
1559
+ {
1560
+ "epoch": 0.36,
1561
+ "learning_rate": 3.195182321427576e-05,
1562
+ "loss": 0.2352,
1563
+ "step": 129500
1564
+ },
1565
+ {
1566
+ "epoch": 0.36,
1567
+ "learning_rate": 3.1882139134021996e-05,
1568
+ "loss": 0.2358,
1569
+ "step": 130000
1570
+ },
1571
+ {
1572
+ "epoch": 0.36,
1573
+ "learning_rate": 3.181245505376824e-05,
1574
+ "loss": 0.2431,
1575
+ "step": 130500
1576
+ },
1577
+ {
1578
+ "epoch": 0.37,
1579
+ "learning_rate": 3.174277097351448e-05,
1580
+ "loss": 0.2312,
1581
+ "step": 131000
1582
+ },
1583
+ {
1584
+ "epoch": 0.37,
1585
+ "learning_rate": 3.1673086893260714e-05,
1586
+ "loss": 0.2384,
1587
+ "step": 131500
1588
+ },
1589
+ {
1590
+ "epoch": 0.37,
1591
+ "learning_rate": 3.160340281300696e-05,
1592
+ "loss": 0.2352,
1593
+ "step": 132000
1594
+ },
1595
+ {
1596
+ "epoch": 0.37,
1597
+ "learning_rate": 3.1533718732753195e-05,
1598
+ "loss": 0.236,
1599
+ "step": 132500
1600
+ },
1601
+ {
1602
+ "epoch": 0.37,
1603
+ "learning_rate": 3.146403465249943e-05,
1604
+ "loss": 0.2491,
1605
+ "step": 133000
1606
+ },
1607
+ {
1608
+ "epoch": 0.37,
1609
+ "learning_rate": 3.139435057224567e-05,
1610
+ "loss": 0.2431,
1611
+ "step": 133500
1612
+ },
1613
+ {
1614
+ "epoch": 0.37,
1615
+ "learning_rate": 3.132466649199191e-05,
1616
+ "loss": 0.2351,
1617
+ "step": 134000
1618
+ },
1619
+ {
1620
+ "epoch": 0.37,
1621
+ "learning_rate": 3.1254982411738144e-05,
1622
+ "loss": 0.2347,
1623
+ "step": 134500
1624
+ },
1625
+ {
1626
+ "epoch": 0.38,
1627
+ "learning_rate": 3.118529833148439e-05,
1628
+ "loss": 0.2349,
1629
+ "step": 135000
1630
+ },
1631
+ {
1632
+ "epoch": 0.38,
1633
+ "learning_rate": 3.1115614251230625e-05,
1634
+ "loss": 0.2296,
1635
+ "step": 135500
1636
+ },
1637
+ {
1638
+ "epoch": 0.38,
1639
+ "learning_rate": 3.104593017097686e-05,
1640
+ "loss": 0.2408,
1641
+ "step": 136000
1642
+ },
1643
+ {
1644
+ "epoch": 0.38,
1645
+ "learning_rate": 3.09762460907231e-05,
1646
+ "loss": 0.2287,
1647
+ "step": 136500
1648
+ },
1649
+ {
1650
+ "epoch": 0.38,
1651
+ "learning_rate": 3.0906562010469337e-05,
1652
+ "loss": 0.2327,
1653
+ "step": 137000
1654
+ },
1655
+ {
1656
+ "epoch": 0.38,
1657
+ "learning_rate": 3.0836877930215574e-05,
1658
+ "loss": 0.2329,
1659
+ "step": 137500
1660
+ },
1661
+ {
1662
+ "epoch": 0.38,
1663
+ "learning_rate": 3.076719384996182e-05,
1664
+ "loss": 0.2401,
1665
+ "step": 138000
1666
+ },
1667
+ {
1668
+ "epoch": 0.39,
1669
+ "learning_rate": 3.0697509769708055e-05,
1670
+ "loss": 0.2366,
1671
+ "step": 138500
1672
+ },
1673
+ {
1674
+ "epoch": 0.39,
1675
+ "learning_rate": 3.062782568945429e-05,
1676
+ "loss": 0.2351,
1677
+ "step": 139000
1678
+ },
1679
+ {
1680
+ "epoch": 0.39,
1681
+ "learning_rate": 3.055814160920053e-05,
1682
+ "loss": 0.236,
1683
+ "step": 139500
1684
+ },
1685
+ {
1686
+ "epoch": 0.39,
1687
+ "learning_rate": 3.0488457528946766e-05,
1688
+ "loss": 0.2282,
1689
+ "step": 140000
1690
+ },
1691
+ {
1692
+ "epoch": 0.39,
1693
+ "learning_rate": 3.0418773448693007e-05,
1694
+ "loss": 0.2339,
1695
+ "step": 140500
1696
+ },
1697
+ {
1698
+ "epoch": 0.39,
1699
+ "learning_rate": 3.0349089368439244e-05,
1700
+ "loss": 0.2336,
1701
+ "step": 141000
1702
+ },
1703
+ {
1704
+ "epoch": 0.39,
1705
+ "learning_rate": 3.0279405288185485e-05,
1706
+ "loss": 0.2384,
1707
+ "step": 141500
1708
+ },
1709
+ {
1710
+ "epoch": 0.4,
1711
+ "learning_rate": 3.0209721207931725e-05,
1712
+ "loss": 0.2247,
1713
+ "step": 142000
1714
+ },
1715
+ {
1716
+ "epoch": 0.4,
1717
+ "learning_rate": 3.0140037127677963e-05,
1718
+ "loss": 0.2345,
1719
+ "step": 142500
1720
+ },
1721
+ {
1722
+ "epoch": 0.4,
1723
+ "learning_rate": 3.00703530474242e-05,
1724
+ "loss": 0.2296,
1725
+ "step": 143000
1726
+ },
1727
+ {
1728
+ "epoch": 0.4,
1729
+ "learning_rate": 3.0000668967170437e-05,
1730
+ "loss": 0.2319,
1731
+ "step": 143500
1732
+ },
1733
+ {
1734
+ "epoch": 0.4,
1735
+ "learning_rate": 2.9930984886916674e-05,
1736
+ "loss": 0.2454,
1737
+ "step": 144000
1738
+ },
1739
+ {
1740
+ "epoch": 0.4,
1741
+ "learning_rate": 2.986130080666291e-05,
1742
+ "loss": 0.234,
1743
+ "step": 144500
1744
+ },
1745
+ {
1746
+ "epoch": 0.4,
1747
+ "learning_rate": 2.9791616726409155e-05,
1748
+ "loss": 0.2305,
1749
+ "step": 145000
1750
+ },
1751
+ {
1752
+ "epoch": 0.41,
1753
+ "learning_rate": 2.9721932646155392e-05,
1754
+ "loss": 0.2346,
1755
+ "step": 145500
1756
+ },
1757
+ {
1758
+ "epoch": 0.41,
1759
+ "learning_rate": 2.965224856590163e-05,
1760
+ "loss": 0.2336,
1761
+ "step": 146000
1762
+ },
1763
+ {
1764
+ "epoch": 0.41,
1765
+ "learning_rate": 2.9582564485647867e-05,
1766
+ "loss": 0.2312,
1767
+ "step": 146500
1768
+ },
1769
+ {
1770
+ "epoch": 0.41,
1771
+ "learning_rate": 2.9512880405394104e-05,
1772
+ "loss": 0.2349,
1773
+ "step": 147000
1774
+ },
1775
+ {
1776
+ "epoch": 0.41,
1777
+ "learning_rate": 2.9443196325140345e-05,
1778
+ "loss": 0.2343,
1779
+ "step": 147500
1780
+ },
1781
+ {
1782
+ "epoch": 0.41,
1783
+ "learning_rate": 2.9373512244886585e-05,
1784
+ "loss": 0.23,
1785
+ "step": 148000
1786
+ },
1787
+ {
1788
+ "epoch": 0.41,
1789
+ "learning_rate": 2.9303828164632822e-05,
1790
+ "loss": 0.2221,
1791
+ "step": 148500
1792
+ },
1793
+ {
1794
+ "epoch": 0.42,
1795
+ "learning_rate": 2.9234144084379063e-05,
1796
+ "loss": 0.2368,
1797
+ "step": 149000
1798
+ },
1799
+ {
1800
+ "epoch": 0.42,
1801
+ "learning_rate": 2.91644600041253e-05,
1802
+ "loss": 0.2262,
1803
+ "step": 149500
1804
+ },
1805
+ {
1806
+ "epoch": 0.42,
1807
+ "learning_rate": 2.9094775923871537e-05,
1808
+ "loss": 0.2276,
1809
+ "step": 150000
1810
+ },
1811
+ {
1812
+ "epoch": 0.42,
1813
+ "learning_rate": 2.9025091843617774e-05,
1814
+ "loss": 0.2368,
1815
+ "step": 150500
1816
+ },
1817
+ {
1818
+ "epoch": 0.42,
1819
+ "learning_rate": 2.895540776336401e-05,
1820
+ "loss": 0.2373,
1821
+ "step": 151000
1822
+ },
1823
+ {
1824
+ "epoch": 0.42,
1825
+ "learning_rate": 2.8885723683110256e-05,
1826
+ "loss": 0.2397,
1827
+ "step": 151500
1828
+ },
1829
+ {
1830
+ "epoch": 0.42,
1831
+ "learning_rate": 2.8816039602856493e-05,
1832
+ "loss": 0.2328,
1833
+ "step": 152000
1834
+ },
1835
+ {
1836
+ "epoch": 0.43,
1837
+ "learning_rate": 2.874635552260273e-05,
1838
+ "loss": 0.2284,
1839
+ "step": 152500
1840
+ },
1841
+ {
1842
+ "epoch": 0.43,
1843
+ "learning_rate": 2.8676671442348967e-05,
1844
+ "loss": 0.229,
1845
+ "step": 153000
1846
+ },
1847
+ {
1848
+ "epoch": 0.43,
1849
+ "learning_rate": 2.8606987362095204e-05,
1850
+ "loss": 0.228,
1851
+ "step": 153500
1852
+ },
1853
+ {
1854
+ "epoch": 0.43,
1855
+ "learning_rate": 2.853730328184144e-05,
1856
+ "loss": 0.2335,
1857
+ "step": 154000
1858
+ },
1859
+ {
1860
+ "epoch": 0.43,
1861
+ "learning_rate": 2.8467619201587682e-05,
1862
+ "loss": 0.2281,
1863
+ "step": 154500
1864
+ },
1865
+ {
1866
+ "epoch": 0.43,
1867
+ "learning_rate": 2.8397935121333923e-05,
1868
+ "loss": 0.2329,
1869
+ "step": 155000
1870
+ },
1871
+ {
1872
+ "epoch": 0.43,
1873
+ "learning_rate": 2.8328251041080163e-05,
1874
+ "loss": 0.2294,
1875
+ "step": 155500
1876
+ },
1877
+ {
1878
+ "epoch": 0.43,
1879
+ "learning_rate": 2.82585669608264e-05,
1880
+ "loss": 0.2292,
1881
+ "step": 156000
1882
+ },
1883
+ {
1884
+ "epoch": 0.44,
1885
+ "learning_rate": 2.8188882880572638e-05,
1886
+ "loss": 0.23,
1887
+ "step": 156500
1888
+ },
1889
+ {
1890
+ "epoch": 0.44,
1891
+ "learning_rate": 2.8119198800318875e-05,
1892
+ "loss": 0.2255,
1893
+ "step": 157000
1894
+ },
1895
+ {
1896
+ "epoch": 0.44,
1897
+ "learning_rate": 2.8049514720065112e-05,
1898
+ "loss": 0.235,
1899
+ "step": 157500
1900
+ },
1901
+ {
1902
+ "epoch": 0.44,
1903
+ "learning_rate": 2.7979830639811356e-05,
1904
+ "loss": 0.2368,
1905
+ "step": 158000
1906
+ },
1907
+ {
1908
+ "epoch": 0.44,
1909
+ "learning_rate": 2.7910146559557593e-05,
1910
+ "loss": 0.2277,
1911
+ "step": 158500
1912
+ },
1913
+ {
1914
+ "epoch": 0.44,
1915
+ "learning_rate": 2.784046247930383e-05,
1916
+ "loss": 0.2381,
1917
+ "step": 159000
1918
+ },
1919
+ {
1920
+ "epoch": 0.44,
1921
+ "learning_rate": 2.7770778399050068e-05,
1922
+ "loss": 0.2302,
1923
+ "step": 159500
1924
+ },
1925
+ {
1926
+ "epoch": 0.45,
1927
+ "learning_rate": 2.7701094318796305e-05,
1928
+ "loss": 0.225,
1929
+ "step": 160000
1930
+ },
1931
+ {
1932
+ "epoch": 0.45,
1933
+ "learning_rate": 2.7631410238542542e-05,
1934
+ "loss": 0.227,
1935
+ "step": 160500
1936
+ },
1937
+ {
1938
+ "epoch": 0.45,
1939
+ "learning_rate": 2.7561726158288782e-05,
1940
+ "loss": 0.2362,
1941
+ "step": 161000
1942
+ },
1943
+ {
1944
+ "epoch": 0.45,
1945
+ "learning_rate": 2.7492042078035023e-05,
1946
+ "loss": 0.2283,
1947
+ "step": 161500
1948
+ },
1949
+ {
1950
+ "epoch": 0.45,
1951
+ "learning_rate": 2.742235799778126e-05,
1952
+ "loss": 0.2201,
1953
+ "step": 162000
1954
+ },
1955
+ {
1956
+ "epoch": 0.45,
1957
+ "learning_rate": 2.73526739175275e-05,
1958
+ "loss": 0.2327,
1959
+ "step": 162500
1960
+ },
1961
+ {
1962
+ "epoch": 0.45,
1963
+ "learning_rate": 2.7282989837273738e-05,
1964
+ "loss": 0.2236,
1965
+ "step": 163000
1966
+ },
1967
+ {
1968
+ "epoch": 0.46,
1969
+ "learning_rate": 2.7213305757019975e-05,
1970
+ "loss": 0.2362,
1971
+ "step": 163500
1972
+ },
1973
+ {
1974
+ "epoch": 0.46,
1975
+ "learning_rate": 2.7143621676766212e-05,
1976
+ "loss": 0.2282,
1977
+ "step": 164000
1978
+ },
1979
+ {
1980
+ "epoch": 0.46,
1981
+ "learning_rate": 2.707393759651245e-05,
1982
+ "loss": 0.2388,
1983
+ "step": 164500
1984
+ },
1985
+ {
1986
+ "epoch": 0.46,
1987
+ "learning_rate": 2.7004253516258693e-05,
1988
+ "loss": 0.2274,
1989
+ "step": 165000
1990
+ },
1991
+ {
1992
+ "epoch": 0.46,
1993
+ "learning_rate": 2.693456943600493e-05,
1994
+ "loss": 0.2294,
1995
+ "step": 165500
1996
+ },
1997
+ {
1998
+ "epoch": 0.46,
1999
+ "learning_rate": 2.6864885355751168e-05,
2000
+ "loss": 0.2206,
2001
+ "step": 166000
2002
+ },
2003
+ {
2004
+ "epoch": 0.46,
2005
+ "learning_rate": 2.6795201275497405e-05,
2006
+ "loss": 0.2253,
2007
+ "step": 166500
2008
+ },
2009
+ {
2010
+ "epoch": 0.47,
2011
+ "learning_rate": 2.6725517195243642e-05,
2012
+ "loss": 0.2387,
2013
+ "step": 167000
2014
+ },
2015
+ {
2016
+ "epoch": 0.47,
2017
+ "learning_rate": 2.665583311498988e-05,
2018
+ "loss": 0.232,
2019
+ "step": 167500
2020
+ },
2021
+ {
2022
+ "epoch": 0.47,
2023
+ "learning_rate": 2.6586149034736123e-05,
2024
+ "loss": 0.2296,
2025
+ "step": 168000
2026
+ },
2027
+ {
2028
+ "epoch": 0.47,
2029
+ "learning_rate": 2.651646495448236e-05,
2030
+ "loss": 0.2168,
2031
+ "step": 168500
2032
+ },
2033
+ {
2034
+ "epoch": 0.47,
2035
+ "learning_rate": 2.6446780874228598e-05,
2036
+ "loss": 0.2217,
2037
+ "step": 169000
2038
+ },
2039
+ {
2040
+ "epoch": 0.47,
2041
+ "learning_rate": 2.637709679397484e-05,
2042
+ "loss": 0.2333,
2043
+ "step": 169500
2044
+ },
2045
+ {
2046
+ "epoch": 0.47,
2047
+ "learning_rate": 2.6307412713721076e-05,
2048
+ "loss": 0.2269,
2049
+ "step": 170000
2050
+ },
2051
+ {
2052
+ "epoch": 0.48,
2053
+ "learning_rate": 2.6237728633467313e-05,
2054
+ "loss": 0.2326,
2055
+ "step": 170500
2056
+ },
2057
+ {
2058
+ "epoch": 0.48,
2059
+ "learning_rate": 2.616804455321355e-05,
2060
+ "loss": 0.2262,
2061
+ "step": 171000
2062
+ },
2063
+ {
2064
+ "epoch": 0.48,
2065
+ "learning_rate": 2.6098360472959794e-05,
2066
+ "loss": 0.2245,
2067
+ "step": 171500
2068
+ },
2069
+ {
2070
+ "epoch": 0.48,
2071
+ "learning_rate": 2.602867639270603e-05,
2072
+ "loss": 0.2269,
2073
+ "step": 172000
2074
+ },
2075
+ {
2076
+ "epoch": 0.48,
2077
+ "learning_rate": 2.5958992312452268e-05,
2078
+ "loss": 0.2313,
2079
+ "step": 172500
2080
+ },
2081
+ {
2082
+ "epoch": 0.48,
2083
+ "learning_rate": 2.5889308232198505e-05,
2084
+ "loss": 0.2239,
2085
+ "step": 173000
2086
+ },
2087
+ {
2088
+ "epoch": 0.48,
2089
+ "learning_rate": 2.5819624151944743e-05,
2090
+ "loss": 0.2328,
2091
+ "step": 173500
2092
+ },
2093
+ {
2094
+ "epoch": 0.49,
2095
+ "learning_rate": 2.574994007169098e-05,
2096
+ "loss": 0.2284,
2097
+ "step": 174000
2098
+ },
2099
+ {
2100
+ "epoch": 0.49,
2101
+ "learning_rate": 2.568025599143722e-05,
2102
+ "loss": 0.2314,
2103
+ "step": 174500
2104
+ },
2105
+ {
2106
+ "epoch": 0.49,
2107
+ "learning_rate": 2.561057191118346e-05,
2108
+ "loss": 0.2284,
2109
+ "step": 175000
2110
+ },
2111
+ {
2112
+ "epoch": 0.49,
2113
+ "learning_rate": 2.5540887830929698e-05,
2114
+ "loss": 0.216,
2115
+ "step": 175500
2116
+ },
2117
+ {
2118
+ "epoch": 0.49,
2119
+ "learning_rate": 2.547120375067594e-05,
2120
+ "loss": 0.2232,
2121
+ "step": 176000
2122
+ },
2123
+ {
2124
+ "epoch": 0.49,
2125
+ "learning_rate": 2.5401519670422176e-05,
2126
+ "loss": 0.2267,
2127
+ "step": 176500
2128
+ },
2129
+ {
2130
+ "epoch": 0.49,
2131
+ "learning_rate": 2.5331835590168413e-05,
2132
+ "loss": 0.227,
2133
+ "step": 177000
2134
+ },
2135
+ {
2136
+ "epoch": 0.49,
2137
+ "learning_rate": 2.526215150991465e-05,
2138
+ "loss": 0.2336,
2139
+ "step": 177500
2140
+ },
2141
+ {
2142
+ "epoch": 0.5,
2143
+ "learning_rate": 2.5192467429660894e-05,
2144
+ "loss": 0.2279,
2145
+ "step": 178000
2146
+ },
2147
+ {
2148
+ "epoch": 0.5,
2149
+ "learning_rate": 2.512278334940713e-05,
2150
+ "loss": 0.2236,
2151
+ "step": 178500
2152
+ },
2153
+ {
2154
+ "epoch": 0.5,
2155
+ "learning_rate": 2.505309926915337e-05,
2156
+ "loss": 0.2309,
2157
+ "step": 179000
2158
+ },
2159
+ {
2160
+ "epoch": 0.5,
2161
+ "learning_rate": 2.4983415188899606e-05,
2162
+ "loss": 0.2229,
2163
+ "step": 179500
2164
+ },
2165
+ {
2166
+ "epoch": 0.5,
2167
+ "learning_rate": 2.4913731108645843e-05,
2168
+ "loss": 0.2181,
2169
+ "step": 180000
2170
+ },
2171
+ {
2172
+ "epoch": 0.5,
2173
+ "learning_rate": 2.4844047028392083e-05,
2174
+ "loss": 0.2312,
2175
+ "step": 180500
2176
+ },
2177
+ {
2178
+ "epoch": 0.5,
2179
+ "learning_rate": 2.477436294813832e-05,
2180
+ "loss": 0.2264,
2181
+ "step": 181000
2182
+ },
2183
+ {
2184
+ "epoch": 0.51,
2185
+ "learning_rate": 2.4704678867884558e-05,
2186
+ "loss": 0.2313,
2187
+ "step": 181500
2188
+ },
2189
+ {
2190
+ "epoch": 0.51,
2191
+ "learning_rate": 2.46349947876308e-05,
2192
+ "loss": 0.2267,
2193
+ "step": 182000
2194
+ },
2195
+ {
2196
+ "epoch": 0.51,
2197
+ "learning_rate": 2.4565310707377036e-05,
2198
+ "loss": 0.234,
2199
+ "step": 182500
2200
+ },
2201
+ {
2202
+ "epoch": 0.51,
2203
+ "learning_rate": 2.4495626627123276e-05,
2204
+ "loss": 0.2303,
2205
+ "step": 183000
2206
+ },
2207
+ {
2208
+ "epoch": 0.51,
2209
+ "learning_rate": 2.4425942546869513e-05,
2210
+ "loss": 0.2201,
2211
+ "step": 183500
2212
+ },
2213
+ {
2214
+ "epoch": 0.51,
2215
+ "learning_rate": 2.4356258466615754e-05,
2216
+ "loss": 0.23,
2217
+ "step": 184000
2218
+ },
2219
+ {
2220
+ "epoch": 0.51,
2221
+ "learning_rate": 2.428657438636199e-05,
2222
+ "loss": 0.2192,
2223
+ "step": 184500
2224
+ },
2225
+ {
2226
+ "epoch": 0.52,
2227
+ "learning_rate": 2.421689030610823e-05,
2228
+ "loss": 0.2319,
2229
+ "step": 185000
2230
+ },
2231
+ {
2232
+ "epoch": 0.52,
2233
+ "learning_rate": 2.414720622585447e-05,
2234
+ "loss": 0.226,
2235
+ "step": 185500
2236
+ },
2237
+ {
2238
+ "epoch": 0.52,
2239
+ "learning_rate": 2.4077522145600706e-05,
2240
+ "loss": 0.232,
2241
+ "step": 186000
2242
+ },
2243
+ {
2244
+ "epoch": 0.52,
2245
+ "learning_rate": 2.4007838065346943e-05,
2246
+ "loss": 0.2327,
2247
+ "step": 186500
2248
+ },
2249
+ {
2250
+ "epoch": 0.52,
2251
+ "learning_rate": 2.3938153985093184e-05,
2252
+ "loss": 0.2231,
2253
+ "step": 187000
2254
+ },
2255
+ {
2256
+ "epoch": 0.52,
2257
+ "learning_rate": 2.386846990483942e-05,
2258
+ "loss": 0.22,
2259
+ "step": 187500
2260
+ },
2261
+ {
2262
+ "epoch": 0.52,
2263
+ "learning_rate": 2.3798785824585658e-05,
2264
+ "loss": 0.2243,
2265
+ "step": 188000
2266
+ },
2267
+ {
2268
+ "epoch": 0.53,
2269
+ "learning_rate": 2.3729101744331895e-05,
2270
+ "loss": 0.2287,
2271
+ "step": 188500
2272
+ },
2273
+ {
2274
+ "epoch": 0.53,
2275
+ "learning_rate": 2.3659417664078136e-05,
2276
+ "loss": 0.2227,
2277
+ "step": 189000
2278
+ },
2279
+ {
2280
+ "epoch": 0.53,
2281
+ "learning_rate": 2.3589733583824377e-05,
2282
+ "loss": 0.2321,
2283
+ "step": 189500
2284
+ },
2285
+ {
2286
+ "epoch": 0.53,
2287
+ "learning_rate": 2.3520049503570614e-05,
2288
+ "loss": 0.22,
2289
+ "step": 190000
2290
+ },
2291
+ {
2292
+ "epoch": 0.53,
2293
+ "learning_rate": 2.3450365423316854e-05,
2294
+ "loss": 0.2247,
2295
+ "step": 190500
2296
+ },
2297
+ {
2298
+ "epoch": 0.53,
2299
+ "learning_rate": 2.338068134306309e-05,
2300
+ "loss": 0.2251,
2301
+ "step": 191000
2302
+ },
2303
+ {
2304
+ "epoch": 0.53,
2305
+ "learning_rate": 2.331099726280933e-05,
2306
+ "loss": 0.2182,
2307
+ "step": 191500
2308
+ },
2309
+ {
2310
+ "epoch": 0.54,
2311
+ "learning_rate": 2.324131318255557e-05,
2312
+ "loss": 0.2244,
2313
+ "step": 192000
2314
+ },
2315
+ {
2316
+ "epoch": 0.54,
2317
+ "learning_rate": 2.3171629102301806e-05,
2318
+ "loss": 0.2236,
2319
+ "step": 192500
2320
+ },
2321
+ {
2322
+ "epoch": 0.54,
2323
+ "learning_rate": 2.3101945022048044e-05,
2324
+ "loss": 0.2228,
2325
+ "step": 193000
2326
+ },
2327
+ {
2328
+ "epoch": 0.54,
2329
+ "learning_rate": 2.303226094179428e-05,
2330
+ "loss": 0.216,
2331
+ "step": 193500
2332
+ },
2333
+ {
2334
+ "epoch": 0.54,
2335
+ "learning_rate": 2.296257686154052e-05,
2336
+ "loss": 0.2137,
2337
+ "step": 194000
2338
+ },
2339
+ {
2340
+ "epoch": 0.54,
2341
+ "learning_rate": 2.289289278128676e-05,
2342
+ "loss": 0.2207,
2343
+ "step": 194500
2344
+ },
2345
+ {
2346
+ "epoch": 0.54,
2347
+ "learning_rate": 2.2823208701032996e-05,
2348
+ "loss": 0.2285,
2349
+ "step": 195000
2350
+ },
2351
+ {
2352
+ "epoch": 0.54,
2353
+ "learning_rate": 2.2753524620779236e-05,
2354
+ "loss": 0.2209,
2355
+ "step": 195500
2356
+ },
2357
+ {
2358
+ "epoch": 0.55,
2359
+ "learning_rate": 2.2683840540525473e-05,
2360
+ "loss": 0.2313,
2361
+ "step": 196000
2362
+ },
2363
+ {
2364
+ "epoch": 0.55,
2365
+ "learning_rate": 2.2614156460271714e-05,
2366
+ "loss": 0.2304,
2367
+ "step": 196500
2368
+ },
2369
+ {
2370
+ "epoch": 0.55,
2371
+ "learning_rate": 2.2544472380017955e-05,
2372
+ "loss": 0.2236,
2373
+ "step": 197000
2374
+ },
2375
+ {
2376
+ "epoch": 0.55,
2377
+ "learning_rate": 2.2474788299764192e-05,
2378
+ "loss": 0.2211,
2379
+ "step": 197500
2380
+ },
2381
+ {
2382
+ "epoch": 0.55,
2383
+ "learning_rate": 2.240510421951043e-05,
2384
+ "loss": 0.2294,
2385
+ "step": 198000
2386
+ },
2387
+ {
2388
+ "epoch": 0.55,
2389
+ "learning_rate": 2.2335420139256666e-05,
2390
+ "loss": 0.221,
2391
+ "step": 198500
2392
+ },
2393
+ {
2394
+ "epoch": 0.55,
2395
+ "learning_rate": 2.2265736059002907e-05,
2396
+ "loss": 0.2227,
2397
+ "step": 199000
2398
+ },
2399
+ {
2400
+ "epoch": 0.56,
2401
+ "learning_rate": 2.2196051978749144e-05,
2402
+ "loss": 0.2204,
2403
+ "step": 199500
2404
+ },
2405
+ {
2406
+ "epoch": 0.56,
2407
+ "learning_rate": 2.212636789849538e-05,
2408
+ "loss": 0.2248,
2409
+ "step": 200000
2410
+ },
2411
+ {
2412
+ "epoch": 0.56,
2413
+ "learning_rate": 2.2056683818241622e-05,
2414
+ "loss": 0.2206,
2415
+ "step": 200500
2416
+ },
2417
+ {
2418
+ "epoch": 0.56,
2419
+ "learning_rate": 2.198699973798786e-05,
2420
+ "loss": 0.2322,
2421
+ "step": 201000
2422
+ },
2423
+ {
2424
+ "epoch": 0.56,
2425
+ "learning_rate": 2.1917315657734096e-05,
2426
+ "loss": 0.2211,
2427
+ "step": 201500
2428
+ },
2429
+ {
2430
+ "epoch": 0.56,
2431
+ "learning_rate": 2.1847631577480337e-05,
2432
+ "loss": 0.2307,
2433
+ "step": 202000
2434
+ },
2435
+ {
2436
+ "epoch": 0.56,
2437
+ "learning_rate": 2.1777947497226574e-05,
2438
+ "loss": 0.2143,
2439
+ "step": 202500
2440
+ },
2441
+ {
2442
+ "epoch": 0.57,
2443
+ "learning_rate": 2.1708263416972814e-05,
2444
+ "loss": 0.2201,
2445
+ "step": 203000
2446
+ },
2447
+ {
2448
+ "epoch": 0.57,
2449
+ "learning_rate": 2.163857933671905e-05,
2450
+ "loss": 0.2269,
2451
+ "step": 203500
2452
+ },
2453
+ {
2454
+ "epoch": 0.57,
2455
+ "learning_rate": 2.1568895256465292e-05,
2456
+ "loss": 0.2222,
2457
+ "step": 204000
2458
+ },
2459
+ {
2460
+ "epoch": 0.57,
2461
+ "learning_rate": 2.149921117621153e-05,
2462
+ "loss": 0.2158,
2463
+ "step": 204500
2464
+ },
2465
+ {
2466
+ "epoch": 0.57,
2467
+ "learning_rate": 2.1429527095957767e-05,
2468
+ "loss": 0.216,
2469
+ "step": 205000
2470
+ },
2471
+ {
2472
+ "epoch": 0.57,
2473
+ "learning_rate": 2.1359843015704007e-05,
2474
+ "loss": 0.2322,
2475
+ "step": 205500
2476
+ },
2477
+ {
2478
+ "epoch": 0.57,
2479
+ "learning_rate": 2.1290158935450244e-05,
2480
+ "loss": 0.2167,
2481
+ "step": 206000
2482
+ },
2483
+ {
2484
+ "epoch": 0.58,
2485
+ "learning_rate": 2.122047485519648e-05,
2486
+ "loss": 0.2231,
2487
+ "step": 206500
2488
+ },
2489
+ {
2490
+ "epoch": 0.58,
2491
+ "learning_rate": 2.1150790774942722e-05,
2492
+ "loss": 0.2242,
2493
+ "step": 207000
2494
+ },
2495
+ {
2496
+ "epoch": 0.58,
2497
+ "learning_rate": 2.108110669468896e-05,
2498
+ "loss": 0.237,
2499
+ "step": 207500
2500
+ },
2501
+ {
2502
+ "epoch": 0.58,
2503
+ "learning_rate": 2.1011422614435196e-05,
2504
+ "loss": 0.221,
2505
+ "step": 208000
2506
+ },
2507
+ {
2508
+ "epoch": 0.58,
2509
+ "learning_rate": 2.0941738534181434e-05,
2510
+ "loss": 0.2251,
2511
+ "step": 208500
2512
+ },
2513
+ {
2514
+ "epoch": 0.58,
2515
+ "learning_rate": 2.0872054453927674e-05,
2516
+ "loss": 0.2272,
2517
+ "step": 209000
2518
+ },
2519
+ {
2520
+ "epoch": 0.58,
2521
+ "learning_rate": 2.080237037367391e-05,
2522
+ "loss": 0.2235,
2523
+ "step": 209500
2524
+ },
2525
+ {
2526
+ "epoch": 0.59,
2527
+ "learning_rate": 2.0732686293420152e-05,
2528
+ "loss": 0.2179,
2529
+ "step": 210000
2530
+ },
2531
+ {
2532
+ "epoch": 0.59,
2533
+ "learning_rate": 2.0663002213166392e-05,
2534
+ "loss": 0.2276,
2535
+ "step": 210500
2536
+ },
2537
+ {
2538
+ "epoch": 0.59,
2539
+ "learning_rate": 2.059331813291263e-05,
2540
+ "loss": 0.2215,
2541
+ "step": 211000
2542
+ },
2543
+ {
2544
+ "epoch": 0.59,
2545
+ "learning_rate": 2.0523634052658867e-05,
2546
+ "loss": 0.2266,
2547
+ "step": 211500
2548
+ },
2549
+ {
2550
+ "epoch": 0.59,
2551
+ "learning_rate": 2.0453949972405107e-05,
2552
+ "loss": 0.2185,
2553
+ "step": 212000
2554
+ },
2555
+ {
2556
+ "epoch": 0.59,
2557
+ "learning_rate": 2.0384265892151345e-05,
2558
+ "loss": 0.2255,
2559
+ "step": 212500
2560
+ },
2561
+ {
2562
+ "epoch": 0.59,
2563
+ "learning_rate": 2.0314581811897582e-05,
2564
+ "loss": 0.2307,
2565
+ "step": 213000
2566
+ },
2567
+ {
2568
+ "epoch": 0.6,
2569
+ "learning_rate": 2.024489773164382e-05,
2570
+ "loss": 0.2201,
2571
+ "step": 213500
2572
+ },
2573
+ {
2574
+ "epoch": 0.6,
2575
+ "learning_rate": 2.017521365139006e-05,
2576
+ "loss": 0.2208,
2577
+ "step": 214000
2578
+ },
2579
+ {
2580
+ "epoch": 0.6,
2581
+ "learning_rate": 2.0105529571136297e-05,
2582
+ "loss": 0.2177,
2583
+ "step": 214500
2584
+ },
2585
+ {
2586
+ "epoch": 0.6,
2587
+ "learning_rate": 2.0035845490882534e-05,
2588
+ "loss": 0.2136,
2589
+ "step": 215000
2590
+ },
2591
+ {
2592
+ "epoch": 0.6,
2593
+ "learning_rate": 1.9966161410628775e-05,
2594
+ "loss": 0.2254,
2595
+ "step": 215500
2596
+ },
2597
+ {
2598
+ "epoch": 0.6,
2599
+ "learning_rate": 1.9896477330375012e-05,
2600
+ "loss": 0.2208,
2601
+ "step": 216000
2602
+ },
2603
+ {
2604
+ "epoch": 0.6,
2605
+ "learning_rate": 1.982679325012125e-05,
2606
+ "loss": 0.2213,
2607
+ "step": 216500
2608
+ },
2609
+ {
2610
+ "epoch": 0.6,
2611
+ "learning_rate": 1.975710916986749e-05,
2612
+ "loss": 0.2253,
2613
+ "step": 217000
2614
+ },
2615
+ {
2616
+ "epoch": 0.61,
2617
+ "learning_rate": 1.968742508961373e-05,
2618
+ "loss": 0.2172,
2619
+ "step": 217500
2620
+ },
2621
+ {
2622
+ "epoch": 0.61,
2623
+ "learning_rate": 1.9617741009359967e-05,
2624
+ "loss": 0.2342,
2625
+ "step": 218000
2626
+ },
2627
+ {
2628
+ "epoch": 0.61,
2629
+ "learning_rate": 1.9548056929106204e-05,
2630
+ "loss": 0.2177,
2631
+ "step": 218500
2632
+ },
2633
+ {
2634
+ "epoch": 0.61,
2635
+ "learning_rate": 1.9478372848852445e-05,
2636
+ "loss": 0.2263,
2637
+ "step": 219000
2638
+ },
2639
+ {
2640
+ "epoch": 0.61,
2641
+ "learning_rate": 1.9408688768598682e-05,
2642
+ "loss": 0.2334,
2643
+ "step": 219500
2644
+ },
2645
+ {
2646
+ "epoch": 0.61,
2647
+ "learning_rate": 1.933900468834492e-05,
2648
+ "loss": 0.2152,
2649
+ "step": 220000
2650
+ },
2651
+ {
2652
+ "epoch": 0.61,
2653
+ "learning_rate": 1.926932060809116e-05,
2654
+ "loss": 0.2154,
2655
+ "step": 220500
2656
+ },
2657
+ {
2658
+ "epoch": 0.62,
2659
+ "learning_rate": 1.9199636527837397e-05,
2660
+ "loss": 0.2123,
2661
+ "step": 221000
2662
+ },
2663
+ {
2664
+ "epoch": 0.62,
2665
+ "learning_rate": 1.9129952447583634e-05,
2666
+ "loss": 0.2204,
2667
+ "step": 221500
2668
+ },
2669
+ {
2670
+ "epoch": 0.62,
2671
+ "learning_rate": 1.9060268367329875e-05,
2672
+ "loss": 0.2167,
2673
+ "step": 222000
2674
+ },
2675
+ {
2676
+ "epoch": 0.62,
2677
+ "learning_rate": 1.8990584287076112e-05,
2678
+ "loss": 0.2234,
2679
+ "step": 222500
2680
+ },
2681
+ {
2682
+ "epoch": 0.62,
2683
+ "learning_rate": 1.892090020682235e-05,
2684
+ "loss": 0.2244,
2685
+ "step": 223000
2686
+ },
2687
+ {
2688
+ "epoch": 0.62,
2689
+ "learning_rate": 1.885121612656859e-05,
2690
+ "loss": 0.2227,
2691
+ "step": 223500
2692
+ },
2693
+ {
2694
+ "epoch": 0.62,
2695
+ "learning_rate": 1.8781532046314827e-05,
2696
+ "loss": 0.2177,
2697
+ "step": 224000
2698
+ },
2699
+ {
2700
+ "epoch": 0.63,
2701
+ "learning_rate": 1.8711847966061068e-05,
2702
+ "loss": 0.2212,
2703
+ "step": 224500
2704
+ },
2705
+ {
2706
+ "epoch": 0.63,
2707
+ "learning_rate": 1.8642163885807305e-05,
2708
+ "loss": 0.2218,
2709
+ "step": 225000
2710
+ },
2711
+ {
2712
+ "epoch": 0.63,
2713
+ "learning_rate": 1.8572479805553545e-05,
2714
+ "loss": 0.2278,
2715
+ "step": 225500
2716
+ },
2717
+ {
2718
+ "epoch": 0.63,
2719
+ "learning_rate": 1.8502795725299782e-05,
2720
+ "loss": 0.2233,
2721
+ "step": 226000
2722
+ },
2723
+ {
2724
+ "epoch": 0.63,
2725
+ "learning_rate": 1.843311164504602e-05,
2726
+ "loss": 0.2242,
2727
+ "step": 226500
2728
+ },
2729
+ {
2730
+ "epoch": 0.63,
2731
+ "learning_rate": 1.836342756479226e-05,
2732
+ "loss": 0.2268,
2733
+ "step": 227000
2734
+ },
2735
+ {
2736
+ "epoch": 0.63,
2737
+ "learning_rate": 1.8293743484538497e-05,
2738
+ "loss": 0.2189,
2739
+ "step": 227500
2740
+ },
2741
+ {
2742
+ "epoch": 0.64,
2743
+ "learning_rate": 1.8224059404284735e-05,
2744
+ "loss": 0.2257,
2745
+ "step": 228000
2746
+ },
2747
+ {
2748
+ "epoch": 0.64,
2749
+ "learning_rate": 1.8154375324030972e-05,
2750
+ "loss": 0.2134,
2751
+ "step": 228500
2752
+ },
2753
+ {
2754
+ "epoch": 0.64,
2755
+ "learning_rate": 1.8084691243777212e-05,
2756
+ "loss": 0.2147,
2757
+ "step": 229000
2758
+ },
2759
+ {
2760
+ "epoch": 0.64,
2761
+ "learning_rate": 1.801500716352345e-05,
2762
+ "loss": 0.2204,
2763
+ "step": 229500
2764
+ },
2765
+ {
2766
+ "epoch": 0.64,
2767
+ "learning_rate": 1.7945323083269687e-05,
2768
+ "loss": 0.215,
2769
+ "step": 230000
2770
+ },
2771
+ {
2772
+ "epoch": 0.64,
2773
+ "learning_rate": 1.7875639003015927e-05,
2774
+ "loss": 0.2182,
2775
+ "step": 230500
2776
+ },
2777
+ {
2778
+ "epoch": 0.64,
2779
+ "learning_rate": 1.7805954922762168e-05,
2780
+ "loss": 0.2259,
2781
+ "step": 231000
2782
+ },
2783
+ {
2784
+ "epoch": 0.65,
2785
+ "learning_rate": 1.7736270842508405e-05,
2786
+ "loss": 0.2207,
2787
+ "step": 231500
2788
+ },
2789
+ {
2790
+ "epoch": 0.65,
2791
+ "learning_rate": 1.7666586762254646e-05,
2792
+ "loss": 0.2197,
2793
+ "step": 232000
2794
+ },
2795
+ {
2796
+ "epoch": 0.65,
2797
+ "learning_rate": 1.7596902682000883e-05,
2798
+ "loss": 0.2229,
2799
+ "step": 232500
2800
+ },
2801
+ {
2802
+ "epoch": 0.65,
2803
+ "learning_rate": 1.752721860174712e-05,
2804
+ "loss": 0.2196,
2805
+ "step": 233000
2806
+ },
2807
+ {
2808
+ "epoch": 0.65,
2809
+ "learning_rate": 1.7457534521493357e-05,
2810
+ "loss": 0.217,
2811
+ "step": 233500
2812
+ },
2813
+ {
2814
+ "epoch": 0.65,
2815
+ "learning_rate": 1.7387850441239598e-05,
2816
+ "loss": 0.2143,
2817
+ "step": 234000
2818
+ },
2819
+ {
2820
+ "epoch": 0.65,
2821
+ "learning_rate": 1.7318166360985835e-05,
2822
+ "loss": 0.2177,
2823
+ "step": 234500
2824
+ },
2825
+ {
2826
+ "epoch": 0.66,
2827
+ "learning_rate": 1.7248482280732072e-05,
2828
+ "loss": 0.2198,
2829
+ "step": 235000
2830
+ },
2831
+ {
2832
+ "epoch": 0.66,
2833
+ "learning_rate": 1.7178798200478313e-05,
2834
+ "loss": 0.2264,
2835
+ "step": 235500
2836
+ },
2837
+ {
2838
+ "epoch": 0.66,
2839
+ "learning_rate": 1.710911412022455e-05,
2840
+ "loss": 0.2157,
2841
+ "step": 236000
2842
+ },
2843
+ {
2844
+ "epoch": 0.66,
2845
+ "learning_rate": 1.7039430039970787e-05,
2846
+ "loss": 0.2275,
2847
+ "step": 236500
2848
+ },
2849
+ {
2850
+ "epoch": 0.66,
2851
+ "learning_rate": 1.6969745959717028e-05,
2852
+ "loss": 0.2197,
2853
+ "step": 237000
2854
+ },
2855
+ {
2856
+ "epoch": 0.66,
2857
+ "learning_rate": 1.6900061879463265e-05,
2858
+ "loss": 0.2203,
2859
+ "step": 237500
2860
+ },
2861
+ {
2862
+ "epoch": 0.66,
2863
+ "learning_rate": 1.6830377799209505e-05,
2864
+ "loss": 0.2242,
2865
+ "step": 238000
2866
+ },
2867
+ {
2868
+ "epoch": 0.66,
2869
+ "learning_rate": 1.6760693718955743e-05,
2870
+ "loss": 0.2214,
2871
+ "step": 238500
2872
+ },
2873
+ {
2874
+ "epoch": 0.67,
2875
+ "learning_rate": 1.6691009638701983e-05,
2876
+ "loss": 0.2226,
2877
+ "step": 239000
2878
+ },
2879
+ {
2880
+ "epoch": 0.67,
2881
+ "learning_rate": 1.662132555844822e-05,
2882
+ "loss": 0.2247,
2883
+ "step": 239500
2884
+ },
2885
+ {
2886
+ "epoch": 0.67,
2887
+ "learning_rate": 1.6551641478194458e-05,
2888
+ "loss": 0.2177,
2889
+ "step": 240000
2890
+ },
2891
+ {
2892
+ "epoch": 0.67,
2893
+ "learning_rate": 1.6481957397940698e-05,
2894
+ "loss": 0.215,
2895
+ "step": 240500
2896
+ },
2897
+ {
2898
+ "epoch": 0.67,
2899
+ "learning_rate": 1.6412273317686935e-05,
2900
+ "loss": 0.211,
2901
+ "step": 241000
2902
+ },
2903
+ {
2904
+ "epoch": 0.67,
2905
+ "learning_rate": 1.6342589237433172e-05,
2906
+ "loss": 0.2142,
2907
+ "step": 241500
2908
+ },
2909
+ {
2910
+ "epoch": 0.67,
2911
+ "learning_rate": 1.6272905157179413e-05,
2912
+ "loss": 0.2119,
2913
+ "step": 242000
2914
+ },
2915
+ {
2916
+ "epoch": 0.68,
2917
+ "learning_rate": 1.620322107692565e-05,
2918
+ "loss": 0.2178,
2919
+ "step": 242500
2920
+ },
2921
+ {
2922
+ "epoch": 0.68,
2923
+ "learning_rate": 1.6133536996671887e-05,
2924
+ "loss": 0.2145,
2925
+ "step": 243000
2926
+ },
2927
+ {
2928
+ "epoch": 0.68,
2929
+ "learning_rate": 1.6063852916418125e-05,
2930
+ "loss": 0.2108,
2931
+ "step": 243500
2932
+ },
2933
+ {
2934
+ "epoch": 0.68,
2935
+ "learning_rate": 1.5994168836164365e-05,
2936
+ "loss": 0.2207,
2937
+ "step": 244000
2938
+ },
2939
+ {
2940
+ "epoch": 0.68,
2941
+ "learning_rate": 1.5924484755910606e-05,
2942
+ "loss": 0.2127,
2943
+ "step": 244500
2944
+ },
2945
+ {
2946
+ "epoch": 0.68,
2947
+ "learning_rate": 1.5854800675656843e-05,
2948
+ "loss": 0.2249,
2949
+ "step": 245000
2950
+ },
2951
+ {
2952
+ "epoch": 0.68,
2953
+ "learning_rate": 1.5785116595403084e-05,
2954
+ "loss": 0.2257,
2955
+ "step": 245500
2956
+ },
2957
+ {
2958
+ "epoch": 0.69,
2959
+ "learning_rate": 1.571543251514932e-05,
2960
+ "loss": 0.2178,
2961
+ "step": 246000
2962
+ },
2963
+ {
2964
+ "epoch": 0.69,
2965
+ "learning_rate": 1.5645748434895558e-05,
2966
+ "loss": 0.2237,
2967
+ "step": 246500
2968
+ },
2969
+ {
2970
+ "epoch": 0.69,
2971
+ "learning_rate": 1.55760643546418e-05,
2972
+ "loss": 0.2141,
2973
+ "step": 247000
2974
+ },
2975
+ {
2976
+ "epoch": 0.69,
2977
+ "learning_rate": 1.5506380274388036e-05,
2978
+ "loss": 0.2165,
2979
+ "step": 247500
2980
+ },
2981
+ {
2982
+ "epoch": 0.69,
2983
+ "learning_rate": 1.5436696194134273e-05,
2984
+ "loss": 0.2096,
2985
+ "step": 248000
2986
+ },
2987
+ {
2988
+ "epoch": 0.69,
2989
+ "learning_rate": 1.5367012113880513e-05,
2990
+ "loss": 0.2096,
2991
+ "step": 248500
2992
+ },
2993
+ {
2994
+ "epoch": 0.69,
2995
+ "learning_rate": 1.529732803362675e-05,
2996
+ "loss": 0.2258,
2997
+ "step": 249000
2998
+ },
2999
+ {
3000
+ "epoch": 0.7,
3001
+ "learning_rate": 1.522764395337299e-05,
3002
+ "loss": 0.2141,
3003
+ "step": 249500
3004
+ },
3005
+ {
3006
+ "epoch": 0.7,
3007
+ "learning_rate": 1.5157959873119227e-05,
3008
+ "loss": 0.2116,
3009
+ "step": 250000
3010
+ },
3011
+ {
3012
+ "epoch": 0.7,
3013
+ "learning_rate": 1.5088275792865467e-05,
3014
+ "loss": 0.2153,
3015
+ "step": 250500
3016
+ },
3017
+ {
3018
+ "epoch": 0.7,
3019
+ "learning_rate": 1.5018591712611704e-05,
3020
+ "loss": 0.218,
3021
+ "step": 251000
3022
+ },
3023
+ {
3024
+ "epoch": 0.7,
3025
+ "learning_rate": 1.4948907632357942e-05,
3026
+ "loss": 0.2172,
3027
+ "step": 251500
3028
+ },
3029
+ {
3030
+ "epoch": 0.7,
3031
+ "learning_rate": 1.4879223552104182e-05,
3032
+ "loss": 0.2281,
3033
+ "step": 252000
3034
+ },
3035
+ {
3036
+ "epoch": 0.7,
3037
+ "learning_rate": 1.480953947185042e-05,
3038
+ "loss": 0.2171,
3039
+ "step": 252500
3040
+ },
3041
+ {
3042
+ "epoch": 0.71,
3043
+ "learning_rate": 1.4739855391596658e-05,
3044
+ "loss": 0.2155,
3045
+ "step": 253000
3046
+ },
3047
+ {
3048
+ "epoch": 0.71,
3049
+ "learning_rate": 1.4670171311342897e-05,
3050
+ "loss": 0.2148,
3051
+ "step": 253500
3052
+ },
3053
+ {
3054
+ "epoch": 0.71,
3055
+ "learning_rate": 1.4600487231089136e-05,
3056
+ "loss": 0.2189,
3057
+ "step": 254000
3058
+ },
3059
+ {
3060
+ "epoch": 0.71,
3061
+ "learning_rate": 1.4530803150835373e-05,
3062
+ "loss": 0.2221,
3063
+ "step": 254500
3064
+ },
3065
+ {
3066
+ "epoch": 0.71,
3067
+ "learning_rate": 1.446111907058161e-05,
3068
+ "loss": 0.2223,
3069
+ "step": 255000
3070
+ },
3071
+ {
3072
+ "epoch": 0.71,
3073
+ "learning_rate": 1.4391434990327851e-05,
3074
+ "loss": 0.2191,
3075
+ "step": 255500
3076
+ },
3077
+ {
3078
+ "epoch": 0.71,
3079
+ "learning_rate": 1.4321750910074088e-05,
3080
+ "loss": 0.2195,
3081
+ "step": 256000
3082
+ },
3083
+ {
3084
+ "epoch": 0.71,
3085
+ "learning_rate": 1.4252066829820327e-05,
3086
+ "loss": 0.2182,
3087
+ "step": 256500
3088
+ },
3089
+ {
3090
+ "epoch": 0.72,
3091
+ "learning_rate": 1.4182382749566568e-05,
3092
+ "loss": 0.2169,
3093
+ "step": 257000
3094
+ },
3095
+ {
3096
+ "epoch": 0.72,
3097
+ "learning_rate": 1.4112698669312805e-05,
3098
+ "loss": 0.2239,
3099
+ "step": 257500
3100
+ },
3101
+ {
3102
+ "epoch": 0.72,
3103
+ "learning_rate": 1.4043014589059042e-05,
3104
+ "loss": 0.2209,
3105
+ "step": 258000
3106
+ },
3107
+ {
3108
+ "epoch": 0.72,
3109
+ "learning_rate": 1.3973330508805283e-05,
3110
+ "loss": 0.2043,
3111
+ "step": 258500
3112
+ },
3113
+ {
3114
+ "epoch": 0.72,
3115
+ "learning_rate": 1.390364642855152e-05,
3116
+ "loss": 0.2211,
3117
+ "step": 259000
3118
+ },
3119
+ {
3120
+ "epoch": 0.72,
3121
+ "learning_rate": 1.3833962348297757e-05,
3122
+ "loss": 0.2159,
3123
+ "step": 259500
3124
+ },
3125
+ {
3126
+ "epoch": 0.72,
3127
+ "learning_rate": 1.3764278268043996e-05,
3128
+ "loss": 0.2114,
3129
+ "step": 260000
3130
+ },
3131
+ {
3132
+ "epoch": 0.73,
3133
+ "learning_rate": 1.3694594187790236e-05,
3134
+ "loss": 0.2229,
3135
+ "step": 260500
3136
+ },
3137
+ {
3138
+ "epoch": 0.73,
3139
+ "learning_rate": 1.3624910107536474e-05,
3140
+ "loss": 0.2158,
3141
+ "step": 261000
3142
+ },
3143
+ {
3144
+ "epoch": 0.73,
3145
+ "learning_rate": 1.355522602728271e-05,
3146
+ "loss": 0.2145,
3147
+ "step": 261500
3148
+ },
3149
+ {
3150
+ "epoch": 0.73,
3151
+ "learning_rate": 1.3485541947028951e-05,
3152
+ "loss": 0.2132,
3153
+ "step": 262000
3154
+ },
3155
+ {
3156
+ "epoch": 0.73,
3157
+ "learning_rate": 1.3415857866775188e-05,
3158
+ "loss": 0.2184,
3159
+ "step": 262500
3160
+ },
3161
+ {
3162
+ "epoch": 0.73,
3163
+ "learning_rate": 1.3346173786521426e-05,
3164
+ "loss": 0.2164,
3165
+ "step": 263000
3166
+ },
3167
+ {
3168
+ "epoch": 0.73,
3169
+ "learning_rate": 1.3276489706267666e-05,
3170
+ "loss": 0.2111,
3171
+ "step": 263500
3172
+ },
3173
+ {
3174
+ "epoch": 0.74,
3175
+ "learning_rate": 1.3206805626013905e-05,
3176
+ "loss": 0.2128,
3177
+ "step": 264000
3178
+ },
3179
+ {
3180
+ "epoch": 0.74,
3181
+ "learning_rate": 1.3137121545760142e-05,
3182
+ "loss": 0.2203,
3183
+ "step": 264500
3184
+ },
3185
+ {
3186
+ "epoch": 0.74,
3187
+ "learning_rate": 1.306743746550638e-05,
3188
+ "loss": 0.2163,
3189
+ "step": 265000
3190
+ },
3191
+ {
3192
+ "epoch": 0.74,
3193
+ "learning_rate": 1.299775338525262e-05,
3194
+ "loss": 0.2169,
3195
+ "step": 265500
3196
+ },
3197
+ {
3198
+ "epoch": 0.74,
3199
+ "learning_rate": 1.2928069304998857e-05,
3200
+ "loss": 0.2157,
3201
+ "step": 266000
3202
+ },
3203
+ {
3204
+ "epoch": 0.74,
3205
+ "learning_rate": 1.2858385224745096e-05,
3206
+ "loss": 0.2246,
3207
+ "step": 266500
3208
+ },
3209
+ {
3210
+ "epoch": 0.74,
3211
+ "learning_rate": 1.2788701144491335e-05,
3212
+ "loss": 0.2204,
3213
+ "step": 267000
3214
+ },
3215
+ {
3216
+ "epoch": 0.75,
3217
+ "learning_rate": 1.2719017064237574e-05,
3218
+ "loss": 0.2137,
3219
+ "step": 267500
3220
+ },
3221
+ {
3222
+ "epoch": 0.75,
3223
+ "learning_rate": 1.2649332983983811e-05,
3224
+ "loss": 0.2195,
3225
+ "step": 268000
3226
+ },
3227
+ {
3228
+ "epoch": 0.75,
3229
+ "learning_rate": 1.2579648903730052e-05,
3230
+ "loss": 0.2189,
3231
+ "step": 268500
3232
+ },
3233
+ {
3234
+ "epoch": 0.75,
3235
+ "learning_rate": 1.2509964823476289e-05,
3236
+ "loss": 0.208,
3237
+ "step": 269000
3238
+ },
3239
+ {
3240
+ "epoch": 0.75,
3241
+ "learning_rate": 1.2440280743222526e-05,
3242
+ "loss": 0.2109,
3243
+ "step": 269500
3244
+ },
3245
+ {
3246
+ "epoch": 0.75,
3247
+ "learning_rate": 1.2370596662968765e-05,
3248
+ "loss": 0.2249,
3249
+ "step": 270000
3250
+ },
3251
+ {
3252
+ "epoch": 0.75,
3253
+ "learning_rate": 1.2300912582715004e-05,
3254
+ "loss": 0.2159,
3255
+ "step": 270500
3256
+ },
3257
+ {
3258
+ "epoch": 0.76,
3259
+ "learning_rate": 1.2231228502461243e-05,
3260
+ "loss": 0.2114,
3261
+ "step": 271000
3262
+ },
3263
+ {
3264
+ "epoch": 0.76,
3265
+ "learning_rate": 1.2161544422207482e-05,
3266
+ "loss": 0.2095,
3267
+ "step": 271500
3268
+ },
3269
+ {
3270
+ "epoch": 0.76,
3271
+ "learning_rate": 1.2091860341953719e-05,
3272
+ "loss": 0.218,
3273
+ "step": 272000
3274
+ },
3275
+ {
3276
+ "epoch": 0.76,
3277
+ "learning_rate": 1.2022176261699958e-05,
3278
+ "loss": 0.2208,
3279
+ "step": 272500
3280
+ },
3281
+ {
3282
+ "epoch": 0.76,
3283
+ "learning_rate": 1.1952492181446195e-05,
3284
+ "loss": 0.2098,
3285
+ "step": 273000
3286
+ },
3287
+ {
3288
+ "epoch": 0.76,
3289
+ "learning_rate": 1.1882808101192434e-05,
3290
+ "loss": 0.2073,
3291
+ "step": 273500
3292
+ },
3293
+ {
3294
+ "epoch": 0.76,
3295
+ "learning_rate": 1.1813124020938674e-05,
3296
+ "loss": 0.2084,
3297
+ "step": 274000
3298
+ },
3299
+ {
3300
+ "epoch": 0.77,
3301
+ "learning_rate": 1.1743439940684911e-05,
3302
+ "loss": 0.2071,
3303
+ "step": 274500
3304
+ },
3305
+ {
3306
+ "epoch": 0.77,
3307
+ "learning_rate": 1.167375586043115e-05,
3308
+ "loss": 0.2207,
3309
+ "step": 275000
3310
+ },
3311
+ {
3312
+ "epoch": 0.77,
3313
+ "learning_rate": 1.1604071780177387e-05,
3314
+ "loss": 0.2186,
3315
+ "step": 275500
3316
+ },
3317
+ {
3318
+ "epoch": 0.77,
3319
+ "learning_rate": 1.1534387699923626e-05,
3320
+ "loss": 0.2132,
3321
+ "step": 276000
3322
+ },
3323
+ {
3324
+ "epoch": 0.77,
3325
+ "learning_rate": 1.1464703619669865e-05,
3326
+ "loss": 0.2107,
3327
+ "step": 276500
3328
+ },
3329
+ {
3330
+ "epoch": 0.77,
3331
+ "learning_rate": 1.1395019539416104e-05,
3332
+ "loss": 0.2157,
3333
+ "step": 277000
3334
+ },
3335
+ {
3336
+ "epoch": 0.77,
3337
+ "learning_rate": 1.1325335459162343e-05,
3338
+ "loss": 0.2138,
3339
+ "step": 277500
3340
+ },
3341
+ {
3342
+ "epoch": 0.77,
3343
+ "learning_rate": 1.1255651378908582e-05,
3344
+ "loss": 0.2157,
3345
+ "step": 278000
3346
+ },
3347
+ {
3348
+ "epoch": 0.78,
3349
+ "learning_rate": 1.1185967298654819e-05,
3350
+ "loss": 0.2101,
3351
+ "step": 278500
3352
+ },
3353
+ {
3354
+ "epoch": 0.78,
3355
+ "learning_rate": 1.1116283218401058e-05,
3356
+ "loss": 0.2177,
3357
+ "step": 279000
3358
+ },
3359
+ {
3360
+ "epoch": 0.78,
3361
+ "learning_rate": 1.1046599138147295e-05,
3362
+ "loss": 0.2142,
3363
+ "step": 279500
3364
+ },
3365
+ {
3366
+ "epoch": 0.78,
3367
+ "learning_rate": 1.0976915057893534e-05,
3368
+ "loss": 0.2122,
3369
+ "step": 280000
3370
+ },
3371
+ {
3372
+ "epoch": 0.78,
3373
+ "learning_rate": 1.0907230977639773e-05,
3374
+ "loss": 0.2212,
3375
+ "step": 280500
3376
+ },
3377
+ {
3378
+ "epoch": 0.78,
3379
+ "learning_rate": 1.0837546897386012e-05,
3380
+ "loss": 0.2096,
3381
+ "step": 281000
3382
+ },
3383
+ {
3384
+ "epoch": 0.78,
3385
+ "learning_rate": 1.076786281713225e-05,
3386
+ "loss": 0.2091,
3387
+ "step": 281500
3388
+ },
3389
+ {
3390
+ "epoch": 0.79,
3391
+ "learning_rate": 1.0698178736878488e-05,
3392
+ "loss": 0.2124,
3393
+ "step": 282000
3394
+ },
3395
+ {
3396
+ "epoch": 0.79,
3397
+ "learning_rate": 1.0628494656624727e-05,
3398
+ "loss": 0.2175,
3399
+ "step": 282500
3400
+ },
3401
+ {
3402
+ "epoch": 0.79,
3403
+ "learning_rate": 1.0558810576370966e-05,
3404
+ "loss": 0.2124,
3405
+ "step": 283000
3406
+ },
3407
+ {
3408
+ "epoch": 0.79,
3409
+ "learning_rate": 1.0489126496117203e-05,
3410
+ "loss": 0.2046,
3411
+ "step": 283500
3412
+ },
3413
+ {
3414
+ "epoch": 0.79,
3415
+ "learning_rate": 1.0419442415863442e-05,
3416
+ "loss": 0.2072,
3417
+ "step": 284000
3418
+ },
3419
+ {
3420
+ "epoch": 0.79,
3421
+ "learning_rate": 1.034975833560968e-05,
3422
+ "loss": 0.2181,
3423
+ "step": 284500
3424
+ },
3425
+ {
3426
+ "epoch": 0.79,
3427
+ "learning_rate": 1.028007425535592e-05,
3428
+ "loss": 0.22,
3429
+ "step": 285000
3430
+ },
3431
+ {
3432
+ "epoch": 0.8,
3433
+ "learning_rate": 1.0210390175102158e-05,
3434
+ "loss": 0.2087,
3435
+ "step": 285500
3436
+ },
3437
+ {
3438
+ "epoch": 0.8,
3439
+ "learning_rate": 1.0140706094848395e-05,
3440
+ "loss": 0.2074,
3441
+ "step": 286000
3442
+ },
3443
+ {
3444
+ "epoch": 0.8,
3445
+ "learning_rate": 1.0071022014594634e-05,
3446
+ "loss": 0.211,
3447
+ "step": 286500
3448
+ },
3449
+ {
3450
+ "epoch": 0.8,
3451
+ "learning_rate": 1.0001337934340872e-05,
3452
+ "loss": 0.2156,
3453
+ "step": 287000
3454
+ },
3455
+ {
3456
+ "epoch": 0.8,
3457
+ "learning_rate": 9.931653854087112e-06,
3458
+ "loss": 0.2081,
3459
+ "step": 287500
3460
+ },
3461
+ {
3462
+ "epoch": 0.8,
3463
+ "learning_rate": 9.861969773833351e-06,
3464
+ "loss": 0.2158,
3465
+ "step": 288000
3466
+ },
3467
+ {
3468
+ "epoch": 0.8,
3469
+ "learning_rate": 9.792285693579588e-06,
3470
+ "loss": 0.2212,
3471
+ "step": 288500
3472
+ },
3473
+ {
3474
+ "epoch": 0.81,
3475
+ "learning_rate": 9.722601613325827e-06,
3476
+ "loss": 0.2113,
3477
+ "step": 289000
3478
+ },
3479
+ {
3480
+ "epoch": 0.81,
3481
+ "learning_rate": 9.652917533072064e-06,
3482
+ "loss": 0.2052,
3483
+ "step": 289500
3484
+ },
3485
+ {
3486
+ "epoch": 0.81,
3487
+ "learning_rate": 9.583233452818303e-06,
3488
+ "loss": 0.2151,
3489
+ "step": 290000
3490
+ },
3491
+ {
3492
+ "epoch": 0.81,
3493
+ "learning_rate": 9.513549372564542e-06,
3494
+ "loss": 0.2147,
3495
+ "step": 290500
3496
+ },
3497
+ {
3498
+ "epoch": 0.81,
3499
+ "learning_rate": 9.44386529231078e-06,
3500
+ "loss": 0.2114,
3501
+ "step": 291000
3502
+ },
3503
+ {
3504
+ "epoch": 0.81,
3505
+ "learning_rate": 9.37418121205702e-06,
3506
+ "loss": 0.2175,
3507
+ "step": 291500
3508
+ },
3509
+ {
3510
+ "epoch": 0.81,
3511
+ "learning_rate": 9.304497131803257e-06,
3512
+ "loss": 0.2145,
3513
+ "step": 292000
3514
+ },
3515
+ {
3516
+ "epoch": 0.82,
3517
+ "learning_rate": 9.234813051549496e-06,
3518
+ "loss": 0.212,
3519
+ "step": 292500
3520
+ },
3521
+ {
3522
+ "epoch": 0.82,
3523
+ "learning_rate": 9.165128971295735e-06,
3524
+ "loss": 0.2117,
3525
+ "step": 293000
3526
+ },
3527
+ {
3528
+ "epoch": 0.82,
3529
+ "learning_rate": 9.095444891041972e-06,
3530
+ "loss": 0.2065,
3531
+ "step": 293500
3532
+ },
3533
+ {
3534
+ "epoch": 0.82,
3535
+ "learning_rate": 9.02576081078821e-06,
3536
+ "loss": 0.2192,
3537
+ "step": 294000
3538
+ },
3539
+ {
3540
+ "epoch": 0.82,
3541
+ "learning_rate": 8.95607673053445e-06,
3542
+ "loss": 0.2074,
3543
+ "step": 294500
3544
+ },
3545
+ {
3546
+ "epoch": 0.82,
3547
+ "learning_rate": 8.886392650280688e-06,
3548
+ "loss": 0.2155,
3549
+ "step": 295000
3550
+ },
3551
+ {
3552
+ "epoch": 0.82,
3553
+ "learning_rate": 8.816708570026927e-06,
3554
+ "loss": 0.2099,
3555
+ "step": 295500
3556
+ },
3557
+ {
3558
+ "epoch": 0.83,
3559
+ "learning_rate": 8.747024489773165e-06,
3560
+ "loss": 0.2192,
3561
+ "step": 296000
3562
+ },
3563
+ {
3564
+ "epoch": 0.83,
3565
+ "learning_rate": 8.677340409519403e-06,
3566
+ "loss": 0.2117,
3567
+ "step": 296500
3568
+ },
3569
+ {
3570
+ "epoch": 0.83,
3571
+ "learning_rate": 8.60765632926564e-06,
3572
+ "loss": 0.2118,
3573
+ "step": 297000
3574
+ },
3575
+ {
3576
+ "epoch": 0.83,
3577
+ "learning_rate": 8.53797224901188e-06,
3578
+ "loss": 0.2074,
3579
+ "step": 297500
3580
+ },
3581
+ {
3582
+ "epoch": 0.83,
3583
+ "learning_rate": 8.468288168758118e-06,
3584
+ "loss": 0.2071,
3585
+ "step": 298000
3586
+ },
3587
+ {
3588
+ "epoch": 0.83,
3589
+ "learning_rate": 8.398604088504357e-06,
3590
+ "loss": 0.2113,
3591
+ "step": 298500
3592
+ },
3593
+ {
3594
+ "epoch": 0.83,
3595
+ "learning_rate": 8.328920008250596e-06,
3596
+ "loss": 0.2072,
3597
+ "step": 299000
3598
+ },
3599
+ {
3600
+ "epoch": 0.83,
3601
+ "learning_rate": 8.259235927996833e-06,
3602
+ "loss": 0.2048,
3603
+ "step": 299500
3604
+ },
3605
+ {
3606
+ "epoch": 0.84,
3607
+ "learning_rate": 8.189551847743072e-06,
3608
+ "loss": 0.2104,
3609
+ "step": 300000
3610
+ },
3611
+ {
3612
+ "epoch": 0.84,
3613
+ "learning_rate": 8.119867767489311e-06,
3614
+ "loss": 0.218,
3615
+ "step": 300500
3616
+ },
3617
+ {
3618
+ "epoch": 0.84,
3619
+ "learning_rate": 8.050183687235548e-06,
3620
+ "loss": 0.2168,
3621
+ "step": 301000
3622
+ },
3623
+ {
3624
+ "epoch": 0.84,
3625
+ "learning_rate": 7.980499606981789e-06,
3626
+ "loss": 0.2217,
3627
+ "step": 301500
3628
+ },
3629
+ {
3630
+ "epoch": 0.84,
3631
+ "learning_rate": 7.910815526728026e-06,
3632
+ "loss": 0.2126,
3633
+ "step": 302000
3634
+ },
3635
+ {
3636
+ "epoch": 0.84,
3637
+ "learning_rate": 7.841131446474265e-06,
3638
+ "loss": 0.2084,
3639
+ "step": 302500
3640
+ },
3641
+ {
3642
+ "epoch": 0.84,
3643
+ "learning_rate": 7.771447366220504e-06,
3644
+ "loss": 0.2098,
3645
+ "step": 303000
3646
+ },
3647
+ {
3648
+ "epoch": 0.85,
3649
+ "learning_rate": 7.701763285966741e-06,
3650
+ "loss": 0.2131,
3651
+ "step": 303500
3652
+ },
3653
+ {
3654
+ "epoch": 0.85,
3655
+ "learning_rate": 7.63207920571298e-06,
3656
+ "loss": 0.2078,
3657
+ "step": 304000
3658
+ },
3659
+ {
3660
+ "epoch": 0.85,
3661
+ "learning_rate": 7.562395125459218e-06,
3662
+ "loss": 0.2164,
3663
+ "step": 304500
3664
+ },
3665
+ {
3666
+ "epoch": 0.85,
3667
+ "learning_rate": 7.492711045205457e-06,
3668
+ "loss": 0.2114,
3669
+ "step": 305000
3670
+ },
3671
+ {
3672
+ "epoch": 0.85,
3673
+ "learning_rate": 7.423026964951696e-06,
3674
+ "loss": 0.206,
3675
+ "step": 305500
3676
+ },
3677
+ {
3678
+ "epoch": 0.85,
3679
+ "learning_rate": 7.353342884697934e-06,
3680
+ "loss": 0.2108,
3681
+ "step": 306000
3682
+ },
3683
+ {
3684
+ "epoch": 0.85,
3685
+ "learning_rate": 7.2836588044441725e-06,
3686
+ "loss": 0.2156,
3687
+ "step": 306500
3688
+ },
3689
+ {
3690
+ "epoch": 0.86,
3691
+ "learning_rate": 7.2139747241904106e-06,
3692
+ "loss": 0.2124,
3693
+ "step": 307000
3694
+ },
3695
+ {
3696
+ "epoch": 0.86,
3697
+ "learning_rate": 7.1442906439366494e-06,
3698
+ "loss": 0.2197,
3699
+ "step": 307500
3700
+ },
3701
+ {
3702
+ "epoch": 0.86,
3703
+ "learning_rate": 7.074606563682888e-06,
3704
+ "loss": 0.2127,
3705
+ "step": 308000
3706
+ },
3707
+ {
3708
+ "epoch": 0.86,
3709
+ "learning_rate": 7.0049224834291255e-06,
3710
+ "loss": 0.2113,
3711
+ "step": 308500
3712
+ },
3713
+ {
3714
+ "epoch": 0.86,
3715
+ "learning_rate": 6.935238403175364e-06,
3716
+ "loss": 0.2093,
3717
+ "step": 309000
3718
+ },
3719
+ {
3720
+ "epoch": 0.86,
3721
+ "learning_rate": 6.865554322921602e-06,
3722
+ "loss": 0.2191,
3723
+ "step": 309500
3724
+ },
3725
+ {
3726
+ "epoch": 0.86,
3727
+ "learning_rate": 6.795870242667841e-06,
3728
+ "loss": 0.2023,
3729
+ "step": 310000
3730
+ },
3731
+ {
3732
+ "epoch": 0.87,
3733
+ "learning_rate": 6.72618616241408e-06,
3734
+ "loss": 0.2092,
3735
+ "step": 310500
3736
+ },
3737
+ {
3738
+ "epoch": 0.87,
3739
+ "learning_rate": 6.656502082160318e-06,
3740
+ "loss": 0.2088,
3741
+ "step": 311000
3742
+ },
3743
+ {
3744
+ "epoch": 0.87,
3745
+ "learning_rate": 6.586818001906557e-06,
3746
+ "loss": 0.2077,
3747
+ "step": 311500
3748
+ },
3749
+ {
3750
+ "epoch": 0.87,
3751
+ "learning_rate": 6.517133921652794e-06,
3752
+ "loss": 0.2075,
3753
+ "step": 312000
3754
+ },
3755
+ {
3756
+ "epoch": 0.87,
3757
+ "learning_rate": 6.447449841399034e-06,
3758
+ "loss": 0.2123,
3759
+ "step": 312500
3760
+ },
3761
+ {
3762
+ "epoch": 0.87,
3763
+ "learning_rate": 6.377765761145273e-06,
3764
+ "loss": 0.2137,
3765
+ "step": 313000
3766
+ },
3767
+ {
3768
+ "epoch": 0.87,
3769
+ "learning_rate": 6.30808168089151e-06,
3770
+ "loss": 0.2154,
3771
+ "step": 313500
3772
+ },
3773
+ {
3774
+ "epoch": 0.88,
3775
+ "learning_rate": 6.238397600637749e-06,
3776
+ "loss": 0.2137,
3777
+ "step": 314000
3778
+ },
3779
+ {
3780
+ "epoch": 0.88,
3781
+ "learning_rate": 6.168713520383988e-06,
3782
+ "loss": 0.2106,
3783
+ "step": 314500
3784
+ },
3785
+ {
3786
+ "epoch": 0.88,
3787
+ "learning_rate": 6.099029440130226e-06,
3788
+ "loss": 0.2101,
3789
+ "step": 315000
3790
+ },
3791
+ {
3792
+ "epoch": 0.88,
3793
+ "learning_rate": 6.029345359876464e-06,
3794
+ "loss": 0.2154,
3795
+ "step": 315500
3796
+ },
3797
+ {
3798
+ "epoch": 0.88,
3799
+ "learning_rate": 5.959661279622703e-06,
3800
+ "loss": 0.2122,
3801
+ "step": 316000
3802
+ },
3803
+ {
3804
+ "epoch": 0.88,
3805
+ "learning_rate": 5.889977199368942e-06,
3806
+ "loss": 0.2119,
3807
+ "step": 316500
3808
+ },
3809
+ {
3810
+ "epoch": 0.88,
3811
+ "learning_rate": 5.82029311911518e-06,
3812
+ "loss": 0.2069,
3813
+ "step": 317000
3814
+ },
3815
+ {
3816
+ "epoch": 0.88,
3817
+ "learning_rate": 5.750609038861418e-06,
3818
+ "loss": 0.2166,
3819
+ "step": 317500
3820
+ },
3821
+ {
3822
+ "epoch": 0.89,
3823
+ "learning_rate": 5.6809249586076566e-06,
3824
+ "loss": 0.2135,
3825
+ "step": 318000
3826
+ },
3827
+ {
3828
+ "epoch": 0.89,
3829
+ "learning_rate": 5.611240878353895e-06,
3830
+ "loss": 0.216,
3831
+ "step": 318500
3832
+ },
3833
+ {
3834
+ "epoch": 0.89,
3835
+ "learning_rate": 5.5415567981001335e-06,
3836
+ "loss": 0.2078,
3837
+ "step": 319000
3838
+ },
3839
+ {
3840
+ "epoch": 0.89,
3841
+ "learning_rate": 5.471872717846372e-06,
3842
+ "loss": 0.2033,
3843
+ "step": 319500
3844
+ },
3845
+ {
3846
+ "epoch": 0.89,
3847
+ "learning_rate": 5.40218863759261e-06,
3848
+ "loss": 0.218,
3849
+ "step": 320000
3850
+ },
3851
+ {
3852
+ "epoch": 0.89,
3853
+ "learning_rate": 5.3325045573388484e-06,
3854
+ "loss": 0.2128,
3855
+ "step": 320500
3856
+ },
3857
+ {
3858
+ "epoch": 0.89,
3859
+ "learning_rate": 5.262820477085087e-06,
3860
+ "loss": 0.2115,
3861
+ "step": 321000
3862
+ },
3863
+ {
3864
+ "epoch": 0.9,
3865
+ "learning_rate": 5.193136396831326e-06,
3866
+ "loss": 0.2094,
3867
+ "step": 321500
3868
+ },
3869
+ {
3870
+ "epoch": 0.9,
3871
+ "learning_rate": 5.123452316577564e-06,
3872
+ "loss": 0.2096,
3873
+ "step": 322000
3874
+ },
3875
+ {
3876
+ "epoch": 0.9,
3877
+ "learning_rate": 5.053768236323802e-06,
3878
+ "loss": 0.2107,
3879
+ "step": 322500
3880
+ },
3881
+ {
3882
+ "epoch": 0.9,
3883
+ "learning_rate": 4.984084156070041e-06,
3884
+ "loss": 0.2123,
3885
+ "step": 323000
3886
+ },
3887
+ {
3888
+ "epoch": 0.9,
3889
+ "learning_rate": 4.91440007581628e-06,
3890
+ "loss": 0.2089,
3891
+ "step": 323500
3892
+ },
3893
+ {
3894
+ "epoch": 0.9,
3895
+ "learning_rate": 4.844715995562518e-06,
3896
+ "loss": 0.2103,
3897
+ "step": 324000
3898
+ },
3899
+ {
3900
+ "epoch": 0.9,
3901
+ "learning_rate": 4.775031915308757e-06,
3902
+ "loss": 0.2089,
3903
+ "step": 324500
3904
+ },
3905
+ {
3906
+ "epoch": 0.91,
3907
+ "learning_rate": 4.705347835054995e-06,
3908
+ "loss": 0.2041,
3909
+ "step": 325000
3910
+ },
3911
+ {
3912
+ "epoch": 0.91,
3913
+ "learning_rate": 4.635663754801233e-06,
3914
+ "loss": 0.2033,
3915
+ "step": 325500
3916
+ },
3917
+ {
3918
+ "epoch": 0.91,
3919
+ "learning_rate": 4.565979674547472e-06,
3920
+ "loss": 0.2161,
3921
+ "step": 326000
3922
+ },
3923
+ {
3924
+ "epoch": 0.91,
3925
+ "learning_rate": 4.496295594293711e-06,
3926
+ "loss": 0.2082,
3927
+ "step": 326500
3928
+ },
3929
+ {
3930
+ "epoch": 0.91,
3931
+ "learning_rate": 4.426611514039949e-06,
3932
+ "loss": 0.2064,
3933
+ "step": 327000
3934
+ },
3935
+ {
3936
+ "epoch": 0.91,
3937
+ "learning_rate": 4.356927433786187e-06,
3938
+ "loss": 0.2146,
3939
+ "step": 327500
3940
+ },
3941
+ {
3942
+ "epoch": 0.91,
3943
+ "learning_rate": 4.287243353532426e-06,
3944
+ "loss": 0.2149,
3945
+ "step": 328000
3946
+ },
3947
+ {
3948
+ "epoch": 0.92,
3949
+ "learning_rate": 4.2175592732786646e-06,
3950
+ "loss": 0.2031,
3951
+ "step": 328500
3952
+ },
3953
+ {
3954
+ "epoch": 0.92,
3955
+ "learning_rate": 4.147875193024903e-06,
3956
+ "loss": 0.2051,
3957
+ "step": 329000
3958
+ },
3959
+ {
3960
+ "epoch": 0.92,
3961
+ "learning_rate": 4.078191112771141e-06,
3962
+ "loss": 0.2138,
3963
+ "step": 329500
3964
+ },
3965
+ {
3966
+ "epoch": 0.92,
3967
+ "learning_rate": 4.0085070325173795e-06,
3968
+ "loss": 0.2039,
3969
+ "step": 330000
3970
+ },
3971
+ {
3972
+ "epoch": 0.92,
3973
+ "learning_rate": 3.9388229522636175e-06,
3974
+ "loss": 0.2155,
3975
+ "step": 330500
3976
+ },
3977
+ {
3978
+ "epoch": 0.92,
3979
+ "learning_rate": 3.869138872009856e-06,
3980
+ "loss": 0.2039,
3981
+ "step": 331000
3982
+ },
3983
+ {
3984
+ "epoch": 0.92,
3985
+ "learning_rate": 3.799454791756095e-06,
3986
+ "loss": 0.2121,
3987
+ "step": 331500
3988
+ },
3989
+ {
3990
+ "epoch": 0.93,
3991
+ "learning_rate": 3.7297707115023333e-06,
3992
+ "loss": 0.2129,
3993
+ "step": 332000
3994
+ },
3995
+ {
3996
+ "epoch": 0.93,
3997
+ "learning_rate": 3.6600866312485714e-06,
3998
+ "loss": 0.2153,
3999
+ "step": 332500
4000
+ },
4001
+ {
4002
+ "epoch": 0.93,
4003
+ "learning_rate": 3.59040255099481e-06,
4004
+ "loss": 0.2072,
4005
+ "step": 333000
4006
+ },
4007
+ {
4008
+ "epoch": 0.93,
4009
+ "learning_rate": 3.5207184707410487e-06,
4010
+ "loss": 0.2059,
4011
+ "step": 333500
4012
+ },
4013
+ {
4014
+ "epoch": 0.93,
4015
+ "learning_rate": 3.451034390487287e-06,
4016
+ "loss": 0.2111,
4017
+ "step": 334000
4018
+ },
4019
+ {
4020
+ "epoch": 0.93,
4021
+ "learning_rate": 3.3813503102335256e-06,
4022
+ "loss": 0.2161,
4023
+ "step": 334500
4024
+ },
4025
+ {
4026
+ "epoch": 0.93,
4027
+ "learning_rate": 3.3116662299797636e-06,
4028
+ "loss": 0.2153,
4029
+ "step": 335000
4030
+ },
4031
+ {
4032
+ "epoch": 0.94,
4033
+ "learning_rate": 3.241982149726002e-06,
4034
+ "loss": 0.2092,
4035
+ "step": 335500
4036
+ },
4037
+ {
4038
+ "epoch": 0.94,
4039
+ "learning_rate": 3.172298069472241e-06,
4040
+ "loss": 0.2151,
4041
+ "step": 336000
4042
+ },
4043
+ {
4044
+ "epoch": 0.94,
4045
+ "learning_rate": 3.1026139892184794e-06,
4046
+ "loss": 0.2036,
4047
+ "step": 336500
4048
+ },
4049
+ {
4050
+ "epoch": 0.94,
4051
+ "learning_rate": 3.032929908964718e-06,
4052
+ "loss": 0.214,
4053
+ "step": 337000
4054
+ },
4055
+ {
4056
+ "epoch": 0.94,
4057
+ "learning_rate": 2.963245828710956e-06,
4058
+ "loss": 0.214,
4059
+ "step": 337500
4060
+ },
4061
+ {
4062
+ "epoch": 0.94,
4063
+ "learning_rate": 2.8935617484571948e-06,
4064
+ "loss": 0.2063,
4065
+ "step": 338000
4066
+ },
4067
+ {
4068
+ "epoch": 0.94,
4069
+ "learning_rate": 2.823877668203433e-06,
4070
+ "loss": 0.2064,
4071
+ "step": 338500
4072
+ },
4073
+ {
4074
+ "epoch": 0.94,
4075
+ "learning_rate": 2.7541935879496717e-06,
4076
+ "loss": 0.2051,
4077
+ "step": 339000
4078
+ },
4079
+ {
4080
+ "epoch": 0.95,
4081
+ "learning_rate": 2.68450950769591e-06,
4082
+ "loss": 0.2069,
4083
+ "step": 339500
4084
+ },
4085
+ {
4086
+ "epoch": 0.95,
4087
+ "learning_rate": 2.614825427442148e-06,
4088
+ "loss": 0.2058,
4089
+ "step": 340000
4090
+ },
4091
+ {
4092
+ "epoch": 0.95,
4093
+ "learning_rate": 2.545141347188387e-06,
4094
+ "loss": 0.2122,
4095
+ "step": 340500
4096
+ },
4097
+ {
4098
+ "epoch": 0.95,
4099
+ "learning_rate": 2.475457266934625e-06,
4100
+ "loss": 0.202,
4101
+ "step": 341000
4102
+ },
4103
+ {
4104
+ "epoch": 0.95,
4105
+ "learning_rate": 2.405773186680864e-06,
4106
+ "loss": 0.2069,
4107
+ "step": 341500
4108
+ },
4109
+ {
4110
+ "epoch": 0.95,
4111
+ "learning_rate": 2.336089106427102e-06,
4112
+ "loss": 0.2081,
4113
+ "step": 342000
4114
+ },
4115
+ {
4116
+ "epoch": 0.95,
4117
+ "learning_rate": 2.2664050261733405e-06,
4118
+ "loss": 0.2072,
4119
+ "step": 342500
4120
+ },
4121
+ {
4122
+ "epoch": 0.96,
4123
+ "learning_rate": 2.1967209459195793e-06,
4124
+ "loss": 0.2092,
4125
+ "step": 343000
4126
+ },
4127
+ {
4128
+ "epoch": 0.96,
4129
+ "learning_rate": 2.1270368656658174e-06,
4130
+ "loss": 0.2121,
4131
+ "step": 343500
4132
+ },
4133
+ {
4134
+ "epoch": 0.96,
4135
+ "learning_rate": 2.0573527854120563e-06,
4136
+ "loss": 0.2068,
4137
+ "step": 344000
4138
+ },
4139
+ {
4140
+ "epoch": 0.96,
4141
+ "learning_rate": 1.9876687051582943e-06,
4142
+ "loss": 0.201,
4143
+ "step": 344500
4144
+ },
4145
+ {
4146
+ "epoch": 0.96,
4147
+ "learning_rate": 1.9179846249045327e-06,
4148
+ "loss": 0.2171,
4149
+ "step": 345000
4150
+ },
4151
+ {
4152
+ "epoch": 0.96,
4153
+ "learning_rate": 1.8483005446507714e-06,
4154
+ "loss": 0.2043,
4155
+ "step": 345500
4156
+ },
4157
+ {
4158
+ "epoch": 0.96,
4159
+ "learning_rate": 1.7786164643970096e-06,
4160
+ "loss": 0.2038,
4161
+ "step": 346000
4162
+ },
4163
+ {
4164
+ "epoch": 0.97,
4165
+ "learning_rate": 1.7089323841432483e-06,
4166
+ "loss": 0.2107,
4167
+ "step": 346500
4168
+ },
4169
+ {
4170
+ "epoch": 0.97,
4171
+ "learning_rate": 1.6392483038894866e-06,
4172
+ "loss": 0.2078,
4173
+ "step": 347000
4174
+ },
4175
+ {
4176
+ "epoch": 0.97,
4177
+ "learning_rate": 1.5695642236357252e-06,
4178
+ "loss": 0.209,
4179
+ "step": 347500
4180
+ },
4181
+ {
4182
+ "epoch": 0.97,
4183
+ "learning_rate": 1.4998801433819637e-06,
4184
+ "loss": 0.2118,
4185
+ "step": 348000
4186
+ },
4187
+ {
4188
+ "epoch": 0.97,
4189
+ "learning_rate": 1.4301960631282021e-06,
4190
+ "loss": 0.2247,
4191
+ "step": 348500
4192
+ },
4193
+ {
4194
+ "epoch": 0.97,
4195
+ "learning_rate": 1.3605119828744404e-06,
4196
+ "loss": 0.2094,
4197
+ "step": 349000
4198
+ },
4199
+ {
4200
+ "epoch": 0.97,
4201
+ "learning_rate": 1.2908279026206788e-06,
4202
+ "loss": 0.2154,
4203
+ "step": 349500
4204
+ },
4205
+ {
4206
+ "epoch": 0.98,
4207
+ "learning_rate": 1.2211438223669173e-06,
4208
+ "loss": 0.2062,
4209
+ "step": 350000
4210
+ },
4211
+ {
4212
+ "epoch": 0.98,
4213
+ "learning_rate": 1.151459742113156e-06,
4214
+ "loss": 0.2105,
4215
+ "step": 350500
4216
+ },
4217
+ {
4218
+ "epoch": 0.98,
4219
+ "learning_rate": 1.0817756618593944e-06,
4220
+ "loss": 0.2177,
4221
+ "step": 351000
4222
+ },
4223
+ {
4224
+ "epoch": 0.98,
4225
+ "learning_rate": 1.0120915816056329e-06,
4226
+ "loss": 0.2062,
4227
+ "step": 351500
4228
+ },
4229
+ {
4230
+ "epoch": 0.98,
4231
+ "learning_rate": 9.424075013518711e-07,
4232
+ "loss": 0.205,
4233
+ "step": 352000
4234
+ },
4235
+ {
4236
+ "epoch": 0.98,
4237
+ "learning_rate": 8.727234210981097e-07,
4238
+ "loss": 0.2032,
4239
+ "step": 352500
4240
+ },
4241
+ {
4242
+ "epoch": 0.98,
4243
+ "learning_rate": 8.030393408443481e-07,
4244
+ "loss": 0.2169,
4245
+ "step": 353000
4246
+ },
4247
+ {
4248
+ "epoch": 0.99,
4249
+ "learning_rate": 7.333552605905866e-07,
4250
+ "loss": 0.2153,
4251
+ "step": 353500
4252
+ },
4253
+ {
4254
+ "epoch": 0.99,
4255
+ "learning_rate": 6.636711803368249e-07,
4256
+ "loss": 0.2046,
4257
+ "step": 354000
4258
+ },
4259
+ {
4260
+ "epoch": 0.99,
4261
+ "learning_rate": 5.939871000830635e-07,
4262
+ "loss": 0.2068,
4263
+ "step": 354500
4264
+ },
4265
+ {
4266
+ "epoch": 0.99,
4267
+ "learning_rate": 5.243030198293018e-07,
4268
+ "loss": 0.2188,
4269
+ "step": 355000
4270
+ },
4271
+ {
4272
+ "epoch": 0.99,
4273
+ "learning_rate": 4.5461893957554035e-07,
4274
+ "loss": 0.2139,
4275
+ "step": 355500
4276
+ },
4277
+ {
4278
+ "epoch": 0.99,
4279
+ "learning_rate": 3.849348593217788e-07,
4280
+ "loss": 0.2085,
4281
+ "step": 356000
4282
+ },
4283
+ {
4284
+ "epoch": 0.99,
4285
+ "learning_rate": 3.1525077906801726e-07,
4286
+ "loss": 0.1994,
4287
+ "step": 356500
4288
+ },
4289
+ {
4290
+ "epoch": 1.0,
4291
+ "learning_rate": 2.455666988142557e-07,
4292
+ "loss": 0.216,
4293
+ "step": 357000
4294
+ },
4295
+ {
4296
+ "epoch": 1.0,
4297
+ "learning_rate": 1.7588261856049415e-07,
4298
+ "loss": 0.2123,
4299
+ "step": 357500
4300
+ },
4301
+ {
4302
+ "epoch": 1.0,
4303
+ "learning_rate": 1.061985383067326e-07,
4304
+ "loss": 0.2063,
4305
+ "step": 358000
4306
+ },
4307
+ {
4308
+ "epoch": 1.0,
4309
+ "learning_rate": 3.651445805297105e-08,
4310
+ "loss": 0.2098,
4311
+ "step": 358500
4312
+ }
4313
+ ],
4314
+ "logging_steps": 500,
4315
+ "max_steps": 358762,
4316
+ "num_train_epochs": 1,
4317
+ "save_steps": 500,
4318
+ "total_flos": 1.26740447502336e+17,
4319
+ "trial_name": null,
4320
+ "trial_params": null
4321
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4857a93b7f495250fd40ee4b4e5310a7d0bd6c67cb62f3d79979dd30c751c2ee
3
+ size 4600