Hugging Face's logo

robzchhangte
/
MizBERT

Fill-Mask
Transformers
Safetensors
bert
Inference Endpoints
Model card Files Community
robzchhangte commited on
Commit
0f74d11
•
1 Parent(s): e6dd110

Upload 12 files

Browse files
Files changed (12) hide show
  1. config.json +26 -0
  2. generation_config.json +5 -0
  3. model.safetensors +3 -0
  4. optimizer.pt +3 -0
  5. rng_state.pth +3 -0
  6. scheduler.pt +3 -0
  7. special_tokens_map.json +37 -0
  8. tokenizer.json +0 -0
  9. tokenizer_config.json +55 -0
  10. trainer_state.json +3225 -0
  11. training_args.bin +3 -0
  12. vocab.txt +0 -0
config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "bert-base-uncased",
3
+ "architectures": [
4
+ "BertForMaskedLM"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.1,
7
+ "classifier_dropout": null,
8
+ "gradient_checkpointing": false,
9
+ "hidden_act": "gelu",
10
+ "hidden_dropout_prob": 0.1,
11
+ "hidden_size": 768,
12
+ "initializer_range": 0.02,
13
+ "intermediate_size": 3072,
14
+ "layer_norm_eps": 1e-12,
15
+ "max_position_embeddings": 512,
16
+ "model_type": "bert",
17
+ "num_attention_heads": 12,
18
+ "num_hidden_layers": 12,
19
+ "pad_token_id": 0,
20
+ "position_embedding_type": "absolute",
21
+ "torch_dtype": "float32",
22
+ "transformers_version": "4.38.0.dev0",
23
+ "type_vocab_size": 2,
24
+ "use_cache": true,
25
+ "vocab_size": 30522
26
+ }
generation_config.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "pad_token_id": 0,
4
+ "transformers_version": "4.38.0.dev0"
5
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:517b1eca84d9acaa3b8e2c90a959942f3dffda565c1ae98d577e5698cbc84176
3
+ size 438080896
optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7f06a611c8f461963216b0292b39a3f96da2c555134e894145e276f69c3cad3a
3
+ size 876283258
rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6c29a8087ff2fdd4581412e1a5084370071c1513e42a5a20e8858a8a45cb3149
3
+ size 14244
scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:46abf0dfa3e244bdfe8423f6807e2fa602164af529deafc67ffbf5286ca4ae12
3
+ size 1064
special_tokens_map.json ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "cls_token": {
3
+ "content": "[CLS]",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "mask_token": {
10
+ "content": "[MASK]",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": {
17
+ "content": "[PAD]",
18
+ "lstrip": false,
19
+ "normalized": false,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ },
23
+ "sep_token": {
24
+ "content": "[SEP]",
25
+ "lstrip": false,
26
+ "normalized": false,
27
+ "rstrip": false,
28
+ "single_word": false
29
+ },
30
+ "unk_token": {
31
+ "content": "[UNK]",
32
+ "lstrip": false,
33
+ "normalized": false,
34
+ "rstrip": false,
35
+ "single_word": false
36
+ }
37
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,55 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": false,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ }
43
+ },
44
+ "clean_up_tokenization_spaces": true,
45
+ "cls_token": "[CLS]",
46
+ "do_lower_case": true,
47
+ "mask_token": "[MASK]",
48
+ "model_max_length": 512,
49
+ "pad_token": "[PAD]",
50
+ "sep_token": "[SEP]",
51
+ "strip_accents": null,
52
+ "tokenize_chinese_chars": true,
53
+ "tokenizer_class": "BertTokenizer",
54
+ "unk_token": "[UNK]"
55
+ }
trainer_state.json ADDED
@@ -0,0 +1,3225 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 24.955603327413776,
5
+ "eval_steps": 500,
6
+ "global_step": 267000,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.05,
13
+ "learning_rate": 4.9906533320871115e-05,
14
+ "loss": 6.5864,
15
+ "step": 500
16
+ },
17
+ {
18
+ "epoch": 0.09,
19
+ "learning_rate": 4.981306664174222e-05,
20
+ "loss": 5.4802,
21
+ "step": 1000
22
+ },
23
+ {
24
+ "epoch": 0.14,
25
+ "learning_rate": 4.9719599962613325e-05,
26
+ "loss": 4.7221,
27
+ "step": 1500
28
+ },
29
+ {
30
+ "epoch": 0.19,
31
+ "learning_rate": 4.962613328348444e-05,
32
+ "loss": 4.2025,
33
+ "step": 2000
34
+ },
35
+ {
36
+ "epoch": 0.23,
37
+ "learning_rate": 4.953266660435555e-05,
38
+ "loss": 3.8489,
39
+ "step": 2500
40
+ },
41
+ {
42
+ "epoch": 0.28,
43
+ "learning_rate": 4.943919992522666e-05,
44
+ "loss": 3.5856,
45
+ "step": 3000
46
+ },
47
+ {
48
+ "epoch": 0.33,
49
+ "learning_rate": 4.934573324609777e-05,
50
+ "loss": 3.3865,
51
+ "step": 3500
52
+ },
53
+ {
54
+ "epoch": 0.37,
55
+ "learning_rate": 4.925226656696888e-05,
56
+ "loss": 3.2125,
57
+ "step": 4000
58
+ },
59
+ {
60
+ "epoch": 0.42,
61
+ "learning_rate": 4.9158799887839984e-05,
62
+ "loss": 3.0721,
63
+ "step": 4500
64
+ },
65
+ {
66
+ "epoch": 0.47,
67
+ "learning_rate": 4.9065333208711096e-05,
68
+ "loss": 2.9681,
69
+ "step": 5000
70
+ },
71
+ {
72
+ "epoch": 0.51,
73
+ "learning_rate": 4.897186652958221e-05,
74
+ "loss": 2.8801,
75
+ "step": 5500
76
+ },
77
+ {
78
+ "epoch": 0.56,
79
+ "learning_rate": 4.8878399850453314e-05,
80
+ "loss": 2.782,
81
+ "step": 6000
82
+ },
83
+ {
84
+ "epoch": 0.61,
85
+ "learning_rate": 4.8784933171324426e-05,
86
+ "loss": 2.7237,
87
+ "step": 6500
88
+ },
89
+ {
90
+ "epoch": 0.65,
91
+ "learning_rate": 4.869146649219554e-05,
92
+ "loss": 2.6715,
93
+ "step": 7000
94
+ },
95
+ {
96
+ "epoch": 0.7,
97
+ "learning_rate": 4.859799981306664e-05,
98
+ "loss": 2.637,
99
+ "step": 7500
100
+ },
101
+ {
102
+ "epoch": 0.75,
103
+ "learning_rate": 4.850453313393775e-05,
104
+ "loss": 2.5817,
105
+ "step": 8000
106
+ },
107
+ {
108
+ "epoch": 0.79,
109
+ "learning_rate": 4.841106645480886e-05,
110
+ "loss": 2.5346,
111
+ "step": 8500
112
+ },
113
+ {
114
+ "epoch": 0.84,
115
+ "learning_rate": 4.831759977567997e-05,
116
+ "loss": 2.49,
117
+ "step": 9000
118
+ },
119
+ {
120
+ "epoch": 0.89,
121
+ "learning_rate": 4.8224133096551085e-05,
122
+ "loss": 2.4597,
123
+ "step": 9500
124
+ },
125
+ {
126
+ "epoch": 0.93,
127
+ "learning_rate": 4.813066641742219e-05,
128
+ "loss": 2.4381,
129
+ "step": 10000
130
+ },
131
+ {
132
+ "epoch": 0.98,
133
+ "learning_rate": 4.8037199738293296e-05,
134
+ "loss": 2.3995,
135
+ "step": 10500
136
+ },
137
+ {
138
+ "epoch": 1.03,
139
+ "learning_rate": 4.794373305916441e-05,
140
+ "loss": 2.385,
141
+ "step": 11000
142
+ },
143
+ {
144
+ "epoch": 1.07,
145
+ "learning_rate": 4.785026638003552e-05,
146
+ "loss": 2.3463,
147
+ "step": 11500
148
+ },
149
+ {
150
+ "epoch": 1.12,
151
+ "learning_rate": 4.775679970090663e-05,
152
+ "loss": 2.3033,
153
+ "step": 12000
154
+ },
155
+ {
156
+ "epoch": 1.17,
157
+ "learning_rate": 4.766333302177774e-05,
158
+ "loss": 2.2915,
159
+ "step": 12500
160
+ },
161
+ {
162
+ "epoch": 1.22,
163
+ "learning_rate": 4.756986634264885e-05,
164
+ "loss": 2.2722,
165
+ "step": 13000
166
+ },
167
+ {
168
+ "epoch": 1.26,
169
+ "learning_rate": 4.7476399663519955e-05,
170
+ "loss": 2.2443,
171
+ "step": 13500
172
+ },
173
+ {
174
+ "epoch": 1.31,
175
+ "learning_rate": 4.738293298439107e-05,
176
+ "loss": 2.2412,
177
+ "step": 14000
178
+ },
179
+ {
180
+ "epoch": 1.36,
181
+ "learning_rate": 4.728946630526218e-05,
182
+ "loss": 2.22,
183
+ "step": 14500
184
+ },
185
+ {
186
+ "epoch": 1.4,
187
+ "learning_rate": 4.7195999626133284e-05,
188
+ "loss": 2.19,
189
+ "step": 15000
190
+ },
191
+ {
192
+ "epoch": 1.45,
193
+ "learning_rate": 4.7102532947004396e-05,
194
+ "loss": 2.1646,
195
+ "step": 15500
196
+ },
197
+ {
198
+ "epoch": 1.5,
199
+ "learning_rate": 4.700906626787551e-05,
200
+ "loss": 2.1665,
201
+ "step": 16000
202
+ },
203
+ {
204
+ "epoch": 1.54,
205
+ "learning_rate": 4.6915599588746614e-05,
206
+ "loss": 2.1406,
207
+ "step": 16500
208
+ },
209
+ {
210
+ "epoch": 1.59,
211
+ "learning_rate": 4.682213290961772e-05,
212
+ "loss": 2.1191,
213
+ "step": 17000
214
+ },
215
+ {
216
+ "epoch": 1.64,
217
+ "learning_rate": 4.672866623048883e-05,
218
+ "loss": 2.1071,
219
+ "step": 17500
220
+ },
221
+ {
222
+ "epoch": 1.68,
223
+ "learning_rate": 4.663519955135994e-05,
224
+ "loss": 2.0976,
225
+ "step": 18000
226
+ },
227
+ {
228
+ "epoch": 1.73,
229
+ "learning_rate": 4.6541732872231055e-05,
230
+ "loss": 2.0935,
231
+ "step": 18500
232
+ },
233
+ {
234
+ "epoch": 1.78,
235
+ "learning_rate": 4.644826619310216e-05,
236
+ "loss": 2.0682,
237
+ "step": 19000
238
+ },
239
+ {
240
+ "epoch": 1.82,
241
+ "learning_rate": 4.6354799513973266e-05,
242
+ "loss": 2.0487,
243
+ "step": 19500
244
+ },
245
+ {
246
+ "epoch": 1.87,
247
+ "learning_rate": 4.626133283484438e-05,
248
+ "loss": 2.0384,
249
+ "step": 20000
250
+ },
251
+ {
252
+ "epoch": 1.92,
253
+ "learning_rate": 4.616786615571549e-05,
254
+ "loss": 2.0184,
255
+ "step": 20500
256
+ },
257
+ {
258
+ "epoch": 1.96,
259
+ "learning_rate": 4.60743994765866e-05,
260
+ "loss": 2.0046,
261
+ "step": 21000
262
+ },
263
+ {
264
+ "epoch": 2.01,
265
+ "learning_rate": 4.598093279745771e-05,
266
+ "loss": 1.9972,
267
+ "step": 21500
268
+ },
269
+ {
270
+ "epoch": 2.06,
271
+ "learning_rate": 4.588746611832882e-05,
272
+ "loss": 1.9796,
273
+ "step": 22000
274
+ },
275
+ {
276
+ "epoch": 2.1,
277
+ "learning_rate": 4.5793999439199925e-05,
278
+ "loss": 1.9909,
279
+ "step": 22500
280
+ },
281
+ {
282
+ "epoch": 2.15,
283
+ "learning_rate": 4.570053276007104e-05,
284
+ "loss": 1.9776,
285
+ "step": 23000
286
+ },
287
+ {
288
+ "epoch": 2.2,
289
+ "learning_rate": 4.560706608094214e-05,
290
+ "loss": 1.9556,
291
+ "step": 23500
292
+ },
293
+ {
294
+ "epoch": 2.24,
295
+ "learning_rate": 4.5513599401813255e-05,
296
+ "loss": 1.9427,
297
+ "step": 24000
298
+ },
299
+ {
300
+ "epoch": 2.29,
301
+ "learning_rate": 4.542013272268437e-05,
302
+ "loss": 1.9452,
303
+ "step": 24500
304
+ },
305
+ {
306
+ "epoch": 2.34,
307
+ "learning_rate": 4.532666604355548e-05,
308
+ "loss": 1.9365,
309
+ "step": 25000
310
+ },
311
+ {
312
+ "epoch": 2.38,
313
+ "learning_rate": 4.5233199364426584e-05,
314
+ "loss": 1.925,
315
+ "step": 25500
316
+ },
317
+ {
318
+ "epoch": 2.43,
319
+ "learning_rate": 4.513973268529769e-05,
320
+ "loss": 1.9045,
321
+ "step": 26000
322
+ },
323
+ {
324
+ "epoch": 2.48,
325
+ "learning_rate": 4.50462660061688e-05,
326
+ "loss": 1.8886,
327
+ "step": 26500
328
+ },
329
+ {
330
+ "epoch": 2.52,
331
+ "learning_rate": 4.4952799327039914e-05,
332
+ "loss": 1.9086,
333
+ "step": 27000
334
+ },
335
+ {
336
+ "epoch": 2.57,
337
+ "learning_rate": 4.4859332647911026e-05,
338
+ "loss": 1.882,
339
+ "step": 27500
340
+ },
341
+ {
342
+ "epoch": 2.62,
343
+ "learning_rate": 4.476586596878213e-05,
344
+ "loss": 1.8831,
345
+ "step": 28000
346
+ },
347
+ {
348
+ "epoch": 2.66,
349
+ "learning_rate": 4.4672399289653237e-05,
350
+ "loss": 1.8745,
351
+ "step": 28500
352
+ },
353
+ {
354
+ "epoch": 2.71,
355
+ "learning_rate": 4.457893261052435e-05,
356
+ "loss": 1.8645,
357
+ "step": 29000
358
+ },
359
+ {
360
+ "epoch": 2.76,
361
+ "learning_rate": 4.448546593139546e-05,
362
+ "loss": 1.8481,
363
+ "step": 29500
364
+ },
365
+ {
366
+ "epoch": 2.8,
367
+ "learning_rate": 4.439199925226657e-05,
368
+ "loss": 1.8524,
369
+ "step": 30000
370
+ },
371
+ {
372
+ "epoch": 2.85,
373
+ "learning_rate": 4.429853257313768e-05,
374
+ "loss": 1.8299,
375
+ "step": 30500
376
+ },
377
+ {
378
+ "epoch": 2.9,
379
+ "learning_rate": 4.420506589400879e-05,
380
+ "loss": 1.835,
381
+ "step": 31000
382
+ },
383
+ {
384
+ "epoch": 2.94,
385
+ "learning_rate": 4.4111599214879896e-05,
386
+ "loss": 1.8246,
387
+ "step": 31500
388
+ },
389
+ {
390
+ "epoch": 2.99,
391
+ "learning_rate": 4.401813253575101e-05,
392
+ "loss": 1.8154,
393
+ "step": 32000
394
+ },
395
+ {
396
+ "epoch": 3.04,
397
+ "learning_rate": 4.392466585662211e-05,
398
+ "loss": 1.8097,
399
+ "step": 32500
400
+ },
401
+ {
402
+ "epoch": 3.08,
403
+ "learning_rate": 4.3831199177493225e-05,
404
+ "loss": 1.7977,
405
+ "step": 33000
406
+ },
407
+ {
408
+ "epoch": 3.13,
409
+ "learning_rate": 4.373773249836434e-05,
410
+ "loss": 1.7857,
411
+ "step": 33500
412
+ },
413
+ {
414
+ "epoch": 3.18,
415
+ "learning_rate": 4.364426581923545e-05,
416
+ "loss": 1.7933,
417
+ "step": 34000
418
+ },
419
+ {
420
+ "epoch": 3.22,
421
+ "learning_rate": 4.3550799140106555e-05,
422
+ "loss": 1.774,
423
+ "step": 34500
424
+ },
425
+ {
426
+ "epoch": 3.27,
427
+ "learning_rate": 4.345733246097766e-05,
428
+ "loss": 1.7747,
429
+ "step": 35000
430
+ },
431
+ {
432
+ "epoch": 3.32,
433
+ "learning_rate": 4.336386578184877e-05,
434
+ "loss": 1.7663,
435
+ "step": 35500
436
+ },
437
+ {
438
+ "epoch": 3.36,
439
+ "learning_rate": 4.3270399102719884e-05,
440
+ "loss": 1.7746,
441
+ "step": 36000
442
+ },
443
+ {
444
+ "epoch": 3.41,
445
+ "learning_rate": 4.3176932423590996e-05,
446
+ "loss": 1.7586,
447
+ "step": 36500
448
+ },
449
+ {
450
+ "epoch": 3.46,
451
+ "learning_rate": 4.30834657444621e-05,
452
+ "loss": 1.7545,
453
+ "step": 37000
454
+ },
455
+ {
456
+ "epoch": 3.51,
457
+ "learning_rate": 4.298999906533321e-05,
458
+ "loss": 1.7424,
459
+ "step": 37500
460
+ },
461
+ {
462
+ "epoch": 3.55,
463
+ "learning_rate": 4.289653238620432e-05,
464
+ "loss": 1.7472,
465
+ "step": 38000
466
+ },
467
+ {
468
+ "epoch": 3.6,
469
+ "learning_rate": 4.280306570707543e-05,
470
+ "loss": 1.7587,
471
+ "step": 38500
472
+ },
473
+ {
474
+ "epoch": 3.65,
475
+ "learning_rate": 4.2709599027946537e-05,
476
+ "loss": 1.7486,
477
+ "step": 39000
478
+ },
479
+ {
480
+ "epoch": 3.69,
481
+ "learning_rate": 4.261613234881765e-05,
482
+ "loss": 1.7288,
483
+ "step": 39500
484
+ },
485
+ {
486
+ "epoch": 3.74,
487
+ "learning_rate": 4.252266566968876e-05,
488
+ "loss": 1.7361,
489
+ "step": 40000
490
+ },
491
+ {
492
+ "epoch": 3.79,
493
+ "learning_rate": 4.2429198990559866e-05,
494
+ "loss": 1.7089,
495
+ "step": 40500
496
+ },
497
+ {
498
+ "epoch": 3.83,
499
+ "learning_rate": 4.233573231143098e-05,
500
+ "loss": 1.7253,
501
+ "step": 41000
502
+ },
503
+ {
504
+ "epoch": 3.88,
505
+ "learning_rate": 4.2242265632302084e-05,
506
+ "loss": 1.7134,
507
+ "step": 41500
508
+ },
509
+ {
510
+ "epoch": 3.93,
511
+ "learning_rate": 4.2148798953173196e-05,
512
+ "loss": 1.6936,
513
+ "step": 42000
514
+ },
515
+ {
516
+ "epoch": 3.97,
517
+ "learning_rate": 4.205533227404431e-05,
518
+ "loss": 1.7115,
519
+ "step": 42500
520
+ },
521
+ {
522
+ "epoch": 4.02,
523
+ "learning_rate": 4.196186559491542e-05,
524
+ "loss": 1.6981,
525
+ "step": 43000
526
+ },
527
+ {
528
+ "epoch": 4.07,
529
+ "learning_rate": 4.1868398915786525e-05,
530
+ "loss": 1.6913,
531
+ "step": 43500
532
+ },
533
+ {
534
+ "epoch": 4.11,
535
+ "learning_rate": 4.177493223665763e-05,
536
+ "loss": 1.6916,
537
+ "step": 44000
538
+ },
539
+ {
540
+ "epoch": 4.16,
541
+ "learning_rate": 4.168146555752874e-05,
542
+ "loss": 1.6953,
543
+ "step": 44500
544
+ },
545
+ {
546
+ "epoch": 4.21,
547
+ "learning_rate": 4.1587998878399855e-05,
548
+ "loss": 1.6778,
549
+ "step": 45000
550
+ },
551
+ {
552
+ "epoch": 4.25,
553
+ "learning_rate": 4.149453219927097e-05,
554
+ "loss": 1.6706,
555
+ "step": 45500
556
+ },
557
+ {
558
+ "epoch": 4.3,
559
+ "learning_rate": 4.140106552014207e-05,
560
+ "loss": 1.6703,
561
+ "step": 46000
562
+ },
563
+ {
564
+ "epoch": 4.35,
565
+ "learning_rate": 4.130759884101318e-05,
566
+ "loss": 1.6639,
567
+ "step": 46500
568
+ },
569
+ {
570
+ "epoch": 4.39,
571
+ "learning_rate": 4.121413216188429e-05,
572
+ "loss": 1.6728,
573
+ "step": 47000
574
+ },
575
+ {
576
+ "epoch": 4.44,
577
+ "learning_rate": 4.11206654827554e-05,
578
+ "loss": 1.6553,
579
+ "step": 47500
580
+ },
581
+ {
582
+ "epoch": 4.49,
583
+ "learning_rate": 4.102719880362651e-05,
584
+ "loss": 1.6433,
585
+ "step": 48000
586
+ },
587
+ {
588
+ "epoch": 4.53,
589
+ "learning_rate": 4.093373212449762e-05,
590
+ "loss": 1.6499,
591
+ "step": 48500
592
+ },
593
+ {
594
+ "epoch": 4.58,
595
+ "learning_rate": 4.084026544536873e-05,
596
+ "loss": 1.6553,
597
+ "step": 49000
598
+ },
599
+ {
600
+ "epoch": 4.63,
601
+ "learning_rate": 4.0746798766239837e-05,
602
+ "loss": 1.6401,
603
+ "step": 49500
604
+ },
605
+ {
606
+ "epoch": 4.67,
607
+ "learning_rate": 4.065333208711095e-05,
608
+ "loss": 1.6444,
609
+ "step": 50000
610
+ },
611
+ {
612
+ "epoch": 4.72,
613
+ "learning_rate": 4.0559865407982054e-05,
614
+ "loss": 1.6398,
615
+ "step": 50500
616
+ },
617
+ {
618
+ "epoch": 4.77,
619
+ "learning_rate": 4.0466398728853166e-05,
620
+ "loss": 1.6338,
621
+ "step": 51000
622
+ },
623
+ {
624
+ "epoch": 4.81,
625
+ "learning_rate": 4.037293204972428e-05,
626
+ "loss": 1.6194,
627
+ "step": 51500
628
+ },
629
+ {
630
+ "epoch": 4.86,
631
+ "learning_rate": 4.027946537059539e-05,
632
+ "loss": 1.6327,
633
+ "step": 52000
634
+ },
635
+ {
636
+ "epoch": 4.91,
637
+ "learning_rate": 4.018599869146649e-05,
638
+ "loss": 1.6232,
639
+ "step": 52500
640
+ },
641
+ {
642
+ "epoch": 4.95,
643
+ "learning_rate": 4.00925320123376e-05,
644
+ "loss": 1.6296,
645
+ "step": 53000
646
+ },
647
+ {
648
+ "epoch": 5.0,
649
+ "learning_rate": 3.999906533320871e-05,
650
+ "loss": 1.6152,
651
+ "step": 53500
652
+ },
653
+ {
654
+ "epoch": 5.05,
655
+ "learning_rate": 3.9905598654079825e-05,
656
+ "loss": 1.6012,
657
+ "step": 54000
658
+ },
659
+ {
660
+ "epoch": 5.09,
661
+ "learning_rate": 3.981213197495093e-05,
662
+ "loss": 1.6087,
663
+ "step": 54500
664
+ },
665
+ {
666
+ "epoch": 5.14,
667
+ "learning_rate": 3.971866529582204e-05,
668
+ "loss": 1.5971,
669
+ "step": 55000
670
+ },
671
+ {
672
+ "epoch": 5.19,
673
+ "learning_rate": 3.962519861669315e-05,
674
+ "loss": 1.5956,
675
+ "step": 55500
676
+ },
677
+ {
678
+ "epoch": 5.23,
679
+ "learning_rate": 3.953173193756426e-05,
680
+ "loss": 1.5947,
681
+ "step": 56000
682
+ },
683
+ {
684
+ "epoch": 5.28,
685
+ "learning_rate": 3.943826525843537e-05,
686
+ "loss": 1.5993,
687
+ "step": 56500
688
+ },
689
+ {
690
+ "epoch": 5.33,
691
+ "learning_rate": 3.934479857930648e-05,
692
+ "loss": 1.5816,
693
+ "step": 57000
694
+ },
695
+ {
696
+ "epoch": 5.37,
697
+ "learning_rate": 3.925133190017759e-05,
698
+ "loss": 1.5837,
699
+ "step": 57500
700
+ },
701
+ {
702
+ "epoch": 5.42,
703
+ "learning_rate": 3.91578652210487e-05,
704
+ "loss": 1.5854,
705
+ "step": 58000
706
+ },
707
+ {
708
+ "epoch": 5.47,
709
+ "learning_rate": 3.906439854191981e-05,
710
+ "loss": 1.5734,
711
+ "step": 58500
712
+ },
713
+ {
714
+ "epoch": 5.51,
715
+ "learning_rate": 3.897093186279092e-05,
716
+ "loss": 1.578,
717
+ "step": 59000
718
+ },
719
+ {
720
+ "epoch": 5.56,
721
+ "learning_rate": 3.8877465183662024e-05,
722
+ "loss": 1.5817,
723
+ "step": 59500
724
+ },
725
+ {
726
+ "epoch": 5.61,
727
+ "learning_rate": 3.8783998504533137e-05,
728
+ "loss": 1.578,
729
+ "step": 60000
730
+ },
731
+ {
732
+ "epoch": 5.65,
733
+ "learning_rate": 3.869053182540425e-05,
734
+ "loss": 1.5732,
735
+ "step": 60500
736
+ },
737
+ {
738
+ "epoch": 5.7,
739
+ "learning_rate": 3.8597065146275354e-05,
740
+ "loss": 1.5606,
741
+ "step": 61000
742
+ },
743
+ {
744
+ "epoch": 5.75,
745
+ "learning_rate": 3.850359846714646e-05,
746
+ "loss": 1.5709,
747
+ "step": 61500
748
+ },
749
+ {
750
+ "epoch": 5.79,
751
+ "learning_rate": 3.841013178801757e-05,
752
+ "loss": 1.5607,
753
+ "step": 62000
754
+ },
755
+ {
756
+ "epoch": 5.84,
757
+ "learning_rate": 3.8316665108888684e-05,
758
+ "loss": 1.5687,
759
+ "step": 62500
760
+ },
761
+ {
762
+ "epoch": 5.89,
763
+ "learning_rate": 3.8223198429759796e-05,
764
+ "loss": 1.5488,
765
+ "step": 63000
766
+ },
767
+ {
768
+ "epoch": 5.94,
769
+ "learning_rate": 3.81297317506309e-05,
770
+ "loss": 1.5601,
771
+ "step": 63500
772
+ },
773
+ {
774
+ "epoch": 5.98,
775
+ "learning_rate": 3.803626507150201e-05,
776
+ "loss": 1.5611,
777
+ "step": 64000
778
+ },
779
+ {
780
+ "epoch": 6.03,
781
+ "learning_rate": 3.794279839237312e-05,
782
+ "loss": 1.5515,
783
+ "step": 64500
784
+ },
785
+ {
786
+ "epoch": 6.08,
787
+ "learning_rate": 3.784933171324423e-05,
788
+ "loss": 1.5412,
789
+ "step": 65000
790
+ },
791
+ {
792
+ "epoch": 6.12,
793
+ "learning_rate": 3.775586503411534e-05,
794
+ "loss": 1.5434,
795
+ "step": 65500
796
+ },
797
+ {
798
+ "epoch": 6.17,
799
+ "learning_rate": 3.766239835498645e-05,
800
+ "loss": 1.5372,
801
+ "step": 66000
802
+ },
803
+ {
804
+ "epoch": 6.22,
805
+ "learning_rate": 3.756893167585756e-05,
806
+ "loss": 1.5293,
807
+ "step": 66500
808
+ },
809
+ {
810
+ "epoch": 6.26,
811
+ "learning_rate": 3.7475464996728665e-05,
812
+ "loss": 1.5335,
813
+ "step": 67000
814
+ },
815
+ {
816
+ "epoch": 6.31,
817
+ "learning_rate": 3.738199831759978e-05,
818
+ "loss": 1.5296,
819
+ "step": 67500
820
+ },
821
+ {
822
+ "epoch": 6.36,
823
+ "learning_rate": 3.728853163847088e-05,
824
+ "loss": 1.5238,
825
+ "step": 68000
826
+ },
827
+ {
828
+ "epoch": 6.4,
829
+ "learning_rate": 3.7195064959341995e-05,
830
+ "loss": 1.5269,
831
+ "step": 68500
832
+ },
833
+ {
834
+ "epoch": 6.45,
835
+ "learning_rate": 3.710159828021311e-05,
836
+ "loss": 1.5233,
837
+ "step": 69000
838
+ },
839
+ {
840
+ "epoch": 6.5,
841
+ "learning_rate": 3.700813160108422e-05,
842
+ "loss": 1.5234,
843
+ "step": 69500
844
+ },
845
+ {
846
+ "epoch": 6.54,
847
+ "learning_rate": 3.6914664921955325e-05,
848
+ "loss": 1.5277,
849
+ "step": 70000
850
+ },
851
+ {
852
+ "epoch": 6.59,
853
+ "learning_rate": 3.682119824282643e-05,
854
+ "loss": 1.5185,
855
+ "step": 70500
856
+ },
857
+ {
858
+ "epoch": 6.64,
859
+ "learning_rate": 3.672773156369754e-05,
860
+ "loss": 1.5185,
861
+ "step": 71000
862
+ },
863
+ {
864
+ "epoch": 6.68,
865
+ "learning_rate": 3.6634264884568654e-05,
866
+ "loss": 1.5316,
867
+ "step": 71500
868
+ },
869
+ {
870
+ "epoch": 6.73,
871
+ "learning_rate": 3.6540798205439766e-05,
872
+ "loss": 1.5165,
873
+ "step": 72000
874
+ },
875
+ {
876
+ "epoch": 6.78,
877
+ "learning_rate": 3.644733152631087e-05,
878
+ "loss": 1.5067,
879
+ "step": 72500
880
+ },
881
+ {
882
+ "epoch": 6.82,
883
+ "learning_rate": 3.635386484718198e-05,
884
+ "loss": 1.5108,
885
+ "step": 73000
886
+ },
887
+ {
888
+ "epoch": 6.87,
889
+ "learning_rate": 3.626039816805309e-05,
890
+ "loss": 1.4999,
891
+ "step": 73500
892
+ },
893
+ {
894
+ "epoch": 6.92,
895
+ "learning_rate": 3.61669314889242e-05,
896
+ "loss": 1.4997,
897
+ "step": 74000
898
+ },
899
+ {
900
+ "epoch": 6.96,
901
+ "learning_rate": 3.607346480979531e-05,
902
+ "loss": 1.513,
903
+ "step": 74500
904
+ },
905
+ {
906
+ "epoch": 7.01,
907
+ "learning_rate": 3.597999813066642e-05,
908
+ "loss": 1.5012,
909
+ "step": 75000
910
+ },
911
+ {
912
+ "epoch": 7.06,
913
+ "learning_rate": 3.588653145153753e-05,
914
+ "loss": 1.4837,
915
+ "step": 75500
916
+ },
917
+ {
918
+ "epoch": 7.1,
919
+ "learning_rate": 3.5793064772408636e-05,
920
+ "loss": 1.4824,
921
+ "step": 76000
922
+ },
923
+ {
924
+ "epoch": 7.15,
925
+ "learning_rate": 3.569959809327975e-05,
926
+ "loss": 1.4896,
927
+ "step": 76500
928
+ },
929
+ {
930
+ "epoch": 7.2,
931
+ "learning_rate": 3.560613141415085e-05,
932
+ "loss": 1.4865,
933
+ "step": 77000
934
+ },
935
+ {
936
+ "epoch": 7.24,
937
+ "learning_rate": 3.5512664735021965e-05,
938
+ "loss": 1.4937,
939
+ "step": 77500
940
+ },
941
+ {
942
+ "epoch": 7.29,
943
+ "learning_rate": 3.541919805589308e-05,
944
+ "loss": 1.4828,
945
+ "step": 78000
946
+ },
947
+ {
948
+ "epoch": 7.34,
949
+ "learning_rate": 3.532573137676419e-05,
950
+ "loss": 1.4767,
951
+ "step": 78500
952
+ },
953
+ {
954
+ "epoch": 7.38,
955
+ "learning_rate": 3.5232264697635295e-05,
956
+ "loss": 1.4878,
957
+ "step": 79000
958
+ },
959
+ {
960
+ "epoch": 7.43,
961
+ "learning_rate": 3.51387980185064e-05,
962
+ "loss": 1.4946,
963
+ "step": 79500
964
+ },
965
+ {
966
+ "epoch": 7.48,
967
+ "learning_rate": 3.504533133937751e-05,
968
+ "loss": 1.4756,
969
+ "step": 80000
970
+ },
971
+ {
972
+ "epoch": 7.52,
973
+ "learning_rate": 3.4951864660248625e-05,
974
+ "loss": 1.464,
975
+ "step": 80500
976
+ },
977
+ {
978
+ "epoch": 7.57,
979
+ "learning_rate": 3.485839798111974e-05,
980
+ "loss": 1.4754,
981
+ "step": 81000
982
+ },
983
+ {
984
+ "epoch": 7.62,
985
+ "learning_rate": 3.476493130199084e-05,
986
+ "loss": 1.472,
987
+ "step": 81500
988
+ },
989
+ {
990
+ "epoch": 7.66,
991
+ "learning_rate": 3.467146462286195e-05,
992
+ "loss": 1.4716,
993
+ "step": 82000
994
+ },
995
+ {
996
+ "epoch": 7.71,
997
+ "learning_rate": 3.457799794373306e-05,
998
+ "loss": 1.4888,
999
+ "step": 82500
1000
+ },
1001
+ {
1002
+ "epoch": 7.76,
1003
+ "learning_rate": 3.448453126460417e-05,
1004
+ "loss": 1.4678,
1005
+ "step": 83000
1006
+ },
1007
+ {
1008
+ "epoch": 7.8,
1009
+ "learning_rate": 3.439106458547528e-05,
1010
+ "loss": 1.4694,
1011
+ "step": 83500
1012
+ },
1013
+ {
1014
+ "epoch": 7.85,
1015
+ "learning_rate": 3.429759790634639e-05,
1016
+ "loss": 1.4729,
1017
+ "step": 84000
1018
+ },
1019
+ {
1020
+ "epoch": 7.9,
1021
+ "learning_rate": 3.42041312272175e-05,
1022
+ "loss": 1.4558,
1023
+ "step": 84500
1024
+ },
1025
+ {
1026
+ "epoch": 7.94,
1027
+ "learning_rate": 3.4110664548088606e-05,
1028
+ "loss": 1.4597,
1029
+ "step": 85000
1030
+ },
1031
+ {
1032
+ "epoch": 7.99,
1033
+ "learning_rate": 3.401719786895972e-05,
1034
+ "loss": 1.449,
1035
+ "step": 85500
1036
+ },
1037
+ {
1038
+ "epoch": 8.04,
1039
+ "learning_rate": 3.3923731189830824e-05,
1040
+ "loss": 1.4543,
1041
+ "step": 86000
1042
+ },
1043
+ {
1044
+ "epoch": 8.08,
1045
+ "learning_rate": 3.3830264510701936e-05,
1046
+ "loss": 1.4437,
1047
+ "step": 86500
1048
+ },
1049
+ {
1050
+ "epoch": 8.13,
1051
+ "learning_rate": 3.373679783157305e-05,
1052
+ "loss": 1.4423,
1053
+ "step": 87000
1054
+ },
1055
+ {
1056
+ "epoch": 8.18,
1057
+ "learning_rate": 3.364333115244416e-05,
1058
+ "loss": 1.4417,
1059
+ "step": 87500
1060
+ },
1061
+ {
1062
+ "epoch": 8.23,
1063
+ "learning_rate": 3.3549864473315265e-05,
1064
+ "loss": 1.4505,
1065
+ "step": 88000
1066
+ },
1067
+ {
1068
+ "epoch": 8.27,
1069
+ "learning_rate": 3.345639779418637e-05,
1070
+ "loss": 1.4426,
1071
+ "step": 88500
1072
+ },
1073
+ {
1074
+ "epoch": 8.32,
1075
+ "learning_rate": 3.336293111505748e-05,
1076
+ "loss": 1.4359,
1077
+ "step": 89000
1078
+ },
1079
+ {
1080
+ "epoch": 8.37,
1081
+ "learning_rate": 3.3269464435928595e-05,
1082
+ "loss": 1.4463,
1083
+ "step": 89500
1084
+ },
1085
+ {
1086
+ "epoch": 8.41,
1087
+ "learning_rate": 3.317599775679971e-05,
1088
+ "loss": 1.4395,
1089
+ "step": 90000
1090
+ },
1091
+ {
1092
+ "epoch": 8.46,
1093
+ "learning_rate": 3.308253107767081e-05,
1094
+ "loss": 1.4388,
1095
+ "step": 90500
1096
+ },
1097
+ {
1098
+ "epoch": 8.51,
1099
+ "learning_rate": 3.298906439854192e-05,
1100
+ "loss": 1.4395,
1101
+ "step": 91000
1102
+ },
1103
+ {
1104
+ "epoch": 8.55,
1105
+ "learning_rate": 3.289559771941303e-05,
1106
+ "loss": 1.4335,
1107
+ "step": 91500
1108
+ },
1109
+ {
1110
+ "epoch": 8.6,
1111
+ "learning_rate": 3.280213104028414e-05,
1112
+ "loss": 1.4334,
1113
+ "step": 92000
1114
+ },
1115
+ {
1116
+ "epoch": 8.65,
1117
+ "learning_rate": 3.270866436115525e-05,
1118
+ "loss": 1.4457,
1119
+ "step": 92500
1120
+ },
1121
+ {
1122
+ "epoch": 8.69,
1123
+ "learning_rate": 3.261519768202636e-05,
1124
+ "loss": 1.4371,
1125
+ "step": 93000
1126
+ },
1127
+ {
1128
+ "epoch": 8.74,
1129
+ "learning_rate": 3.252173100289747e-05,
1130
+ "loss": 1.4299,
1131
+ "step": 93500
1132
+ },
1133
+ {
1134
+ "epoch": 8.79,
1135
+ "learning_rate": 3.242826432376858e-05,
1136
+ "loss": 1.4414,
1137
+ "step": 94000
1138
+ },
1139
+ {
1140
+ "epoch": 8.83,
1141
+ "learning_rate": 3.233479764463969e-05,
1142
+ "loss": 1.4191,
1143
+ "step": 94500
1144
+ },
1145
+ {
1146
+ "epoch": 8.88,
1147
+ "learning_rate": 3.2241330965510794e-05,
1148
+ "loss": 1.4261,
1149
+ "step": 95000
1150
+ },
1151
+ {
1152
+ "epoch": 8.93,
1153
+ "learning_rate": 3.2147864286381906e-05,
1154
+ "loss": 1.4292,
1155
+ "step": 95500
1156
+ },
1157
+ {
1158
+ "epoch": 8.97,
1159
+ "learning_rate": 3.205439760725302e-05,
1160
+ "loss": 1.4245,
1161
+ "step": 96000
1162
+ },
1163
+ {
1164
+ "epoch": 9.02,
1165
+ "learning_rate": 3.196093092812413e-05,
1166
+ "loss": 1.4169,
1167
+ "step": 96500
1168
+ },
1169
+ {
1170
+ "epoch": 9.07,
1171
+ "learning_rate": 3.186746424899523e-05,
1172
+ "loss": 1.4099,
1173
+ "step": 97000
1174
+ },
1175
+ {
1176
+ "epoch": 9.11,
1177
+ "learning_rate": 3.177399756986634e-05,
1178
+ "loss": 1.4087,
1179
+ "step": 97500
1180
+ },
1181
+ {
1182
+ "epoch": 9.16,
1183
+ "learning_rate": 3.168053089073745e-05,
1184
+ "loss": 1.4289,
1185
+ "step": 98000
1186
+ },
1187
+ {
1188
+ "epoch": 9.21,
1189
+ "learning_rate": 3.1587064211608565e-05,
1190
+ "loss": 1.4251,
1191
+ "step": 98500
1192
+ },
1193
+ {
1194
+ "epoch": 9.25,
1195
+ "learning_rate": 3.149359753247967e-05,
1196
+ "loss": 1.4228,
1197
+ "step": 99000
1198
+ },
1199
+ {
1200
+ "epoch": 9.3,
1201
+ "learning_rate": 3.140013085335078e-05,
1202
+ "loss": 1.4062,
1203
+ "step": 99500
1204
+ },
1205
+ {
1206
+ "epoch": 9.35,
1207
+ "learning_rate": 3.130666417422189e-05,
1208
+ "loss": 1.4032,
1209
+ "step": 100000
1210
+ },
1211
+ {
1212
+ "epoch": 9.39,
1213
+ "learning_rate": 3.1213197495093e-05,
1214
+ "loss": 1.4143,
1215
+ "step": 100500
1216
+ },
1217
+ {
1218
+ "epoch": 9.44,
1219
+ "learning_rate": 3.111973081596411e-05,
1220
+ "loss": 1.4038,
1221
+ "step": 101000
1222
+ },
1223
+ {
1224
+ "epoch": 9.49,
1225
+ "learning_rate": 3.102626413683522e-05,
1226
+ "loss": 1.3984,
1227
+ "step": 101500
1228
+ },
1229
+ {
1230
+ "epoch": 9.53,
1231
+ "learning_rate": 3.093279745770633e-05,
1232
+ "loss": 1.4098,
1233
+ "step": 102000
1234
+ },
1235
+ {
1236
+ "epoch": 9.58,
1237
+ "learning_rate": 3.083933077857744e-05,
1238
+ "loss": 1.4021,
1239
+ "step": 102500
1240
+ },
1241
+ {
1242
+ "epoch": 9.63,
1243
+ "learning_rate": 3.074586409944855e-05,
1244
+ "loss": 1.4041,
1245
+ "step": 103000
1246
+ },
1247
+ {
1248
+ "epoch": 9.67,
1249
+ "learning_rate": 3.065239742031966e-05,
1250
+ "loss": 1.3972,
1251
+ "step": 103500
1252
+ },
1253
+ {
1254
+ "epoch": 9.72,
1255
+ "learning_rate": 3.0558930741190765e-05,
1256
+ "loss": 1.3955,
1257
+ "step": 104000
1258
+ },
1259
+ {
1260
+ "epoch": 9.77,
1261
+ "learning_rate": 3.0465464062061877e-05,
1262
+ "loss": 1.4066,
1263
+ "step": 104500
1264
+ },
1265
+ {
1266
+ "epoch": 9.81,
1267
+ "learning_rate": 3.037199738293299e-05,
1268
+ "loss": 1.4019,
1269
+ "step": 105000
1270
+ },
1271
+ {
1272
+ "epoch": 9.86,
1273
+ "learning_rate": 3.0278530703804098e-05,
1274
+ "loss": 1.3893,
1275
+ "step": 105500
1276
+ },
1277
+ {
1278
+ "epoch": 9.91,
1279
+ "learning_rate": 3.0185064024675203e-05,
1280
+ "loss": 1.3995,
1281
+ "step": 106000
1282
+ },
1283
+ {
1284
+ "epoch": 9.95,
1285
+ "learning_rate": 3.0091597345546312e-05,
1286
+ "loss": 1.3958,
1287
+ "step": 106500
1288
+ },
1289
+ {
1290
+ "epoch": 10.0,
1291
+ "learning_rate": 2.9998130666417424e-05,
1292
+ "loss": 1.3839,
1293
+ "step": 107000
1294
+ },
1295
+ {
1296
+ "epoch": 10.05,
1297
+ "learning_rate": 2.9904663987288533e-05,
1298
+ "loss": 1.3878,
1299
+ "step": 107500
1300
+ },
1301
+ {
1302
+ "epoch": 10.09,
1303
+ "learning_rate": 2.9811197308159645e-05,
1304
+ "loss": 1.3848,
1305
+ "step": 108000
1306
+ },
1307
+ {
1308
+ "epoch": 10.14,
1309
+ "learning_rate": 2.9717730629030753e-05,
1310
+ "loss": 1.3804,
1311
+ "step": 108500
1312
+ },
1313
+ {
1314
+ "epoch": 10.19,
1315
+ "learning_rate": 2.962426394990186e-05,
1316
+ "loss": 1.3841,
1317
+ "step": 109000
1318
+ },
1319
+ {
1320
+ "epoch": 10.23,
1321
+ "learning_rate": 2.953079727077297e-05,
1322
+ "loss": 1.3878,
1323
+ "step": 109500
1324
+ },
1325
+ {
1326
+ "epoch": 10.28,
1327
+ "learning_rate": 2.943733059164408e-05,
1328
+ "loss": 1.3662,
1329
+ "step": 110000
1330
+ },
1331
+ {
1332
+ "epoch": 10.33,
1333
+ "learning_rate": 2.934386391251519e-05,
1334
+ "loss": 1.3775,
1335
+ "step": 110500
1336
+ },
1337
+ {
1338
+ "epoch": 10.37,
1339
+ "learning_rate": 2.92503972333863e-05,
1340
+ "loss": 1.3757,
1341
+ "step": 111000
1342
+ },
1343
+ {
1344
+ "epoch": 10.42,
1345
+ "learning_rate": 2.9156930554257412e-05,
1346
+ "loss": 1.3816,
1347
+ "step": 111500
1348
+ },
1349
+ {
1350
+ "epoch": 10.47,
1351
+ "learning_rate": 2.9063463875128514e-05,
1352
+ "loss": 1.3769,
1353
+ "step": 112000
1354
+ },
1355
+ {
1356
+ "epoch": 10.52,
1357
+ "learning_rate": 2.8969997195999626e-05,
1358
+ "loss": 1.3824,
1359
+ "step": 112500
1360
+ },
1361
+ {
1362
+ "epoch": 10.56,
1363
+ "learning_rate": 2.8876530516870735e-05,
1364
+ "loss": 1.3753,
1365
+ "step": 113000
1366
+ },
1367
+ {
1368
+ "epoch": 10.61,
1369
+ "learning_rate": 2.8783063837741847e-05,
1370
+ "loss": 1.3728,
1371
+ "step": 113500
1372
+ },
1373
+ {
1374
+ "epoch": 10.66,
1375
+ "learning_rate": 2.8689597158612956e-05,
1376
+ "loss": 1.3699,
1377
+ "step": 114000
1378
+ },
1379
+ {
1380
+ "epoch": 10.7,
1381
+ "learning_rate": 2.8596130479484068e-05,
1382
+ "loss": 1.3758,
1383
+ "step": 114500
1384
+ },
1385
+ {
1386
+ "epoch": 10.75,
1387
+ "learning_rate": 2.8502663800355173e-05,
1388
+ "loss": 1.3763,
1389
+ "step": 115000
1390
+ },
1391
+ {
1392
+ "epoch": 10.8,
1393
+ "learning_rate": 2.8409197121226282e-05,
1394
+ "loss": 1.3677,
1395
+ "step": 115500
1396
+ },
1397
+ {
1398
+ "epoch": 10.84,
1399
+ "learning_rate": 2.8315730442097394e-05,
1400
+ "loss": 1.3578,
1401
+ "step": 116000
1402
+ },
1403
+ {
1404
+ "epoch": 10.89,
1405
+ "learning_rate": 2.8222263762968503e-05,
1406
+ "loss": 1.3693,
1407
+ "step": 116500
1408
+ },
1409
+ {
1410
+ "epoch": 10.94,
1411
+ "learning_rate": 2.8128797083839615e-05,
1412
+ "loss": 1.3726,
1413
+ "step": 117000
1414
+ },
1415
+ {
1416
+ "epoch": 10.98,
1417
+ "learning_rate": 2.8035330404710724e-05,
1418
+ "loss": 1.3772,
1419
+ "step": 117500
1420
+ },
1421
+ {
1422
+ "epoch": 11.03,
1423
+ "learning_rate": 2.794186372558183e-05,
1424
+ "loss": 1.3472,
1425
+ "step": 118000
1426
+ },
1427
+ {
1428
+ "epoch": 11.08,
1429
+ "learning_rate": 2.784839704645294e-05,
1430
+ "loss": 1.3638,
1431
+ "step": 118500
1432
+ },
1433
+ {
1434
+ "epoch": 11.12,
1435
+ "learning_rate": 2.775493036732405e-05,
1436
+ "loss": 1.3556,
1437
+ "step": 119000
1438
+ },
1439
+ {
1440
+ "epoch": 11.17,
1441
+ "learning_rate": 2.7661463688195162e-05,
1442
+ "loss": 1.3598,
1443
+ "step": 119500
1444
+ },
1445
+ {
1446
+ "epoch": 11.22,
1447
+ "learning_rate": 2.756799700906627e-05,
1448
+ "loss": 1.344,
1449
+ "step": 120000
1450
+ },
1451
+ {
1452
+ "epoch": 11.26,
1453
+ "learning_rate": 2.7474530329937383e-05,
1454
+ "loss": 1.3532,
1455
+ "step": 120500
1456
+ },
1457
+ {
1458
+ "epoch": 11.31,
1459
+ "learning_rate": 2.7381063650808485e-05,
1460
+ "loss": 1.351,
1461
+ "step": 121000
1462
+ },
1463
+ {
1464
+ "epoch": 11.36,
1465
+ "learning_rate": 2.7287596971679597e-05,
1466
+ "loss": 1.3555,
1467
+ "step": 121500
1468
+ },
1469
+ {
1470
+ "epoch": 11.4,
1471
+ "learning_rate": 2.7194130292550706e-05,
1472
+ "loss": 1.361,
1473
+ "step": 122000
1474
+ },
1475
+ {
1476
+ "epoch": 11.45,
1477
+ "learning_rate": 2.7100663613421818e-05,
1478
+ "loss": 1.3472,
1479
+ "step": 122500
1480
+ },
1481
+ {
1482
+ "epoch": 11.5,
1483
+ "learning_rate": 2.7007196934292926e-05,
1484
+ "loss": 1.3462,
1485
+ "step": 123000
1486
+ },
1487
+ {
1488
+ "epoch": 11.54,
1489
+ "learning_rate": 2.691373025516404e-05,
1490
+ "loss": 1.3539,
1491
+ "step": 123500
1492
+ },
1493
+ {
1494
+ "epoch": 11.59,
1495
+ "learning_rate": 2.6820263576035144e-05,
1496
+ "loss": 1.3493,
1497
+ "step": 124000
1498
+ },
1499
+ {
1500
+ "epoch": 11.64,
1501
+ "learning_rate": 2.6726796896906253e-05,
1502
+ "loss": 1.3504,
1503
+ "step": 124500
1504
+ },
1505
+ {
1506
+ "epoch": 11.68,
1507
+ "learning_rate": 2.6633330217777365e-05,
1508
+ "loss": 1.3548,
1509
+ "step": 125000
1510
+ },
1511
+ {
1512
+ "epoch": 11.73,
1513
+ "learning_rate": 2.6539863538648473e-05,
1514
+ "loss": 1.3373,
1515
+ "step": 125500
1516
+ },
1517
+ {
1518
+ "epoch": 11.78,
1519
+ "learning_rate": 2.6446396859519586e-05,
1520
+ "loss": 1.3506,
1521
+ "step": 126000
1522
+ },
1523
+ {
1524
+ "epoch": 11.82,
1525
+ "learning_rate": 2.6352930180390694e-05,
1526
+ "loss": 1.3431,
1527
+ "step": 126500
1528
+ },
1529
+ {
1530
+ "epoch": 11.87,
1531
+ "learning_rate": 2.62594635012618e-05,
1532
+ "loss": 1.3458,
1533
+ "step": 127000
1534
+ },
1535
+ {
1536
+ "epoch": 11.92,
1537
+ "learning_rate": 2.616599682213291e-05,
1538
+ "loss": 1.345,
1539
+ "step": 127500
1540
+ },
1541
+ {
1542
+ "epoch": 11.96,
1543
+ "learning_rate": 2.607253014300402e-05,
1544
+ "loss": 1.3478,
1545
+ "step": 128000
1546
+ },
1547
+ {
1548
+ "epoch": 12.01,
1549
+ "learning_rate": 2.597906346387513e-05,
1550
+ "loss": 1.3453,
1551
+ "step": 128500
1552
+ },
1553
+ {
1554
+ "epoch": 12.06,
1555
+ "learning_rate": 2.588559678474624e-05,
1556
+ "loss": 1.3339,
1557
+ "step": 129000
1558
+ },
1559
+ {
1560
+ "epoch": 12.1,
1561
+ "learning_rate": 2.5792130105617347e-05,
1562
+ "loss": 1.3325,
1563
+ "step": 129500
1564
+ },
1565
+ {
1566
+ "epoch": 12.15,
1567
+ "learning_rate": 2.5698663426488455e-05,
1568
+ "loss": 1.339,
1569
+ "step": 130000
1570
+ },
1571
+ {
1572
+ "epoch": 12.2,
1573
+ "learning_rate": 2.5605196747359567e-05,
1574
+ "loss": 1.3329,
1575
+ "step": 130500
1576
+ },
1577
+ {
1578
+ "epoch": 12.24,
1579
+ "learning_rate": 2.5511730068230676e-05,
1580
+ "loss": 1.3341,
1581
+ "step": 131000
1582
+ },
1583
+ {
1584
+ "epoch": 12.29,
1585
+ "learning_rate": 2.5418263389101788e-05,
1586
+ "loss": 1.3396,
1587
+ "step": 131500
1588
+ },
1589
+ {
1590
+ "epoch": 12.34,
1591
+ "learning_rate": 2.5324796709972897e-05,
1592
+ "loss": 1.3341,
1593
+ "step": 132000
1594
+ },
1595
+ {
1596
+ "epoch": 12.38,
1597
+ "learning_rate": 2.5231330030844002e-05,
1598
+ "loss": 1.3358,
1599
+ "step": 132500
1600
+ },
1601
+ {
1602
+ "epoch": 12.43,
1603
+ "learning_rate": 2.5137863351715114e-05,
1604
+ "loss": 1.3294,
1605
+ "step": 133000
1606
+ },
1607
+ {
1608
+ "epoch": 12.48,
1609
+ "learning_rate": 2.5044396672586223e-05,
1610
+ "loss": 1.3339,
1611
+ "step": 133500
1612
+ },
1613
+ {
1614
+ "epoch": 12.52,
1615
+ "learning_rate": 2.4950929993457335e-05,
1616
+ "loss": 1.3338,
1617
+ "step": 134000
1618
+ },
1619
+ {
1620
+ "epoch": 12.57,
1621
+ "learning_rate": 2.4857463314328444e-05,
1622
+ "loss": 1.324,
1623
+ "step": 134500
1624
+ },
1625
+ {
1626
+ "epoch": 12.62,
1627
+ "learning_rate": 2.4763996635199553e-05,
1628
+ "loss": 1.3188,
1629
+ "step": 135000
1630
+ },
1631
+ {
1632
+ "epoch": 12.66,
1633
+ "learning_rate": 2.467052995607066e-05,
1634
+ "loss": 1.3244,
1635
+ "step": 135500
1636
+ },
1637
+ {
1638
+ "epoch": 12.71,
1639
+ "learning_rate": 2.4577063276941773e-05,
1640
+ "loss": 1.3296,
1641
+ "step": 136000
1642
+ },
1643
+ {
1644
+ "epoch": 12.76,
1645
+ "learning_rate": 2.448359659781288e-05,
1646
+ "loss": 1.3148,
1647
+ "step": 136500
1648
+ },
1649
+ {
1650
+ "epoch": 12.8,
1651
+ "learning_rate": 2.439012991868399e-05,
1652
+ "loss": 1.3261,
1653
+ "step": 137000
1654
+ },
1655
+ {
1656
+ "epoch": 12.85,
1657
+ "learning_rate": 2.42966632395551e-05,
1658
+ "loss": 1.3166,
1659
+ "step": 137500
1660
+ },
1661
+ {
1662
+ "epoch": 12.9,
1663
+ "learning_rate": 2.420319656042621e-05,
1664
+ "loss": 1.3137,
1665
+ "step": 138000
1666
+ },
1667
+ {
1668
+ "epoch": 12.95,
1669
+ "learning_rate": 2.410972988129732e-05,
1670
+ "loss": 1.3217,
1671
+ "step": 138500
1672
+ },
1673
+ {
1674
+ "epoch": 12.99,
1675
+ "learning_rate": 2.401626320216843e-05,
1676
+ "loss": 1.3341,
1677
+ "step": 139000
1678
+ },
1679
+ {
1680
+ "epoch": 13.04,
1681
+ "learning_rate": 2.3922796523039538e-05,
1682
+ "loss": 1.3248,
1683
+ "step": 139500
1684
+ },
1685
+ {
1686
+ "epoch": 13.09,
1687
+ "learning_rate": 2.3829329843910647e-05,
1688
+ "loss": 1.3087,
1689
+ "step": 140000
1690
+ },
1691
+ {
1692
+ "epoch": 13.13,
1693
+ "learning_rate": 2.373586316478176e-05,
1694
+ "loss": 1.3049,
1695
+ "step": 140500
1696
+ },
1697
+ {
1698
+ "epoch": 13.18,
1699
+ "learning_rate": 2.3642396485652864e-05,
1700
+ "loss": 1.3074,
1701
+ "step": 141000
1702
+ },
1703
+ {
1704
+ "epoch": 13.23,
1705
+ "learning_rate": 2.3548929806523976e-05,
1706
+ "loss": 1.3133,
1707
+ "step": 141500
1708
+ },
1709
+ {
1710
+ "epoch": 13.27,
1711
+ "learning_rate": 2.3455463127395085e-05,
1712
+ "loss": 1.3221,
1713
+ "step": 142000
1714
+ },
1715
+ {
1716
+ "epoch": 13.32,
1717
+ "learning_rate": 2.3361996448266194e-05,
1718
+ "loss": 1.3113,
1719
+ "step": 142500
1720
+ },
1721
+ {
1722
+ "epoch": 13.37,
1723
+ "learning_rate": 2.3268529769137302e-05,
1724
+ "loss": 1.3138,
1725
+ "step": 143000
1726
+ },
1727
+ {
1728
+ "epoch": 13.41,
1729
+ "learning_rate": 2.3175063090008414e-05,
1730
+ "loss": 1.3091,
1731
+ "step": 143500
1732
+ },
1733
+ {
1734
+ "epoch": 13.46,
1735
+ "learning_rate": 2.3081596410879523e-05,
1736
+ "loss": 1.3132,
1737
+ "step": 144000
1738
+ },
1739
+ {
1740
+ "epoch": 13.51,
1741
+ "learning_rate": 2.2988129731750632e-05,
1742
+ "loss": 1.3095,
1743
+ "step": 144500
1744
+ },
1745
+ {
1746
+ "epoch": 13.55,
1747
+ "learning_rate": 2.2894663052621744e-05,
1748
+ "loss": 1.3046,
1749
+ "step": 145000
1750
+ },
1751
+ {
1752
+ "epoch": 13.6,
1753
+ "learning_rate": 2.280119637349285e-05,
1754
+ "loss": 1.3136,
1755
+ "step": 145500
1756
+ },
1757
+ {
1758
+ "epoch": 13.65,
1759
+ "learning_rate": 2.270772969436396e-05,
1760
+ "loss": 1.3067,
1761
+ "step": 146000
1762
+ },
1763
+ {
1764
+ "epoch": 13.69,
1765
+ "learning_rate": 2.261426301523507e-05,
1766
+ "loss": 1.3025,
1767
+ "step": 146500
1768
+ },
1769
+ {
1770
+ "epoch": 13.74,
1771
+ "learning_rate": 2.252079633610618e-05,
1772
+ "loss": 1.3085,
1773
+ "step": 147000
1774
+ },
1775
+ {
1776
+ "epoch": 13.79,
1777
+ "learning_rate": 2.2427329656977288e-05,
1778
+ "loss": 1.2976,
1779
+ "step": 147500
1780
+ },
1781
+ {
1782
+ "epoch": 13.83,
1783
+ "learning_rate": 2.23338629778484e-05,
1784
+ "loss": 1.3007,
1785
+ "step": 148000
1786
+ },
1787
+ {
1788
+ "epoch": 13.88,
1789
+ "learning_rate": 2.224039629871951e-05,
1790
+ "loss": 1.3138,
1791
+ "step": 148500
1792
+ },
1793
+ {
1794
+ "epoch": 13.93,
1795
+ "learning_rate": 2.2146929619590617e-05,
1796
+ "loss": 1.3143,
1797
+ "step": 149000
1798
+ },
1799
+ {
1800
+ "epoch": 13.97,
1801
+ "learning_rate": 2.205346294046173e-05,
1802
+ "loss": 1.3029,
1803
+ "step": 149500
1804
+ },
1805
+ {
1806
+ "epoch": 14.02,
1807
+ "learning_rate": 2.1959996261332835e-05,
1808
+ "loss": 1.2919,
1809
+ "step": 150000
1810
+ },
1811
+ {
1812
+ "epoch": 14.07,
1813
+ "learning_rate": 2.1866529582203947e-05,
1814
+ "loss": 1.2982,
1815
+ "step": 150500
1816
+ },
1817
+ {
1818
+ "epoch": 14.11,
1819
+ "learning_rate": 2.1773062903075055e-05,
1820
+ "loss": 1.3012,
1821
+ "step": 151000
1822
+ },
1823
+ {
1824
+ "epoch": 14.16,
1825
+ "learning_rate": 2.1679596223946164e-05,
1826
+ "loss": 1.2841,
1827
+ "step": 151500
1828
+ },
1829
+ {
1830
+ "epoch": 14.21,
1831
+ "learning_rate": 2.1586129544817273e-05,
1832
+ "loss": 1.3044,
1833
+ "step": 152000
1834
+ },
1835
+ {
1836
+ "epoch": 14.25,
1837
+ "learning_rate": 2.1492662865688385e-05,
1838
+ "loss": 1.2973,
1839
+ "step": 152500
1840
+ },
1841
+ {
1842
+ "epoch": 14.3,
1843
+ "learning_rate": 2.1399196186559494e-05,
1844
+ "loss": 1.2884,
1845
+ "step": 153000
1846
+ },
1847
+ {
1848
+ "epoch": 14.35,
1849
+ "learning_rate": 2.1305729507430602e-05,
1850
+ "loss": 1.2883,
1851
+ "step": 153500
1852
+ },
1853
+ {
1854
+ "epoch": 14.39,
1855
+ "learning_rate": 2.1212262828301714e-05,
1856
+ "loss": 1.2993,
1857
+ "step": 154000
1858
+ },
1859
+ {
1860
+ "epoch": 14.44,
1861
+ "learning_rate": 2.111879614917282e-05,
1862
+ "loss": 1.2919,
1863
+ "step": 154500
1864
+ },
1865
+ {
1866
+ "epoch": 14.49,
1867
+ "learning_rate": 2.1025329470043932e-05,
1868
+ "loss": 1.3026,
1869
+ "step": 155000
1870
+ },
1871
+ {
1872
+ "epoch": 14.53,
1873
+ "learning_rate": 2.093186279091504e-05,
1874
+ "loss": 1.2882,
1875
+ "step": 155500
1876
+ },
1877
+ {
1878
+ "epoch": 14.58,
1879
+ "learning_rate": 2.083839611178615e-05,
1880
+ "loss": 1.289,
1881
+ "step": 156000
1882
+ },
1883
+ {
1884
+ "epoch": 14.63,
1885
+ "learning_rate": 2.0744929432657258e-05,
1886
+ "loss": 1.2917,
1887
+ "step": 156500
1888
+ },
1889
+ {
1890
+ "epoch": 14.67,
1891
+ "learning_rate": 2.065146275352837e-05,
1892
+ "loss": 1.2897,
1893
+ "step": 157000
1894
+ },
1895
+ {
1896
+ "epoch": 14.72,
1897
+ "learning_rate": 2.0557996074399475e-05,
1898
+ "loss": 1.2859,
1899
+ "step": 157500
1900
+ },
1901
+ {
1902
+ "epoch": 14.77,
1903
+ "learning_rate": 2.0464529395270588e-05,
1904
+ "loss": 1.2924,
1905
+ "step": 158000
1906
+ },
1907
+ {
1908
+ "epoch": 14.81,
1909
+ "learning_rate": 2.0371062716141696e-05,
1910
+ "loss": 1.2873,
1911
+ "step": 158500
1912
+ },
1913
+ {
1914
+ "epoch": 14.86,
1915
+ "learning_rate": 2.0277596037012805e-05,
1916
+ "loss": 1.29,
1917
+ "step": 159000
1918
+ },
1919
+ {
1920
+ "epoch": 14.91,
1921
+ "learning_rate": 2.0184129357883917e-05,
1922
+ "loss": 1.2848,
1923
+ "step": 159500
1924
+ },
1925
+ {
1926
+ "epoch": 14.95,
1927
+ "learning_rate": 2.0090662678755026e-05,
1928
+ "loss": 1.2831,
1929
+ "step": 160000
1930
+ },
1931
+ {
1932
+ "epoch": 15.0,
1933
+ "learning_rate": 1.9997195999626135e-05,
1934
+ "loss": 1.2841,
1935
+ "step": 160500
1936
+ },
1937
+ {
1938
+ "epoch": 15.05,
1939
+ "learning_rate": 1.9903729320497243e-05,
1940
+ "loss": 1.2716,
1941
+ "step": 161000
1942
+ },
1943
+ {
1944
+ "epoch": 15.09,
1945
+ "learning_rate": 1.9810262641368352e-05,
1946
+ "loss": 1.2795,
1947
+ "step": 161500
1948
+ },
1949
+ {
1950
+ "epoch": 15.14,
1951
+ "learning_rate": 1.971679596223946e-05,
1952
+ "loss": 1.2836,
1953
+ "step": 162000
1954
+ },
1955
+ {
1956
+ "epoch": 15.19,
1957
+ "learning_rate": 1.9623329283110573e-05,
1958
+ "loss": 1.2854,
1959
+ "step": 162500
1960
+ },
1961
+ {
1962
+ "epoch": 15.24,
1963
+ "learning_rate": 1.952986260398168e-05,
1964
+ "loss": 1.2819,
1965
+ "step": 163000
1966
+ },
1967
+ {
1968
+ "epoch": 15.28,
1969
+ "learning_rate": 1.943639592485279e-05,
1970
+ "loss": 1.2762,
1971
+ "step": 163500
1972
+ },
1973
+ {
1974
+ "epoch": 15.33,
1975
+ "learning_rate": 1.9342929245723902e-05,
1976
+ "loss": 1.2638,
1977
+ "step": 164000
1978
+ },
1979
+ {
1980
+ "epoch": 15.38,
1981
+ "learning_rate": 1.9249462566595008e-05,
1982
+ "loss": 1.269,
1983
+ "step": 164500
1984
+ },
1985
+ {
1986
+ "epoch": 15.42,
1987
+ "learning_rate": 1.915599588746612e-05,
1988
+ "loss": 1.2691,
1989
+ "step": 165000
1990
+ },
1991
+ {
1992
+ "epoch": 15.47,
1993
+ "learning_rate": 1.906252920833723e-05,
1994
+ "loss": 1.2802,
1995
+ "step": 165500
1996
+ },
1997
+ {
1998
+ "epoch": 15.52,
1999
+ "learning_rate": 1.8969062529208337e-05,
2000
+ "loss": 1.275,
2001
+ "step": 166000
2002
+ },
2003
+ {
2004
+ "epoch": 15.56,
2005
+ "learning_rate": 1.8875595850079446e-05,
2006
+ "loss": 1.278,
2007
+ "step": 166500
2008
+ },
2009
+ {
2010
+ "epoch": 15.61,
2011
+ "learning_rate": 1.8782129170950558e-05,
2012
+ "loss": 1.2768,
2013
+ "step": 167000
2014
+ },
2015
+ {
2016
+ "epoch": 15.66,
2017
+ "learning_rate": 1.8688662491821667e-05,
2018
+ "loss": 1.2761,
2019
+ "step": 167500
2020
+ },
2021
+ {
2022
+ "epoch": 15.7,
2023
+ "learning_rate": 1.8595195812692775e-05,
2024
+ "loss": 1.271,
2025
+ "step": 168000
2026
+ },
2027
+ {
2028
+ "epoch": 15.75,
2029
+ "learning_rate": 1.8501729133563888e-05,
2030
+ "loss": 1.2687,
2031
+ "step": 168500
2032
+ },
2033
+ {
2034
+ "epoch": 15.8,
2035
+ "learning_rate": 1.8408262454434993e-05,
2036
+ "loss": 1.2644,
2037
+ "step": 169000
2038
+ },
2039
+ {
2040
+ "epoch": 15.84,
2041
+ "learning_rate": 1.8314795775306105e-05,
2042
+ "loss": 1.2732,
2043
+ "step": 169500
2044
+ },
2045
+ {
2046
+ "epoch": 15.89,
2047
+ "learning_rate": 1.8221329096177214e-05,
2048
+ "loss": 1.2742,
2049
+ "step": 170000
2050
+ },
2051
+ {
2052
+ "epoch": 15.94,
2053
+ "learning_rate": 1.8127862417048322e-05,
2054
+ "loss": 1.266,
2055
+ "step": 170500
2056
+ },
2057
+ {
2058
+ "epoch": 15.98,
2059
+ "learning_rate": 1.803439573791943e-05,
2060
+ "loss": 1.27,
2061
+ "step": 171000
2062
+ },
2063
+ {
2064
+ "epoch": 16.03,
2065
+ "learning_rate": 1.7940929058790543e-05,
2066
+ "loss": 1.2682,
2067
+ "step": 171500
2068
+ },
2069
+ {
2070
+ "epoch": 16.08,
2071
+ "learning_rate": 1.784746237966165e-05,
2072
+ "loss": 1.2584,
2073
+ "step": 172000
2074
+ },
2075
+ {
2076
+ "epoch": 16.12,
2077
+ "learning_rate": 1.775399570053276e-05,
2078
+ "loss": 1.2702,
2079
+ "step": 172500
2080
+ },
2081
+ {
2082
+ "epoch": 16.17,
2083
+ "learning_rate": 1.766052902140387e-05,
2084
+ "loss": 1.2602,
2085
+ "step": 173000
2086
+ },
2087
+ {
2088
+ "epoch": 16.22,
2089
+ "learning_rate": 1.7567062342274978e-05,
2090
+ "loss": 1.2595,
2091
+ "step": 173500
2092
+ },
2093
+ {
2094
+ "epoch": 16.26,
2095
+ "learning_rate": 1.747359566314609e-05,
2096
+ "loss": 1.261,
2097
+ "step": 174000
2098
+ },
2099
+ {
2100
+ "epoch": 16.31,
2101
+ "learning_rate": 1.73801289840172e-05,
2102
+ "loss": 1.2556,
2103
+ "step": 174500
2104
+ },
2105
+ {
2106
+ "epoch": 16.36,
2107
+ "learning_rate": 1.7286662304888308e-05,
2108
+ "loss": 1.2722,
2109
+ "step": 175000
2110
+ },
2111
+ {
2112
+ "epoch": 16.4,
2113
+ "learning_rate": 1.7193195625759416e-05,
2114
+ "loss": 1.2553,
2115
+ "step": 175500
2116
+ },
2117
+ {
2118
+ "epoch": 16.45,
2119
+ "learning_rate": 1.709972894663053e-05,
2120
+ "loss": 1.2577,
2121
+ "step": 176000
2122
+ },
2123
+ {
2124
+ "epoch": 16.5,
2125
+ "learning_rate": 1.7006262267501634e-05,
2126
+ "loss": 1.2607,
2127
+ "step": 176500
2128
+ },
2129
+ {
2130
+ "epoch": 16.54,
2131
+ "learning_rate": 1.6912795588372746e-05,
2132
+ "loss": 1.2646,
2133
+ "step": 177000
2134
+ },
2135
+ {
2136
+ "epoch": 16.59,
2137
+ "learning_rate": 1.6819328909243855e-05,
2138
+ "loss": 1.267,
2139
+ "step": 177500
2140
+ },
2141
+ {
2142
+ "epoch": 16.64,
2143
+ "learning_rate": 1.6725862230114963e-05,
2144
+ "loss": 1.2596,
2145
+ "step": 178000
2146
+ },
2147
+ {
2148
+ "epoch": 16.68,
2149
+ "learning_rate": 1.6632395550986075e-05,
2150
+ "loss": 1.2553,
2151
+ "step": 178500
2152
+ },
2153
+ {
2154
+ "epoch": 16.73,
2155
+ "learning_rate": 1.6538928871857184e-05,
2156
+ "loss": 1.2538,
2157
+ "step": 179000
2158
+ },
2159
+ {
2160
+ "epoch": 16.78,
2161
+ "learning_rate": 1.6445462192728293e-05,
2162
+ "loss": 1.2626,
2163
+ "step": 179500
2164
+ },
2165
+ {
2166
+ "epoch": 16.82,
2167
+ "learning_rate": 1.63519955135994e-05,
2168
+ "loss": 1.2551,
2169
+ "step": 180000
2170
+ },
2171
+ {
2172
+ "epoch": 16.87,
2173
+ "learning_rate": 1.6258528834470514e-05,
2174
+ "loss": 1.2569,
2175
+ "step": 180500
2176
+ },
2177
+ {
2178
+ "epoch": 16.92,
2179
+ "learning_rate": 1.616506215534162e-05,
2180
+ "loss": 1.2535,
2181
+ "step": 181000
2182
+ },
2183
+ {
2184
+ "epoch": 16.96,
2185
+ "learning_rate": 1.607159547621273e-05,
2186
+ "loss": 1.2591,
2187
+ "step": 181500
2188
+ },
2189
+ {
2190
+ "epoch": 17.01,
2191
+ "learning_rate": 1.597812879708384e-05,
2192
+ "loss": 1.2508,
2193
+ "step": 182000
2194
+ },
2195
+ {
2196
+ "epoch": 17.06,
2197
+ "learning_rate": 1.588466211795495e-05,
2198
+ "loss": 1.2517,
2199
+ "step": 182500
2200
+ },
2201
+ {
2202
+ "epoch": 17.1,
2203
+ "learning_rate": 1.579119543882606e-05,
2204
+ "loss": 1.2546,
2205
+ "step": 183000
2206
+ },
2207
+ {
2208
+ "epoch": 17.15,
2209
+ "learning_rate": 1.569772875969717e-05,
2210
+ "loss": 1.241,
2211
+ "step": 183500
2212
+ },
2213
+ {
2214
+ "epoch": 17.2,
2215
+ "learning_rate": 1.5604262080568278e-05,
2216
+ "loss": 1.2421,
2217
+ "step": 184000
2218
+ },
2219
+ {
2220
+ "epoch": 17.24,
2221
+ "learning_rate": 1.5510795401439387e-05,
2222
+ "loss": 1.243,
2223
+ "step": 184500
2224
+ },
2225
+ {
2226
+ "epoch": 17.29,
2227
+ "learning_rate": 1.54173287223105e-05,
2228
+ "loss": 1.2459,
2229
+ "step": 185000
2230
+ },
2231
+ {
2232
+ "epoch": 17.34,
2233
+ "learning_rate": 1.5323862043181604e-05,
2234
+ "loss": 1.2497,
2235
+ "step": 185500
2236
+ },
2237
+ {
2238
+ "epoch": 17.38,
2239
+ "learning_rate": 1.5230395364052716e-05,
2240
+ "loss": 1.2433,
2241
+ "step": 186000
2242
+ },
2243
+ {
2244
+ "epoch": 17.43,
2245
+ "learning_rate": 1.5136928684923827e-05,
2246
+ "loss": 1.2497,
2247
+ "step": 186500
2248
+ },
2249
+ {
2250
+ "epoch": 17.48,
2251
+ "learning_rate": 1.5043462005794934e-05,
2252
+ "loss": 1.2416,
2253
+ "step": 187000
2254
+ },
2255
+ {
2256
+ "epoch": 17.53,
2257
+ "learning_rate": 1.4949995326666044e-05,
2258
+ "loss": 1.2457,
2259
+ "step": 187500
2260
+ },
2261
+ {
2262
+ "epoch": 17.57,
2263
+ "learning_rate": 1.4856528647537155e-05,
2264
+ "loss": 1.2516,
2265
+ "step": 188000
2266
+ },
2267
+ {
2268
+ "epoch": 17.62,
2269
+ "learning_rate": 1.4763061968408262e-05,
2270
+ "loss": 1.2492,
2271
+ "step": 188500
2272
+ },
2273
+ {
2274
+ "epoch": 17.67,
2275
+ "learning_rate": 1.4669595289279372e-05,
2276
+ "loss": 1.2462,
2277
+ "step": 189000
2278
+ },
2279
+ {
2280
+ "epoch": 17.71,
2281
+ "learning_rate": 1.4576128610150482e-05,
2282
+ "loss": 1.2485,
2283
+ "step": 189500
2284
+ },
2285
+ {
2286
+ "epoch": 17.76,
2287
+ "learning_rate": 1.4482661931021591e-05,
2288
+ "loss": 1.2398,
2289
+ "step": 190000
2290
+ },
2291
+ {
2292
+ "epoch": 17.81,
2293
+ "learning_rate": 1.4389195251892702e-05,
2294
+ "loss": 1.2464,
2295
+ "step": 190500
2296
+ },
2297
+ {
2298
+ "epoch": 17.85,
2299
+ "learning_rate": 1.4295728572763812e-05,
2300
+ "loss": 1.2444,
2301
+ "step": 191000
2302
+ },
2303
+ {
2304
+ "epoch": 17.9,
2305
+ "learning_rate": 1.4202261893634919e-05,
2306
+ "loss": 1.2464,
2307
+ "step": 191500
2308
+ },
2309
+ {
2310
+ "epoch": 17.95,
2311
+ "learning_rate": 1.410879521450603e-05,
2312
+ "loss": 1.2507,
2313
+ "step": 192000
2314
+ },
2315
+ {
2316
+ "epoch": 17.99,
2317
+ "learning_rate": 1.401532853537714e-05,
2318
+ "loss": 1.2374,
2319
+ "step": 192500
2320
+ },
2321
+ {
2322
+ "epoch": 18.04,
2323
+ "learning_rate": 1.3921861856248247e-05,
2324
+ "loss": 1.2439,
2325
+ "step": 193000
2326
+ },
2327
+ {
2328
+ "epoch": 18.09,
2329
+ "learning_rate": 1.3828395177119357e-05,
2330
+ "loss": 1.2455,
2331
+ "step": 193500
2332
+ },
2333
+ {
2334
+ "epoch": 18.13,
2335
+ "learning_rate": 1.3734928497990468e-05,
2336
+ "loss": 1.2368,
2337
+ "step": 194000
2338
+ },
2339
+ {
2340
+ "epoch": 18.18,
2341
+ "learning_rate": 1.3641461818861575e-05,
2342
+ "loss": 1.2434,
2343
+ "step": 194500
2344
+ },
2345
+ {
2346
+ "epoch": 18.23,
2347
+ "learning_rate": 1.3547995139732685e-05,
2348
+ "loss": 1.2292,
2349
+ "step": 195000
2350
+ },
2351
+ {
2352
+ "epoch": 18.27,
2353
+ "learning_rate": 1.3454528460603796e-05,
2354
+ "loss": 1.2316,
2355
+ "step": 195500
2356
+ },
2357
+ {
2358
+ "epoch": 18.32,
2359
+ "learning_rate": 1.3361061781474904e-05,
2360
+ "loss": 1.2312,
2361
+ "step": 196000
2362
+ },
2363
+ {
2364
+ "epoch": 18.37,
2365
+ "learning_rate": 1.3267595102346015e-05,
2366
+ "loss": 1.229,
2367
+ "step": 196500
2368
+ },
2369
+ {
2370
+ "epoch": 18.41,
2371
+ "learning_rate": 1.3174128423217125e-05,
2372
+ "loss": 1.2375,
2373
+ "step": 197000
2374
+ },
2375
+ {
2376
+ "epoch": 18.46,
2377
+ "learning_rate": 1.3080661744088232e-05,
2378
+ "loss": 1.2346,
2379
+ "step": 197500
2380
+ },
2381
+ {
2382
+ "epoch": 18.51,
2383
+ "learning_rate": 1.2987195064959343e-05,
2384
+ "loss": 1.2334,
2385
+ "step": 198000
2386
+ },
2387
+ {
2388
+ "epoch": 18.55,
2389
+ "learning_rate": 1.2893728385830453e-05,
2390
+ "loss": 1.2317,
2391
+ "step": 198500
2392
+ },
2393
+ {
2394
+ "epoch": 18.6,
2395
+ "learning_rate": 1.280026170670156e-05,
2396
+ "loss": 1.2342,
2397
+ "step": 199000
2398
+ },
2399
+ {
2400
+ "epoch": 18.65,
2401
+ "learning_rate": 1.270679502757267e-05,
2402
+ "loss": 1.2327,
2403
+ "step": 199500
2404
+ },
2405
+ {
2406
+ "epoch": 18.69,
2407
+ "learning_rate": 1.261332834844378e-05,
2408
+ "loss": 1.2303,
2409
+ "step": 200000
2410
+ },
2411
+ {
2412
+ "epoch": 18.74,
2413
+ "learning_rate": 1.251986166931489e-05,
2414
+ "loss": 1.2406,
2415
+ "step": 200500
2416
+ },
2417
+ {
2418
+ "epoch": 18.79,
2419
+ "learning_rate": 1.2426394990186e-05,
2420
+ "loss": 1.2306,
2421
+ "step": 201000
2422
+ },
2423
+ {
2424
+ "epoch": 18.83,
2425
+ "learning_rate": 1.2332928311057109e-05,
2426
+ "loss": 1.2289,
2427
+ "step": 201500
2428
+ },
2429
+ {
2430
+ "epoch": 18.88,
2431
+ "learning_rate": 1.2239461631928219e-05,
2432
+ "loss": 1.231,
2433
+ "step": 202000
2434
+ },
2435
+ {
2436
+ "epoch": 18.93,
2437
+ "learning_rate": 1.2145994952799328e-05,
2438
+ "loss": 1.227,
2439
+ "step": 202500
2440
+ },
2441
+ {
2442
+ "epoch": 18.97,
2443
+ "learning_rate": 1.2052528273670437e-05,
2444
+ "loss": 1.2329,
2445
+ "step": 203000
2446
+ },
2447
+ {
2448
+ "epoch": 19.02,
2449
+ "learning_rate": 1.1959061594541547e-05,
2450
+ "loss": 1.2278,
2451
+ "step": 203500
2452
+ },
2453
+ {
2454
+ "epoch": 19.07,
2455
+ "learning_rate": 1.1865594915412656e-05,
2456
+ "loss": 1.2342,
2457
+ "step": 204000
2458
+ },
2459
+ {
2460
+ "epoch": 19.11,
2461
+ "learning_rate": 1.1772128236283764e-05,
2462
+ "loss": 1.2174,
2463
+ "step": 204500
2464
+ },
2465
+ {
2466
+ "epoch": 19.16,
2467
+ "learning_rate": 1.1678661557154875e-05,
2468
+ "loss": 1.2299,
2469
+ "step": 205000
2470
+ },
2471
+ {
2472
+ "epoch": 19.21,
2473
+ "learning_rate": 1.1585194878025985e-05,
2474
+ "loss": 1.2276,
2475
+ "step": 205500
2476
+ },
2477
+ {
2478
+ "epoch": 19.25,
2479
+ "learning_rate": 1.1491728198897094e-05,
2480
+ "loss": 1.2266,
2481
+ "step": 206000
2482
+ },
2483
+ {
2484
+ "epoch": 19.3,
2485
+ "learning_rate": 1.1398261519768204e-05,
2486
+ "loss": 1.2229,
2487
+ "step": 206500
2488
+ },
2489
+ {
2490
+ "epoch": 19.35,
2491
+ "learning_rate": 1.1304794840639313e-05,
2492
+ "loss": 1.2258,
2493
+ "step": 207000
2494
+ },
2495
+ {
2496
+ "epoch": 19.39,
2497
+ "learning_rate": 1.1211328161510422e-05,
2498
+ "loss": 1.2275,
2499
+ "step": 207500
2500
+ },
2501
+ {
2502
+ "epoch": 19.44,
2503
+ "learning_rate": 1.1117861482381532e-05,
2504
+ "loss": 1.2148,
2505
+ "step": 208000
2506
+ },
2507
+ {
2508
+ "epoch": 19.49,
2509
+ "learning_rate": 1.1024394803252641e-05,
2510
+ "loss": 1.2229,
2511
+ "step": 208500
2512
+ },
2513
+ {
2514
+ "epoch": 19.53,
2515
+ "learning_rate": 1.093092812412375e-05,
2516
+ "loss": 1.2218,
2517
+ "step": 209000
2518
+ },
2519
+ {
2520
+ "epoch": 19.58,
2521
+ "learning_rate": 1.083746144499486e-05,
2522
+ "loss": 1.2114,
2523
+ "step": 209500
2524
+ },
2525
+ {
2526
+ "epoch": 19.63,
2527
+ "learning_rate": 1.0743994765865969e-05,
2528
+ "loss": 1.2243,
2529
+ "step": 210000
2530
+ },
2531
+ {
2532
+ "epoch": 19.67,
2533
+ "learning_rate": 1.0650528086737079e-05,
2534
+ "loss": 1.2211,
2535
+ "step": 210500
2536
+ },
2537
+ {
2538
+ "epoch": 19.72,
2539
+ "learning_rate": 1.0557061407608188e-05,
2540
+ "loss": 1.2223,
2541
+ "step": 211000
2542
+ },
2543
+ {
2544
+ "epoch": 19.77,
2545
+ "learning_rate": 1.0463594728479298e-05,
2546
+ "loss": 1.2263,
2547
+ "step": 211500
2548
+ },
2549
+ {
2550
+ "epoch": 19.81,
2551
+ "learning_rate": 1.0370128049350407e-05,
2552
+ "loss": 1.2184,
2553
+ "step": 212000
2554
+ },
2555
+ {
2556
+ "epoch": 19.86,
2557
+ "learning_rate": 1.0276661370221516e-05,
2558
+ "loss": 1.2193,
2559
+ "step": 212500
2560
+ },
2561
+ {
2562
+ "epoch": 19.91,
2563
+ "learning_rate": 1.0183194691092626e-05,
2564
+ "loss": 1.2147,
2565
+ "step": 213000
2566
+ },
2567
+ {
2568
+ "epoch": 19.96,
2569
+ "learning_rate": 1.0089728011963735e-05,
2570
+ "loss": 1.2179,
2571
+ "step": 213500
2572
+ },
2573
+ {
2574
+ "epoch": 20.0,
2575
+ "learning_rate": 9.996261332834844e-06,
2576
+ "loss": 1.2093,
2577
+ "step": 214000
2578
+ },
2579
+ {
2580
+ "epoch": 20.05,
2581
+ "learning_rate": 9.902794653705954e-06,
2582
+ "loss": 1.2069,
2583
+ "step": 214500
2584
+ },
2585
+ {
2586
+ "epoch": 20.1,
2587
+ "learning_rate": 9.809327974577064e-06,
2588
+ "loss": 1.2147,
2589
+ "step": 215000
2590
+ },
2591
+ {
2592
+ "epoch": 20.14,
2593
+ "learning_rate": 9.715861295448173e-06,
2594
+ "loss": 1.2125,
2595
+ "step": 215500
2596
+ },
2597
+ {
2598
+ "epoch": 20.19,
2599
+ "learning_rate": 9.622394616319283e-06,
2600
+ "loss": 1.2221,
2601
+ "step": 216000
2602
+ },
2603
+ {
2604
+ "epoch": 20.24,
2605
+ "learning_rate": 9.528927937190392e-06,
2606
+ "loss": 1.2056,
2607
+ "step": 216500
2608
+ },
2609
+ {
2610
+ "epoch": 20.28,
2611
+ "learning_rate": 9.435461258061501e-06,
2612
+ "loss": 1.211,
2613
+ "step": 217000
2614
+ },
2615
+ {
2616
+ "epoch": 20.33,
2617
+ "learning_rate": 9.341994578932611e-06,
2618
+ "loss": 1.2146,
2619
+ "step": 217500
2620
+ },
2621
+ {
2622
+ "epoch": 20.38,
2623
+ "learning_rate": 9.24852789980372e-06,
2624
+ "loss": 1.2152,
2625
+ "step": 218000
2626
+ },
2627
+ {
2628
+ "epoch": 20.42,
2629
+ "learning_rate": 9.155061220674829e-06,
2630
+ "loss": 1.2146,
2631
+ "step": 218500
2632
+ },
2633
+ {
2634
+ "epoch": 20.47,
2635
+ "learning_rate": 9.06159454154594e-06,
2636
+ "loss": 1.2256,
2637
+ "step": 219000
2638
+ },
2639
+ {
2640
+ "epoch": 20.52,
2641
+ "learning_rate": 8.968127862417048e-06,
2642
+ "loss": 1.2058,
2643
+ "step": 219500
2644
+ },
2645
+ {
2646
+ "epoch": 20.56,
2647
+ "learning_rate": 8.874661183288158e-06,
2648
+ "loss": 1.2128,
2649
+ "step": 220000
2650
+ },
2651
+ {
2652
+ "epoch": 20.61,
2653
+ "learning_rate": 8.781194504159269e-06,
2654
+ "loss": 1.2137,
2655
+ "step": 220500
2656
+ },
2657
+ {
2658
+ "epoch": 20.66,
2659
+ "learning_rate": 8.687727825030377e-06,
2660
+ "loss": 1.2129,
2661
+ "step": 221000
2662
+ },
2663
+ {
2664
+ "epoch": 20.7,
2665
+ "learning_rate": 8.594261145901486e-06,
2666
+ "loss": 1.2181,
2667
+ "step": 221500
2668
+ },
2669
+ {
2670
+ "epoch": 20.75,
2671
+ "learning_rate": 8.500794466772597e-06,
2672
+ "loss": 1.201,
2673
+ "step": 222000
2674
+ },
2675
+ {
2676
+ "epoch": 20.8,
2677
+ "learning_rate": 8.407327787643705e-06,
2678
+ "loss": 1.2162,
2679
+ "step": 222500
2680
+ },
2681
+ {
2682
+ "epoch": 20.84,
2683
+ "learning_rate": 8.313861108514814e-06,
2684
+ "loss": 1.2077,
2685
+ "step": 223000
2686
+ },
2687
+ {
2688
+ "epoch": 20.89,
2689
+ "learning_rate": 8.220394429385924e-06,
2690
+ "loss": 1.2098,
2691
+ "step": 223500
2692
+ },
2693
+ {
2694
+ "epoch": 20.94,
2695
+ "learning_rate": 8.126927750257033e-06,
2696
+ "loss": 1.2092,
2697
+ "step": 224000
2698
+ },
2699
+ {
2700
+ "epoch": 20.98,
2701
+ "learning_rate": 8.033461071128144e-06,
2702
+ "loss": 1.2132,
2703
+ "step": 224500
2704
+ },
2705
+ {
2706
+ "epoch": 21.03,
2707
+ "learning_rate": 7.939994391999252e-06,
2708
+ "loss": 1.2166,
2709
+ "step": 225000
2710
+ },
2711
+ {
2712
+ "epoch": 21.08,
2713
+ "learning_rate": 7.846527712870363e-06,
2714
+ "loss": 1.2063,
2715
+ "step": 225500
2716
+ },
2717
+ {
2718
+ "epoch": 21.12,
2719
+ "learning_rate": 7.753061033741471e-06,
2720
+ "loss": 1.2029,
2721
+ "step": 226000
2722
+ },
2723
+ {
2724
+ "epoch": 21.17,
2725
+ "learning_rate": 7.659594354612582e-06,
2726
+ "loss": 1.201,
2727
+ "step": 226500
2728
+ },
2729
+ {
2730
+ "epoch": 21.22,
2731
+ "learning_rate": 7.5661276754836905e-06,
2732
+ "loss": 1.2006,
2733
+ "step": 227000
2734
+ },
2735
+ {
2736
+ "epoch": 21.26,
2737
+ "learning_rate": 7.472660996354799e-06,
2738
+ "loss": 1.2028,
2739
+ "step": 227500
2740
+ },
2741
+ {
2742
+ "epoch": 21.31,
2743
+ "learning_rate": 7.37919431722591e-06,
2744
+ "loss": 1.1959,
2745
+ "step": 228000
2746
+ },
2747
+ {
2748
+ "epoch": 21.36,
2749
+ "learning_rate": 7.285727638097019e-06,
2750
+ "loss": 1.2126,
2751
+ "step": 228500
2752
+ },
2753
+ {
2754
+ "epoch": 21.4,
2755
+ "learning_rate": 7.192260958968128e-06,
2756
+ "loss": 1.208,
2757
+ "step": 229000
2758
+ },
2759
+ {
2760
+ "epoch": 21.45,
2761
+ "learning_rate": 7.098794279839238e-06,
2762
+ "loss": 1.2026,
2763
+ "step": 229500
2764
+ },
2765
+ {
2766
+ "epoch": 21.5,
2767
+ "learning_rate": 7.005327600710347e-06,
2768
+ "loss": 1.2069,
2769
+ "step": 230000
2770
+ },
2771
+ {
2772
+ "epoch": 21.54,
2773
+ "learning_rate": 6.911860921581456e-06,
2774
+ "loss": 1.2042,
2775
+ "step": 230500
2776
+ },
2777
+ {
2778
+ "epoch": 21.59,
2779
+ "learning_rate": 6.818394242452566e-06,
2780
+ "loss": 1.2057,
2781
+ "step": 231000
2782
+ },
2783
+ {
2784
+ "epoch": 21.64,
2785
+ "learning_rate": 6.724927563323676e-06,
2786
+ "loss": 1.212,
2787
+ "step": 231500
2788
+ },
2789
+ {
2790
+ "epoch": 21.68,
2791
+ "learning_rate": 6.6314608841947845e-06,
2792
+ "loss": 1.2012,
2793
+ "step": 232000
2794
+ },
2795
+ {
2796
+ "epoch": 21.73,
2797
+ "learning_rate": 6.537994205065895e-06,
2798
+ "loss": 1.2066,
2799
+ "step": 232500
2800
+ },
2801
+ {
2802
+ "epoch": 21.78,
2803
+ "learning_rate": 6.444527525937004e-06,
2804
+ "loss": 1.2042,
2805
+ "step": 233000
2806
+ },
2807
+ {
2808
+ "epoch": 21.82,
2809
+ "learning_rate": 6.351060846808113e-06,
2810
+ "loss": 1.2101,
2811
+ "step": 233500
2812
+ },
2813
+ {
2814
+ "epoch": 21.87,
2815
+ "learning_rate": 6.257594167679224e-06,
2816
+ "loss": 1.1961,
2817
+ "step": 234000
2818
+ },
2819
+ {
2820
+ "epoch": 21.92,
2821
+ "learning_rate": 6.164127488550332e-06,
2822
+ "loss": 1.1994,
2823
+ "step": 234500
2824
+ },
2825
+ {
2826
+ "epoch": 21.96,
2827
+ "learning_rate": 6.070660809421441e-06,
2828
+ "loss": 1.2034,
2829
+ "step": 235000
2830
+ },
2831
+ {
2832
+ "epoch": 22.01,
2833
+ "learning_rate": 5.977194130292551e-06,
2834
+ "loss": 1.1977,
2835
+ "step": 235500
2836
+ },
2837
+ {
2838
+ "epoch": 22.06,
2839
+ "learning_rate": 5.883727451163661e-06,
2840
+ "loss": 1.1961,
2841
+ "step": 236000
2842
+ },
2843
+ {
2844
+ "epoch": 22.1,
2845
+ "learning_rate": 5.79026077203477e-06,
2846
+ "loss": 1.1929,
2847
+ "step": 236500
2848
+ },
2849
+ {
2850
+ "epoch": 22.15,
2851
+ "learning_rate": 5.696794092905879e-06,
2852
+ "loss": 1.1968,
2853
+ "step": 237000
2854
+ },
2855
+ {
2856
+ "epoch": 22.2,
2857
+ "learning_rate": 5.603327413776989e-06,
2858
+ "loss": 1.1946,
2859
+ "step": 237500
2860
+ },
2861
+ {
2862
+ "epoch": 22.25,
2863
+ "learning_rate": 5.5098607346480976e-06,
2864
+ "loss": 1.1971,
2865
+ "step": 238000
2866
+ },
2867
+ {
2868
+ "epoch": 22.29,
2869
+ "learning_rate": 5.416394055519208e-06,
2870
+ "loss": 1.2018,
2871
+ "step": 238500
2872
+ },
2873
+ {
2874
+ "epoch": 22.34,
2875
+ "learning_rate": 5.3229273763903175e-06,
2876
+ "loss": 1.1986,
2877
+ "step": 239000
2878
+ },
2879
+ {
2880
+ "epoch": 22.39,
2881
+ "learning_rate": 5.229460697261426e-06,
2882
+ "loss": 1.1877,
2883
+ "step": 239500
2884
+ },
2885
+ {
2886
+ "epoch": 22.43,
2887
+ "learning_rate": 5.135994018132536e-06,
2888
+ "loss": 1.1955,
2889
+ "step": 240000
2890
+ },
2891
+ {
2892
+ "epoch": 22.48,
2893
+ "learning_rate": 5.042527339003645e-06,
2894
+ "loss": 1.1989,
2895
+ "step": 240500
2896
+ },
2897
+ {
2898
+ "epoch": 22.53,
2899
+ "learning_rate": 4.949060659874755e-06,
2900
+ "loss": 1.1888,
2901
+ "step": 241000
2902
+ },
2903
+ {
2904
+ "epoch": 22.57,
2905
+ "learning_rate": 4.8555939807458645e-06,
2906
+ "loss": 1.2003,
2907
+ "step": 241500
2908
+ },
2909
+ {
2910
+ "epoch": 22.62,
2911
+ "learning_rate": 4.762127301616974e-06,
2912
+ "loss": 1.1945,
2913
+ "step": 242000
2914
+ },
2915
+ {
2916
+ "epoch": 22.67,
2917
+ "learning_rate": 4.668660622488083e-06,
2918
+ "loss": 1.1833,
2919
+ "step": 242500
2920
+ },
2921
+ {
2922
+ "epoch": 22.71,
2923
+ "learning_rate": 4.575193943359192e-06,
2924
+ "loss": 1.1866,
2925
+ "step": 243000
2926
+ },
2927
+ {
2928
+ "epoch": 22.76,
2929
+ "learning_rate": 4.481727264230303e-06,
2930
+ "loss": 1.1924,
2931
+ "step": 243500
2932
+ },
2933
+ {
2934
+ "epoch": 22.81,
2935
+ "learning_rate": 4.3882605851014115e-06,
2936
+ "loss": 1.1924,
2937
+ "step": 244000
2938
+ },
2939
+ {
2940
+ "epoch": 22.85,
2941
+ "learning_rate": 4.294793905972521e-06,
2942
+ "loss": 1.1894,
2943
+ "step": 244500
2944
+ },
2945
+ {
2946
+ "epoch": 22.9,
2947
+ "learning_rate": 4.201327226843631e-06,
2948
+ "loss": 1.2008,
2949
+ "step": 245000
2950
+ },
2951
+ {
2952
+ "epoch": 22.95,
2953
+ "learning_rate": 4.107860547714739e-06,
2954
+ "loss": 1.1959,
2955
+ "step": 245500
2956
+ },
2957
+ {
2958
+ "epoch": 22.99,
2959
+ "learning_rate": 4.01439386858585e-06,
2960
+ "loss": 1.1996,
2961
+ "step": 246000
2962
+ },
2963
+ {
2964
+ "epoch": 23.04,
2965
+ "learning_rate": 3.920927189456959e-06,
2966
+ "loss": 1.1928,
2967
+ "step": 246500
2968
+ },
2969
+ {
2970
+ "epoch": 23.09,
2971
+ "learning_rate": 3.827460510328068e-06,
2972
+ "loss": 1.195,
2973
+ "step": 247000
2974
+ },
2975
+ {
2976
+ "epoch": 23.13,
2977
+ "learning_rate": 3.7339938311991776e-06,
2978
+ "loss": 1.1873,
2979
+ "step": 247500
2980
+ },
2981
+ {
2982
+ "epoch": 23.18,
2983
+ "learning_rate": 3.6405271520702876e-06,
2984
+ "loss": 1.1977,
2985
+ "step": 248000
2986
+ },
2987
+ {
2988
+ "epoch": 23.23,
2989
+ "learning_rate": 3.5470604729413963e-06,
2990
+ "loss": 1.1909,
2991
+ "step": 248500
2992
+ },
2993
+ {
2994
+ "epoch": 23.27,
2995
+ "learning_rate": 3.4535937938125063e-06,
2996
+ "loss": 1.1881,
2997
+ "step": 249000
2998
+ },
2999
+ {
3000
+ "epoch": 23.32,
3001
+ "learning_rate": 3.360127114683616e-06,
3002
+ "loss": 1.1914,
3003
+ "step": 249500
3004
+ },
3005
+ {
3006
+ "epoch": 23.37,
3007
+ "learning_rate": 3.266660435554725e-06,
3008
+ "loss": 1.1813,
3009
+ "step": 250000
3010
+ },
3011
+ {
3012
+ "epoch": 23.41,
3013
+ "learning_rate": 3.1731937564258346e-06,
3014
+ "loss": 1.2004,
3015
+ "step": 250500
3016
+ },
3017
+ {
3018
+ "epoch": 23.46,
3019
+ "learning_rate": 3.0797270772969437e-06,
3020
+ "loss": 1.1867,
3021
+ "step": 251000
3022
+ },
3023
+ {
3024
+ "epoch": 23.51,
3025
+ "learning_rate": 2.9862603981680533e-06,
3026
+ "loss": 1.1933,
3027
+ "step": 251500
3028
+ },
3029
+ {
3030
+ "epoch": 23.55,
3031
+ "learning_rate": 2.892793719039163e-06,
3032
+ "loss": 1.1834,
3033
+ "step": 252000
3034
+ },
3035
+ {
3036
+ "epoch": 23.6,
3037
+ "learning_rate": 2.799327039910272e-06,
3038
+ "loss": 1.18,
3039
+ "step": 252500
3040
+ },
3041
+ {
3042
+ "epoch": 23.65,
3043
+ "learning_rate": 2.7058603607813815e-06,
3044
+ "loss": 1.1923,
3045
+ "step": 253000
3046
+ },
3047
+ {
3048
+ "epoch": 23.69,
3049
+ "learning_rate": 2.612393681652491e-06,
3050
+ "loss": 1.1912,
3051
+ "step": 253500
3052
+ },
3053
+ {
3054
+ "epoch": 23.74,
3055
+ "learning_rate": 2.5189270025236007e-06,
3056
+ "loss": 1.1888,
3057
+ "step": 254000
3058
+ },
3059
+ {
3060
+ "epoch": 23.79,
3061
+ "learning_rate": 2.42546032339471e-06,
3062
+ "loss": 1.1843,
3063
+ "step": 254500
3064
+ },
3065
+ {
3066
+ "epoch": 23.83,
3067
+ "learning_rate": 2.3319936442658194e-06,
3068
+ "loss": 1.1911,
3069
+ "step": 255000
3070
+ },
3071
+ {
3072
+ "epoch": 23.88,
3073
+ "learning_rate": 2.238526965136929e-06,
3074
+ "loss": 1.1943,
3075
+ "step": 255500
3076
+ },
3077
+ {
3078
+ "epoch": 23.93,
3079
+ "learning_rate": 2.145060286008038e-06,
3080
+ "loss": 1.1811,
3081
+ "step": 256000
3082
+ },
3083
+ {
3084
+ "epoch": 23.97,
3085
+ "learning_rate": 2.0515936068791476e-06,
3086
+ "loss": 1.1904,
3087
+ "step": 256500
3088
+ },
3089
+ {
3090
+ "epoch": 24.02,
3091
+ "learning_rate": 1.958126927750257e-06,
3092
+ "loss": 1.1752,
3093
+ "step": 257000
3094
+ },
3095
+ {
3096
+ "epoch": 24.07,
3097
+ "learning_rate": 1.8646602486213666e-06,
3098
+ "loss": 1.1851,
3099
+ "step": 257500
3100
+ },
3101
+ {
3102
+ "epoch": 24.11,
3103
+ "learning_rate": 1.771193569492476e-06,
3104
+ "loss": 1.1874,
3105
+ "step": 258000
3106
+ },
3107
+ {
3108
+ "epoch": 24.16,
3109
+ "learning_rate": 1.6777268903635857e-06,
3110
+ "loss": 1.1904,
3111
+ "step": 258500
3112
+ },
3113
+ {
3114
+ "epoch": 24.21,
3115
+ "learning_rate": 1.5842602112346948e-06,
3116
+ "loss": 1.188,
3117
+ "step": 259000
3118
+ },
3119
+ {
3120
+ "epoch": 24.25,
3121
+ "learning_rate": 1.4907935321058044e-06,
3122
+ "loss": 1.1808,
3123
+ "step": 259500
3124
+ },
3125
+ {
3126
+ "epoch": 24.3,
3127
+ "learning_rate": 1.3973268529769137e-06,
3128
+ "loss": 1.1855,
3129
+ "step": 260000
3130
+ },
3131
+ {
3132
+ "epoch": 24.35,
3133
+ "learning_rate": 1.3038601738480233e-06,
3134
+ "loss": 1.1856,
3135
+ "step": 260500
3136
+ },
3137
+ {
3138
+ "epoch": 24.39,
3139
+ "learning_rate": 1.2103934947191327e-06,
3140
+ "loss": 1.1943,
3141
+ "step": 261000
3142
+ },
3143
+ {
3144
+ "epoch": 24.44,
3145
+ "learning_rate": 1.1169268155902422e-06,
3146
+ "loss": 1.19,
3147
+ "step": 261500
3148
+ },
3149
+ {
3150
+ "epoch": 24.49,
3151
+ "learning_rate": 1.0234601364613516e-06,
3152
+ "loss": 1.1819,
3153
+ "step": 262000
3154
+ },
3155
+ {
3156
+ "epoch": 24.54,
3157
+ "learning_rate": 9.299934573324609e-07,
3158
+ "loss": 1.1836,
3159
+ "step": 262500
3160
+ },
3161
+ {
3162
+ "epoch": 24.58,
3163
+ "learning_rate": 8.365267782035705e-07,
3164
+ "loss": 1.182,
3165
+ "step": 263000
3166
+ },
3167
+ {
3168
+ "epoch": 24.63,
3169
+ "learning_rate": 7.4306009907468e-07,
3170
+ "loss": 1.1844,
3171
+ "step": 263500
3172
+ },
3173
+ {
3174
+ "epoch": 24.68,
3175
+ "learning_rate": 6.495934199457893e-07,
3176
+ "loss": 1.1928,
3177
+ "step": 264000
3178
+ },
3179
+ {
3180
+ "epoch": 24.72,
3181
+ "learning_rate": 5.561267408168989e-07,
3182
+ "loss": 1.1838,
3183
+ "step": 264500
3184
+ },
3185
+ {
3186
+ "epoch": 24.77,
3187
+ "learning_rate": 4.626600616880083e-07,
3188
+ "loss": 1.1849,
3189
+ "step": 265000
3190
+ },
3191
+ {
3192
+ "epoch": 24.82,
3193
+ "learning_rate": 3.691933825591177e-07,
3194
+ "loss": 1.1851,
3195
+ "step": 265500
3196
+ },
3197
+ {
3198
+ "epoch": 24.86,
3199
+ "learning_rate": 2.7572670343022714e-07,
3200
+ "loss": 1.1919,
3201
+ "step": 266000
3202
+ },
3203
+ {
3204
+ "epoch": 24.91,
3205
+ "learning_rate": 1.8226002430133658e-07,
3206
+ "loss": 1.1752,
3207
+ "step": 266500
3208
+ },
3209
+ {
3210
+ "epoch": 24.96,
3211
+ "learning_rate": 8.879334517244602e-08,
3212
+ "loss": 1.1807,
3213
+ "step": 267000
3214
+ }
3215
+ ],
3216
+ "logging_steps": 500,
3217
+ "max_steps": 267475,
3218
+ "num_input_tokens_seen": 0,
3219
+ "num_train_epochs": 25,
3220
+ "save_steps": 500,
3221
+ "total_flos": 1.1243414742110208e+18,
3222
+ "train_batch_size": 16,
3223
+ "trial_name": null,
3224
+ "trial_params": null
3225
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:71da81dd5c0baa1a02c9de1e1961c4247f951fd48acbea8964cf43d4aee1558a
3
+ size 4728
vocab.txt ADDED
The diff for this file is too large to render. See raw diff