Hugging Face's logo

cx-olquinjica
/
AngOFA-SYN

Fill-Mask
Transformers
PyTorch
xlm-roberta
Inference Endpoints
Model card Files Community
cx-olquinjica commited on
Commit
7ceeaa1
•
1 Parent(s): a15a572

Upload 7 files

Browse files
Files changed (8) hide show
  1. .gitattributes +1 -0
  2. config.json +28 -0
  3. pytorch_model.bin +3 -0
  4. special_tokens_map.json +15 -0
  5. tokenizer.json +3 -0
  6. tokenizer_config.json +21 -0
  7. trainer_state.json +4048 -0
  8. 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
config.json ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "xlm-roberta-base",
3
+ "architectures": [
4
+ "XLMRobertaForMaskedLM"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.1,
7
+ "bos_token_id": 0,
8
+ "classifier_dropout": null,
9
+ "eos_token_id": 2,
10
+ "hidden_act": "gelu",
11
+ "hidden_dropout_prob": 0.1,
12
+ "hidden_size": 768,
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 3072,
15
+ "layer_norm_eps": 1e-05,
16
+ "max_position_embeddings": 514,
17
+ "model_type": "xlm-roberta",
18
+ "num_attention_heads": 12,
19
+ "num_hidden_layers": 12,
20
+ "output_past": true,
21
+ "pad_token_id": 1,
22
+ "position_embedding_type": "absolute",
23
+ "torch_dtype": "float32",
24
+ "transformers_version": "4.32.1",
25
+ "type_vocab_size": 1,
26
+ "use_cache": true,
27
+ "vocab_size": 401145
28
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b2dad26deeedcf4bcf7cb387d598cc9c1d0d80c080d152ee4d69538f1b4798ca
3
+ size 1578166514
special_tokens_map.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "cls_token": "<s>",
4
+ "eos_token": "</s>",
5
+ "mask_token": {
6
+ "content": "<mask>",
7
+ "lstrip": true,
8
+ "normalized": true,
9
+ "rstrip": false,
10
+ "single_word": false
11
+ },
12
+ "pad_token": "<pad>",
13
+ "sep_token": "</s>",
14
+ "unk_token": "<unk>"
15
+ }
tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1e5bc2b9d6651c5444453dfab4c296455ea0279dfafb5e49283f23cc94af0846
3
+ size 26910120
tokenizer_config.json ADDED
@@ -0,0 +1,21 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "clean_up_tokenization_spaces": true,
4
+ "cls_token": "<s>",
5
+ "eos_token": "</s>",
6
+ "mask_token": {
7
+ "__type": "AddedToken",
8
+ "content": "<mask>",
9
+ "lstrip": true,
10
+ "normalized": true,
11
+ "rstrip": false,
12
+ "single_word": false
13
+ },
14
+ "model_max_length": 512,
15
+ "pad_token": "<pad>",
16
+ "sep_token": "</s>",
17
+ "sp_model_kwargs": {},
18
+ "tokenizer_class": "XLMRobertaTokenizer",
19
+ "unk_token": "<unk>",
20
+ "use_fast": true
21
+ }
trainer_state.json ADDED
@@ -0,0 +1,4048 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 24.999766967123723,
5
+ "eval_steps": 500,
6
+ "global_step": 335250,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.04,
13
+ "learning_rate": 4.992647278150634e-05,
14
+ "loss": 5.3605,
15
+ "step": 500
16
+ },
17
+ {
18
+ "epoch": 0.07,
19
+ "learning_rate": 4.985190156599553e-05,
20
+ "loss": 3.4845,
21
+ "step": 1000
22
+ },
23
+ {
24
+ "epoch": 0.11,
25
+ "learning_rate": 4.9777330350484714e-05,
26
+ "loss": 2.9795,
27
+ "step": 1500
28
+ },
29
+ {
30
+ "epoch": 0.15,
31
+ "learning_rate": 4.97027591349739e-05,
32
+ "loss": 2.6753,
33
+ "step": 2000
34
+ },
35
+ {
36
+ "epoch": 0.19,
37
+ "learning_rate": 4.9628187919463085e-05,
38
+ "loss": 2.485,
39
+ "step": 2500
40
+ },
41
+ {
42
+ "epoch": 0.22,
43
+ "learning_rate": 4.9553616703952274e-05,
44
+ "loss": 2.3458,
45
+ "step": 3000
46
+ },
47
+ {
48
+ "epoch": 0.26,
49
+ "learning_rate": 4.947904548844147e-05,
50
+ "loss": 2.2333,
51
+ "step": 3500
52
+ },
53
+ {
54
+ "epoch": 0.3,
55
+ "learning_rate": 4.940447427293065e-05,
56
+ "loss": 2.1338,
57
+ "step": 4000
58
+ },
59
+ {
60
+ "epoch": 0.34,
61
+ "learning_rate": 4.932990305741984e-05,
62
+ "loss": 2.0561,
63
+ "step": 4500
64
+ },
65
+ {
66
+ "epoch": 0.37,
67
+ "learning_rate": 4.925533184190902e-05,
68
+ "loss": 2.0012,
69
+ "step": 5000
70
+ },
71
+ {
72
+ "epoch": 0.41,
73
+ "learning_rate": 4.918076062639822e-05,
74
+ "loss": 1.9287,
75
+ "step": 5500
76
+ },
77
+ {
78
+ "epoch": 0.45,
79
+ "learning_rate": 4.91061894108874e-05,
80
+ "loss": 1.8906,
81
+ "step": 6000
82
+ },
83
+ {
84
+ "epoch": 0.48,
85
+ "learning_rate": 4.903161819537659e-05,
86
+ "loss": 1.8421,
87
+ "step": 6500
88
+ },
89
+ {
90
+ "epoch": 0.52,
91
+ "learning_rate": 4.895704697986577e-05,
92
+ "loss": 1.8052,
93
+ "step": 7000
94
+ },
95
+ {
96
+ "epoch": 0.56,
97
+ "learning_rate": 4.8882475764354966e-05,
98
+ "loss": 1.7832,
99
+ "step": 7500
100
+ },
101
+ {
102
+ "epoch": 0.6,
103
+ "learning_rate": 4.880790454884415e-05,
104
+ "loss": 1.7501,
105
+ "step": 8000
106
+ },
107
+ {
108
+ "epoch": 0.63,
109
+ "learning_rate": 4.8733333333333337e-05,
110
+ "loss": 1.7011,
111
+ "step": 8500
112
+ },
113
+ {
114
+ "epoch": 0.67,
115
+ "learning_rate": 4.865876211782252e-05,
116
+ "loss": 1.6742,
117
+ "step": 9000
118
+ },
119
+ {
120
+ "epoch": 0.71,
121
+ "learning_rate": 4.858419090231171e-05,
122
+ "loss": 1.6682,
123
+ "step": 9500
124
+ },
125
+ {
126
+ "epoch": 0.75,
127
+ "learning_rate": 4.8509619686800896e-05,
128
+ "loss": 1.6518,
129
+ "step": 10000
130
+ },
131
+ {
132
+ "epoch": 0.78,
133
+ "learning_rate": 4.8435048471290085e-05,
134
+ "loss": 1.6028,
135
+ "step": 10500
136
+ },
137
+ {
138
+ "epoch": 0.82,
139
+ "learning_rate": 4.8360477255779274e-05,
140
+ "loss": 1.5983,
141
+ "step": 11000
142
+ },
143
+ {
144
+ "epoch": 0.86,
145
+ "learning_rate": 4.8285906040268456e-05,
146
+ "loss": 1.5778,
147
+ "step": 11500
148
+ },
149
+ {
150
+ "epoch": 0.89,
151
+ "learning_rate": 4.821133482475765e-05,
152
+ "loss": 1.5532,
153
+ "step": 12000
154
+ },
155
+ {
156
+ "epoch": 0.93,
157
+ "learning_rate": 4.813691275167786e-05,
158
+ "loss": 1.5414,
159
+ "step": 12500
160
+ },
161
+ {
162
+ "epoch": 0.97,
163
+ "learning_rate": 4.8062341536167046e-05,
164
+ "loss": 1.5212,
165
+ "step": 13000
166
+ },
167
+ {
168
+ "epoch": 1.01,
169
+ "learning_rate": 4.798791946308725e-05,
170
+ "loss": 1.5105,
171
+ "step": 13500
172
+ },
173
+ {
174
+ "epoch": 1.04,
175
+ "learning_rate": 4.7913497390007464e-05,
176
+ "loss": 1.4837,
177
+ "step": 14000
178
+ },
179
+ {
180
+ "epoch": 1.08,
181
+ "learning_rate": 4.7838926174496646e-05,
182
+ "loss": 1.4661,
183
+ "step": 14500
184
+ },
185
+ {
186
+ "epoch": 1.12,
187
+ "learning_rate": 4.7764354958985834e-05,
188
+ "loss": 1.4747,
189
+ "step": 15000
190
+ },
191
+ {
192
+ "epoch": 1.16,
193
+ "learning_rate": 4.7689783743475016e-05,
194
+ "loss": 1.4458,
195
+ "step": 15500
196
+ },
197
+ {
198
+ "epoch": 1.19,
199
+ "learning_rate": 4.761521252796421e-05,
200
+ "loss": 1.4461,
201
+ "step": 16000
202
+ },
203
+ {
204
+ "epoch": 1.23,
205
+ "learning_rate": 4.7540641312453394e-05,
206
+ "loss": 1.4218,
207
+ "step": 16500
208
+ },
209
+ {
210
+ "epoch": 1.27,
211
+ "learning_rate": 4.746607009694258e-05,
212
+ "loss": 1.3968,
213
+ "step": 17000
214
+ },
215
+ {
216
+ "epoch": 1.3,
217
+ "learning_rate": 4.7391498881431765e-05,
218
+ "loss": 1.3959,
219
+ "step": 17500
220
+ },
221
+ {
222
+ "epoch": 1.34,
223
+ "learning_rate": 4.731692766592096e-05,
224
+ "loss": 1.3846,
225
+ "step": 18000
226
+ },
227
+ {
228
+ "epoch": 1.38,
229
+ "learning_rate": 4.724235645041014e-05,
230
+ "loss": 1.3882,
231
+ "step": 18500
232
+ },
233
+ {
234
+ "epoch": 1.42,
235
+ "learning_rate": 4.716778523489933e-05,
236
+ "loss": 1.3611,
237
+ "step": 19000
238
+ },
239
+ {
240
+ "epoch": 1.45,
241
+ "learning_rate": 4.709321401938852e-05,
242
+ "loss": 1.3609,
243
+ "step": 19500
244
+ },
245
+ {
246
+ "epoch": 1.49,
247
+ "learning_rate": 4.70186428038777e-05,
248
+ "loss": 1.3447,
249
+ "step": 20000
250
+ },
251
+ {
252
+ "epoch": 1.53,
253
+ "learning_rate": 4.69440715883669e-05,
254
+ "loss": 1.3426,
255
+ "step": 20500
256
+ },
257
+ {
258
+ "epoch": 1.57,
259
+ "learning_rate": 4.68696495152871e-05,
260
+ "loss": 1.3421,
261
+ "step": 21000
262
+ },
263
+ {
264
+ "epoch": 1.6,
265
+ "learning_rate": 4.679507829977629e-05,
266
+ "loss": 1.3309,
267
+ "step": 21500
268
+ },
269
+ {
270
+ "epoch": 1.64,
271
+ "learning_rate": 4.6720507084265474e-05,
272
+ "loss": 1.2993,
273
+ "step": 22000
274
+ },
275
+ {
276
+ "epoch": 1.68,
277
+ "learning_rate": 4.6646085011185686e-05,
278
+ "loss": 1.3166,
279
+ "step": 22500
280
+ },
281
+ {
282
+ "epoch": 1.72,
283
+ "learning_rate": 4.657151379567487e-05,
284
+ "loss": 1.3053,
285
+ "step": 23000
286
+ },
287
+ {
288
+ "epoch": 1.75,
289
+ "learning_rate": 4.649694258016406e-05,
290
+ "loss": 1.3013,
291
+ "step": 23500
292
+ },
293
+ {
294
+ "epoch": 1.79,
295
+ "learning_rate": 4.6422371364653246e-05,
296
+ "loss": 1.2956,
297
+ "step": 24000
298
+ },
299
+ {
300
+ "epoch": 1.83,
301
+ "learning_rate": 4.6347800149142434e-05,
302
+ "loss": 1.2732,
303
+ "step": 24500
304
+ },
305
+ {
306
+ "epoch": 1.86,
307
+ "learning_rate": 4.6273228933631616e-05,
308
+ "loss": 1.2753,
309
+ "step": 25000
310
+ },
311
+ {
312
+ "epoch": 1.9,
313
+ "learning_rate": 4.6198657718120805e-05,
314
+ "loss": 1.2672,
315
+ "step": 25500
316
+ },
317
+ {
318
+ "epoch": 1.94,
319
+ "learning_rate": 4.6124086502609994e-05,
320
+ "loss": 1.261,
321
+ "step": 26000
322
+ },
323
+ {
324
+ "epoch": 1.98,
325
+ "learning_rate": 4.604951528709918e-05,
326
+ "loss": 1.2585,
327
+ "step": 26500
328
+ },
329
+ {
330
+ "epoch": 2.01,
331
+ "learning_rate": 4.597509321401939e-05,
332
+ "loss": 1.2504,
333
+ "step": 27000
334
+ },
335
+ {
336
+ "epoch": 2.05,
337
+ "learning_rate": 4.590052199850858e-05,
338
+ "loss": 1.2506,
339
+ "step": 27500
340
+ },
341
+ {
342
+ "epoch": 2.09,
343
+ "learning_rate": 4.5825950782997766e-05,
344
+ "loss": 1.2363,
345
+ "step": 28000
346
+ },
347
+ {
348
+ "epoch": 2.13,
349
+ "learning_rate": 4.5751379567486955e-05,
350
+ "loss": 1.222,
351
+ "step": 28500
352
+ },
353
+ {
354
+ "epoch": 2.16,
355
+ "learning_rate": 4.5676808351976143e-05,
356
+ "loss": 1.2189,
357
+ "step": 29000
358
+ },
359
+ {
360
+ "epoch": 2.2,
361
+ "learning_rate": 4.560238627889635e-05,
362
+ "loss": 1.2014,
363
+ "step": 29500
364
+ },
365
+ {
366
+ "epoch": 2.24,
367
+ "learning_rate": 4.552781506338554e-05,
368
+ "loss": 1.2192,
369
+ "step": 30000
370
+ },
371
+ {
372
+ "epoch": 2.27,
373
+ "learning_rate": 4.545324384787472e-05,
374
+ "loss": 1.1989,
375
+ "step": 30500
376
+ },
377
+ {
378
+ "epoch": 2.31,
379
+ "learning_rate": 4.537867263236391e-05,
380
+ "loss": 1.1974,
381
+ "step": 31000
382
+ },
383
+ {
384
+ "epoch": 2.35,
385
+ "learning_rate": 4.53041014168531e-05,
386
+ "loss": 1.2056,
387
+ "step": 31500
388
+ },
389
+ {
390
+ "epoch": 2.39,
391
+ "learning_rate": 4.5229530201342286e-05,
392
+ "loss": 1.1954,
393
+ "step": 32000
394
+ },
395
+ {
396
+ "epoch": 2.42,
397
+ "learning_rate": 4.515510812826249e-05,
398
+ "loss": 1.1942,
399
+ "step": 32500
400
+ },
401
+ {
402
+ "epoch": 2.46,
403
+ "learning_rate": 4.508053691275168e-05,
404
+ "loss": 1.1946,
405
+ "step": 33000
406
+ },
407
+ {
408
+ "epoch": 2.5,
409
+ "learning_rate": 4.500596569724086e-05,
410
+ "loss": 1.174,
411
+ "step": 33500
412
+ },
413
+ {
414
+ "epoch": 2.54,
415
+ "learning_rate": 4.493139448173005e-05,
416
+ "loss": 1.1633,
417
+ "step": 34000
418
+ },
419
+ {
420
+ "epoch": 2.57,
421
+ "learning_rate": 4.485682326621924e-05,
422
+ "loss": 1.1667,
423
+ "step": 34500
424
+ },
425
+ {
426
+ "epoch": 2.61,
427
+ "learning_rate": 4.478240119313945e-05,
428
+ "loss": 1.1628,
429
+ "step": 35000
430
+ },
431
+ {
432
+ "epoch": 2.65,
433
+ "learning_rate": 4.470797912005966e-05,
434
+ "loss": 1.1526,
435
+ "step": 35500
436
+ },
437
+ {
438
+ "epoch": 2.68,
439
+ "learning_rate": 4.463340790454885e-05,
440
+ "loss": 1.1463,
441
+ "step": 36000
442
+ },
443
+ {
444
+ "epoch": 2.72,
445
+ "learning_rate": 4.4558836689038036e-05,
446
+ "loss": 1.1465,
447
+ "step": 36500
448
+ },
449
+ {
450
+ "epoch": 2.76,
451
+ "learning_rate": 4.448426547352722e-05,
452
+ "loss": 1.1445,
453
+ "step": 37000
454
+ },
455
+ {
456
+ "epoch": 2.8,
457
+ "learning_rate": 4.440984340044743e-05,
458
+ "loss": 1.1454,
459
+ "step": 37500
460
+ },
461
+ {
462
+ "epoch": 2.83,
463
+ "learning_rate": 4.433527218493662e-05,
464
+ "loss": 1.1362,
465
+ "step": 38000
466
+ },
467
+ {
468
+ "epoch": 2.87,
469
+ "learning_rate": 4.426070096942581e-05,
470
+ "loss": 1.1308,
471
+ "step": 38500
472
+ },
473
+ {
474
+ "epoch": 2.91,
475
+ "learning_rate": 4.418612975391499e-05,
476
+ "loss": 1.1211,
477
+ "step": 39000
478
+ },
479
+ {
480
+ "epoch": 2.95,
481
+ "learning_rate": 4.411155853840418e-05,
482
+ "loss": 1.1337,
483
+ "step": 39500
484
+ },
485
+ {
486
+ "epoch": 2.98,
487
+ "learning_rate": 4.403698732289336e-05,
488
+ "loss": 1.1272,
489
+ "step": 40000
490
+ },
491
+ {
492
+ "epoch": 3.02,
493
+ "learning_rate": 4.396256524981357e-05,
494
+ "loss": 1.1187,
495
+ "step": 40500
496
+ },
497
+ {
498
+ "epoch": 3.06,
499
+ "learning_rate": 4.388799403430276e-05,
500
+ "loss": 1.1082,
501
+ "step": 41000
502
+ },
503
+ {
504
+ "epoch": 3.09,
505
+ "learning_rate": 4.381342281879195e-05,
506
+ "loss": 1.1056,
507
+ "step": 41500
508
+ },
509
+ {
510
+ "epoch": 3.13,
511
+ "learning_rate": 4.373885160328113e-05,
512
+ "loss": 1.1042,
513
+ "step": 42000
514
+ },
515
+ {
516
+ "epoch": 3.17,
517
+ "learning_rate": 4.366428038777032e-05,
518
+ "loss": 1.1029,
519
+ "step": 42500
520
+ },
521
+ {
522
+ "epoch": 3.21,
523
+ "learning_rate": 4.358970917225951e-05,
524
+ "loss": 1.1082,
525
+ "step": 43000
526
+ },
527
+ {
528
+ "epoch": 3.24,
529
+ "learning_rate": 4.35151379567487e-05,
530
+ "loss": 1.1002,
531
+ "step": 43500
532
+ },
533
+ {
534
+ "epoch": 3.28,
535
+ "learning_rate": 4.344056674123789e-05,
536
+ "loss": 1.0933,
537
+ "step": 44000
538
+ },
539
+ {
540
+ "epoch": 3.32,
541
+ "learning_rate": 4.336599552572707e-05,
542
+ "loss": 1.0962,
543
+ "step": 44500
544
+ },
545
+ {
546
+ "epoch": 3.36,
547
+ "learning_rate": 4.329142431021626e-05,
548
+ "loss": 1.0848,
549
+ "step": 45000
550
+ },
551
+ {
552
+ "epoch": 3.39,
553
+ "learning_rate": 4.321685309470545e-05,
554
+ "loss": 1.0868,
555
+ "step": 45500
556
+ },
557
+ {
558
+ "epoch": 3.43,
559
+ "learning_rate": 4.3142281879194636e-05,
560
+ "loss": 1.0955,
561
+ "step": 46000
562
+ },
563
+ {
564
+ "epoch": 3.47,
565
+ "learning_rate": 4.306771066368382e-05,
566
+ "loss": 1.075,
567
+ "step": 46500
568
+ },
569
+ {
570
+ "epoch": 3.5,
571
+ "learning_rate": 4.299328859060403e-05,
572
+ "loss": 1.0769,
573
+ "step": 47000
574
+ },
575
+ {
576
+ "epoch": 3.54,
577
+ "learning_rate": 4.291871737509321e-05,
578
+ "loss": 1.0804,
579
+ "step": 47500
580
+ },
581
+ {
582
+ "epoch": 3.58,
583
+ "learning_rate": 4.284414615958241e-05,
584
+ "loss": 1.0595,
585
+ "step": 48000
586
+ },
587
+ {
588
+ "epoch": 3.62,
589
+ "learning_rate": 4.2769724086502613e-05,
590
+ "loss": 1.0628,
591
+ "step": 48500
592
+ },
593
+ {
594
+ "epoch": 3.65,
595
+ "learning_rate": 4.26951528709918e-05,
596
+ "loss": 1.0656,
597
+ "step": 49000
598
+ },
599
+ {
600
+ "epoch": 3.69,
601
+ "learning_rate": 4.2620581655480984e-05,
602
+ "loss": 1.0656,
603
+ "step": 49500
604
+ },
605
+ {
606
+ "epoch": 3.73,
607
+ "learning_rate": 4.254601043997017e-05,
608
+ "loss": 1.0523,
609
+ "step": 50000
610
+ },
611
+ {
612
+ "epoch": 3.77,
613
+ "learning_rate": 4.247143922445936e-05,
614
+ "loss": 1.0538,
615
+ "step": 50500
616
+ },
617
+ {
618
+ "epoch": 3.8,
619
+ "learning_rate": 4.239686800894855e-05,
620
+ "loss": 1.0577,
621
+ "step": 51000
622
+ },
623
+ {
624
+ "epoch": 3.84,
625
+ "learning_rate": 4.232229679343773e-05,
626
+ "loss": 1.0548,
627
+ "step": 51500
628
+ },
629
+ {
630
+ "epoch": 3.88,
631
+ "learning_rate": 4.2247874720357945e-05,
632
+ "loss": 1.0555,
633
+ "step": 52000
634
+ },
635
+ {
636
+ "epoch": 3.91,
637
+ "learning_rate": 4.2173303504847134e-05,
638
+ "loss": 1.0373,
639
+ "step": 52500
640
+ },
641
+ {
642
+ "epoch": 3.95,
643
+ "learning_rate": 4.2098732289336316e-05,
644
+ "loss": 1.0473,
645
+ "step": 53000
646
+ },
647
+ {
648
+ "epoch": 3.99,
649
+ "learning_rate": 4.2024161073825504e-05,
650
+ "loss": 1.0423,
651
+ "step": 53500
652
+ },
653
+ {
654
+ "epoch": 4.03,
655
+ "learning_rate": 4.194973900074572e-05,
656
+ "loss": 1.0484,
657
+ "step": 54000
658
+ },
659
+ {
660
+ "epoch": 4.06,
661
+ "learning_rate": 4.1875167785234906e-05,
662
+ "loss": 1.0332,
663
+ "step": 54500
664
+ },
665
+ {
666
+ "epoch": 4.1,
667
+ "learning_rate": 4.180059656972409e-05,
668
+ "loss": 1.0329,
669
+ "step": 55000
670
+ },
671
+ {
672
+ "epoch": 4.14,
673
+ "learning_rate": 4.1726025354213276e-05,
674
+ "loss": 1.0273,
675
+ "step": 55500
676
+ },
677
+ {
678
+ "epoch": 4.18,
679
+ "learning_rate": 4.165160328113348e-05,
680
+ "loss": 1.0288,
681
+ "step": 56000
682
+ },
683
+ {
684
+ "epoch": 4.21,
685
+ "learning_rate": 4.157703206562267e-05,
686
+ "loss": 1.0206,
687
+ "step": 56500
688
+ },
689
+ {
690
+ "epoch": 4.25,
691
+ "learning_rate": 4.150246085011186e-05,
692
+ "loss": 1.023,
693
+ "step": 57000
694
+ },
695
+ {
696
+ "epoch": 4.29,
697
+ "learning_rate": 4.142788963460105e-05,
698
+ "loss": 1.0277,
699
+ "step": 57500
700
+ },
701
+ {
702
+ "epoch": 4.33,
703
+ "learning_rate": 4.135331841909023e-05,
704
+ "loss": 1.0213,
705
+ "step": 58000
706
+ },
707
+ {
708
+ "epoch": 4.36,
709
+ "learning_rate": 4.127874720357942e-05,
710
+ "loss": 1.017,
711
+ "step": 58500
712
+ },
713
+ {
714
+ "epoch": 4.4,
715
+ "learning_rate": 4.120417598806861e-05,
716
+ "loss": 1.0195,
717
+ "step": 59000
718
+ },
719
+ {
720
+ "epoch": 4.44,
721
+ "learning_rate": 4.1129604772557797e-05,
722
+ "loss": 1.0132,
723
+ "step": 59500
724
+ },
725
+ {
726
+ "epoch": 4.47,
727
+ "learning_rate": 4.1055182699478e-05,
728
+ "loss": 1.0123,
729
+ "step": 60000
730
+ },
731
+ {
732
+ "epoch": 4.51,
733
+ "learning_rate": 4.098061148396719e-05,
734
+ "loss": 0.9978,
735
+ "step": 60500
736
+ },
737
+ {
738
+ "epoch": 4.55,
739
+ "learning_rate": 4.090604026845638e-05,
740
+ "loss": 1.0106,
741
+ "step": 61000
742
+ },
743
+ {
744
+ "epoch": 4.59,
745
+ "learning_rate": 4.083146905294556e-05,
746
+ "loss": 0.9945,
747
+ "step": 61500
748
+ },
749
+ {
750
+ "epoch": 4.62,
751
+ "learning_rate": 4.075689783743476e-05,
752
+ "loss": 1.0011,
753
+ "step": 62000
754
+ },
755
+ {
756
+ "epoch": 4.66,
757
+ "learning_rate": 4.068232662192394e-05,
758
+ "loss": 0.9933,
759
+ "step": 62500
760
+ },
761
+ {
762
+ "epoch": 4.7,
763
+ "learning_rate": 4.060790454884415e-05,
764
+ "loss": 1.0049,
765
+ "step": 63000
766
+ },
767
+ {
768
+ "epoch": 4.74,
769
+ "learning_rate": 4.053348247576436e-05,
770
+ "loss": 0.9991,
771
+ "step": 63500
772
+ },
773
+ {
774
+ "epoch": 4.77,
775
+ "learning_rate": 4.0458911260253546e-05,
776
+ "loss": 0.9994,
777
+ "step": 64000
778
+ },
779
+ {
780
+ "epoch": 4.81,
781
+ "learning_rate": 4.038434004474273e-05,
782
+ "loss": 0.9937,
783
+ "step": 64500
784
+ },
785
+ {
786
+ "epoch": 4.85,
787
+ "learning_rate": 4.030976882923192e-05,
788
+ "loss": 0.9941,
789
+ "step": 65000
790
+ },
791
+ {
792
+ "epoch": 4.88,
793
+ "learning_rate": 4.0235197613721106e-05,
794
+ "loss": 0.9905,
795
+ "step": 65500
796
+ },
797
+ {
798
+ "epoch": 4.92,
799
+ "learning_rate": 4.0160626398210294e-05,
800
+ "loss": 1.0044,
801
+ "step": 66000
802
+ },
803
+ {
804
+ "epoch": 4.96,
805
+ "learning_rate": 4.0086055182699476e-05,
806
+ "loss": 0.9899,
807
+ "step": 66500
808
+ },
809
+ {
810
+ "epoch": 5.0,
811
+ "learning_rate": 4.0011483967188665e-05,
812
+ "loss": 0.9891,
813
+ "step": 67000
814
+ },
815
+ {
816
+ "epoch": 5.03,
817
+ "learning_rate": 3.9936912751677854e-05,
818
+ "loss": 0.9818,
819
+ "step": 67500
820
+ },
821
+ {
822
+ "epoch": 5.07,
823
+ "learning_rate": 3.9862490678598066e-05,
824
+ "loss": 0.9858,
825
+ "step": 68000
826
+ },
827
+ {
828
+ "epoch": 5.11,
829
+ "learning_rate": 3.978791946308725e-05,
830
+ "loss": 0.9752,
831
+ "step": 68500
832
+ },
833
+ {
834
+ "epoch": 5.15,
835
+ "learning_rate": 3.971334824757644e-05,
836
+ "loss": 0.9735,
837
+ "step": 69000
838
+ },
839
+ {
840
+ "epoch": 5.18,
841
+ "learning_rate": 3.9638777032065626e-05,
842
+ "loss": 0.9742,
843
+ "step": 69500
844
+ },
845
+ {
846
+ "epoch": 5.22,
847
+ "learning_rate": 3.956420581655481e-05,
848
+ "loss": 0.9745,
849
+ "step": 70000
850
+ },
851
+ {
852
+ "epoch": 5.26,
853
+ "learning_rate": 3.9489634601044003e-05,
854
+ "loss": 0.9675,
855
+ "step": 70500
856
+ },
857
+ {
858
+ "epoch": 5.29,
859
+ "learning_rate": 3.9415063385533185e-05,
860
+ "loss": 0.9649,
861
+ "step": 71000
862
+ },
863
+ {
864
+ "epoch": 5.33,
865
+ "learning_rate": 3.93406413124534e-05,
866
+ "loss": 0.983,
867
+ "step": 71500
868
+ },
869
+ {
870
+ "epoch": 5.37,
871
+ "learning_rate": 3.926607009694258e-05,
872
+ "loss": 0.9622,
873
+ "step": 72000
874
+ },
875
+ {
876
+ "epoch": 5.41,
877
+ "learning_rate": 3.919149888143177e-05,
878
+ "loss": 0.9659,
879
+ "step": 72500
880
+ },
881
+ {
882
+ "epoch": 5.44,
883
+ "learning_rate": 3.911692766592096e-05,
884
+ "loss": 0.9717,
885
+ "step": 73000
886
+ },
887
+ {
888
+ "epoch": 5.48,
889
+ "learning_rate": 3.9042356450410146e-05,
890
+ "loss": 0.9503,
891
+ "step": 73500
892
+ },
893
+ {
894
+ "epoch": 5.52,
895
+ "learning_rate": 3.896793437733035e-05,
896
+ "loss": 0.9699,
897
+ "step": 74000
898
+ },
899
+ {
900
+ "epoch": 5.56,
901
+ "learning_rate": 3.889336316181954e-05,
902
+ "loss": 0.962,
903
+ "step": 74500
904
+ },
905
+ {
906
+ "epoch": 5.59,
907
+ "learning_rate": 3.881879194630872e-05,
908
+ "loss": 0.9561,
909
+ "step": 75000
910
+ },
911
+ {
912
+ "epoch": 5.63,
913
+ "learning_rate": 3.874422073079791e-05,
914
+ "loss": 0.9587,
915
+ "step": 75500
916
+ },
917
+ {
918
+ "epoch": 5.67,
919
+ "learning_rate": 3.86696495152871e-05,
920
+ "loss": 0.9576,
921
+ "step": 76000
922
+ },
923
+ {
924
+ "epoch": 5.7,
925
+ "learning_rate": 3.859507829977629e-05,
926
+ "loss": 0.9509,
927
+ "step": 76500
928
+ },
929
+ {
930
+ "epoch": 5.74,
931
+ "learning_rate": 3.8520656226696495e-05,
932
+ "loss": 0.9565,
933
+ "step": 77000
934
+ },
935
+ {
936
+ "epoch": 5.78,
937
+ "learning_rate": 3.844608501118568e-05,
938
+ "loss": 0.9555,
939
+ "step": 77500
940
+ },
941
+ {
942
+ "epoch": 5.82,
943
+ "learning_rate": 3.837151379567487e-05,
944
+ "loss": 0.9573,
945
+ "step": 78000
946
+ },
947
+ {
948
+ "epoch": 5.85,
949
+ "learning_rate": 3.829694258016406e-05,
950
+ "loss": 0.9459,
951
+ "step": 78500
952
+ },
953
+ {
954
+ "epoch": 5.89,
955
+ "learning_rate": 3.822237136465325e-05,
956
+ "loss": 0.9394,
957
+ "step": 79000
958
+ },
959
+ {
960
+ "epoch": 5.93,
961
+ "learning_rate": 3.814780014914243e-05,
962
+ "loss": 0.9417,
963
+ "step": 79500
964
+ },
965
+ {
966
+ "epoch": 5.97,
967
+ "learning_rate": 3.8073378076062644e-05,
968
+ "loss": 0.9459,
969
+ "step": 80000
970
+ },
971
+ {
972
+ "epoch": 6.0,
973
+ "learning_rate": 3.7998806860551826e-05,
974
+ "loss": 0.9329,
975
+ "step": 80500
976
+ },
977
+ {
978
+ "epoch": 6.04,
979
+ "learning_rate": 3.7924235645041015e-05,
980
+ "loss": 0.9297,
981
+ "step": 81000
982
+ },
983
+ {
984
+ "epoch": 6.08,
985
+ "learning_rate": 3.7849664429530204e-05,
986
+ "loss": 0.9379,
987
+ "step": 81500
988
+ },
989
+ {
990
+ "epoch": 6.11,
991
+ "learning_rate": 3.777509321401939e-05,
992
+ "loss": 0.9363,
993
+ "step": 82000
994
+ },
995
+ {
996
+ "epoch": 6.15,
997
+ "learning_rate": 3.77006711409396e-05,
998
+ "loss": 0.9273,
999
+ "step": 82500
1000
+ },
1001
+ {
1002
+ "epoch": 6.19,
1003
+ "learning_rate": 3.762609992542879e-05,
1004
+ "loss": 0.9367,
1005
+ "step": 83000
1006
+ },
1007
+ {
1008
+ "epoch": 6.23,
1009
+ "learning_rate": 3.755152870991797e-05,
1010
+ "loss": 0.9395,
1011
+ "step": 83500
1012
+ },
1013
+ {
1014
+ "epoch": 6.26,
1015
+ "learning_rate": 3.7476957494407164e-05,
1016
+ "loss": 0.9302,
1017
+ "step": 84000
1018
+ },
1019
+ {
1020
+ "epoch": 6.3,
1021
+ "learning_rate": 3.7402386278896346e-05,
1022
+ "loss": 0.9259,
1023
+ "step": 84500
1024
+ },
1025
+ {
1026
+ "epoch": 6.34,
1027
+ "learning_rate": 3.732796420581656e-05,
1028
+ "loss": 0.9315,
1029
+ "step": 85000
1030
+ },
1031
+ {
1032
+ "epoch": 6.38,
1033
+ "learning_rate": 3.725339299030574e-05,
1034
+ "loss": 0.9245,
1035
+ "step": 85500
1036
+ },
1037
+ {
1038
+ "epoch": 6.41,
1039
+ "learning_rate": 3.717882177479493e-05,
1040
+ "loss": 0.9289,
1041
+ "step": 86000
1042
+ },
1043
+ {
1044
+ "epoch": 6.45,
1045
+ "learning_rate": 3.710425055928412e-05,
1046
+ "loss": 0.9199,
1047
+ "step": 86500
1048
+ },
1049
+ {
1050
+ "epoch": 6.49,
1051
+ "learning_rate": 3.702967934377331e-05,
1052
+ "loss": 0.9177,
1053
+ "step": 87000
1054
+ },
1055
+ {
1056
+ "epoch": 6.52,
1057
+ "learning_rate": 3.6955108128262496e-05,
1058
+ "loss": 0.9263,
1059
+ "step": 87500
1060
+ },
1061
+ {
1062
+ "epoch": 6.56,
1063
+ "learning_rate": 3.68806860551827e-05,
1064
+ "loss": 0.9239,
1065
+ "step": 88000
1066
+ },
1067
+ {
1068
+ "epoch": 6.6,
1069
+ "learning_rate": 3.680611483967189e-05,
1070
+ "loss": 0.924,
1071
+ "step": 88500
1072
+ },
1073
+ {
1074
+ "epoch": 6.64,
1075
+ "learning_rate": 3.673154362416107e-05,
1076
+ "loss": 0.918,
1077
+ "step": 89000
1078
+ },
1079
+ {
1080
+ "epoch": 6.67,
1081
+ "learning_rate": 3.665697240865026e-05,
1082
+ "loss": 0.9203,
1083
+ "step": 89500
1084
+ },
1085
+ {
1086
+ "epoch": 6.71,
1087
+ "learning_rate": 3.658269947800149e-05,
1088
+ "loss": 0.9224,
1089
+ "step": 90000
1090
+ },
1091
+ {
1092
+ "epoch": 6.75,
1093
+ "learning_rate": 3.650812826249068e-05,
1094
+ "loss": 0.9149,
1095
+ "step": 90500
1096
+ },
1097
+ {
1098
+ "epoch": 6.79,
1099
+ "learning_rate": 3.643355704697987e-05,
1100
+ "loss": 0.9134,
1101
+ "step": 91000
1102
+ },
1103
+ {
1104
+ "epoch": 6.82,
1105
+ "learning_rate": 3.635898583146906e-05,
1106
+ "loss": 0.9185,
1107
+ "step": 91500
1108
+ },
1109
+ {
1110
+ "epoch": 6.86,
1111
+ "learning_rate": 3.628441461595824e-05,
1112
+ "loss": 0.9132,
1113
+ "step": 92000
1114
+ },
1115
+ {
1116
+ "epoch": 6.9,
1117
+ "learning_rate": 3.620984340044743e-05,
1118
+ "loss": 0.9159,
1119
+ "step": 92500
1120
+ },
1121
+ {
1122
+ "epoch": 6.94,
1123
+ "learning_rate": 3.6135272184936616e-05,
1124
+ "loss": 0.9081,
1125
+ "step": 93000
1126
+ },
1127
+ {
1128
+ "epoch": 6.97,
1129
+ "learning_rate": 3.6060700969425805e-05,
1130
+ "loss": 0.9047,
1131
+ "step": 93500
1132
+ },
1133
+ {
1134
+ "epoch": 7.01,
1135
+ "learning_rate": 3.5986129753914994e-05,
1136
+ "loss": 0.9107,
1137
+ "step": 94000
1138
+ },
1139
+ {
1140
+ "epoch": 7.05,
1141
+ "learning_rate": 3.5911558538404176e-05,
1142
+ "loss": 0.9004,
1143
+ "step": 94500
1144
+ },
1145
+ {
1146
+ "epoch": 7.08,
1147
+ "learning_rate": 3.5836987322893364e-05,
1148
+ "loss": 0.8956,
1149
+ "step": 95000
1150
+ },
1151
+ {
1152
+ "epoch": 7.12,
1153
+ "learning_rate": 3.576256524981357e-05,
1154
+ "loss": 0.9009,
1155
+ "step": 95500
1156
+ },
1157
+ {
1158
+ "epoch": 7.16,
1159
+ "learning_rate": 3.5687994034302766e-05,
1160
+ "loss": 0.9023,
1161
+ "step": 96000
1162
+ },
1163
+ {
1164
+ "epoch": 7.2,
1165
+ "learning_rate": 3.561342281879195e-05,
1166
+ "loss": 0.8972,
1167
+ "step": 96500
1168
+ },
1169
+ {
1170
+ "epoch": 7.23,
1171
+ "learning_rate": 3.5538851603281136e-05,
1172
+ "loss": 0.8967,
1173
+ "step": 97000
1174
+ },
1175
+ {
1176
+ "epoch": 7.27,
1177
+ "learning_rate": 3.546428038777032e-05,
1178
+ "loss": 0.8867,
1179
+ "step": 97500
1180
+ },
1181
+ {
1182
+ "epoch": 7.31,
1183
+ "learning_rate": 3.538985831469053e-05,
1184
+ "loss": 0.9029,
1185
+ "step": 98000
1186
+ },
1187
+ {
1188
+ "epoch": 7.35,
1189
+ "learning_rate": 3.5315436241610737e-05,
1190
+ "loss": 0.901,
1191
+ "step": 98500
1192
+ },
1193
+ {
1194
+ "epoch": 7.38,
1195
+ "learning_rate": 3.5240865026099925e-05,
1196
+ "loss": 0.889,
1197
+ "step": 99000
1198
+ },
1199
+ {
1200
+ "epoch": 7.42,
1201
+ "learning_rate": 3.5166293810589114e-05,
1202
+ "loss": 0.8931,
1203
+ "step": 99500
1204
+ },
1205
+ {
1206
+ "epoch": 7.46,
1207
+ "learning_rate": 3.50917225950783e-05,
1208
+ "loss": 0.892,
1209
+ "step": 100000
1210
+ },
1211
+ {
1212
+ "epoch": 7.49,
1213
+ "learning_rate": 3.5017151379567485e-05,
1214
+ "loss": 0.8955,
1215
+ "step": 100500
1216
+ },
1217
+ {
1218
+ "epoch": 7.53,
1219
+ "learning_rate": 3.4942580164056674e-05,
1220
+ "loss": 0.891,
1221
+ "step": 101000
1222
+ },
1223
+ {
1224
+ "epoch": 7.57,
1225
+ "learning_rate": 3.486800894854586e-05,
1226
+ "loss": 0.8959,
1227
+ "step": 101500
1228
+ },
1229
+ {
1230
+ "epoch": 7.61,
1231
+ "learning_rate": 3.4793586875466075e-05,
1232
+ "loss": 0.8927,
1233
+ "step": 102000
1234
+ },
1235
+ {
1236
+ "epoch": 7.64,
1237
+ "learning_rate": 3.471901565995526e-05,
1238
+ "loss": 0.886,
1239
+ "step": 102500
1240
+ },
1241
+ {
1242
+ "epoch": 7.68,
1243
+ "learning_rate": 3.4644444444444446e-05,
1244
+ "loss": 0.8811,
1245
+ "step": 103000
1246
+ },
1247
+ {
1248
+ "epoch": 7.72,
1249
+ "learning_rate": 3.4569873228933634e-05,
1250
+ "loss": 0.8788,
1251
+ "step": 103500
1252
+ },
1253
+ {
1254
+ "epoch": 7.76,
1255
+ "learning_rate": 3.449530201342282e-05,
1256
+ "loss": 0.8892,
1257
+ "step": 104000
1258
+ },
1259
+ {
1260
+ "epoch": 7.79,
1261
+ "learning_rate": 3.442073079791201e-05,
1262
+ "loss": 0.8826,
1263
+ "step": 104500
1264
+ },
1265
+ {
1266
+ "epoch": 7.83,
1267
+ "learning_rate": 3.4346159582401194e-05,
1268
+ "loss": 0.8872,
1269
+ "step": 105000
1270
+ },
1271
+ {
1272
+ "epoch": 7.87,
1273
+ "learning_rate": 3.427158836689038e-05,
1274
+ "loss": 0.8826,
1275
+ "step": 105500
1276
+ },
1277
+ {
1278
+ "epoch": 7.9,
1279
+ "learning_rate": 3.4197017151379565e-05,
1280
+ "loss": 0.8824,
1281
+ "step": 106000
1282
+ },
1283
+ {
1284
+ "epoch": 7.94,
1285
+ "learning_rate": 3.412244593586876e-05,
1286
+ "loss": 0.8805,
1287
+ "step": 106500
1288
+ },
1289
+ {
1290
+ "epoch": 7.98,
1291
+ "learning_rate": 3.404787472035794e-05,
1292
+ "loss": 0.8835,
1293
+ "step": 107000
1294
+ },
1295
+ {
1296
+ "epoch": 8.02,
1297
+ "learning_rate": 3.397330350484713e-05,
1298
+ "loss": 0.8835,
1299
+ "step": 107500
1300
+ },
1301
+ {
1302
+ "epoch": 8.05,
1303
+ "learning_rate": 3.3898881431767337e-05,
1304
+ "loss": 0.877,
1305
+ "step": 108000
1306
+ },
1307
+ {
1308
+ "epoch": 8.09,
1309
+ "learning_rate": 3.3824310216256525e-05,
1310
+ "loss": 0.8645,
1311
+ "step": 108500
1312
+ },
1313
+ {
1314
+ "epoch": 8.13,
1315
+ "learning_rate": 3.3749739000745714e-05,
1316
+ "loss": 0.8742,
1317
+ "step": 109000
1318
+ },
1319
+ {
1320
+ "epoch": 8.17,
1321
+ "learning_rate": 3.36751677852349e-05,
1322
+ "loss": 0.8655,
1323
+ "step": 109500
1324
+ },
1325
+ {
1326
+ "epoch": 8.2,
1327
+ "learning_rate": 3.3600596569724085e-05,
1328
+ "loss": 0.8711,
1329
+ "step": 110000
1330
+ },
1331
+ {
1332
+ "epoch": 8.24,
1333
+ "learning_rate": 3.352632363907532e-05,
1334
+ "loss": 0.8715,
1335
+ "step": 110500
1336
+ },
1337
+ {
1338
+ "epoch": 8.28,
1339
+ "learning_rate": 3.34517524235645e-05,
1340
+ "loss": 0.8717,
1341
+ "step": 111000
1342
+ },
1343
+ {
1344
+ "epoch": 8.31,
1345
+ "learning_rate": 3.337718120805369e-05,
1346
+ "loss": 0.8632,
1347
+ "step": 111500
1348
+ },
1349
+ {
1350
+ "epoch": 8.35,
1351
+ "learning_rate": 3.330260999254288e-05,
1352
+ "loss": 0.8666,
1353
+ "step": 112000
1354
+ },
1355
+ {
1356
+ "epoch": 8.39,
1357
+ "learning_rate": 3.3228187919463086e-05,
1358
+ "loss": 0.8669,
1359
+ "step": 112500
1360
+ },
1361
+ {
1362
+ "epoch": 8.43,
1363
+ "learning_rate": 3.3153616703952275e-05,
1364
+ "loss": 0.8653,
1365
+ "step": 113000
1366
+ },
1367
+ {
1368
+ "epoch": 8.46,
1369
+ "learning_rate": 3.3079045488441464e-05,
1370
+ "loss": 0.8602,
1371
+ "step": 113500
1372
+ },
1373
+ {
1374
+ "epoch": 8.5,
1375
+ "learning_rate": 3.300447427293065e-05,
1376
+ "loss": 0.8666,
1377
+ "step": 114000
1378
+ },
1379
+ {
1380
+ "epoch": 8.54,
1381
+ "learning_rate": 3.2929903057419834e-05,
1382
+ "loss": 0.8565,
1383
+ "step": 114500
1384
+ },
1385
+ {
1386
+ "epoch": 8.58,
1387
+ "learning_rate": 3.285533184190902e-05,
1388
+ "loss": 0.8628,
1389
+ "step": 115000
1390
+ },
1391
+ {
1392
+ "epoch": 8.61,
1393
+ "learning_rate": 3.278076062639821e-05,
1394
+ "loss": 0.8666,
1395
+ "step": 115500
1396
+ },
1397
+ {
1398
+ "epoch": 8.65,
1399
+ "learning_rate": 3.27061894108874e-05,
1400
+ "loss": 0.8649,
1401
+ "step": 116000
1402
+ },
1403
+ {
1404
+ "epoch": 8.69,
1405
+ "learning_rate": 3.2631767337807606e-05,
1406
+ "loss": 0.8605,
1407
+ "step": 116500
1408
+ },
1409
+ {
1410
+ "epoch": 8.72,
1411
+ "learning_rate": 3.2557196122296795e-05,
1412
+ "loss": 0.8594,
1413
+ "step": 117000
1414
+ },
1415
+ {
1416
+ "epoch": 8.76,
1417
+ "learning_rate": 3.248262490678598e-05,
1418
+ "loss": 0.8624,
1419
+ "step": 117500
1420
+ },
1421
+ {
1422
+ "epoch": 8.8,
1423
+ "learning_rate": 3.240805369127517e-05,
1424
+ "loss": 0.8604,
1425
+ "step": 118000
1426
+ },
1427
+ {
1428
+ "epoch": 8.84,
1429
+ "learning_rate": 3.2333482475764355e-05,
1430
+ "loss": 0.8592,
1431
+ "step": 118500
1432
+ },
1433
+ {
1434
+ "epoch": 8.87,
1435
+ "learning_rate": 3.225906040268457e-05,
1436
+ "loss": 0.8643,
1437
+ "step": 119000
1438
+ },
1439
+ {
1440
+ "epoch": 8.91,
1441
+ "learning_rate": 3.218448918717375e-05,
1442
+ "loss": 0.8638,
1443
+ "step": 119500
1444
+ },
1445
+ {
1446
+ "epoch": 8.95,
1447
+ "learning_rate": 3.210991797166294e-05,
1448
+ "loss": 0.8563,
1449
+ "step": 120000
1450
+ },
1451
+ {
1452
+ "epoch": 8.99,
1453
+ "learning_rate": 3.203534675615213e-05,
1454
+ "loss": 0.8497,
1455
+ "step": 120500
1456
+ },
1457
+ {
1458
+ "epoch": 9.02,
1459
+ "learning_rate": 3.1960775540641315e-05,
1460
+ "loss": 0.8588,
1461
+ "step": 121000
1462
+ },
1463
+ {
1464
+ "epoch": 9.06,
1465
+ "learning_rate": 3.188635346756153e-05,
1466
+ "loss": 0.8427,
1467
+ "step": 121500
1468
+ },
1469
+ {
1470
+ "epoch": 9.1,
1471
+ "learning_rate": 3.181178225205071e-05,
1472
+ "loss": 0.8571,
1473
+ "step": 122000
1474
+ },
1475
+ {
1476
+ "epoch": 9.13,
1477
+ "learning_rate": 3.17372110365399e-05,
1478
+ "loss": 0.8445,
1479
+ "step": 122500
1480
+ },
1481
+ {
1482
+ "epoch": 9.17,
1483
+ "learning_rate": 3.166263982102908e-05,
1484
+ "loss": 0.8491,
1485
+ "step": 123000
1486
+ },
1487
+ {
1488
+ "epoch": 9.21,
1489
+ "learning_rate": 3.1588068605518276e-05,
1490
+ "loss": 0.8394,
1491
+ "step": 123500
1492
+ },
1493
+ {
1494
+ "epoch": 9.25,
1495
+ "learning_rate": 3.151364653243848e-05,
1496
+ "loss": 0.8485,
1497
+ "step": 124000
1498
+ },
1499
+ {
1500
+ "epoch": 9.28,
1501
+ "learning_rate": 3.143907531692767e-05,
1502
+ "loss": 0.8426,
1503
+ "step": 124500
1504
+ },
1505
+ {
1506
+ "epoch": 9.32,
1507
+ "learning_rate": 3.136450410141685e-05,
1508
+ "loss": 0.847,
1509
+ "step": 125000
1510
+ },
1511
+ {
1512
+ "epoch": 9.36,
1513
+ "learning_rate": 3.128993288590604e-05,
1514
+ "loss": 0.8453,
1515
+ "step": 125500
1516
+ },
1517
+ {
1518
+ "epoch": 9.4,
1519
+ "learning_rate": 3.121536167039523e-05,
1520
+ "loss": 0.8464,
1521
+ "step": 126000
1522
+ },
1523
+ {
1524
+ "epoch": 9.43,
1525
+ "learning_rate": 3.114079045488442e-05,
1526
+ "loss": 0.8477,
1527
+ "step": 126500
1528
+ },
1529
+ {
1530
+ "epoch": 9.47,
1531
+ "learning_rate": 3.1066368381804625e-05,
1532
+ "loss": 0.84,
1533
+ "step": 127000
1534
+ },
1535
+ {
1536
+ "epoch": 9.51,
1537
+ "learning_rate": 3.099179716629381e-05,
1538
+ "loss": 0.8442,
1539
+ "step": 127500
1540
+ },
1541
+ {
1542
+ "epoch": 9.55,
1543
+ "learning_rate": 3.091737509321402e-05,
1544
+ "loss": 0.8431,
1545
+ "step": 128000
1546
+ },
1547
+ {
1548
+ "epoch": 9.58,
1549
+ "learning_rate": 3.084280387770321e-05,
1550
+ "loss": 0.842,
1551
+ "step": 128500
1552
+ },
1553
+ {
1554
+ "epoch": 9.62,
1555
+ "learning_rate": 3.0768232662192396e-05,
1556
+ "loss": 0.8428,
1557
+ "step": 129000
1558
+ },
1559
+ {
1560
+ "epoch": 9.66,
1561
+ "learning_rate": 3.0693661446681585e-05,
1562
+ "loss": 0.8397,
1563
+ "step": 129500
1564
+ },
1565
+ {
1566
+ "epoch": 9.69,
1567
+ "learning_rate": 3.0619090231170774e-05,
1568
+ "loss": 0.8439,
1569
+ "step": 130000
1570
+ },
1571
+ {
1572
+ "epoch": 9.73,
1573
+ "learning_rate": 3.0544519015659956e-05,
1574
+ "loss": 0.8376,
1575
+ "step": 130500
1576
+ },
1577
+ {
1578
+ "epoch": 9.77,
1579
+ "learning_rate": 3.047009694258017e-05,
1580
+ "loss": 0.835,
1581
+ "step": 131000
1582
+ },
1583
+ {
1584
+ "epoch": 9.81,
1585
+ "learning_rate": 3.039552572706935e-05,
1586
+ "loss": 0.8321,
1587
+ "step": 131500
1588
+ },
1589
+ {
1590
+ "epoch": 9.84,
1591
+ "learning_rate": 3.0320954511558543e-05,
1592
+ "loss": 0.8346,
1593
+ "step": 132000
1594
+ },
1595
+ {
1596
+ "epoch": 9.88,
1597
+ "learning_rate": 3.0246383296047725e-05,
1598
+ "loss": 0.8362,
1599
+ "step": 132500
1600
+ },
1601
+ {
1602
+ "epoch": 9.92,
1603
+ "learning_rate": 3.0171961222967937e-05,
1604
+ "loss": 0.8297,
1605
+ "step": 133000
1606
+ },
1607
+ {
1608
+ "epoch": 9.96,
1609
+ "learning_rate": 3.0097539149888143e-05,
1610
+ "loss": 0.841,
1611
+ "step": 133500
1612
+ },
1613
+ {
1614
+ "epoch": 9.99,
1615
+ "learning_rate": 3.002296793437733e-05,
1616
+ "loss": 0.8359,
1617
+ "step": 134000
1618
+ },
1619
+ {
1620
+ "epoch": 10.03,
1621
+ "learning_rate": 2.9948396718866517e-05,
1622
+ "loss": 0.8275,
1623
+ "step": 134500
1624
+ },
1625
+ {
1626
+ "epoch": 10.07,
1627
+ "learning_rate": 2.9873825503355706e-05,
1628
+ "loss": 0.8226,
1629
+ "step": 135000
1630
+ },
1631
+ {
1632
+ "epoch": 10.1,
1633
+ "learning_rate": 2.979925428784489e-05,
1634
+ "loss": 0.8322,
1635
+ "step": 135500
1636
+ },
1637
+ {
1638
+ "epoch": 10.14,
1639
+ "learning_rate": 2.972468307233408e-05,
1640
+ "loss": 0.8213,
1641
+ "step": 136000
1642
+ },
1643
+ {
1644
+ "epoch": 10.18,
1645
+ "learning_rate": 2.9650111856823265e-05,
1646
+ "loss": 0.8237,
1647
+ "step": 136500
1648
+ },
1649
+ {
1650
+ "epoch": 10.22,
1651
+ "learning_rate": 2.9575540641312454e-05,
1652
+ "loss": 0.8273,
1653
+ "step": 137000
1654
+ },
1655
+ {
1656
+ "epoch": 10.25,
1657
+ "learning_rate": 2.9500969425801646e-05,
1658
+ "loss": 0.8243,
1659
+ "step": 137500
1660
+ },
1661
+ {
1662
+ "epoch": 10.29,
1663
+ "learning_rate": 2.9426398210290828e-05,
1664
+ "loss": 0.82,
1665
+ "step": 138000
1666
+ },
1667
+ {
1668
+ "epoch": 10.33,
1669
+ "learning_rate": 2.9351826994780017e-05,
1670
+ "loss": 0.8232,
1671
+ "step": 138500
1672
+ },
1673
+ {
1674
+ "epoch": 10.37,
1675
+ "learning_rate": 2.9277255779269202e-05,
1676
+ "loss": 0.8242,
1677
+ "step": 139000
1678
+ },
1679
+ {
1680
+ "epoch": 10.4,
1681
+ "learning_rate": 2.9202833706189415e-05,
1682
+ "loss": 0.8229,
1683
+ "step": 139500
1684
+ },
1685
+ {
1686
+ "epoch": 10.44,
1687
+ "learning_rate": 2.91282624906786e-05,
1688
+ "loss": 0.8172,
1689
+ "step": 140000
1690
+ },
1691
+ {
1692
+ "epoch": 10.48,
1693
+ "learning_rate": 2.905369127516779e-05,
1694
+ "loss": 0.823,
1695
+ "step": 140500
1696
+ },
1697
+ {
1698
+ "epoch": 10.51,
1699
+ "learning_rate": 2.897912005965697e-05,
1700
+ "loss": 0.8224,
1701
+ "step": 141000
1702
+ },
1703
+ {
1704
+ "epoch": 10.55,
1705
+ "learning_rate": 2.8904548844146163e-05,
1706
+ "loss": 0.8265,
1707
+ "step": 141500
1708
+ },
1709
+ {
1710
+ "epoch": 10.59,
1711
+ "learning_rate": 2.8829977628635345e-05,
1712
+ "loss": 0.8191,
1713
+ "step": 142000
1714
+ },
1715
+ {
1716
+ "epoch": 10.63,
1717
+ "learning_rate": 2.8755406413124537e-05,
1718
+ "loss": 0.8236,
1719
+ "step": 142500
1720
+ },
1721
+ {
1722
+ "epoch": 10.66,
1723
+ "learning_rate": 2.868083519761372e-05,
1724
+ "loss": 0.8158,
1725
+ "step": 143000
1726
+ },
1727
+ {
1728
+ "epoch": 10.7,
1729
+ "learning_rate": 2.860626398210291e-05,
1730
+ "loss": 0.8221,
1731
+ "step": 143500
1732
+ },
1733
+ {
1734
+ "epoch": 10.74,
1735
+ "learning_rate": 2.8531841909023117e-05,
1736
+ "loss": 0.8212,
1737
+ "step": 144000
1738
+ },
1739
+ {
1740
+ "epoch": 10.78,
1741
+ "learning_rate": 2.8457419835943326e-05,
1742
+ "loss": 0.825,
1743
+ "step": 144500
1744
+ },
1745
+ {
1746
+ "epoch": 10.81,
1747
+ "learning_rate": 2.838284862043251e-05,
1748
+ "loss": 0.8255,
1749
+ "step": 145000
1750
+ },
1751
+ {
1752
+ "epoch": 10.85,
1753
+ "learning_rate": 2.83082774049217e-05,
1754
+ "loss": 0.8139,
1755
+ "step": 145500
1756
+ },
1757
+ {
1758
+ "epoch": 10.89,
1759
+ "learning_rate": 2.8233706189410892e-05,
1760
+ "loss": 0.8157,
1761
+ "step": 146000
1762
+ },
1763
+ {
1764
+ "epoch": 10.92,
1765
+ "learning_rate": 2.8159134973900074e-05,
1766
+ "loss": 0.8177,
1767
+ "step": 146500
1768
+ },
1769
+ {
1770
+ "epoch": 10.96,
1771
+ "learning_rate": 2.8084563758389266e-05,
1772
+ "loss": 0.8097,
1773
+ "step": 147000
1774
+ },
1775
+ {
1776
+ "epoch": 11.0,
1777
+ "learning_rate": 2.8009992542878448e-05,
1778
+ "loss": 0.8206,
1779
+ "step": 147500
1780
+ },
1781
+ {
1782
+ "epoch": 11.04,
1783
+ "learning_rate": 2.793557046979866e-05,
1784
+ "loss": 0.8084,
1785
+ "step": 148000
1786
+ },
1787
+ {
1788
+ "epoch": 11.07,
1789
+ "learning_rate": 2.7860999254287846e-05,
1790
+ "loss": 0.8082,
1791
+ "step": 148500
1792
+ },
1793
+ {
1794
+ "epoch": 11.11,
1795
+ "learning_rate": 2.7786428038777035e-05,
1796
+ "loss": 0.8034,
1797
+ "step": 149000
1798
+ },
1799
+ {
1800
+ "epoch": 11.15,
1801
+ "learning_rate": 2.771185682326622e-05,
1802
+ "loss": 0.8013,
1803
+ "step": 149500
1804
+ },
1805
+ {
1806
+ "epoch": 11.19,
1807
+ "learning_rate": 2.763728560775541e-05,
1808
+ "loss": 0.8098,
1809
+ "step": 150000
1810
+ },
1811
+ {
1812
+ "epoch": 11.22,
1813
+ "learning_rate": 2.7562714392244594e-05,
1814
+ "loss": 0.8062,
1815
+ "step": 150500
1816
+ },
1817
+ {
1818
+ "epoch": 11.26,
1819
+ "learning_rate": 2.7488143176733783e-05,
1820
+ "loss": 0.8112,
1821
+ "step": 151000
1822
+ },
1823
+ {
1824
+ "epoch": 11.3,
1825
+ "learning_rate": 2.741372110365399e-05,
1826
+ "loss": 0.8068,
1827
+ "step": 151500
1828
+ },
1829
+ {
1830
+ "epoch": 11.33,
1831
+ "learning_rate": 2.7339149888143178e-05,
1832
+ "loss": 0.8069,
1833
+ "step": 152000
1834
+ },
1835
+ {
1836
+ "epoch": 11.37,
1837
+ "learning_rate": 2.7264578672632363e-05,
1838
+ "loss": 0.8089,
1839
+ "step": 152500
1840
+ },
1841
+ {
1842
+ "epoch": 11.41,
1843
+ "learning_rate": 2.7190156599552575e-05,
1844
+ "loss": 0.8094,
1845
+ "step": 153000
1846
+ },
1847
+ {
1848
+ "epoch": 11.45,
1849
+ "learning_rate": 2.7115585384041757e-05,
1850
+ "loss": 0.8012,
1851
+ "step": 153500
1852
+ },
1853
+ {
1854
+ "epoch": 11.48,
1855
+ "learning_rate": 2.704101416853095e-05,
1856
+ "loss": 0.8105,
1857
+ "step": 154000
1858
+ },
1859
+ {
1860
+ "epoch": 11.52,
1861
+ "learning_rate": 2.696644295302014e-05,
1862
+ "loss": 0.8066,
1863
+ "step": 154500
1864
+ },
1865
+ {
1866
+ "epoch": 11.56,
1867
+ "learning_rate": 2.6891871737509324e-05,
1868
+ "loss": 0.8094,
1869
+ "step": 155000
1870
+ },
1871
+ {
1872
+ "epoch": 11.6,
1873
+ "learning_rate": 2.6817300521998512e-05,
1874
+ "loss": 0.809,
1875
+ "step": 155500
1876
+ },
1877
+ {
1878
+ "epoch": 11.63,
1879
+ "learning_rate": 2.6742729306487694e-05,
1880
+ "loss": 0.7998,
1881
+ "step": 156000
1882
+ },
1883
+ {
1884
+ "epoch": 11.67,
1885
+ "learning_rate": 2.6668158090976887e-05,
1886
+ "loss": 0.8112,
1887
+ "step": 156500
1888
+ },
1889
+ {
1890
+ "epoch": 11.71,
1891
+ "learning_rate": 2.6593736017897092e-05,
1892
+ "loss": 0.8068,
1893
+ "step": 157000
1894
+ },
1895
+ {
1896
+ "epoch": 11.74,
1897
+ "learning_rate": 2.651916480238628e-05,
1898
+ "loss": 0.8042,
1899
+ "step": 157500
1900
+ },
1901
+ {
1902
+ "epoch": 11.78,
1903
+ "learning_rate": 2.6444593586875466e-05,
1904
+ "loss": 0.7972,
1905
+ "step": 158000
1906
+ },
1907
+ {
1908
+ "epoch": 11.82,
1909
+ "learning_rate": 2.6370022371364655e-05,
1910
+ "loss": 0.8119,
1911
+ "step": 158500
1912
+ },
1913
+ {
1914
+ "epoch": 11.86,
1915
+ "learning_rate": 2.629545115585384e-05,
1916
+ "loss": 0.7982,
1917
+ "step": 159000
1918
+ },
1919
+ {
1920
+ "epoch": 11.89,
1921
+ "learning_rate": 2.622102908277405e-05,
1922
+ "loss": 0.7973,
1923
+ "step": 159500
1924
+ },
1925
+ {
1926
+ "epoch": 11.93,
1927
+ "learning_rate": 2.6146457867263235e-05,
1928
+ "loss": 0.7919,
1929
+ "step": 160000
1930
+ },
1931
+ {
1932
+ "epoch": 11.97,
1933
+ "learning_rate": 2.6071886651752424e-05,
1934
+ "loss": 0.8013,
1935
+ "step": 160500
1936
+ },
1937
+ {
1938
+ "epoch": 12.01,
1939
+ "learning_rate": 2.599731543624161e-05,
1940
+ "loss": 0.7986,
1941
+ "step": 161000
1942
+ },
1943
+ {
1944
+ "epoch": 12.04,
1945
+ "learning_rate": 2.592289336316182e-05,
1946
+ "loss": 0.7927,
1947
+ "step": 161500
1948
+ },
1949
+ {
1950
+ "epoch": 12.08,
1951
+ "learning_rate": 2.584832214765101e-05,
1952
+ "loss": 0.7983,
1953
+ "step": 162000
1954
+ },
1955
+ {
1956
+ "epoch": 12.12,
1957
+ "learning_rate": 2.5773750932140196e-05,
1958
+ "loss": 0.7913,
1959
+ "step": 162500
1960
+ },
1961
+ {
1962
+ "epoch": 12.15,
1963
+ "learning_rate": 2.5699179716629384e-05,
1964
+ "loss": 0.7979,
1965
+ "step": 163000
1966
+ },
1967
+ {
1968
+ "epoch": 12.19,
1969
+ "learning_rate": 2.562460850111857e-05,
1970
+ "loss": 0.7941,
1971
+ "step": 163500
1972
+ },
1973
+ {
1974
+ "epoch": 12.23,
1975
+ "learning_rate": 2.555003728560776e-05,
1976
+ "loss": 0.7871,
1977
+ "step": 164000
1978
+ },
1979
+ {
1980
+ "epoch": 12.27,
1981
+ "learning_rate": 2.5475466070096944e-05,
1982
+ "loss": 0.7927,
1983
+ "step": 164500
1984
+ },
1985
+ {
1986
+ "epoch": 12.3,
1987
+ "learning_rate": 2.5400894854586133e-05,
1988
+ "loss": 0.7908,
1989
+ "step": 165000
1990
+ },
1991
+ {
1992
+ "epoch": 12.34,
1993
+ "learning_rate": 2.532647278150634e-05,
1994
+ "loss": 0.7872,
1995
+ "step": 165500
1996
+ },
1997
+ {
1998
+ "epoch": 12.38,
1999
+ "learning_rate": 2.5251901565995527e-05,
2000
+ "loss": 0.7883,
2001
+ "step": 166000
2002
+ },
2003
+ {
2004
+ "epoch": 12.42,
2005
+ "learning_rate": 2.5177330350484713e-05,
2006
+ "loss": 0.785,
2007
+ "step": 166500
2008
+ },
2009
+ {
2010
+ "epoch": 12.45,
2011
+ "learning_rate": 2.51027591349739e-05,
2012
+ "loss": 0.7894,
2013
+ "step": 167000
2014
+ },
2015
+ {
2016
+ "epoch": 12.49,
2017
+ "learning_rate": 2.5028337061894107e-05,
2018
+ "loss": 0.7822,
2019
+ "step": 167500
2020
+ },
2021
+ {
2022
+ "epoch": 12.53,
2023
+ "learning_rate": 2.49537658463833e-05,
2024
+ "loss": 0.7952,
2025
+ "step": 168000
2026
+ },
2027
+ {
2028
+ "epoch": 12.57,
2029
+ "learning_rate": 2.4879194630872485e-05,
2030
+ "loss": 0.787,
2031
+ "step": 168500
2032
+ },
2033
+ {
2034
+ "epoch": 12.6,
2035
+ "learning_rate": 2.4804623415361673e-05,
2036
+ "loss": 0.7909,
2037
+ "step": 169000
2038
+ },
2039
+ {
2040
+ "epoch": 12.64,
2041
+ "learning_rate": 2.473005219985086e-05,
2042
+ "loss": 0.7913,
2043
+ "step": 169500
2044
+ },
2045
+ {
2046
+ "epoch": 12.68,
2047
+ "learning_rate": 2.4655480984340047e-05,
2048
+ "loss": 0.7831,
2049
+ "step": 170000
2050
+ },
2051
+ {
2052
+ "epoch": 12.71,
2053
+ "learning_rate": 2.4580909768829233e-05,
2054
+ "loss": 0.7899,
2055
+ "step": 170500
2056
+ },
2057
+ {
2058
+ "epoch": 12.75,
2059
+ "learning_rate": 2.4506338553318418e-05,
2060
+ "loss": 0.79,
2061
+ "step": 171000
2062
+ },
2063
+ {
2064
+ "epoch": 12.79,
2065
+ "learning_rate": 2.4431767337807607e-05,
2066
+ "loss": 0.7897,
2067
+ "step": 171500
2068
+ },
2069
+ {
2070
+ "epoch": 12.83,
2071
+ "learning_rate": 2.4357345264727816e-05,
2072
+ "loss": 0.7813,
2073
+ "step": 172000
2074
+ },
2075
+ {
2076
+ "epoch": 12.86,
2077
+ "learning_rate": 2.4282774049217e-05,
2078
+ "loss": 0.7852,
2079
+ "step": 172500
2080
+ },
2081
+ {
2082
+ "epoch": 12.9,
2083
+ "learning_rate": 2.420820283370619e-05,
2084
+ "loss": 0.7856,
2085
+ "step": 173000
2086
+ },
2087
+ {
2088
+ "epoch": 12.94,
2089
+ "learning_rate": 2.4133631618195376e-05,
2090
+ "loss": 0.7859,
2091
+ "step": 173500
2092
+ },
2093
+ {
2094
+ "epoch": 12.98,
2095
+ "learning_rate": 2.4059209545115585e-05,
2096
+ "loss": 0.7854,
2097
+ "step": 174000
2098
+ },
2099
+ {
2100
+ "epoch": 13.01,
2101
+ "learning_rate": 2.3984787472035794e-05,
2102
+ "loss": 0.7768,
2103
+ "step": 174500
2104
+ },
2105
+ {
2106
+ "epoch": 13.05,
2107
+ "learning_rate": 2.3910216256524982e-05,
2108
+ "loss": 0.7776,
2109
+ "step": 175000
2110
+ },
2111
+ {
2112
+ "epoch": 13.09,
2113
+ "learning_rate": 2.383564504101417e-05,
2114
+ "loss": 0.7801,
2115
+ "step": 175500
2116
+ },
2117
+ {
2118
+ "epoch": 13.12,
2119
+ "learning_rate": 2.3761073825503357e-05,
2120
+ "loss": 0.7801,
2121
+ "step": 176000
2122
+ },
2123
+ {
2124
+ "epoch": 13.16,
2125
+ "learning_rate": 2.3686502609992545e-05,
2126
+ "loss": 0.7767,
2127
+ "step": 176500
2128
+ },
2129
+ {
2130
+ "epoch": 13.2,
2131
+ "learning_rate": 2.3612080536912754e-05,
2132
+ "loss": 0.7765,
2133
+ "step": 177000
2134
+ },
2135
+ {
2136
+ "epoch": 13.24,
2137
+ "learning_rate": 2.353750932140194e-05,
2138
+ "loss": 0.7761,
2139
+ "step": 177500
2140
+ },
2141
+ {
2142
+ "epoch": 13.27,
2143
+ "learning_rate": 2.346293810589113e-05,
2144
+ "loss": 0.7767,
2145
+ "step": 178000
2146
+ },
2147
+ {
2148
+ "epoch": 13.31,
2149
+ "learning_rate": 2.3388366890380314e-05,
2150
+ "loss": 0.7699,
2151
+ "step": 178500
2152
+ },
2153
+ {
2154
+ "epoch": 13.35,
2155
+ "learning_rate": 2.3313795674869503e-05,
2156
+ "loss": 0.7793,
2157
+ "step": 179000
2158
+ },
2159
+ {
2160
+ "epoch": 13.39,
2161
+ "learning_rate": 2.3239373601789712e-05,
2162
+ "loss": 0.7755,
2163
+ "step": 179500
2164
+ },
2165
+ {
2166
+ "epoch": 13.42,
2167
+ "learning_rate": 2.3164802386278897e-05,
2168
+ "loss": 0.767,
2169
+ "step": 180000
2170
+ },
2171
+ {
2172
+ "epoch": 13.46,
2173
+ "learning_rate": 2.3090231170768082e-05,
2174
+ "loss": 0.7779,
2175
+ "step": 180500
2176
+ },
2177
+ {
2178
+ "epoch": 13.5,
2179
+ "learning_rate": 2.301565995525727e-05,
2180
+ "loss": 0.7851,
2181
+ "step": 181000
2182
+ },
2183
+ {
2184
+ "epoch": 13.53,
2185
+ "learning_rate": 2.2941088739746457e-05,
2186
+ "loss": 0.7793,
2187
+ "step": 181500
2188
+ },
2189
+ {
2190
+ "epoch": 13.57,
2191
+ "learning_rate": 2.2866517524235645e-05,
2192
+ "loss": 0.7716,
2193
+ "step": 182000
2194
+ },
2195
+ {
2196
+ "epoch": 13.61,
2197
+ "learning_rate": 2.279194630872483e-05,
2198
+ "loss": 0.7818,
2199
+ "step": 182500
2200
+ },
2201
+ {
2202
+ "epoch": 13.65,
2203
+ "learning_rate": 2.2717375093214023e-05,
2204
+ "loss": 0.7779,
2205
+ "step": 183000
2206
+ },
2207
+ {
2208
+ "epoch": 13.68,
2209
+ "learning_rate": 2.2642953020134232e-05,
2210
+ "loss": 0.7739,
2211
+ "step": 183500
2212
+ },
2213
+ {
2214
+ "epoch": 13.72,
2215
+ "learning_rate": 2.2568381804623417e-05,
2216
+ "loss": 0.766,
2217
+ "step": 184000
2218
+ },
2219
+ {
2220
+ "epoch": 13.76,
2221
+ "learning_rate": 2.2493810589112603e-05,
2222
+ "loss": 0.7742,
2223
+ "step": 184500
2224
+ },
2225
+ {
2226
+ "epoch": 13.8,
2227
+ "learning_rate": 2.2419388516032812e-05,
2228
+ "loss": 0.767,
2229
+ "step": 185000
2230
+ },
2231
+ {
2232
+ "epoch": 13.83,
2233
+ "learning_rate": 2.2344817300522e-05,
2234
+ "loss": 0.7672,
2235
+ "step": 185500
2236
+ },
2237
+ {
2238
+ "epoch": 13.87,
2239
+ "learning_rate": 2.2270246085011186e-05,
2240
+ "loss": 0.7774,
2241
+ "step": 186000
2242
+ },
2243
+ {
2244
+ "epoch": 13.91,
2245
+ "learning_rate": 2.2195674869500375e-05,
2246
+ "loss": 0.7696,
2247
+ "step": 186500
2248
+ },
2249
+ {
2250
+ "epoch": 13.94,
2251
+ "learning_rate": 2.2121252796420584e-05,
2252
+ "loss": 0.7736,
2253
+ "step": 187000
2254
+ },
2255
+ {
2256
+ "epoch": 13.98,
2257
+ "learning_rate": 2.204668158090977e-05,
2258
+ "loss": 0.7705,
2259
+ "step": 187500
2260
+ },
2261
+ {
2262
+ "epoch": 14.02,
2263
+ "learning_rate": 2.1972110365398958e-05,
2264
+ "loss": 0.7768,
2265
+ "step": 188000
2266
+ },
2267
+ {
2268
+ "epoch": 14.06,
2269
+ "learning_rate": 2.1897539149888143e-05,
2270
+ "loss": 0.7629,
2271
+ "step": 188500
2272
+ },
2273
+ {
2274
+ "epoch": 14.09,
2275
+ "learning_rate": 2.1822967934377332e-05,
2276
+ "loss": 0.758,
2277
+ "step": 189000
2278
+ },
2279
+ {
2280
+ "epoch": 14.13,
2281
+ "learning_rate": 2.174854586129754e-05,
2282
+ "loss": 0.7726,
2283
+ "step": 189500
2284
+ },
2285
+ {
2286
+ "epoch": 14.17,
2287
+ "learning_rate": 2.1673974645786726e-05,
2288
+ "loss": 0.765,
2289
+ "step": 190000
2290
+ },
2291
+ {
2292
+ "epoch": 14.21,
2293
+ "learning_rate": 2.1599403430275912e-05,
2294
+ "loss": 0.7639,
2295
+ "step": 190500
2296
+ },
2297
+ {
2298
+ "epoch": 14.24,
2299
+ "learning_rate": 2.15248322147651e-05,
2300
+ "loss": 0.7613,
2301
+ "step": 191000
2302
+ },
2303
+ {
2304
+ "epoch": 14.28,
2305
+ "learning_rate": 2.145026099925429e-05,
2306
+ "loss": 0.7624,
2307
+ "step": 191500
2308
+ },
2309
+ {
2310
+ "epoch": 14.32,
2311
+ "learning_rate": 2.13758389261745e-05,
2312
+ "loss": 0.7682,
2313
+ "step": 192000
2314
+ },
2315
+ {
2316
+ "epoch": 14.35,
2317
+ "learning_rate": 2.1301267710663687e-05,
2318
+ "loss": 0.7683,
2319
+ "step": 192500
2320
+ },
2321
+ {
2322
+ "epoch": 14.39,
2323
+ "learning_rate": 2.1226696495152873e-05,
2324
+ "loss": 0.7684,
2325
+ "step": 193000
2326
+ },
2327
+ {
2328
+ "epoch": 14.43,
2329
+ "learning_rate": 2.115212527964206e-05,
2330
+ "loss": 0.7561,
2331
+ "step": 193500
2332
+ },
2333
+ {
2334
+ "epoch": 14.47,
2335
+ "learning_rate": 2.1077554064131247e-05,
2336
+ "loss": 0.7672,
2337
+ "step": 194000
2338
+ },
2339
+ {
2340
+ "epoch": 14.5,
2341
+ "learning_rate": 2.1002982848620435e-05,
2342
+ "loss": 0.7658,
2343
+ "step": 194500
2344
+ },
2345
+ {
2346
+ "epoch": 14.54,
2347
+ "learning_rate": 2.092841163310962e-05,
2348
+ "loss": 0.7625,
2349
+ "step": 195000
2350
+ },
2351
+ {
2352
+ "epoch": 14.58,
2353
+ "learning_rate": 2.0853840417598806e-05,
2354
+ "loss": 0.757,
2355
+ "step": 195500
2356
+ },
2357
+ {
2358
+ "epoch": 14.62,
2359
+ "learning_rate": 2.0779269202087995e-05,
2360
+ "loss": 0.7596,
2361
+ "step": 196000
2362
+ },
2363
+ {
2364
+ "epoch": 14.65,
2365
+ "learning_rate": 2.0704847129008204e-05,
2366
+ "loss": 0.7603,
2367
+ "step": 196500
2368
+ },
2369
+ {
2370
+ "epoch": 14.69,
2371
+ "learning_rate": 2.0630425055928413e-05,
2372
+ "loss": 0.7666,
2373
+ "step": 197000
2374
+ },
2375
+ {
2376
+ "epoch": 14.73,
2377
+ "learning_rate": 2.05558538404176e-05,
2378
+ "loss": 0.7602,
2379
+ "step": 197500
2380
+ },
2381
+ {
2382
+ "epoch": 14.76,
2383
+ "learning_rate": 2.0481282624906787e-05,
2384
+ "loss": 0.7575,
2385
+ "step": 198000
2386
+ },
2387
+ {
2388
+ "epoch": 14.8,
2389
+ "learning_rate": 2.0406711409395973e-05,
2390
+ "loss": 0.7615,
2391
+ "step": 198500
2392
+ },
2393
+ {
2394
+ "epoch": 14.84,
2395
+ "learning_rate": 2.033214019388516e-05,
2396
+ "loss": 0.761,
2397
+ "step": 199000
2398
+ },
2399
+ {
2400
+ "epoch": 14.88,
2401
+ "learning_rate": 2.025771812080537e-05,
2402
+ "loss": 0.7628,
2403
+ "step": 199500
2404
+ },
2405
+ {
2406
+ "epoch": 14.91,
2407
+ "learning_rate": 2.0183296047725576e-05,
2408
+ "loss": 0.7629,
2409
+ "step": 200000
2410
+ },
2411
+ {
2412
+ "epoch": 14.95,
2413
+ "learning_rate": 2.0108724832214765e-05,
2414
+ "loss": 0.7623,
2415
+ "step": 200500
2416
+ },
2417
+ {
2418
+ "epoch": 14.99,
2419
+ "learning_rate": 2.0034153616703954e-05,
2420
+ "loss": 0.7528,
2421
+ "step": 201000
2422
+ },
2423
+ {
2424
+ "epoch": 15.03,
2425
+ "learning_rate": 1.9959582401193142e-05,
2426
+ "loss": 0.7458,
2427
+ "step": 201500
2428
+ },
2429
+ {
2430
+ "epoch": 15.06,
2431
+ "learning_rate": 1.9885011185682328e-05,
2432
+ "loss": 0.7596,
2433
+ "step": 202000
2434
+ },
2435
+ {
2436
+ "epoch": 15.1,
2437
+ "learning_rate": 1.9810439970171517e-05,
2438
+ "loss": 0.7547,
2439
+ "step": 202500
2440
+ },
2441
+ {
2442
+ "epoch": 15.14,
2443
+ "learning_rate": 1.9735868754660702e-05,
2444
+ "loss": 0.7581,
2445
+ "step": 203000
2446
+ },
2447
+ {
2448
+ "epoch": 15.18,
2449
+ "learning_rate": 1.966129753914989e-05,
2450
+ "loss": 0.7527,
2451
+ "step": 203500
2452
+ },
2453
+ {
2454
+ "epoch": 15.21,
2455
+ "learning_rate": 1.9586726323639076e-05,
2456
+ "loss": 0.7592,
2457
+ "step": 204000
2458
+ },
2459
+ {
2460
+ "epoch": 15.25,
2461
+ "learning_rate": 1.9512304250559285e-05,
2462
+ "loss": 0.7538,
2463
+ "step": 204500
2464
+ },
2465
+ {
2466
+ "epoch": 15.29,
2467
+ "learning_rate": 1.943773303504847e-05,
2468
+ "loss": 0.7466,
2469
+ "step": 205000
2470
+ },
2471
+ {
2472
+ "epoch": 15.32,
2473
+ "learning_rate": 1.936316181953766e-05,
2474
+ "loss": 0.7529,
2475
+ "step": 205500
2476
+ },
2477
+ {
2478
+ "epoch": 15.36,
2479
+ "learning_rate": 1.9288590604026845e-05,
2480
+ "loss": 0.7533,
2481
+ "step": 206000
2482
+ },
2483
+ {
2484
+ "epoch": 15.4,
2485
+ "learning_rate": 1.9214019388516033e-05,
2486
+ "loss": 0.7532,
2487
+ "step": 206500
2488
+ },
2489
+ {
2490
+ "epoch": 15.44,
2491
+ "learning_rate": 1.9139597315436243e-05,
2492
+ "loss": 0.7627,
2493
+ "step": 207000
2494
+ },
2495
+ {
2496
+ "epoch": 15.47,
2497
+ "learning_rate": 1.906517524235645e-05,
2498
+ "loss": 0.7514,
2499
+ "step": 207500
2500
+ },
2501
+ {
2502
+ "epoch": 15.51,
2503
+ "learning_rate": 1.8990604026845637e-05,
2504
+ "loss": 0.7564,
2505
+ "step": 208000
2506
+ },
2507
+ {
2508
+ "epoch": 15.55,
2509
+ "learning_rate": 1.8916032811334826e-05,
2510
+ "loss": 0.7587,
2511
+ "step": 208500
2512
+ },
2513
+ {
2514
+ "epoch": 15.59,
2515
+ "learning_rate": 1.884146159582401e-05,
2516
+ "loss": 0.7508,
2517
+ "step": 209000
2518
+ },
2519
+ {
2520
+ "epoch": 15.62,
2521
+ "learning_rate": 1.87668903803132e-05,
2522
+ "loss": 0.7443,
2523
+ "step": 209500
2524
+ },
2525
+ {
2526
+ "epoch": 15.66,
2527
+ "learning_rate": 1.869231916480239e-05,
2528
+ "loss": 0.7514,
2529
+ "step": 210000
2530
+ },
2531
+ {
2532
+ "epoch": 15.7,
2533
+ "learning_rate": 1.8617747949291574e-05,
2534
+ "loss": 0.7461,
2535
+ "step": 210500
2536
+ },
2537
+ {
2538
+ "epoch": 15.73,
2539
+ "learning_rate": 1.8543176733780763e-05,
2540
+ "loss": 0.7537,
2541
+ "step": 211000
2542
+ },
2543
+ {
2544
+ "epoch": 15.77,
2545
+ "learning_rate": 1.8468605518269948e-05,
2546
+ "loss": 0.7532,
2547
+ "step": 211500
2548
+ },
2549
+ {
2550
+ "epoch": 15.81,
2551
+ "learning_rate": 1.8394034302759137e-05,
2552
+ "loss": 0.7416,
2553
+ "step": 212000
2554
+ },
2555
+ {
2556
+ "epoch": 15.85,
2557
+ "learning_rate": 1.8319463087248322e-05,
2558
+ "loss": 0.7456,
2559
+ "step": 212500
2560
+ },
2561
+ {
2562
+ "epoch": 15.88,
2563
+ "learning_rate": 1.824489187173751e-05,
2564
+ "loss": 0.7477,
2565
+ "step": 213000
2566
+ },
2567
+ {
2568
+ "epoch": 15.92,
2569
+ "learning_rate": 1.8170320656226696e-05,
2570
+ "loss": 0.749,
2571
+ "step": 213500
2572
+ },
2573
+ {
2574
+ "epoch": 15.96,
2575
+ "learning_rate": 1.8095749440715885e-05,
2576
+ "loss": 0.7512,
2577
+ "step": 214000
2578
+ },
2579
+ {
2580
+ "epoch": 16.0,
2581
+ "learning_rate": 1.8021327367636094e-05,
2582
+ "loss": 0.7422,
2583
+ "step": 214500
2584
+ },
2585
+ {
2586
+ "epoch": 16.03,
2587
+ "learning_rate": 1.794675615212528e-05,
2588
+ "loss": 0.749,
2589
+ "step": 215000
2590
+ },
2591
+ {
2592
+ "epoch": 16.07,
2593
+ "learning_rate": 1.787218493661447e-05,
2594
+ "loss": 0.7409,
2595
+ "step": 215500
2596
+ },
2597
+ {
2598
+ "epoch": 16.11,
2599
+ "learning_rate": 1.7797613721103654e-05,
2600
+ "loss": 0.7411,
2601
+ "step": 216000
2602
+ },
2603
+ {
2604
+ "epoch": 16.14,
2605
+ "learning_rate": 1.7723191648023863e-05,
2606
+ "loss": 0.7397,
2607
+ "step": 216500
2608
+ },
2609
+ {
2610
+ "epoch": 16.18,
2611
+ "learning_rate": 1.764862043251305e-05,
2612
+ "loss": 0.7448,
2613
+ "step": 217000
2614
+ },
2615
+ {
2616
+ "epoch": 16.22,
2617
+ "learning_rate": 1.757404921700224e-05,
2618
+ "loss": 0.7482,
2619
+ "step": 217500
2620
+ },
2621
+ {
2622
+ "epoch": 16.26,
2623
+ "learning_rate": 1.7499478001491426e-05,
2624
+ "loss": 0.7399,
2625
+ "step": 218000
2626
+ },
2627
+ {
2628
+ "epoch": 16.29,
2629
+ "learning_rate": 1.7424906785980614e-05,
2630
+ "loss": 0.7347,
2631
+ "step": 218500
2632
+ },
2633
+ {
2634
+ "epoch": 16.33,
2635
+ "learning_rate": 1.7350633855331844e-05,
2636
+ "loss": 0.7431,
2637
+ "step": 219000
2638
+ },
2639
+ {
2640
+ "epoch": 16.37,
2641
+ "learning_rate": 1.727606263982103e-05,
2642
+ "loss": 0.7426,
2643
+ "step": 219500
2644
+ },
2645
+ {
2646
+ "epoch": 16.41,
2647
+ "learning_rate": 1.7201491424310218e-05,
2648
+ "loss": 0.7437,
2649
+ "step": 220000
2650
+ },
2651
+ {
2652
+ "epoch": 16.44,
2653
+ "learning_rate": 1.7126920208799403e-05,
2654
+ "loss": 0.7407,
2655
+ "step": 220500
2656
+ },
2657
+ {
2658
+ "epoch": 16.48,
2659
+ "learning_rate": 1.7052348993288592e-05,
2660
+ "loss": 0.7389,
2661
+ "step": 221000
2662
+ },
2663
+ {
2664
+ "epoch": 16.52,
2665
+ "learning_rate": 1.6977777777777777e-05,
2666
+ "loss": 0.7369,
2667
+ "step": 221500
2668
+ },
2669
+ {
2670
+ "epoch": 16.55,
2671
+ "learning_rate": 1.6903206562266966e-05,
2672
+ "loss": 0.7444,
2673
+ "step": 222000
2674
+ },
2675
+ {
2676
+ "epoch": 16.59,
2677
+ "learning_rate": 1.682863534675615e-05,
2678
+ "loss": 0.7385,
2679
+ "step": 222500
2680
+ },
2681
+ {
2682
+ "epoch": 16.63,
2683
+ "learning_rate": 1.675421327367636e-05,
2684
+ "loss": 0.7439,
2685
+ "step": 223000
2686
+ },
2687
+ {
2688
+ "epoch": 16.67,
2689
+ "learning_rate": 1.667964205816555e-05,
2690
+ "loss": 0.743,
2691
+ "step": 223500
2692
+ },
2693
+ {
2694
+ "epoch": 16.7,
2695
+ "learning_rate": 1.6605070842654735e-05,
2696
+ "loss": 0.7422,
2697
+ "step": 224000
2698
+ },
2699
+ {
2700
+ "epoch": 16.74,
2701
+ "learning_rate": 1.6530499627143924e-05,
2702
+ "loss": 0.7384,
2703
+ "step": 224500
2704
+ },
2705
+ {
2706
+ "epoch": 16.78,
2707
+ "learning_rate": 1.645592841163311e-05,
2708
+ "loss": 0.7388,
2709
+ "step": 225000
2710
+ },
2711
+ {
2712
+ "epoch": 16.82,
2713
+ "learning_rate": 1.6381357196122298e-05,
2714
+ "loss": 0.7384,
2715
+ "step": 225500
2716
+ },
2717
+ {
2718
+ "epoch": 16.85,
2719
+ "learning_rate": 1.6306785980611486e-05,
2720
+ "loss": 0.7387,
2721
+ "step": 226000
2722
+ },
2723
+ {
2724
+ "epoch": 16.89,
2725
+ "learning_rate": 1.6232363907531696e-05,
2726
+ "loss": 0.7306,
2727
+ "step": 226500
2728
+ },
2729
+ {
2730
+ "epoch": 16.93,
2731
+ "learning_rate": 1.615779269202088e-05,
2732
+ "loss": 0.7398,
2733
+ "step": 227000
2734
+ },
2735
+ {
2736
+ "epoch": 16.96,
2737
+ "learning_rate": 1.608322147651007e-05,
2738
+ "loss": 0.7371,
2739
+ "step": 227500
2740
+ },
2741
+ {
2742
+ "epoch": 17.0,
2743
+ "learning_rate": 1.6008650260999255e-05,
2744
+ "loss": 0.7357,
2745
+ "step": 228000
2746
+ },
2747
+ {
2748
+ "epoch": 17.04,
2749
+ "learning_rate": 1.5934377330350488e-05,
2750
+ "loss": 0.7357,
2751
+ "step": 228500
2752
+ },
2753
+ {
2754
+ "epoch": 17.08,
2755
+ "learning_rate": 1.5859806114839673e-05,
2756
+ "loss": 0.7397,
2757
+ "step": 229000
2758
+ },
2759
+ {
2760
+ "epoch": 17.11,
2761
+ "learning_rate": 1.578523489932886e-05,
2762
+ "loss": 0.735,
2763
+ "step": 229500
2764
+ },
2765
+ {
2766
+ "epoch": 17.15,
2767
+ "learning_rate": 1.5710663683818047e-05,
2768
+ "loss": 0.7345,
2769
+ "step": 230000
2770
+ },
2771
+ {
2772
+ "epoch": 17.19,
2773
+ "learning_rate": 1.5636092468307233e-05,
2774
+ "loss": 0.7339,
2775
+ "step": 230500
2776
+ },
2777
+ {
2778
+ "epoch": 17.23,
2779
+ "learning_rate": 1.556152125279642e-05,
2780
+ "loss": 0.7317,
2781
+ "step": 231000
2782
+ },
2783
+ {
2784
+ "epoch": 17.26,
2785
+ "learning_rate": 1.5486950037285607e-05,
2786
+ "loss": 0.7325,
2787
+ "step": 231500
2788
+ },
2789
+ {
2790
+ "epoch": 17.3,
2791
+ "learning_rate": 1.5412378821774796e-05,
2792
+ "loss": 0.7328,
2793
+ "step": 232000
2794
+ },
2795
+ {
2796
+ "epoch": 17.34,
2797
+ "learning_rate": 1.5337956748695005e-05,
2798
+ "loss": 0.7326,
2799
+ "step": 232500
2800
+ },
2801
+ {
2802
+ "epoch": 17.37,
2803
+ "learning_rate": 1.526338553318419e-05,
2804
+ "loss": 0.734,
2805
+ "step": 233000
2806
+ },
2807
+ {
2808
+ "epoch": 17.41,
2809
+ "learning_rate": 1.5188814317673377e-05,
2810
+ "loss": 0.7369,
2811
+ "step": 233500
2812
+ },
2813
+ {
2814
+ "epoch": 17.45,
2815
+ "learning_rate": 1.5114243102162564e-05,
2816
+ "loss": 0.7334,
2817
+ "step": 234000
2818
+ },
2819
+ {
2820
+ "epoch": 17.49,
2821
+ "learning_rate": 1.5039671886651755e-05,
2822
+ "loss": 0.73,
2823
+ "step": 234500
2824
+ },
2825
+ {
2826
+ "epoch": 17.52,
2827
+ "learning_rate": 1.4965100671140942e-05,
2828
+ "loss": 0.7294,
2829
+ "step": 235000
2830
+ },
2831
+ {
2832
+ "epoch": 17.56,
2833
+ "learning_rate": 1.4890529455630129e-05,
2834
+ "loss": 0.7238,
2835
+ "step": 235500
2836
+ },
2837
+ {
2838
+ "epoch": 17.6,
2839
+ "learning_rate": 1.4815958240119316e-05,
2840
+ "loss": 0.7297,
2841
+ "step": 236000
2842
+ },
2843
+ {
2844
+ "epoch": 17.64,
2845
+ "learning_rate": 1.4741536167039525e-05,
2846
+ "loss": 0.7347,
2847
+ "step": 236500
2848
+ },
2849
+ {
2850
+ "epoch": 17.67,
2851
+ "learning_rate": 1.466696495152871e-05,
2852
+ "loss": 0.7336,
2853
+ "step": 237000
2854
+ },
2855
+ {
2856
+ "epoch": 17.71,
2857
+ "learning_rate": 1.459254287844892e-05,
2858
+ "loss": 0.7238,
2859
+ "step": 237500
2860
+ },
2861
+ {
2862
+ "epoch": 17.75,
2863
+ "learning_rate": 1.4517971662938106e-05,
2864
+ "loss": 0.7348,
2865
+ "step": 238000
2866
+ },
2867
+ {
2868
+ "epoch": 17.79,
2869
+ "learning_rate": 1.4443549589858315e-05,
2870
+ "loss": 0.7348,
2871
+ "step": 238500
2872
+ },
2873
+ {
2874
+ "epoch": 17.82,
2875
+ "learning_rate": 1.4368978374347503e-05,
2876
+ "loss": 0.7221,
2877
+ "step": 239000
2878
+ },
2879
+ {
2880
+ "epoch": 17.86,
2881
+ "learning_rate": 1.429440715883669e-05,
2882
+ "loss": 0.7276,
2883
+ "step": 239500
2884
+ },
2885
+ {
2886
+ "epoch": 17.9,
2887
+ "learning_rate": 1.4219835943325877e-05,
2888
+ "loss": 0.725,
2889
+ "step": 240000
2890
+ },
2891
+ {
2892
+ "epoch": 17.93,
2893
+ "learning_rate": 1.4145264727815064e-05,
2894
+ "loss": 0.7271,
2895
+ "step": 240500
2896
+ },
2897
+ {
2898
+ "epoch": 17.97,
2899
+ "learning_rate": 1.407069351230425e-05,
2900
+ "loss": 0.7253,
2901
+ "step": 241000
2902
+ },
2903
+ {
2904
+ "epoch": 18.01,
2905
+ "learning_rate": 1.3996122296793438e-05,
2906
+ "loss": 0.7249,
2907
+ "step": 241500
2908
+ },
2909
+ {
2910
+ "epoch": 18.05,
2911
+ "learning_rate": 1.3921551081282625e-05,
2912
+ "loss": 0.7248,
2913
+ "step": 242000
2914
+ },
2915
+ {
2916
+ "epoch": 18.08,
2917
+ "learning_rate": 1.3846979865771812e-05,
2918
+ "loss": 0.7279,
2919
+ "step": 242500
2920
+ },
2921
+ {
2922
+ "epoch": 18.12,
2923
+ "learning_rate": 1.3772408650261e-05,
2924
+ "loss": 0.7313,
2925
+ "step": 243000
2926
+ },
2927
+ {
2928
+ "epoch": 18.16,
2929
+ "learning_rate": 1.3697837434750188e-05,
2930
+ "loss": 0.7249,
2931
+ "step": 243500
2932
+ },
2933
+ {
2934
+ "epoch": 18.2,
2935
+ "learning_rate": 1.3623266219239375e-05,
2936
+ "loss": 0.7154,
2937
+ "step": 244000
2938
+ },
2939
+ {
2940
+ "epoch": 18.23,
2941
+ "learning_rate": 1.3548695003728562e-05,
2942
+ "loss": 0.7206,
2943
+ "step": 244500
2944
+ },
2945
+ {
2946
+ "epoch": 18.27,
2947
+ "learning_rate": 1.3474272930648771e-05,
2948
+ "loss": 0.7222,
2949
+ "step": 245000
2950
+ },
2951
+ {
2952
+ "epoch": 18.31,
2953
+ "learning_rate": 1.3399701715137958e-05,
2954
+ "loss": 0.7212,
2955
+ "step": 245500
2956
+ },
2957
+ {
2958
+ "epoch": 18.34,
2959
+ "learning_rate": 1.3325130499627145e-05,
2960
+ "loss": 0.7207,
2961
+ "step": 246000
2962
+ },
2963
+ {
2964
+ "epoch": 18.38,
2965
+ "learning_rate": 1.3250559284116332e-05,
2966
+ "loss": 0.7236,
2967
+ "step": 246500
2968
+ },
2969
+ {
2970
+ "epoch": 18.42,
2971
+ "learning_rate": 1.3176137211036541e-05,
2972
+ "loss": 0.7235,
2973
+ "step": 247000
2974
+ },
2975
+ {
2976
+ "epoch": 18.46,
2977
+ "learning_rate": 1.3101715137956749e-05,
2978
+ "loss": 0.7207,
2979
+ "step": 247500
2980
+ },
2981
+ {
2982
+ "epoch": 18.49,
2983
+ "learning_rate": 1.3027143922445936e-05,
2984
+ "loss": 0.7201,
2985
+ "step": 248000
2986
+ },
2987
+ {
2988
+ "epoch": 18.53,
2989
+ "learning_rate": 1.2952572706935123e-05,
2990
+ "loss": 0.7212,
2991
+ "step": 248500
2992
+ },
2993
+ {
2994
+ "epoch": 18.57,
2995
+ "learning_rate": 1.287800149142431e-05,
2996
+ "loss": 0.7328,
2997
+ "step": 249000
2998
+ },
2999
+ {
3000
+ "epoch": 18.61,
3001
+ "learning_rate": 1.2803430275913497e-05,
3002
+ "loss": 0.7278,
3003
+ "step": 249500
3004
+ },
3005
+ {
3006
+ "epoch": 18.64,
3007
+ "learning_rate": 1.2728859060402684e-05,
3008
+ "loss": 0.7188,
3009
+ "step": 250000
3010
+ },
3011
+ {
3012
+ "epoch": 18.68,
3013
+ "learning_rate": 1.2654287844891871e-05,
3014
+ "loss": 0.7212,
3015
+ "step": 250500
3016
+ },
3017
+ {
3018
+ "epoch": 18.72,
3019
+ "learning_rate": 1.2579716629381062e-05,
3020
+ "loss": 0.7273,
3021
+ "step": 251000
3022
+ },
3023
+ {
3024
+ "epoch": 18.75,
3025
+ "learning_rate": 1.2505294556301269e-05,
3026
+ "loss": 0.718,
3027
+ "step": 251500
3028
+ },
3029
+ {
3030
+ "epoch": 18.79,
3031
+ "learning_rate": 1.2430723340790456e-05,
3032
+ "loss": 0.7256,
3033
+ "step": 252000
3034
+ },
3035
+ {
3036
+ "epoch": 18.83,
3037
+ "learning_rate": 1.2356152125279641e-05,
3038
+ "loss": 0.7269,
3039
+ "step": 252500
3040
+ },
3041
+ {
3042
+ "epoch": 18.87,
3043
+ "learning_rate": 1.228173005219985e-05,
3044
+ "loss": 0.7144,
3045
+ "step": 253000
3046
+ },
3047
+ {
3048
+ "epoch": 18.9,
3049
+ "learning_rate": 1.220730797912006e-05,
3050
+ "loss": 0.7169,
3051
+ "step": 253500
3052
+ },
3053
+ {
3054
+ "epoch": 18.94,
3055
+ "learning_rate": 1.2132736763609247e-05,
3056
+ "loss": 0.7225,
3057
+ "step": 254000
3058
+ },
3059
+ {
3060
+ "epoch": 18.98,
3061
+ "learning_rate": 1.2058165548098435e-05,
3062
+ "loss": 0.7198,
3063
+ "step": 254500
3064
+ },
3065
+ {
3066
+ "epoch": 19.02,
3067
+ "learning_rate": 1.1983594332587622e-05,
3068
+ "loss": 0.7165,
3069
+ "step": 255000
3070
+ },
3071
+ {
3072
+ "epoch": 19.05,
3073
+ "learning_rate": 1.190902311707681e-05,
3074
+ "loss": 0.7114,
3075
+ "step": 255500
3076
+ },
3077
+ {
3078
+ "epoch": 19.09,
3079
+ "learning_rate": 1.1834451901565997e-05,
3080
+ "loss": 0.7191,
3081
+ "step": 256000
3082
+ },
3083
+ {
3084
+ "epoch": 19.13,
3085
+ "learning_rate": 1.1759880686055184e-05,
3086
+ "loss": 0.7164,
3087
+ "step": 256500
3088
+ },
3089
+ {
3090
+ "epoch": 19.16,
3091
+ "learning_rate": 1.168530947054437e-05,
3092
+ "loss": 0.7177,
3093
+ "step": 257000
3094
+ },
3095
+ {
3096
+ "epoch": 19.2,
3097
+ "learning_rate": 1.1610738255033558e-05,
3098
+ "loss": 0.7125,
3099
+ "step": 257500
3100
+ },
3101
+ {
3102
+ "epoch": 19.24,
3103
+ "learning_rate": 1.1536167039522745e-05,
3104
+ "loss": 0.7177,
3105
+ "step": 258000
3106
+ },
3107
+ {
3108
+ "epoch": 19.28,
3109
+ "learning_rate": 1.1461595824011932e-05,
3110
+ "loss": 0.7107,
3111
+ "step": 258500
3112
+ },
3113
+ {
3114
+ "epoch": 19.31,
3115
+ "learning_rate": 1.1387024608501119e-05,
3116
+ "loss": 0.7199,
3117
+ "step": 259000
3118
+ },
3119
+ {
3120
+ "epoch": 19.35,
3121
+ "learning_rate": 1.1312453392990306e-05,
3122
+ "loss": 0.7197,
3123
+ "step": 259500
3124
+ },
3125
+ {
3126
+ "epoch": 19.39,
3127
+ "learning_rate": 1.1238031319910515e-05,
3128
+ "loss": 0.7166,
3129
+ "step": 260000
3130
+ },
3131
+ {
3132
+ "epoch": 19.43,
3133
+ "learning_rate": 1.1163460104399702e-05,
3134
+ "loss": 0.7153,
3135
+ "step": 260500
3136
+ },
3137
+ {
3138
+ "epoch": 19.46,
3139
+ "learning_rate": 1.108888888888889e-05,
3140
+ "loss": 0.7176,
3141
+ "step": 261000
3142
+ },
3143
+ {
3144
+ "epoch": 19.5,
3145
+ "learning_rate": 1.1014466815809098e-05,
3146
+ "loss": 0.7143,
3147
+ "step": 261500
3148
+ },
3149
+ {
3150
+ "epoch": 19.54,
3151
+ "learning_rate": 1.0940044742729306e-05,
3152
+ "loss": 0.7167,
3153
+ "step": 262000
3154
+ },
3155
+ {
3156
+ "epoch": 19.57,
3157
+ "learning_rate": 1.0865473527218494e-05,
3158
+ "loss": 0.7157,
3159
+ "step": 262500
3160
+ },
3161
+ {
3162
+ "epoch": 19.61,
3163
+ "learning_rate": 1.0790902311707682e-05,
3164
+ "loss": 0.7107,
3165
+ "step": 263000
3166
+ },
3167
+ {
3168
+ "epoch": 19.65,
3169
+ "learning_rate": 1.0716331096196869e-05,
3170
+ "loss": 0.7156,
3171
+ "step": 263500
3172
+ },
3173
+ {
3174
+ "epoch": 19.69,
3175
+ "learning_rate": 1.0641909023117078e-05,
3176
+ "loss": 0.7164,
3177
+ "step": 264000
3178
+ },
3179
+ {
3180
+ "epoch": 19.72,
3181
+ "learning_rate": 1.0567337807606265e-05,
3182
+ "loss": 0.7131,
3183
+ "step": 264500
3184
+ },
3185
+ {
3186
+ "epoch": 19.76,
3187
+ "learning_rate": 1.0492766592095452e-05,
3188
+ "loss": 0.7158,
3189
+ "step": 265000
3190
+ },
3191
+ {
3192
+ "epoch": 19.8,
3193
+ "learning_rate": 1.0418195376584639e-05,
3194
+ "loss": 0.7129,
3195
+ "step": 265500
3196
+ },
3197
+ {
3198
+ "epoch": 19.84,
3199
+ "learning_rate": 1.0343624161073826e-05,
3200
+ "loss": 0.7119,
3201
+ "step": 266000
3202
+ },
3203
+ {
3204
+ "epoch": 19.87,
3205
+ "learning_rate": 1.0269052945563013e-05,
3206
+ "loss": 0.7208,
3207
+ "step": 266500
3208
+ },
3209
+ {
3210
+ "epoch": 19.91,
3211
+ "learning_rate": 1.01944817300522e-05,
3212
+ "loss": 0.7132,
3213
+ "step": 267000
3214
+ },
3215
+ {
3216
+ "epoch": 19.95,
3217
+ "learning_rate": 1.0119910514541387e-05,
3218
+ "loss": 0.715,
3219
+ "step": 267500
3220
+ },
3221
+ {
3222
+ "epoch": 19.98,
3223
+ "learning_rate": 1.0045339299030574e-05,
3224
+ "loss": 0.7125,
3225
+ "step": 268000
3226
+ },
3227
+ {
3228
+ "epoch": 20.02,
3229
+ "learning_rate": 9.970768083519761e-06,
3230
+ "loss": 0.7126,
3231
+ "step": 268500
3232
+ },
3233
+ {
3234
+ "epoch": 20.06,
3235
+ "learning_rate": 9.89634601043997e-06,
3236
+ "loss": 0.7056,
3237
+ "step": 269000
3238
+ },
3239
+ {
3240
+ "epoch": 20.1,
3241
+ "learning_rate": 9.821774794929157e-06,
3242
+ "loss": 0.7101,
3243
+ "step": 269500
3244
+ },
3245
+ {
3246
+ "epoch": 20.13,
3247
+ "learning_rate": 9.747203579418344e-06,
3248
+ "loss": 0.7103,
3249
+ "step": 270000
3250
+ },
3251
+ {
3252
+ "epoch": 20.17,
3253
+ "learning_rate": 9.672632363907533e-06,
3254
+ "loss": 0.7109,
3255
+ "step": 270500
3256
+ },
3257
+ {
3258
+ "epoch": 20.21,
3259
+ "learning_rate": 9.59806114839672e-06,
3260
+ "loss": 0.707,
3261
+ "step": 271000
3262
+ },
3263
+ {
3264
+ "epoch": 20.25,
3265
+ "learning_rate": 9.52363907531693e-06,
3266
+ "loss": 0.7121,
3267
+ "step": 271500
3268
+ },
3269
+ {
3270
+ "epoch": 20.28,
3271
+ "learning_rate": 9.449067859806115e-06,
3272
+ "loss": 0.7088,
3273
+ "step": 272000
3274
+ },
3275
+ {
3276
+ "epoch": 20.32,
3277
+ "learning_rate": 9.374496644295302e-06,
3278
+ "loss": 0.7078,
3279
+ "step": 272500
3280
+ },
3281
+ {
3282
+ "epoch": 20.36,
3283
+ "learning_rate": 9.299925428784489e-06,
3284
+ "loss": 0.7094,
3285
+ "step": 273000
3286
+ },
3287
+ {
3288
+ "epoch": 20.4,
3289
+ "learning_rate": 9.225354213273676e-06,
3290
+ "loss": 0.7103,
3291
+ "step": 273500
3292
+ },
3293
+ {
3294
+ "epoch": 20.43,
3295
+ "learning_rate": 9.150932140193885e-06,
3296
+ "loss": 0.7097,
3297
+ "step": 274000
3298
+ },
3299
+ {
3300
+ "epoch": 20.47,
3301
+ "learning_rate": 9.076360924683072e-06,
3302
+ "loss": 0.7092,
3303
+ "step": 274500
3304
+ },
3305
+ {
3306
+ "epoch": 20.51,
3307
+ "learning_rate": 9.001938851603281e-06,
3308
+ "loss": 0.7092,
3309
+ "step": 275000
3310
+ },
3311
+ {
3312
+ "epoch": 20.54,
3313
+ "learning_rate": 8.92736763609247e-06,
3314
+ "loss": 0.7061,
3315
+ "step": 275500
3316
+ },
3317
+ {
3318
+ "epoch": 20.58,
3319
+ "learning_rate": 8.852796420581657e-06,
3320
+ "loss": 0.7076,
3321
+ "step": 276000
3322
+ },
3323
+ {
3324
+ "epoch": 20.62,
3325
+ "learning_rate": 8.778225205070844e-06,
3326
+ "loss": 0.7124,
3327
+ "step": 276500
3328
+ },
3329
+ {
3330
+ "epoch": 20.66,
3331
+ "learning_rate": 8.70365398956003e-06,
3332
+ "loss": 0.7051,
3333
+ "step": 277000
3334
+ },
3335
+ {
3336
+ "epoch": 20.69,
3337
+ "learning_rate": 8.629082774049217e-06,
3338
+ "loss": 0.706,
3339
+ "step": 277500
3340
+ },
3341
+ {
3342
+ "epoch": 20.73,
3343
+ "learning_rate": 8.554511558538404e-06,
3344
+ "loss": 0.7043,
3345
+ "step": 278000
3346
+ },
3347
+ {
3348
+ "epoch": 20.77,
3349
+ "learning_rate": 8.47994034302759e-06,
3350
+ "loss": 0.7081,
3351
+ "step": 278500
3352
+ },
3353
+ {
3354
+ "epoch": 20.81,
3355
+ "learning_rate": 8.40536912751678e-06,
3356
+ "loss": 0.7048,
3357
+ "step": 279000
3358
+ },
3359
+ {
3360
+ "epoch": 20.84,
3361
+ "learning_rate": 8.330797912005966e-06,
3362
+ "loss": 0.715,
3363
+ "step": 279500
3364
+ },
3365
+ {
3366
+ "epoch": 20.88,
3367
+ "learning_rate": 8.256226696495154e-06,
3368
+ "loss": 0.7035,
3369
+ "step": 280000
3370
+ },
3371
+ {
3372
+ "epoch": 20.92,
3373
+ "learning_rate": 8.18165548098434e-06,
3374
+ "loss": 0.7097,
3375
+ "step": 280500
3376
+ },
3377
+ {
3378
+ "epoch": 20.95,
3379
+ "learning_rate": 8.107084265473528e-06,
3380
+ "loss": 0.706,
3381
+ "step": 281000
3382
+ },
3383
+ {
3384
+ "epoch": 20.99,
3385
+ "learning_rate": 8.032662192393737e-06,
3386
+ "loss": 0.7003,
3387
+ "step": 281500
3388
+ },
3389
+ {
3390
+ "epoch": 21.03,
3391
+ "learning_rate": 7.958090976882924e-06,
3392
+ "loss": 0.7037,
3393
+ "step": 282000
3394
+ },
3395
+ {
3396
+ "epoch": 21.07,
3397
+ "learning_rate": 7.883668903803131e-06,
3398
+ "loss": 0.7066,
3399
+ "step": 282500
3400
+ },
3401
+ {
3402
+ "epoch": 21.1,
3403
+ "learning_rate": 7.809097688292318e-06,
3404
+ "loss": 0.7031,
3405
+ "step": 283000
3406
+ },
3407
+ {
3408
+ "epoch": 21.14,
3409
+ "learning_rate": 7.734526472781507e-06,
3410
+ "loss": 0.6978,
3411
+ "step": 283500
3412
+ },
3413
+ {
3414
+ "epoch": 21.18,
3415
+ "learning_rate": 7.659955257270694e-06,
3416
+ "loss": 0.7041,
3417
+ "step": 284000
3418
+ },
3419
+ {
3420
+ "epoch": 21.22,
3421
+ "learning_rate": 7.585533184190903e-06,
3422
+ "loss": 0.7044,
3423
+ "step": 284500
3424
+ },
3425
+ {
3426
+ "epoch": 21.25,
3427
+ "learning_rate": 7.51096196868009e-06,
3428
+ "loss": 0.6957,
3429
+ "step": 285000
3430
+ },
3431
+ {
3432
+ "epoch": 21.29,
3433
+ "learning_rate": 7.436390753169277e-06,
3434
+ "loss": 0.702,
3435
+ "step": 285500
3436
+ },
3437
+ {
3438
+ "epoch": 21.33,
3439
+ "learning_rate": 7.3619686800894855e-06,
3440
+ "loss": 0.7002,
3441
+ "step": 286000
3442
+ },
3443
+ {
3444
+ "epoch": 21.36,
3445
+ "learning_rate": 7.287397464578673e-06,
3446
+ "loss": 0.7031,
3447
+ "step": 286500
3448
+ },
3449
+ {
3450
+ "epoch": 21.4,
3451
+ "learning_rate": 7.21282624906786e-06,
3452
+ "loss": 0.7073,
3453
+ "step": 287000
3454
+ },
3455
+ {
3456
+ "epoch": 21.44,
3457
+ "learning_rate": 7.138255033557047e-06,
3458
+ "loss": 0.706,
3459
+ "step": 287500
3460
+ },
3461
+ {
3462
+ "epoch": 21.48,
3463
+ "learning_rate": 7.063683818046235e-06,
3464
+ "loss": 0.7052,
3465
+ "step": 288000
3466
+ },
3467
+ {
3468
+ "epoch": 21.51,
3469
+ "learning_rate": 6.989112602535422e-06,
3470
+ "loss": 0.7019,
3471
+ "step": 288500
3472
+ },
3473
+ {
3474
+ "epoch": 21.55,
3475
+ "learning_rate": 6.914690529455631e-06,
3476
+ "loss": 0.6996,
3477
+ "step": 289000
3478
+ },
3479
+ {
3480
+ "epoch": 21.59,
3481
+ "learning_rate": 6.840119313944818e-06,
3482
+ "loss": 0.7,
3483
+ "step": 289500
3484
+ },
3485
+ {
3486
+ "epoch": 21.63,
3487
+ "learning_rate": 6.765548098434005e-06,
3488
+ "loss": 0.7037,
3489
+ "step": 290000
3490
+ },
3491
+ {
3492
+ "epoch": 21.66,
3493
+ "learning_rate": 6.690976882923192e-06,
3494
+ "loss": 0.7007,
3495
+ "step": 290500
3496
+ },
3497
+ {
3498
+ "epoch": 21.7,
3499
+ "learning_rate": 6.616405667412379e-06,
3500
+ "loss": 0.7031,
3501
+ "step": 291000
3502
+ },
3503
+ {
3504
+ "epoch": 21.74,
3505
+ "learning_rate": 6.541834451901565e-06,
3506
+ "loss": 0.7011,
3507
+ "step": 291500
3508
+ },
3509
+ {
3510
+ "epoch": 21.77,
3511
+ "learning_rate": 6.467263236390754e-06,
3512
+ "loss": 0.7008,
3513
+ "step": 292000
3514
+ },
3515
+ {
3516
+ "epoch": 21.81,
3517
+ "learning_rate": 6.392692020879941e-06,
3518
+ "loss": 0.696,
3519
+ "step": 292500
3520
+ },
3521
+ {
3522
+ "epoch": 21.85,
3523
+ "learning_rate": 6.31826994780015e-06,
3524
+ "loss": 0.6973,
3525
+ "step": 293000
3526
+ },
3527
+ {
3528
+ "epoch": 21.89,
3529
+ "learning_rate": 6.243698732289336e-06,
3530
+ "loss": 0.6995,
3531
+ "step": 293500
3532
+ },
3533
+ {
3534
+ "epoch": 21.92,
3535
+ "learning_rate": 6.1691275167785235e-06,
3536
+ "loss": 0.6985,
3537
+ "step": 294000
3538
+ },
3539
+ {
3540
+ "epoch": 21.96,
3541
+ "learning_rate": 6.0945563012677105e-06,
3542
+ "loss": 0.703,
3543
+ "step": 294500
3544
+ },
3545
+ {
3546
+ "epoch": 22.0,
3547
+ "learning_rate": 6.0199850857568984e-06,
3548
+ "loss": 0.6983,
3549
+ "step": 295000
3550
+ },
3551
+ {
3552
+ "epoch": 22.04,
3553
+ "learning_rate": 5.9454138702460855e-06,
3554
+ "loss": 0.6993,
3555
+ "step": 295500
3556
+ },
3557
+ {
3558
+ "epoch": 22.07,
3559
+ "learning_rate": 5.8708426547352726e-06,
3560
+ "loss": 0.6923,
3561
+ "step": 296000
3562
+ },
3563
+ {
3564
+ "epoch": 22.11,
3565
+ "learning_rate": 5.79627143922446e-06,
3566
+ "loss": 0.6993,
3567
+ "step": 296500
3568
+ },
3569
+ {
3570
+ "epoch": 22.15,
3571
+ "learning_rate": 5.721700223713647e-06,
3572
+ "loss": 0.6927,
3573
+ "step": 297000
3574
+ },
3575
+ {
3576
+ "epoch": 22.18,
3577
+ "learning_rate": 5.647129008202834e-06,
3578
+ "loss": 0.6942,
3579
+ "step": 297500
3580
+ },
3581
+ {
3582
+ "epoch": 22.22,
3583
+ "learning_rate": 5.572706935123043e-06,
3584
+ "loss": 0.6976,
3585
+ "step": 298000
3586
+ },
3587
+ {
3588
+ "epoch": 22.26,
3589
+ "learning_rate": 5.49813571961223e-06,
3590
+ "loss": 0.6943,
3591
+ "step": 298500
3592
+ },
3593
+ {
3594
+ "epoch": 22.3,
3595
+ "learning_rate": 5.423564504101417e-06,
3596
+ "loss": 0.7016,
3597
+ "step": 299000
3598
+ },
3599
+ {
3600
+ "epoch": 22.33,
3601
+ "learning_rate": 5.348993288590604e-06,
3602
+ "loss": 0.6978,
3603
+ "step": 299500
3604
+ },
3605
+ {
3606
+ "epoch": 22.37,
3607
+ "learning_rate": 5.274422073079791e-06,
3608
+ "loss": 0.6952,
3609
+ "step": 300000
3610
+ },
3611
+ {
3612
+ "epoch": 22.41,
3613
+ "learning_rate": 5.199850857568979e-06,
3614
+ "loss": 0.6971,
3615
+ "step": 300500
3616
+ },
3617
+ {
3618
+ "epoch": 22.45,
3619
+ "learning_rate": 5.125279642058166e-06,
3620
+ "loss": 0.6997,
3621
+ "step": 301000
3622
+ },
3623
+ {
3624
+ "epoch": 22.48,
3625
+ "learning_rate": 5.050708426547353e-06,
3626
+ "loss": 0.6925,
3627
+ "step": 301500
3628
+ },
3629
+ {
3630
+ "epoch": 22.52,
3631
+ "learning_rate": 4.976137211036539e-06,
3632
+ "loss": 0.692,
3633
+ "step": 302000
3634
+ },
3635
+ {
3636
+ "epoch": 22.56,
3637
+ "learning_rate": 4.901715137956749e-06,
3638
+ "loss": 0.7017,
3639
+ "step": 302500
3640
+ },
3641
+ {
3642
+ "epoch": 22.59,
3643
+ "learning_rate": 4.827143922445936e-06,
3644
+ "loss": 0.6978,
3645
+ "step": 303000
3646
+ },
3647
+ {
3648
+ "epoch": 22.63,
3649
+ "learning_rate": 4.7525727069351234e-06,
3650
+ "loss": 0.698,
3651
+ "step": 303500
3652
+ },
3653
+ {
3654
+ "epoch": 22.67,
3655
+ "learning_rate": 4.6780014914243105e-06,
3656
+ "loss": 0.6987,
3657
+ "step": 304000
3658
+ },
3659
+ {
3660
+ "epoch": 22.71,
3661
+ "learning_rate": 4.603430275913498e-06,
3662
+ "loss": 0.6922,
3663
+ "step": 304500
3664
+ },
3665
+ {
3666
+ "epoch": 22.74,
3667
+ "learning_rate": 4.529008202833707e-06,
3668
+ "loss": 0.693,
3669
+ "step": 305000
3670
+ },
3671
+ {
3672
+ "epoch": 22.78,
3673
+ "learning_rate": 4.454436987322894e-06,
3674
+ "loss": 0.6959,
3675
+ "step": 305500
3676
+ },
3677
+ {
3678
+ "epoch": 22.82,
3679
+ "learning_rate": 4.379865771812081e-06,
3680
+ "loss": 0.6993,
3681
+ "step": 306000
3682
+ },
3683
+ {
3684
+ "epoch": 22.86,
3685
+ "learning_rate": 4.305294556301268e-06,
3686
+ "loss": 0.6949,
3687
+ "step": 306500
3688
+ },
3689
+ {
3690
+ "epoch": 22.89,
3691
+ "learning_rate": 4.230872483221477e-06,
3692
+ "loss": 0.6936,
3693
+ "step": 307000
3694
+ },
3695
+ {
3696
+ "epoch": 22.93,
3697
+ "learning_rate": 4.156301267710664e-06,
3698
+ "loss": 0.6956,
3699
+ "step": 307500
3700
+ },
3701
+ {
3702
+ "epoch": 22.97,
3703
+ "learning_rate": 4.081730052199851e-06,
3704
+ "loss": 0.6929,
3705
+ "step": 308000
3706
+ },
3707
+ {
3708
+ "epoch": 23.01,
3709
+ "learning_rate": 4.007158836689038e-06,
3710
+ "loss": 0.6897,
3711
+ "step": 308500
3712
+ },
3713
+ {
3714
+ "epoch": 23.04,
3715
+ "learning_rate": 3.932736763609246e-06,
3716
+ "loss": 0.6975,
3717
+ "step": 309000
3718
+ },
3719
+ {
3720
+ "epoch": 23.08,
3721
+ "learning_rate": 3.858165548098434e-06,
3722
+ "loss": 0.6971,
3723
+ "step": 309500
3724
+ },
3725
+ {
3726
+ "epoch": 23.12,
3727
+ "learning_rate": 3.7835943325876213e-06,
3728
+ "loss": 0.6965,
3729
+ "step": 310000
3730
+ },
3731
+ {
3732
+ "epoch": 23.15,
3733
+ "learning_rate": 3.70917225950783e-06,
3734
+ "loss": 0.697,
3735
+ "step": 310500
3736
+ },
3737
+ {
3738
+ "epoch": 23.19,
3739
+ "learning_rate": 3.634601043997017e-06,
3740
+ "loss": 0.6885,
3741
+ "step": 311000
3742
+ },
3743
+ {
3744
+ "epoch": 23.23,
3745
+ "learning_rate": 3.560029828486205e-06,
3746
+ "loss": 0.6854,
3747
+ "step": 311500
3748
+ },
3749
+ {
3750
+ "epoch": 23.27,
3751
+ "learning_rate": 3.4854586129753916e-06,
3752
+ "loss": 0.6938,
3753
+ "step": 312000
3754
+ },
3755
+ {
3756
+ "epoch": 23.3,
3757
+ "learning_rate": 3.4108873974645787e-06,
3758
+ "loss": 0.6905,
3759
+ "step": 312500
3760
+ },
3761
+ {
3762
+ "epoch": 23.34,
3763
+ "learning_rate": 3.3364653243847873e-06,
3764
+ "loss": 0.6927,
3765
+ "step": 313000
3766
+ },
3767
+ {
3768
+ "epoch": 23.38,
3769
+ "learning_rate": 3.2618941088739744e-06,
3770
+ "loss": 0.693,
3771
+ "step": 313500
3772
+ },
3773
+ {
3774
+ "epoch": 23.42,
3775
+ "learning_rate": 3.1873228933631623e-06,
3776
+ "loss": 0.6894,
3777
+ "step": 314000
3778
+ },
3779
+ {
3780
+ "epoch": 23.45,
3781
+ "learning_rate": 3.112751677852349e-06,
3782
+ "loss": 0.6899,
3783
+ "step": 314500
3784
+ },
3785
+ {
3786
+ "epoch": 23.49,
3787
+ "learning_rate": 3.0381804623415364e-06,
3788
+ "loss": 0.6924,
3789
+ "step": 315000
3790
+ },
3791
+ {
3792
+ "epoch": 23.53,
3793
+ "learning_rate": 2.9636092468307235e-06,
3794
+ "loss": 0.6936,
3795
+ "step": 315500
3796
+ },
3797
+ {
3798
+ "epoch": 23.56,
3799
+ "learning_rate": 2.8890380313199106e-06,
3800
+ "loss": 0.6939,
3801
+ "step": 316000
3802
+ },
3803
+ {
3804
+ "epoch": 23.6,
3805
+ "learning_rate": 2.8144668158090976e-06,
3806
+ "loss": 0.6975,
3807
+ "step": 316500
3808
+ },
3809
+ {
3810
+ "epoch": 23.64,
3811
+ "learning_rate": 2.7400447427293063e-06,
3812
+ "loss": 0.6905,
3813
+ "step": 317000
3814
+ },
3815
+ {
3816
+ "epoch": 23.68,
3817
+ "learning_rate": 2.6654735272184938e-06,
3818
+ "loss": 0.6908,
3819
+ "step": 317500
3820
+ },
3821
+ {
3822
+ "epoch": 23.71,
3823
+ "learning_rate": 2.5909023117076813e-06,
3824
+ "loss": 0.6891,
3825
+ "step": 318000
3826
+ },
3827
+ {
3828
+ "epoch": 23.75,
3829
+ "learning_rate": 2.516331096196868e-06,
3830
+ "loss": 0.6911,
3831
+ "step": 318500
3832
+ },
3833
+ {
3834
+ "epoch": 23.79,
3835
+ "learning_rate": 2.4417598806860554e-06,
3836
+ "loss": 0.6956,
3837
+ "step": 319000
3838
+ },
3839
+ {
3840
+ "epoch": 23.83,
3841
+ "learning_rate": 2.3671886651752425e-06,
3842
+ "loss": 0.6941,
3843
+ "step": 319500
3844
+ },
3845
+ {
3846
+ "epoch": 23.86,
3847
+ "learning_rate": 2.292766592095451e-06,
3848
+ "loss": 0.6864,
3849
+ "step": 320000
3850
+ },
3851
+ {
3852
+ "epoch": 23.9,
3853
+ "learning_rate": 2.2181953765846386e-06,
3854
+ "loss": 0.6959,
3855
+ "step": 320500
3856
+ },
3857
+ {
3858
+ "epoch": 23.94,
3859
+ "learning_rate": 2.1436241610738253e-06,
3860
+ "loss": 0.689,
3861
+ "step": 321000
3862
+ },
3863
+ {
3864
+ "epoch": 23.97,
3865
+ "learning_rate": 2.0690529455630127e-06,
3866
+ "loss": 0.6906,
3867
+ "step": 321500
3868
+ },
3869
+ {
3870
+ "epoch": 24.01,
3871
+ "learning_rate": 1.9944817300522002e-06,
3872
+ "loss": 0.6921,
3873
+ "step": 322000
3874
+ },
3875
+ {
3876
+ "epoch": 24.05,
3877
+ "learning_rate": 1.919910514541387e-06,
3878
+ "loss": 0.6915,
3879
+ "step": 322500
3880
+ },
3881
+ {
3882
+ "epoch": 24.09,
3883
+ "learning_rate": 1.8454884414615957e-06,
3884
+ "loss": 0.6957,
3885
+ "step": 323000
3886
+ },
3887
+ {
3888
+ "epoch": 24.12,
3889
+ "learning_rate": 1.7709172259507832e-06,
3890
+ "loss": 0.6892,
3891
+ "step": 323500
3892
+ },
3893
+ {
3894
+ "epoch": 24.16,
3895
+ "learning_rate": 1.69634601043997e-06,
3896
+ "loss": 0.6875,
3897
+ "step": 324000
3898
+ },
3899
+ {
3900
+ "epoch": 24.2,
3901
+ "learning_rate": 1.6217747949291574e-06,
3902
+ "loss": 0.6899,
3903
+ "step": 324500
3904
+ },
3905
+ {
3906
+ "epoch": 24.24,
3907
+ "learning_rate": 1.5472035794183446e-06,
3908
+ "loss": 0.6888,
3909
+ "step": 325000
3910
+ },
3911
+ {
3912
+ "epoch": 24.27,
3913
+ "learning_rate": 1.4726323639075317e-06,
3914
+ "loss": 0.6921,
3915
+ "step": 325500
3916
+ },
3917
+ {
3918
+ "epoch": 24.31,
3919
+ "learning_rate": 1.3980611483967188e-06,
3920
+ "loss": 0.6866,
3921
+ "step": 326000
3922
+ },
3923
+ {
3924
+ "epoch": 24.35,
3925
+ "learning_rate": 1.3234899328859063e-06,
3926
+ "loss": 0.6873,
3927
+ "step": 326500
3928
+ },
3929
+ {
3930
+ "epoch": 24.38,
3931
+ "learning_rate": 1.249067859806115e-06,
3932
+ "loss": 0.6859,
3933
+ "step": 327000
3934
+ },
3935
+ {
3936
+ "epoch": 24.42,
3937
+ "learning_rate": 1.1744966442953022e-06,
3938
+ "loss": 0.6825,
3939
+ "step": 327500
3940
+ },
3941
+ {
3942
+ "epoch": 24.46,
3943
+ "learning_rate": 1.0999254287844893e-06,
3944
+ "loss": 0.6904,
3945
+ "step": 328000
3946
+ },
3947
+ {
3948
+ "epoch": 24.5,
3949
+ "learning_rate": 1.0253542132736763e-06,
3950
+ "loss": 0.6941,
3951
+ "step": 328500
3952
+ },
3953
+ {
3954
+ "epoch": 24.53,
3955
+ "learning_rate": 9.507829977628635e-07,
3956
+ "loss": 0.69,
3957
+ "step": 329000
3958
+ },
3959
+ {
3960
+ "epoch": 24.57,
3961
+ "learning_rate": 8.762117822520507e-07,
3962
+ "loss": 0.6842,
3963
+ "step": 329500
3964
+ },
3965
+ {
3966
+ "epoch": 24.61,
3967
+ "learning_rate": 8.01640566741238e-07,
3968
+ "loss": 0.6931,
3969
+ "step": 330000
3970
+ },
3971
+ {
3972
+ "epoch": 24.65,
3973
+ "learning_rate": 7.27069351230425e-07,
3974
+ "loss": 0.6869,
3975
+ "step": 330500
3976
+ },
3977
+ {
3978
+ "epoch": 24.68,
3979
+ "learning_rate": 6.527964205816555e-07,
3980
+ "loss": 0.6927,
3981
+ "step": 331000
3982
+ },
3983
+ {
3984
+ "epoch": 24.72,
3985
+ "learning_rate": 5.782252050708427e-07,
3986
+ "loss": 0.6846,
3987
+ "step": 331500
3988
+ },
3989
+ {
3990
+ "epoch": 24.76,
3991
+ "learning_rate": 5.036539895600298e-07,
3992
+ "loss": 0.6932,
3993
+ "step": 332000
3994
+ },
3995
+ {
3996
+ "epoch": 24.79,
3997
+ "learning_rate": 4.29082774049217e-07,
3998
+ "loss": 0.6867,
3999
+ "step": 332500
4000
+ },
4001
+ {
4002
+ "epoch": 24.83,
4003
+ "learning_rate": 3.545115585384042e-07,
4004
+ "loss": 0.6883,
4005
+ "step": 333000
4006
+ },
4007
+ {
4008
+ "epoch": 24.87,
4009
+ "learning_rate": 2.799403430275914e-07,
4010
+ "loss": 0.6826,
4011
+ "step": 333500
4012
+ },
4013
+ {
4014
+ "epoch": 24.91,
4015
+ "learning_rate": 2.053691275167785e-07,
4016
+ "loss": 0.6882,
4017
+ "step": 334000
4018
+ },
4019
+ {
4020
+ "epoch": 24.94,
4021
+ "learning_rate": 1.3109619686800894e-07,
4022
+ "loss": 0.6854,
4023
+ "step": 334500
4024
+ },
4025
+ {
4026
+ "epoch": 24.98,
4027
+ "learning_rate": 5.652498135719612e-08,
4028
+ "loss": 0.6878,
4029
+ "step": 335000
4030
+ },
4031
+ {
4032
+ "epoch": 25.0,
4033
+ "step": 335250,
4034
+ "total_flos": 2.835875553205027e+18,
4035
+ "train_loss": 0.891669009442653,
4036
+ "train_runtime": 467828.9629,
4037
+ "train_samples_per_second": 22.932,
4038
+ "train_steps_per_second": 0.717
4039
+ }
4040
+ ],
4041
+ "logging_steps": 500,
4042
+ "max_steps": 335250,
4043
+ "num_train_epochs": 25,
4044
+ "save_steps": 10000,
4045
+ "total_flos": 2.835875553205027e+18,
4046
+ "trial_name": null,
4047
+ "trial_params": null
4048
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9d684ff271acc2b63d51dc1503c79b5b8ed3ea670cbe0d2f090d6732b3d1c19b
3
+ size 4472