Hugging Face's logo

KM4STfulltext
/
CSSCI_ABS_BERT

Fill-Mask
Transformers
PyTorch
bert
Inference Endpoints
Model card Files Community
1
Gogs commited on
Commit
bac98ed
·
1 Parent(s): 0b732c2

commit from turning

Browse files
Files changed (12) hide show
  1. all_results.json +14 -0
  2. bert_config.json +31 -0
  3. config.json +31 -0
  4. eval_results.json +9 -0
  5. pytorch_model.bin +3 -0
  6. special_tokens_map.json +1 -0
  7. tokenizer.json +0 -0
  8. tokenizer_config.json +1 -0
  9. train_results.json +8 -0
  10. trainer_state.json +3295 -0
  11. training_args.bin +3 -0
  12. vocab.txt +0 -0
all_results.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 3.0,
3
+ "eval_loss": 0.6053370237350464,
4
+ "eval_runtime": 230.4858,
5
+ "eval_samples": 14683,
6
+ "eval_samples_per_second": 63.705,
7
+ "eval_steps_per_second": 3.983,
8
+ "perplexity": 1.8318694883580977,
9
+ "train_loss": 0.7392136628871132,
10
+ "train_runtime": 175400.2766,
11
+ "train_samples": 1453570,
12
+ "train_samples_per_second": 24.861,
13
+ "train_steps_per_second": 1.554
14
+ }
bert_config.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "/home/admin/pretrain_models/bert-base-chinese",
3
+ "architectures": [
4
+ "BertForMaskedLM"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.1,
7
+ "classifier_dropout": null,
8
+ "directionality": "bidi",
9
+ "hidden_act": "gelu",
10
+ "hidden_dropout_prob": 0.1,
11
+ "hidden_size": 768,
12
+ "initializer_range": 0.02,
13
+ "intermediate_size": 3072,
14
+ "layer_norm_eps": 1e-12,
15
+ "max_position_embeddings": 512,
16
+ "model_type": "bert",
17
+ "num_attention_heads": 12,
18
+ "num_hidden_layers": 12,
19
+ "pad_token_id": 0,
20
+ "pooler_fc_size": 768,
21
+ "pooler_num_attention_heads": 12,
22
+ "pooler_num_fc_layers": 3,
23
+ "pooler_size_per_head": 128,
24
+ "pooler_type": "first_token_transform",
25
+ "position_embedding_type": "absolute",
26
+ "torch_dtype": "float32",
27
+ "transformers_version": "4.16.0",
28
+ "type_vocab_size": 2,
29
+ "use_cache": true,
30
+ "vocab_size": 21128
31
+ }
config.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "/home/admin/pretrain_models/bert-base-chinese",
3
+ "architectures": [
4
+ "BertForMaskedLM"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.1,
7
+ "classifier_dropout": null,
8
+ "directionality": "bidi",
9
+ "hidden_act": "gelu",
10
+ "hidden_dropout_prob": 0.1,
11
+ "hidden_size": 768,
12
+ "initializer_range": 0.02,
13
+ "intermediate_size": 3072,
14
+ "layer_norm_eps": 1e-12,
15
+ "max_position_embeddings": 512,
16
+ "model_type": "bert",
17
+ "num_attention_heads": 12,
18
+ "num_hidden_layers": 12,
19
+ "pad_token_id": 0,
20
+ "pooler_fc_size": 768,
21
+ "pooler_num_attention_heads": 12,
22
+ "pooler_num_fc_layers": 3,
23
+ "pooler_size_per_head": 128,
24
+ "pooler_type": "first_token_transform",
25
+ "position_embedding_type": "absolute",
26
+ "torch_dtype": "float32",
27
+ "transformers_version": "4.16.0",
28
+ "type_vocab_size": 2,
29
+ "use_cache": true,
30
+ "vocab_size": 21128
31
+ }
eval_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 3.0,
3
+ "eval_loss": 0.6053370237350464,
4
+ "eval_runtime": 230.4858,
5
+ "eval_samples": 14683,
6
+ "eval_samples_per_second": 63.705,
7
+ "eval_steps_per_second": 3.983,
8
+ "perplexity": 1.8318694883580977
9
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:21fa9c8e48200e0e4e4efb54720c6f37d74c7a32fe7117972616dca38b6d2ca9
3
+ size 409216422
special_tokens_map.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]"}
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"do_lower_case": true, "unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]", "tokenize_chinese_chars": true, "strip_accents": null, "special_tokens_map_file": null, "name_or_path": "/home/admin/pretrain_models/bert-base-chinese", "do_basic_tokenize": true, "never_split": null, "tokenizer_class": "BertTokenizer"}
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 3.0,
3
+ "train_loss": 0.7392136628871132,
4
+ "train_runtime": 175400.2766,
5
+ "train_samples": 1453570,
6
+ "train_samples_per_second": 24.861,
7
+ "train_steps_per_second": 1.554
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,3295 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 3.0,
5
+ "global_step": 272547,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.01,
12
+ "learning_rate": 4.9908272701589084e-05,
13
+ "loss": 0.8705,
14
+ "step": 500
15
+ },
16
+ {
17
+ "epoch": 0.01,
18
+ "learning_rate": 4.981654540317817e-05,
19
+ "loss": 0.877,
20
+ "step": 1000
21
+ },
22
+ {
23
+ "epoch": 0.02,
24
+ "learning_rate": 4.9724818104767253e-05,
25
+ "loss": 0.8699,
26
+ "step": 1500
27
+ },
28
+ {
29
+ "epoch": 0.02,
30
+ "learning_rate": 4.9633090806356335e-05,
31
+ "loss": 0.8724,
32
+ "step": 2000
33
+ },
34
+ {
35
+ "epoch": 0.03,
36
+ "learning_rate": 4.954136350794542e-05,
37
+ "loss": 0.875,
38
+ "step": 2500
39
+ },
40
+ {
41
+ "epoch": 0.03,
42
+ "learning_rate": 4.9449636209534504e-05,
43
+ "loss": 0.8736,
44
+ "step": 3000
45
+ },
46
+ {
47
+ "epoch": 0.04,
48
+ "learning_rate": 4.935790891112359e-05,
49
+ "loss": 0.8723,
50
+ "step": 3500
51
+ },
52
+ {
53
+ "epoch": 0.04,
54
+ "learning_rate": 4.926618161271267e-05,
55
+ "loss": 0.8776,
56
+ "step": 4000
57
+ },
58
+ {
59
+ "epoch": 0.05,
60
+ "learning_rate": 4.9174454314301755e-05,
61
+ "loss": 0.8679,
62
+ "step": 4500
63
+ },
64
+ {
65
+ "epoch": 0.06,
66
+ "learning_rate": 4.908272701589084e-05,
67
+ "loss": 0.8635,
68
+ "step": 5000
69
+ },
70
+ {
71
+ "epoch": 0.06,
72
+ "learning_rate": 4.8990999717479925e-05,
73
+ "loss": 0.863,
74
+ "step": 5500
75
+ },
76
+ {
77
+ "epoch": 0.07,
78
+ "learning_rate": 4.8899272419069006e-05,
79
+ "loss": 0.8591,
80
+ "step": 6000
81
+ },
82
+ {
83
+ "epoch": 0.07,
84
+ "learning_rate": 4.880754512065809e-05,
85
+ "loss": 0.8687,
86
+ "step": 6500
87
+ },
88
+ {
89
+ "epoch": 0.08,
90
+ "learning_rate": 4.8715817822247176e-05,
91
+ "loss": 0.8716,
92
+ "step": 7000
93
+ },
94
+ {
95
+ "epoch": 0.08,
96
+ "learning_rate": 4.862409052383626e-05,
97
+ "loss": 0.8662,
98
+ "step": 7500
99
+ },
100
+ {
101
+ "epoch": 0.09,
102
+ "learning_rate": 4.8532363225425346e-05,
103
+ "loss": 0.8644,
104
+ "step": 8000
105
+ },
106
+ {
107
+ "epoch": 0.09,
108
+ "learning_rate": 4.844063592701443e-05,
109
+ "loss": 0.8674,
110
+ "step": 8500
111
+ },
112
+ {
113
+ "epoch": 0.1,
114
+ "learning_rate": 4.8348908628603515e-05,
115
+ "loss": 0.8583,
116
+ "step": 9000
117
+ },
118
+ {
119
+ "epoch": 0.1,
120
+ "learning_rate": 4.825718133019259e-05,
121
+ "loss": 0.8552,
122
+ "step": 9500
123
+ },
124
+ {
125
+ "epoch": 0.11,
126
+ "learning_rate": 4.816545403178168e-05,
127
+ "loss": 0.8537,
128
+ "step": 10000
129
+ },
130
+ {
131
+ "epoch": 0.12,
132
+ "learning_rate": 4.807372673337076e-05,
133
+ "loss": 0.8591,
134
+ "step": 10500
135
+ },
136
+ {
137
+ "epoch": 0.12,
138
+ "learning_rate": 4.798199943495985e-05,
139
+ "loss": 0.8513,
140
+ "step": 11000
141
+ },
142
+ {
143
+ "epoch": 0.13,
144
+ "learning_rate": 4.789027213654893e-05,
145
+ "loss": 0.8655,
146
+ "step": 11500
147
+ },
148
+ {
149
+ "epoch": 0.13,
150
+ "learning_rate": 4.779854483813801e-05,
151
+ "loss": 0.8571,
152
+ "step": 12000
153
+ },
154
+ {
155
+ "epoch": 0.14,
156
+ "learning_rate": 4.770681753972709e-05,
157
+ "loss": 0.8558,
158
+ "step": 12500
159
+ },
160
+ {
161
+ "epoch": 0.14,
162
+ "learning_rate": 4.761509024131618e-05,
163
+ "loss": 0.8371,
164
+ "step": 13000
165
+ },
166
+ {
167
+ "epoch": 0.15,
168
+ "learning_rate": 4.752336294290526e-05,
169
+ "loss": 0.8426,
170
+ "step": 13500
171
+ },
172
+ {
173
+ "epoch": 0.15,
174
+ "learning_rate": 4.743163564449435e-05,
175
+ "loss": 0.8498,
176
+ "step": 14000
177
+ },
178
+ {
179
+ "epoch": 0.16,
180
+ "learning_rate": 4.733990834608343e-05,
181
+ "loss": 0.8612,
182
+ "step": 14500
183
+ },
184
+ {
185
+ "epoch": 0.17,
186
+ "learning_rate": 4.724818104767251e-05,
187
+ "loss": 0.8525,
188
+ "step": 15000
189
+ },
190
+ {
191
+ "epoch": 0.17,
192
+ "learning_rate": 4.71564537492616e-05,
193
+ "loss": 0.844,
194
+ "step": 15500
195
+ },
196
+ {
197
+ "epoch": 0.18,
198
+ "learning_rate": 4.706472645085068e-05,
199
+ "loss": 0.8655,
200
+ "step": 16000
201
+ },
202
+ {
203
+ "epoch": 0.18,
204
+ "learning_rate": 4.697299915243977e-05,
205
+ "loss": 0.8448,
206
+ "step": 16500
207
+ },
208
+ {
209
+ "epoch": 0.19,
210
+ "learning_rate": 4.6881271854028845e-05,
211
+ "loss": 0.8461,
212
+ "step": 17000
213
+ },
214
+ {
215
+ "epoch": 0.19,
216
+ "learning_rate": 4.678954455561793e-05,
217
+ "loss": 0.8478,
218
+ "step": 17500
219
+ },
220
+ {
221
+ "epoch": 0.2,
222
+ "learning_rate": 4.6697817257207015e-05,
223
+ "loss": 0.8457,
224
+ "step": 18000
225
+ },
226
+ {
227
+ "epoch": 0.2,
228
+ "learning_rate": 4.66060899587961e-05,
229
+ "loss": 0.8446,
230
+ "step": 18500
231
+ },
232
+ {
233
+ "epoch": 0.21,
234
+ "learning_rate": 4.6514362660385184e-05,
235
+ "loss": 0.8412,
236
+ "step": 19000
237
+ },
238
+ {
239
+ "epoch": 0.21,
240
+ "learning_rate": 4.6422635361974266e-05,
241
+ "loss": 0.8327,
242
+ "step": 19500
243
+ },
244
+ {
245
+ "epoch": 0.22,
246
+ "learning_rate": 4.633090806356335e-05,
247
+ "loss": 0.8374,
248
+ "step": 20000
249
+ },
250
+ {
251
+ "epoch": 0.23,
252
+ "learning_rate": 4.6239180765152435e-05,
253
+ "loss": 0.8418,
254
+ "step": 20500
255
+ },
256
+ {
257
+ "epoch": 0.23,
258
+ "learning_rate": 4.614745346674152e-05,
259
+ "loss": 0.8402,
260
+ "step": 21000
261
+ },
262
+ {
263
+ "epoch": 0.24,
264
+ "learning_rate": 4.6055726168330605e-05,
265
+ "loss": 0.8362,
266
+ "step": 21500
267
+ },
268
+ {
269
+ "epoch": 0.24,
270
+ "learning_rate": 4.5963998869919686e-05,
271
+ "loss": 0.8493,
272
+ "step": 22000
273
+ },
274
+ {
275
+ "epoch": 0.25,
276
+ "learning_rate": 4.587227157150877e-05,
277
+ "loss": 0.8413,
278
+ "step": 22500
279
+ },
280
+ {
281
+ "epoch": 0.25,
282
+ "learning_rate": 4.5780544273097856e-05,
283
+ "loss": 0.842,
284
+ "step": 23000
285
+ },
286
+ {
287
+ "epoch": 0.26,
288
+ "learning_rate": 4.568881697468694e-05,
289
+ "loss": 0.8386,
290
+ "step": 23500
291
+ },
292
+ {
293
+ "epoch": 0.26,
294
+ "learning_rate": 4.5597089676276025e-05,
295
+ "loss": 0.8402,
296
+ "step": 24000
297
+ },
298
+ {
299
+ "epoch": 0.27,
300
+ "learning_rate": 4.55053623778651e-05,
301
+ "loss": 0.8378,
302
+ "step": 24500
303
+ },
304
+ {
305
+ "epoch": 0.28,
306
+ "learning_rate": 4.541363507945419e-05,
307
+ "loss": 0.8254,
308
+ "step": 25000
309
+ },
310
+ {
311
+ "epoch": 0.28,
312
+ "learning_rate": 4.532190778104327e-05,
313
+ "loss": 0.8344,
314
+ "step": 25500
315
+ },
316
+ {
317
+ "epoch": 0.29,
318
+ "learning_rate": 4.523018048263236e-05,
319
+ "loss": 0.8399,
320
+ "step": 26000
321
+ },
322
+ {
323
+ "epoch": 0.29,
324
+ "learning_rate": 4.513845318422144e-05,
325
+ "loss": 0.8303,
326
+ "step": 26500
327
+ },
328
+ {
329
+ "epoch": 0.3,
330
+ "learning_rate": 4.504672588581053e-05,
331
+ "loss": 0.8334,
332
+ "step": 27000
333
+ },
334
+ {
335
+ "epoch": 0.3,
336
+ "learning_rate": 4.49549985873996e-05,
337
+ "loss": 0.8334,
338
+ "step": 27500
339
+ },
340
+ {
341
+ "epoch": 0.31,
342
+ "learning_rate": 4.486327128898869e-05,
343
+ "loss": 0.8407,
344
+ "step": 28000
345
+ },
346
+ {
347
+ "epoch": 0.31,
348
+ "learning_rate": 4.477154399057777e-05,
349
+ "loss": 0.8347,
350
+ "step": 28500
351
+ },
352
+ {
353
+ "epoch": 0.32,
354
+ "learning_rate": 4.467981669216686e-05,
355
+ "loss": 0.8354,
356
+ "step": 29000
357
+ },
358
+ {
359
+ "epoch": 0.32,
360
+ "learning_rate": 4.458808939375594e-05,
361
+ "loss": 0.8304,
362
+ "step": 29500
363
+ },
364
+ {
365
+ "epoch": 0.33,
366
+ "learning_rate": 4.449636209534502e-05,
367
+ "loss": 0.8282,
368
+ "step": 30000
369
+ },
370
+ {
371
+ "epoch": 0.34,
372
+ "learning_rate": 4.440463479693411e-05,
373
+ "loss": 0.8293,
374
+ "step": 30500
375
+ },
376
+ {
377
+ "epoch": 0.34,
378
+ "learning_rate": 4.431290749852319e-05,
379
+ "loss": 0.8267,
380
+ "step": 31000
381
+ },
382
+ {
383
+ "epoch": 0.35,
384
+ "learning_rate": 4.422118020011228e-05,
385
+ "loss": 0.8226,
386
+ "step": 31500
387
+ },
388
+ {
389
+ "epoch": 0.35,
390
+ "learning_rate": 4.412945290170136e-05,
391
+ "loss": 0.834,
392
+ "step": 32000
393
+ },
394
+ {
395
+ "epoch": 0.36,
396
+ "learning_rate": 4.403772560329044e-05,
397
+ "loss": 0.8332,
398
+ "step": 32500
399
+ },
400
+ {
401
+ "epoch": 0.36,
402
+ "learning_rate": 4.3945998304879525e-05,
403
+ "loss": 0.8248,
404
+ "step": 33000
405
+ },
406
+ {
407
+ "epoch": 0.37,
408
+ "learning_rate": 4.385427100646861e-05,
409
+ "loss": 0.8275,
410
+ "step": 33500
411
+ },
412
+ {
413
+ "epoch": 0.37,
414
+ "learning_rate": 4.3762543708057694e-05,
415
+ "loss": 0.822,
416
+ "step": 34000
417
+ },
418
+ {
419
+ "epoch": 0.38,
420
+ "learning_rate": 4.367081640964678e-05,
421
+ "loss": 0.8222,
422
+ "step": 34500
423
+ },
424
+ {
425
+ "epoch": 0.39,
426
+ "learning_rate": 4.357908911123586e-05,
427
+ "loss": 0.8194,
428
+ "step": 35000
429
+ },
430
+ {
431
+ "epoch": 0.39,
432
+ "learning_rate": 4.3487361812824945e-05,
433
+ "loss": 0.8213,
434
+ "step": 35500
435
+ },
436
+ {
437
+ "epoch": 0.4,
438
+ "learning_rate": 4.339563451441403e-05,
439
+ "loss": 0.8315,
440
+ "step": 36000
441
+ },
442
+ {
443
+ "epoch": 0.4,
444
+ "learning_rate": 4.3303907216003115e-05,
445
+ "loss": 0.8329,
446
+ "step": 36500
447
+ },
448
+ {
449
+ "epoch": 0.41,
450
+ "learning_rate": 4.3212179917592196e-05,
451
+ "loss": 0.8219,
452
+ "step": 37000
453
+ },
454
+ {
455
+ "epoch": 0.41,
456
+ "learning_rate": 4.312045261918128e-05,
457
+ "loss": 0.8298,
458
+ "step": 37500
459
+ },
460
+ {
461
+ "epoch": 0.42,
462
+ "learning_rate": 4.3028725320770366e-05,
463
+ "loss": 0.8178,
464
+ "step": 38000
465
+ },
466
+ {
467
+ "epoch": 0.42,
468
+ "learning_rate": 4.293699802235945e-05,
469
+ "loss": 0.8185,
470
+ "step": 38500
471
+ },
472
+ {
473
+ "epoch": 0.43,
474
+ "learning_rate": 4.2845270723948536e-05,
475
+ "loss": 0.8076,
476
+ "step": 39000
477
+ },
478
+ {
479
+ "epoch": 0.43,
480
+ "learning_rate": 4.275354342553762e-05,
481
+ "loss": 0.8174,
482
+ "step": 39500
483
+ },
484
+ {
485
+ "epoch": 0.44,
486
+ "learning_rate": 4.26618161271267e-05,
487
+ "loss": 0.8256,
488
+ "step": 40000
489
+ },
490
+ {
491
+ "epoch": 0.45,
492
+ "learning_rate": 4.257008882871578e-05,
493
+ "loss": 0.8134,
494
+ "step": 40500
495
+ },
496
+ {
497
+ "epoch": 0.45,
498
+ "learning_rate": 4.247836153030487e-05,
499
+ "loss": 0.8197,
500
+ "step": 41000
501
+ },
502
+ {
503
+ "epoch": 0.46,
504
+ "learning_rate": 4.238663423189395e-05,
505
+ "loss": 0.8144,
506
+ "step": 41500
507
+ },
508
+ {
509
+ "epoch": 0.46,
510
+ "learning_rate": 4.229490693348304e-05,
511
+ "loss": 0.8135,
512
+ "step": 42000
513
+ },
514
+ {
515
+ "epoch": 0.47,
516
+ "learning_rate": 4.220317963507211e-05,
517
+ "loss": 0.8183,
518
+ "step": 42500
519
+ },
520
+ {
521
+ "epoch": 0.47,
522
+ "learning_rate": 4.21114523366612e-05,
523
+ "loss": 0.8145,
524
+ "step": 43000
525
+ },
526
+ {
527
+ "epoch": 0.48,
528
+ "learning_rate": 4.201972503825028e-05,
529
+ "loss": 0.8237,
530
+ "step": 43500
531
+ },
532
+ {
533
+ "epoch": 0.48,
534
+ "learning_rate": 4.192799773983937e-05,
535
+ "loss": 0.8147,
536
+ "step": 44000
537
+ },
538
+ {
539
+ "epoch": 0.49,
540
+ "learning_rate": 4.183627044142845e-05,
541
+ "loss": 0.815,
542
+ "step": 44500
543
+ },
544
+ {
545
+ "epoch": 0.5,
546
+ "learning_rate": 4.174454314301754e-05,
547
+ "loss": 0.8131,
548
+ "step": 45000
549
+ },
550
+ {
551
+ "epoch": 0.5,
552
+ "learning_rate": 4.165281584460662e-05,
553
+ "loss": 0.8129,
554
+ "step": 45500
555
+ },
556
+ {
557
+ "epoch": 0.51,
558
+ "learning_rate": 4.15610885461957e-05,
559
+ "loss": 0.814,
560
+ "step": 46000
561
+ },
562
+ {
563
+ "epoch": 0.51,
564
+ "learning_rate": 4.146936124778479e-05,
565
+ "loss": 0.8066,
566
+ "step": 46500
567
+ },
568
+ {
569
+ "epoch": 0.52,
570
+ "learning_rate": 4.137763394937387e-05,
571
+ "loss": 0.8027,
572
+ "step": 47000
573
+ },
574
+ {
575
+ "epoch": 0.52,
576
+ "learning_rate": 4.128590665096296e-05,
577
+ "loss": 0.8055,
578
+ "step": 47500
579
+ },
580
+ {
581
+ "epoch": 0.53,
582
+ "learning_rate": 4.1194179352552035e-05,
583
+ "loss": 0.8104,
584
+ "step": 48000
585
+ },
586
+ {
587
+ "epoch": 0.53,
588
+ "learning_rate": 4.110245205414112e-05,
589
+ "loss": 0.798,
590
+ "step": 48500
591
+ },
592
+ {
593
+ "epoch": 0.54,
594
+ "learning_rate": 4.1010724755730205e-05,
595
+ "loss": 0.8099,
596
+ "step": 49000
597
+ },
598
+ {
599
+ "epoch": 0.54,
600
+ "learning_rate": 4.091899745731929e-05,
601
+ "loss": 0.8088,
602
+ "step": 49500
603
+ },
604
+ {
605
+ "epoch": 0.55,
606
+ "learning_rate": 4.0827270158908374e-05,
607
+ "loss": 0.8061,
608
+ "step": 50000
609
+ },
610
+ {
611
+ "epoch": 0.56,
612
+ "learning_rate": 4.0735542860497456e-05,
613
+ "loss": 0.8075,
614
+ "step": 50500
615
+ },
616
+ {
617
+ "epoch": 0.56,
618
+ "learning_rate": 4.064381556208654e-05,
619
+ "loss": 0.7942,
620
+ "step": 51000
621
+ },
622
+ {
623
+ "epoch": 0.57,
624
+ "learning_rate": 4.0552088263675625e-05,
625
+ "loss": 0.8103,
626
+ "step": 51500
627
+ },
628
+ {
629
+ "epoch": 0.57,
630
+ "learning_rate": 4.0460360965264707e-05,
631
+ "loss": 0.8129,
632
+ "step": 52000
633
+ },
634
+ {
635
+ "epoch": 0.58,
636
+ "learning_rate": 4.0368633666853795e-05,
637
+ "loss": 0.8072,
638
+ "step": 52500
639
+ },
640
+ {
641
+ "epoch": 0.58,
642
+ "learning_rate": 4.0276906368442876e-05,
643
+ "loss": 0.8036,
644
+ "step": 53000
645
+ },
646
+ {
647
+ "epoch": 0.59,
648
+ "learning_rate": 4.018517907003196e-05,
649
+ "loss": 0.8049,
650
+ "step": 53500
651
+ },
652
+ {
653
+ "epoch": 0.59,
654
+ "learning_rate": 4.0093451771621046e-05,
655
+ "loss": 0.8079,
656
+ "step": 54000
657
+ },
658
+ {
659
+ "epoch": 0.6,
660
+ "learning_rate": 4.000172447321013e-05,
661
+ "loss": 0.8005,
662
+ "step": 54500
663
+ },
664
+ {
665
+ "epoch": 0.61,
666
+ "learning_rate": 3.9909997174799215e-05,
667
+ "loss": 0.8023,
668
+ "step": 55000
669
+ },
670
+ {
671
+ "epoch": 0.61,
672
+ "learning_rate": 3.981826987638829e-05,
673
+ "loss": 0.8062,
674
+ "step": 55500
675
+ },
676
+ {
677
+ "epoch": 0.62,
678
+ "learning_rate": 3.972654257797738e-05,
679
+ "loss": 0.8024,
680
+ "step": 56000
681
+ },
682
+ {
683
+ "epoch": 0.62,
684
+ "learning_rate": 3.963481527956646e-05,
685
+ "loss": 0.7952,
686
+ "step": 56500
687
+ },
688
+ {
689
+ "epoch": 0.63,
690
+ "learning_rate": 3.954308798115555e-05,
691
+ "loss": 0.8003,
692
+ "step": 57000
693
+ },
694
+ {
695
+ "epoch": 0.63,
696
+ "learning_rate": 3.945136068274463e-05,
697
+ "loss": 0.7995,
698
+ "step": 57500
699
+ },
700
+ {
701
+ "epoch": 0.64,
702
+ "learning_rate": 3.935963338433371e-05,
703
+ "loss": 0.8014,
704
+ "step": 58000
705
+ },
706
+ {
707
+ "epoch": 0.64,
708
+ "learning_rate": 3.926790608592279e-05,
709
+ "loss": 0.795,
710
+ "step": 58500
711
+ },
712
+ {
713
+ "epoch": 0.65,
714
+ "learning_rate": 3.917617878751188e-05,
715
+ "loss": 0.8074,
716
+ "step": 59000
717
+ },
718
+ {
719
+ "epoch": 0.65,
720
+ "learning_rate": 3.908445148910096e-05,
721
+ "loss": 0.7955,
722
+ "step": 59500
723
+ },
724
+ {
725
+ "epoch": 0.66,
726
+ "learning_rate": 3.899272419069005e-05,
727
+ "loss": 0.7974,
728
+ "step": 60000
729
+ },
730
+ {
731
+ "epoch": 0.67,
732
+ "learning_rate": 3.890099689227913e-05,
733
+ "loss": 0.8003,
734
+ "step": 60500
735
+ },
736
+ {
737
+ "epoch": 0.67,
738
+ "learning_rate": 3.880926959386821e-05,
739
+ "loss": 0.7929,
740
+ "step": 61000
741
+ },
742
+ {
743
+ "epoch": 0.68,
744
+ "learning_rate": 3.87175422954573e-05,
745
+ "loss": 0.7987,
746
+ "step": 61500
747
+ },
748
+ {
749
+ "epoch": 0.68,
750
+ "learning_rate": 3.862581499704638e-05,
751
+ "loss": 0.7923,
752
+ "step": 62000
753
+ },
754
+ {
755
+ "epoch": 0.69,
756
+ "learning_rate": 3.853408769863547e-05,
757
+ "loss": 0.7888,
758
+ "step": 62500
759
+ },
760
+ {
761
+ "epoch": 0.69,
762
+ "learning_rate": 3.844236040022455e-05,
763
+ "loss": 0.7967,
764
+ "step": 63000
765
+ },
766
+ {
767
+ "epoch": 0.7,
768
+ "learning_rate": 3.835063310181363e-05,
769
+ "loss": 0.7941,
770
+ "step": 63500
771
+ },
772
+ {
773
+ "epoch": 0.7,
774
+ "learning_rate": 3.8258905803402715e-05,
775
+ "loss": 0.7964,
776
+ "step": 64000
777
+ },
778
+ {
779
+ "epoch": 0.71,
780
+ "learning_rate": 3.81671785049918e-05,
781
+ "loss": 0.7906,
782
+ "step": 64500
783
+ },
784
+ {
785
+ "epoch": 0.72,
786
+ "learning_rate": 3.8075451206580884e-05,
787
+ "loss": 0.8005,
788
+ "step": 65000
789
+ },
790
+ {
791
+ "epoch": 0.72,
792
+ "learning_rate": 3.798372390816997e-05,
793
+ "loss": 0.8007,
794
+ "step": 65500
795
+ },
796
+ {
797
+ "epoch": 0.73,
798
+ "learning_rate": 3.7891996609759054e-05,
799
+ "loss": 0.7897,
800
+ "step": 66000
801
+ },
802
+ {
803
+ "epoch": 0.73,
804
+ "learning_rate": 3.7800269311348135e-05,
805
+ "loss": 0.797,
806
+ "step": 66500
807
+ },
808
+ {
809
+ "epoch": 0.74,
810
+ "learning_rate": 3.7708542012937224e-05,
811
+ "loss": 0.7811,
812
+ "step": 67000
813
+ },
814
+ {
815
+ "epoch": 0.74,
816
+ "learning_rate": 3.7616814714526305e-05,
817
+ "loss": 0.7853,
818
+ "step": 67500
819
+ },
820
+ {
821
+ "epoch": 0.75,
822
+ "learning_rate": 3.752508741611539e-05,
823
+ "loss": 0.7939,
824
+ "step": 68000
825
+ },
826
+ {
827
+ "epoch": 0.75,
828
+ "learning_rate": 3.743336011770447e-05,
829
+ "loss": 0.7885,
830
+ "step": 68500
831
+ },
832
+ {
833
+ "epoch": 0.76,
834
+ "learning_rate": 3.7341632819293556e-05,
835
+ "loss": 0.7942,
836
+ "step": 69000
837
+ },
838
+ {
839
+ "epoch": 0.77,
840
+ "learning_rate": 3.724990552088264e-05,
841
+ "loss": 0.7861,
842
+ "step": 69500
843
+ },
844
+ {
845
+ "epoch": 0.77,
846
+ "learning_rate": 3.7158178222471726e-05,
847
+ "loss": 0.7837,
848
+ "step": 70000
849
+ },
850
+ {
851
+ "epoch": 0.78,
852
+ "learning_rate": 3.706645092406081e-05,
853
+ "loss": 0.7832,
854
+ "step": 70500
855
+ },
856
+ {
857
+ "epoch": 0.78,
858
+ "learning_rate": 3.697472362564989e-05,
859
+ "loss": 0.7834,
860
+ "step": 71000
861
+ },
862
+ {
863
+ "epoch": 0.79,
864
+ "learning_rate": 3.688299632723897e-05,
865
+ "loss": 0.7818,
866
+ "step": 71500
867
+ },
868
+ {
869
+ "epoch": 0.79,
870
+ "learning_rate": 3.679126902882806e-05,
871
+ "loss": 0.795,
872
+ "step": 72000
873
+ },
874
+ {
875
+ "epoch": 0.8,
876
+ "learning_rate": 3.669954173041714e-05,
877
+ "loss": 0.7933,
878
+ "step": 72500
879
+ },
880
+ {
881
+ "epoch": 0.8,
882
+ "learning_rate": 3.660781443200623e-05,
883
+ "loss": 0.7812,
884
+ "step": 73000
885
+ },
886
+ {
887
+ "epoch": 0.81,
888
+ "learning_rate": 3.651608713359531e-05,
889
+ "loss": 0.7944,
890
+ "step": 73500
891
+ },
892
+ {
893
+ "epoch": 0.81,
894
+ "learning_rate": 3.642435983518439e-05,
895
+ "loss": 0.7979,
896
+ "step": 74000
897
+ },
898
+ {
899
+ "epoch": 0.82,
900
+ "learning_rate": 3.633263253677348e-05,
901
+ "loss": 0.7912,
902
+ "step": 74500
903
+ },
904
+ {
905
+ "epoch": 0.83,
906
+ "learning_rate": 3.624090523836256e-05,
907
+ "loss": 0.791,
908
+ "step": 75000
909
+ },
910
+ {
911
+ "epoch": 0.83,
912
+ "learning_rate": 3.614917793995165e-05,
913
+ "loss": 0.7825,
914
+ "step": 75500
915
+ },
916
+ {
917
+ "epoch": 0.84,
918
+ "learning_rate": 3.605745064154072e-05,
919
+ "loss": 0.7833,
920
+ "step": 76000
921
+ },
922
+ {
923
+ "epoch": 0.84,
924
+ "learning_rate": 3.596572334312981e-05,
925
+ "loss": 0.7707,
926
+ "step": 76500
927
+ },
928
+ {
929
+ "epoch": 0.85,
930
+ "learning_rate": 3.587399604471889e-05,
931
+ "loss": 0.7888,
932
+ "step": 77000
933
+ },
934
+ {
935
+ "epoch": 0.85,
936
+ "learning_rate": 3.578226874630798e-05,
937
+ "loss": 0.7792,
938
+ "step": 77500
939
+ },
940
+ {
941
+ "epoch": 0.86,
942
+ "learning_rate": 3.569054144789706e-05,
943
+ "loss": 0.7874,
944
+ "step": 78000
945
+ },
946
+ {
947
+ "epoch": 0.86,
948
+ "learning_rate": 3.5598814149486144e-05,
949
+ "loss": 0.7805,
950
+ "step": 78500
951
+ },
952
+ {
953
+ "epoch": 0.87,
954
+ "learning_rate": 3.5507086851075225e-05,
955
+ "loss": 0.782,
956
+ "step": 79000
957
+ },
958
+ {
959
+ "epoch": 0.88,
960
+ "learning_rate": 3.541535955266431e-05,
961
+ "loss": 0.7809,
962
+ "step": 79500
963
+ },
964
+ {
965
+ "epoch": 0.88,
966
+ "learning_rate": 3.5323632254253395e-05,
967
+ "loss": 0.7795,
968
+ "step": 80000
969
+ },
970
+ {
971
+ "epoch": 0.89,
972
+ "learning_rate": 3.523190495584248e-05,
973
+ "loss": 0.7777,
974
+ "step": 80500
975
+ },
976
+ {
977
+ "epoch": 0.89,
978
+ "learning_rate": 3.5140177657431564e-05,
979
+ "loss": 0.7728,
980
+ "step": 81000
981
+ },
982
+ {
983
+ "epoch": 0.9,
984
+ "learning_rate": 3.5048450359020646e-05,
985
+ "loss": 0.7776,
986
+ "step": 81500
987
+ },
988
+ {
989
+ "epoch": 0.9,
990
+ "learning_rate": 3.4956723060609734e-05,
991
+ "loss": 0.7854,
992
+ "step": 82000
993
+ },
994
+ {
995
+ "epoch": 0.91,
996
+ "learning_rate": 3.4864995762198815e-05,
997
+ "loss": 0.7795,
998
+ "step": 82500
999
+ },
1000
+ {
1001
+ "epoch": 0.91,
1002
+ "learning_rate": 3.47732684637879e-05,
1003
+ "loss": 0.7808,
1004
+ "step": 83000
1005
+ },
1006
+ {
1007
+ "epoch": 0.92,
1008
+ "learning_rate": 3.4681541165376985e-05,
1009
+ "loss": 0.7779,
1010
+ "step": 83500
1011
+ },
1012
+ {
1013
+ "epoch": 0.92,
1014
+ "learning_rate": 3.4589813866966066e-05,
1015
+ "loss": 0.7738,
1016
+ "step": 84000
1017
+ },
1018
+ {
1019
+ "epoch": 0.93,
1020
+ "learning_rate": 3.449808656855515e-05,
1021
+ "loss": 0.7725,
1022
+ "step": 84500
1023
+ },
1024
+ {
1025
+ "epoch": 0.94,
1026
+ "learning_rate": 3.4406359270144236e-05,
1027
+ "loss": 0.7697,
1028
+ "step": 85000
1029
+ },
1030
+ {
1031
+ "epoch": 0.94,
1032
+ "learning_rate": 3.431463197173332e-05,
1033
+ "loss": 0.7743,
1034
+ "step": 85500
1035
+ },
1036
+ {
1037
+ "epoch": 0.95,
1038
+ "learning_rate": 3.4222904673322405e-05,
1039
+ "loss": 0.7807,
1040
+ "step": 86000
1041
+ },
1042
+ {
1043
+ "epoch": 0.95,
1044
+ "learning_rate": 3.413117737491148e-05,
1045
+ "loss": 0.7714,
1046
+ "step": 86500
1047
+ },
1048
+ {
1049
+ "epoch": 0.96,
1050
+ "learning_rate": 3.403945007650057e-05,
1051
+ "loss": 0.7727,
1052
+ "step": 87000
1053
+ },
1054
+ {
1055
+ "epoch": 0.96,
1056
+ "learning_rate": 3.394772277808965e-05,
1057
+ "loss": 0.772,
1058
+ "step": 87500
1059
+ },
1060
+ {
1061
+ "epoch": 0.97,
1062
+ "learning_rate": 3.385599547967874e-05,
1063
+ "loss": 0.7763,
1064
+ "step": 88000
1065
+ },
1066
+ {
1067
+ "epoch": 0.97,
1068
+ "learning_rate": 3.376426818126782e-05,
1069
+ "loss": 0.7669,
1070
+ "step": 88500
1071
+ },
1072
+ {
1073
+ "epoch": 0.98,
1074
+ "learning_rate": 3.36725408828569e-05,
1075
+ "loss": 0.7712,
1076
+ "step": 89000
1077
+ },
1078
+ {
1079
+ "epoch": 0.99,
1080
+ "learning_rate": 3.358081358444599e-05,
1081
+ "loss": 0.7717,
1082
+ "step": 89500
1083
+ },
1084
+ {
1085
+ "epoch": 0.99,
1086
+ "learning_rate": 3.348908628603507e-05,
1087
+ "loss": 0.7715,
1088
+ "step": 90000
1089
+ },
1090
+ {
1091
+ "epoch": 1.0,
1092
+ "learning_rate": 3.339735898762416e-05,
1093
+ "loss": 0.765,
1094
+ "step": 90500
1095
+ },
1096
+ {
1097
+ "epoch": 1.0,
1098
+ "learning_rate": 3.330563168921324e-05,
1099
+ "loss": 0.7689,
1100
+ "step": 91000
1101
+ },
1102
+ {
1103
+ "epoch": 1.01,
1104
+ "learning_rate": 3.321390439080232e-05,
1105
+ "loss": 0.7605,
1106
+ "step": 91500
1107
+ },
1108
+ {
1109
+ "epoch": 1.01,
1110
+ "learning_rate": 3.31221770923914e-05,
1111
+ "loss": 0.7668,
1112
+ "step": 92000
1113
+ },
1114
+ {
1115
+ "epoch": 1.02,
1116
+ "learning_rate": 3.303044979398049e-05,
1117
+ "loss": 0.7567,
1118
+ "step": 92500
1119
+ },
1120
+ {
1121
+ "epoch": 1.02,
1122
+ "learning_rate": 3.293872249556957e-05,
1123
+ "loss": 0.7614,
1124
+ "step": 93000
1125
+ },
1126
+ {
1127
+ "epoch": 1.03,
1128
+ "learning_rate": 3.284699519715866e-05,
1129
+ "loss": 0.7624,
1130
+ "step": 93500
1131
+ },
1132
+ {
1133
+ "epoch": 1.03,
1134
+ "learning_rate": 3.2755267898747735e-05,
1135
+ "loss": 0.7622,
1136
+ "step": 94000
1137
+ },
1138
+ {
1139
+ "epoch": 1.04,
1140
+ "learning_rate": 3.266354060033682e-05,
1141
+ "loss": 0.7649,
1142
+ "step": 94500
1143
+ },
1144
+ {
1145
+ "epoch": 1.05,
1146
+ "learning_rate": 3.2571813301925905e-05,
1147
+ "loss": 0.7574,
1148
+ "step": 95000
1149
+ },
1150
+ {
1151
+ "epoch": 1.05,
1152
+ "learning_rate": 3.248008600351499e-05,
1153
+ "loss": 0.7709,
1154
+ "step": 95500
1155
+ },
1156
+ {
1157
+ "epoch": 1.06,
1158
+ "learning_rate": 3.2388358705104074e-05,
1159
+ "loss": 0.7627,
1160
+ "step": 96000
1161
+ },
1162
+ {
1163
+ "epoch": 1.06,
1164
+ "learning_rate": 3.2296631406693156e-05,
1165
+ "loss": 0.7617,
1166
+ "step": 96500
1167
+ },
1168
+ {
1169
+ "epoch": 1.07,
1170
+ "learning_rate": 3.2204904108282244e-05,
1171
+ "loss": 0.7571,
1172
+ "step": 97000
1173
+ },
1174
+ {
1175
+ "epoch": 1.07,
1176
+ "learning_rate": 3.2113176809871325e-05,
1177
+ "loss": 0.7492,
1178
+ "step": 97500
1179
+ },
1180
+ {
1181
+ "epoch": 1.08,
1182
+ "learning_rate": 3.2021449511460414e-05,
1183
+ "loss": 0.764,
1184
+ "step": 98000
1185
+ },
1186
+ {
1187
+ "epoch": 1.08,
1188
+ "learning_rate": 3.1929722213049495e-05,
1189
+ "loss": 0.772,
1190
+ "step": 98500
1191
+ },
1192
+ {
1193
+ "epoch": 1.09,
1194
+ "learning_rate": 3.183799491463858e-05,
1195
+ "loss": 0.7645,
1196
+ "step": 99000
1197
+ },
1198
+ {
1199
+ "epoch": 1.1,
1200
+ "learning_rate": 3.174626761622766e-05,
1201
+ "loss": 0.7609,
1202
+ "step": 99500
1203
+ },
1204
+ {
1205
+ "epoch": 1.1,
1206
+ "learning_rate": 3.1654540317816746e-05,
1207
+ "loss": 0.7576,
1208
+ "step": 100000
1209
+ },
1210
+ {
1211
+ "epoch": 1.11,
1212
+ "learning_rate": 3.156281301940583e-05,
1213
+ "loss": 0.7515,
1214
+ "step": 100500
1215
+ },
1216
+ {
1217
+ "epoch": 1.11,
1218
+ "learning_rate": 3.1471085720994916e-05,
1219
+ "loss": 0.7522,
1220
+ "step": 101000
1221
+ },
1222
+ {
1223
+ "epoch": 1.12,
1224
+ "learning_rate": 3.1379358422584e-05,
1225
+ "loss": 0.7556,
1226
+ "step": 101500
1227
+ },
1228
+ {
1229
+ "epoch": 1.12,
1230
+ "learning_rate": 3.128763112417308e-05,
1231
+ "loss": 0.7622,
1232
+ "step": 102000
1233
+ },
1234
+ {
1235
+ "epoch": 1.13,
1236
+ "learning_rate": 3.119590382576216e-05,
1237
+ "loss": 0.7582,
1238
+ "step": 102500
1239
+ },
1240
+ {
1241
+ "epoch": 1.13,
1242
+ "learning_rate": 3.110417652735125e-05,
1243
+ "loss": 0.7507,
1244
+ "step": 103000
1245
+ },
1246
+ {
1247
+ "epoch": 1.14,
1248
+ "learning_rate": 3.101244922894033e-05,
1249
+ "loss": 0.7548,
1250
+ "step": 103500
1251
+ },
1252
+ {
1253
+ "epoch": 1.14,
1254
+ "learning_rate": 3.092072193052942e-05,
1255
+ "loss": 0.7588,
1256
+ "step": 104000
1257
+ },
1258
+ {
1259
+ "epoch": 1.15,
1260
+ "learning_rate": 3.08289946321185e-05,
1261
+ "loss": 0.7466,
1262
+ "step": 104500
1263
+ },
1264
+ {
1265
+ "epoch": 1.16,
1266
+ "learning_rate": 3.073726733370758e-05,
1267
+ "loss": 0.7546,
1268
+ "step": 105000
1269
+ },
1270
+ {
1271
+ "epoch": 1.16,
1272
+ "learning_rate": 3.064554003529667e-05,
1273
+ "loss": 0.7601,
1274
+ "step": 105500
1275
+ },
1276
+ {
1277
+ "epoch": 1.17,
1278
+ "learning_rate": 3.055381273688575e-05,
1279
+ "loss": 0.7544,
1280
+ "step": 106000
1281
+ },
1282
+ {
1283
+ "epoch": 1.17,
1284
+ "learning_rate": 3.0462085438474835e-05,
1285
+ "loss": 0.7531,
1286
+ "step": 106500
1287
+ },
1288
+ {
1289
+ "epoch": 1.18,
1290
+ "learning_rate": 3.0370358140063916e-05,
1291
+ "loss": 0.7519,
1292
+ "step": 107000
1293
+ },
1294
+ {
1295
+ "epoch": 1.18,
1296
+ "learning_rate": 3.0278630841653e-05,
1297
+ "loss": 0.7508,
1298
+ "step": 107500
1299
+ },
1300
+ {
1301
+ "epoch": 1.19,
1302
+ "learning_rate": 3.0186903543242086e-05,
1303
+ "loss": 0.7528,
1304
+ "step": 108000
1305
+ },
1306
+ {
1307
+ "epoch": 1.19,
1308
+ "learning_rate": 3.009517624483117e-05,
1309
+ "loss": 0.746,
1310
+ "step": 108500
1311
+ },
1312
+ {
1313
+ "epoch": 1.2,
1314
+ "learning_rate": 3.0003448946420255e-05,
1315
+ "loss": 0.7567,
1316
+ "step": 109000
1317
+ },
1318
+ {
1319
+ "epoch": 1.21,
1320
+ "learning_rate": 2.9911721648009333e-05,
1321
+ "loss": 0.7482,
1322
+ "step": 109500
1323
+ },
1324
+ {
1325
+ "epoch": 1.21,
1326
+ "learning_rate": 2.9819994349598418e-05,
1327
+ "loss": 0.7518,
1328
+ "step": 110000
1329
+ },
1330
+ {
1331
+ "epoch": 1.22,
1332
+ "learning_rate": 2.9728267051187503e-05,
1333
+ "loss": 0.7525,
1334
+ "step": 110500
1335
+ },
1336
+ {
1337
+ "epoch": 1.22,
1338
+ "learning_rate": 2.9636539752776588e-05,
1339
+ "loss": 0.7455,
1340
+ "step": 111000
1341
+ },
1342
+ {
1343
+ "epoch": 1.23,
1344
+ "learning_rate": 2.9544812454365673e-05,
1345
+ "loss": 0.7542,
1346
+ "step": 111500
1347
+ },
1348
+ {
1349
+ "epoch": 1.23,
1350
+ "learning_rate": 2.945308515595475e-05,
1351
+ "loss": 0.7445,
1352
+ "step": 112000
1353
+ },
1354
+ {
1355
+ "epoch": 1.24,
1356
+ "learning_rate": 2.9361357857543836e-05,
1357
+ "loss": 0.7586,
1358
+ "step": 112500
1359
+ },
1360
+ {
1361
+ "epoch": 1.24,
1362
+ "learning_rate": 2.926963055913292e-05,
1363
+ "loss": 0.7586,
1364
+ "step": 113000
1365
+ },
1366
+ {
1367
+ "epoch": 1.25,
1368
+ "learning_rate": 2.9177903260722005e-05,
1369
+ "loss": 0.7467,
1370
+ "step": 113500
1371
+ },
1372
+ {
1373
+ "epoch": 1.25,
1374
+ "learning_rate": 2.908617596231109e-05,
1375
+ "loss": 0.7477,
1376
+ "step": 114000
1377
+ },
1378
+ {
1379
+ "epoch": 1.26,
1380
+ "learning_rate": 2.899444866390017e-05,
1381
+ "loss": 0.7552,
1382
+ "step": 114500
1383
+ },
1384
+ {
1385
+ "epoch": 1.27,
1386
+ "learning_rate": 2.8902721365489256e-05,
1387
+ "loss": 0.7335,
1388
+ "step": 115000
1389
+ },
1390
+ {
1391
+ "epoch": 1.27,
1392
+ "learning_rate": 2.881099406707834e-05,
1393
+ "loss": 0.76,
1394
+ "step": 115500
1395
+ },
1396
+ {
1397
+ "epoch": 1.28,
1398
+ "learning_rate": 2.8719266768667426e-05,
1399
+ "loss": 0.7453,
1400
+ "step": 116000
1401
+ },
1402
+ {
1403
+ "epoch": 1.28,
1404
+ "learning_rate": 2.862753947025651e-05,
1405
+ "loss": 0.7414,
1406
+ "step": 116500
1407
+ },
1408
+ {
1409
+ "epoch": 1.29,
1410
+ "learning_rate": 2.8535812171845595e-05,
1411
+ "loss": 0.7515,
1412
+ "step": 117000
1413
+ },
1414
+ {
1415
+ "epoch": 1.29,
1416
+ "learning_rate": 2.8444084873434673e-05,
1417
+ "loss": 0.7482,
1418
+ "step": 117500
1419
+ },
1420
+ {
1421
+ "epoch": 1.3,
1422
+ "learning_rate": 2.8352357575023758e-05,
1423
+ "loss": 0.7425,
1424
+ "step": 118000
1425
+ },
1426
+ {
1427
+ "epoch": 1.3,
1428
+ "learning_rate": 2.8260630276612843e-05,
1429
+ "loss": 0.7414,
1430
+ "step": 118500
1431
+ },
1432
+ {
1433
+ "epoch": 1.31,
1434
+ "learning_rate": 2.8168902978201928e-05,
1435
+ "loss": 0.7389,
1436
+ "step": 119000
1437
+ },
1438
+ {
1439
+ "epoch": 1.32,
1440
+ "learning_rate": 2.8077175679791013e-05,
1441
+ "loss": 0.7531,
1442
+ "step": 119500
1443
+ },
1444
+ {
1445
+ "epoch": 1.32,
1446
+ "learning_rate": 2.798544838138009e-05,
1447
+ "loss": 0.7423,
1448
+ "step": 120000
1449
+ },
1450
+ {
1451
+ "epoch": 1.33,
1452
+ "learning_rate": 2.7893721082969175e-05,
1453
+ "loss": 0.7497,
1454
+ "step": 120500
1455
+ },
1456
+ {
1457
+ "epoch": 1.33,
1458
+ "learning_rate": 2.780199378455826e-05,
1459
+ "loss": 0.7427,
1460
+ "step": 121000
1461
+ },
1462
+ {
1463
+ "epoch": 1.34,
1464
+ "learning_rate": 2.7710266486147345e-05,
1465
+ "loss": 0.7469,
1466
+ "step": 121500
1467
+ },
1468
+ {
1469
+ "epoch": 1.34,
1470
+ "learning_rate": 2.761853918773643e-05,
1471
+ "loss": 0.7423,
1472
+ "step": 122000
1473
+ },
1474
+ {
1475
+ "epoch": 1.35,
1476
+ "learning_rate": 2.752681188932551e-05,
1477
+ "loss": 0.7419,
1478
+ "step": 122500
1479
+ },
1480
+ {
1481
+ "epoch": 1.35,
1482
+ "learning_rate": 2.7435084590914596e-05,
1483
+ "loss": 0.7465,
1484
+ "step": 123000
1485
+ },
1486
+ {
1487
+ "epoch": 1.36,
1488
+ "learning_rate": 2.734335729250368e-05,
1489
+ "loss": 0.7327,
1490
+ "step": 123500
1491
+ },
1492
+ {
1493
+ "epoch": 1.36,
1494
+ "learning_rate": 2.7251629994092766e-05,
1495
+ "loss": 0.745,
1496
+ "step": 124000
1497
+ },
1498
+ {
1499
+ "epoch": 1.37,
1500
+ "learning_rate": 2.715990269568185e-05,
1501
+ "loss": 0.7372,
1502
+ "step": 124500
1503
+ },
1504
+ {
1505
+ "epoch": 1.38,
1506
+ "learning_rate": 2.706817539727093e-05,
1507
+ "loss": 0.7294,
1508
+ "step": 125000
1509
+ },
1510
+ {
1511
+ "epoch": 1.38,
1512
+ "learning_rate": 2.6976448098860013e-05,
1513
+ "loss": 0.7336,
1514
+ "step": 125500
1515
+ },
1516
+ {
1517
+ "epoch": 1.39,
1518
+ "learning_rate": 2.6884720800449098e-05,
1519
+ "loss": 0.7389,
1520
+ "step": 126000
1521
+ },
1522
+ {
1523
+ "epoch": 1.39,
1524
+ "learning_rate": 2.6792993502038183e-05,
1525
+ "loss": 0.7399,
1526
+ "step": 126500
1527
+ },
1528
+ {
1529
+ "epoch": 1.4,
1530
+ "learning_rate": 2.6701266203627268e-05,
1531
+ "loss": 0.7394,
1532
+ "step": 127000
1533
+ },
1534
+ {
1535
+ "epoch": 1.4,
1536
+ "learning_rate": 2.6609538905216346e-05,
1537
+ "loss": 0.7394,
1538
+ "step": 127500
1539
+ },
1540
+ {
1541
+ "epoch": 1.41,
1542
+ "learning_rate": 2.651781160680543e-05,
1543
+ "loss": 0.7357,
1544
+ "step": 128000
1545
+ },
1546
+ {
1547
+ "epoch": 1.41,
1548
+ "learning_rate": 2.6426084308394515e-05,
1549
+ "loss": 0.7528,
1550
+ "step": 128500
1551
+ },
1552
+ {
1553
+ "epoch": 1.42,
1554
+ "learning_rate": 2.63343570099836e-05,
1555
+ "loss": 0.7406,
1556
+ "step": 129000
1557
+ },
1558
+ {
1559
+ "epoch": 1.43,
1560
+ "learning_rate": 2.6242629711572685e-05,
1561
+ "loss": 0.7347,
1562
+ "step": 129500
1563
+ },
1564
+ {
1565
+ "epoch": 1.43,
1566
+ "learning_rate": 2.6150902413161766e-05,
1567
+ "loss": 0.7312,
1568
+ "step": 130000
1569
+ },
1570
+ {
1571
+ "epoch": 1.44,
1572
+ "learning_rate": 2.605917511475085e-05,
1573
+ "loss": 0.7334,
1574
+ "step": 130500
1575
+ },
1576
+ {
1577
+ "epoch": 1.44,
1578
+ "learning_rate": 2.5967447816339936e-05,
1579
+ "loss": 0.7403,
1580
+ "step": 131000
1581
+ },
1582
+ {
1583
+ "epoch": 1.45,
1584
+ "learning_rate": 2.587572051792902e-05,
1585
+ "loss": 0.7366,
1586
+ "step": 131500
1587
+ },
1588
+ {
1589
+ "epoch": 1.45,
1590
+ "learning_rate": 2.5783993219518106e-05,
1591
+ "loss": 0.7271,
1592
+ "step": 132000
1593
+ },
1594
+ {
1595
+ "epoch": 1.46,
1596
+ "learning_rate": 2.5692265921107184e-05,
1597
+ "loss": 0.721,
1598
+ "step": 132500
1599
+ },
1600
+ {
1601
+ "epoch": 1.46,
1602
+ "learning_rate": 2.560053862269627e-05,
1603
+ "loss": 0.7259,
1604
+ "step": 133000
1605
+ },
1606
+ {
1607
+ "epoch": 1.47,
1608
+ "learning_rate": 2.5508811324285353e-05,
1609
+ "loss": 0.7371,
1610
+ "step": 133500
1611
+ },
1612
+ {
1613
+ "epoch": 1.47,
1614
+ "learning_rate": 2.5417084025874438e-05,
1615
+ "loss": 0.7276,
1616
+ "step": 134000
1617
+ },
1618
+ {
1619
+ "epoch": 1.48,
1620
+ "learning_rate": 2.5325356727463523e-05,
1621
+ "loss": 0.732,
1622
+ "step": 134500
1623
+ },
1624
+ {
1625
+ "epoch": 1.49,
1626
+ "learning_rate": 2.5233629429052608e-05,
1627
+ "loss": 0.7427,
1628
+ "step": 135000
1629
+ },
1630
+ {
1631
+ "epoch": 1.49,
1632
+ "learning_rate": 2.5141902130641686e-05,
1633
+ "loss": 0.7362,
1634
+ "step": 135500
1635
+ },
1636
+ {
1637
+ "epoch": 1.5,
1638
+ "learning_rate": 2.505017483223077e-05,
1639
+ "loss": 0.7344,
1640
+ "step": 136000
1641
+ },
1642
+ {
1643
+ "epoch": 1.5,
1644
+ "learning_rate": 2.4958447533819855e-05,
1645
+ "loss": 0.7248,
1646
+ "step": 136500
1647
+ },
1648
+ {
1649
+ "epoch": 1.51,
1650
+ "learning_rate": 2.486672023540894e-05,
1651
+ "loss": 0.7405,
1652
+ "step": 137000
1653
+ },
1654
+ {
1655
+ "epoch": 1.51,
1656
+ "learning_rate": 2.4774992936998025e-05,
1657
+ "loss": 0.7319,
1658
+ "step": 137500
1659
+ },
1660
+ {
1661
+ "epoch": 1.52,
1662
+ "learning_rate": 2.468326563858711e-05,
1663
+ "loss": 0.7354,
1664
+ "step": 138000
1665
+ },
1666
+ {
1667
+ "epoch": 1.52,
1668
+ "learning_rate": 2.459153834017619e-05,
1669
+ "loss": 0.733,
1670
+ "step": 138500
1671
+ },
1672
+ {
1673
+ "epoch": 1.53,
1674
+ "learning_rate": 2.4499811041765276e-05,
1675
+ "loss": 0.7373,
1676
+ "step": 139000
1677
+ },
1678
+ {
1679
+ "epoch": 1.54,
1680
+ "learning_rate": 2.440808374335436e-05,
1681
+ "loss": 0.7303,
1682
+ "step": 139500
1683
+ },
1684
+ {
1685
+ "epoch": 1.54,
1686
+ "learning_rate": 2.4316356444943442e-05,
1687
+ "loss": 0.721,
1688
+ "step": 140000
1689
+ },
1690
+ {
1691
+ "epoch": 1.55,
1692
+ "learning_rate": 2.4224629146532527e-05,
1693
+ "loss": 0.7325,
1694
+ "step": 140500
1695
+ },
1696
+ {
1697
+ "epoch": 1.55,
1698
+ "learning_rate": 2.4132901848121608e-05,
1699
+ "loss": 0.7225,
1700
+ "step": 141000
1701
+ },
1702
+ {
1703
+ "epoch": 1.56,
1704
+ "learning_rate": 2.4041174549710693e-05,
1705
+ "loss": 0.7313,
1706
+ "step": 141500
1707
+ },
1708
+ {
1709
+ "epoch": 1.56,
1710
+ "learning_rate": 2.3949447251299778e-05,
1711
+ "loss": 0.7372,
1712
+ "step": 142000
1713
+ },
1714
+ {
1715
+ "epoch": 1.57,
1716
+ "learning_rate": 2.385771995288886e-05,
1717
+ "loss": 0.7266,
1718
+ "step": 142500
1719
+ },
1720
+ {
1721
+ "epoch": 1.57,
1722
+ "learning_rate": 2.3765992654477944e-05,
1723
+ "loss": 0.7259,
1724
+ "step": 143000
1725
+ },
1726
+ {
1727
+ "epoch": 1.58,
1728
+ "learning_rate": 2.3674265356067025e-05,
1729
+ "loss": 0.725,
1730
+ "step": 143500
1731
+ },
1732
+ {
1733
+ "epoch": 1.59,
1734
+ "learning_rate": 2.358253805765611e-05,
1735
+ "loss": 0.7242,
1736
+ "step": 144000
1737
+ },
1738
+ {
1739
+ "epoch": 1.59,
1740
+ "learning_rate": 2.3490810759245195e-05,
1741
+ "loss": 0.729,
1742
+ "step": 144500
1743
+ },
1744
+ {
1745
+ "epoch": 1.6,
1746
+ "learning_rate": 2.339908346083428e-05,
1747
+ "loss": 0.7183,
1748
+ "step": 145000
1749
+ },
1750
+ {
1751
+ "epoch": 1.6,
1752
+ "learning_rate": 2.3307356162423365e-05,
1753
+ "loss": 0.7232,
1754
+ "step": 145500
1755
+ },
1756
+ {
1757
+ "epoch": 1.61,
1758
+ "learning_rate": 2.3215628864012446e-05,
1759
+ "loss": 0.722,
1760
+ "step": 146000
1761
+ },
1762
+ {
1763
+ "epoch": 1.61,
1764
+ "learning_rate": 2.312390156560153e-05,
1765
+ "loss": 0.7255,
1766
+ "step": 146500
1767
+ },
1768
+ {
1769
+ "epoch": 1.62,
1770
+ "learning_rate": 2.3032174267190616e-05,
1771
+ "loss": 0.7248,
1772
+ "step": 147000
1773
+ },
1774
+ {
1775
+ "epoch": 1.62,
1776
+ "learning_rate": 2.2940446968779697e-05,
1777
+ "loss": 0.7202,
1778
+ "step": 147500
1779
+ },
1780
+ {
1781
+ "epoch": 1.63,
1782
+ "learning_rate": 2.2848719670368782e-05,
1783
+ "loss": 0.7292,
1784
+ "step": 148000
1785
+ },
1786
+ {
1787
+ "epoch": 1.63,
1788
+ "learning_rate": 2.2756992371957867e-05,
1789
+ "loss": 0.7305,
1790
+ "step": 148500
1791
+ },
1792
+ {
1793
+ "epoch": 1.64,
1794
+ "learning_rate": 2.2665265073546948e-05,
1795
+ "loss": 0.7222,
1796
+ "step": 149000
1797
+ },
1798
+ {
1799
+ "epoch": 1.65,
1800
+ "learning_rate": 2.2573537775136033e-05,
1801
+ "loss": 0.7262,
1802
+ "step": 149500
1803
+ },
1804
+ {
1805
+ "epoch": 1.65,
1806
+ "learning_rate": 2.2481810476725114e-05,
1807
+ "loss": 0.7189,
1808
+ "step": 150000
1809
+ },
1810
+ {
1811
+ "epoch": 1.66,
1812
+ "learning_rate": 2.23900831783142e-05,
1813
+ "loss": 0.7277,
1814
+ "step": 150500
1815
+ },
1816
+ {
1817
+ "epoch": 1.66,
1818
+ "learning_rate": 2.2298355879903284e-05,
1819
+ "loss": 0.7338,
1820
+ "step": 151000
1821
+ },
1822
+ {
1823
+ "epoch": 1.67,
1824
+ "learning_rate": 2.2206628581492365e-05,
1825
+ "loss": 0.7227,
1826
+ "step": 151500
1827
+ },
1828
+ {
1829
+ "epoch": 1.67,
1830
+ "learning_rate": 2.211490128308145e-05,
1831
+ "loss": 0.7247,
1832
+ "step": 152000
1833
+ },
1834
+ {
1835
+ "epoch": 1.68,
1836
+ "learning_rate": 2.2023173984670535e-05,
1837
+ "loss": 0.7201,
1838
+ "step": 152500
1839
+ },
1840
+ {
1841
+ "epoch": 1.68,
1842
+ "learning_rate": 2.193144668625962e-05,
1843
+ "loss": 0.7294,
1844
+ "step": 153000
1845
+ },
1846
+ {
1847
+ "epoch": 1.69,
1848
+ "learning_rate": 2.1839719387848705e-05,
1849
+ "loss": 0.7258,
1850
+ "step": 153500
1851
+ },
1852
+ {
1853
+ "epoch": 1.7,
1854
+ "learning_rate": 2.1747992089437786e-05,
1855
+ "loss": 0.7239,
1856
+ "step": 154000
1857
+ },
1858
+ {
1859
+ "epoch": 1.7,
1860
+ "learning_rate": 2.165626479102687e-05,
1861
+ "loss": 0.7181,
1862
+ "step": 154500
1863
+ },
1864
+ {
1865
+ "epoch": 1.71,
1866
+ "learning_rate": 2.1564537492615952e-05,
1867
+ "loss": 0.7127,
1868
+ "step": 155000
1869
+ },
1870
+ {
1871
+ "epoch": 1.71,
1872
+ "learning_rate": 2.1472810194205037e-05,
1873
+ "loss": 0.7214,
1874
+ "step": 155500
1875
+ },
1876
+ {
1877
+ "epoch": 1.72,
1878
+ "learning_rate": 2.1381082895794122e-05,
1879
+ "loss": 0.7173,
1880
+ "step": 156000
1881
+ },
1882
+ {
1883
+ "epoch": 1.72,
1884
+ "learning_rate": 2.1289355597383203e-05,
1885
+ "loss": 0.7225,
1886
+ "step": 156500
1887
+ },
1888
+ {
1889
+ "epoch": 1.73,
1890
+ "learning_rate": 2.1197628298972288e-05,
1891
+ "loss": 0.7164,
1892
+ "step": 157000
1893
+ },
1894
+ {
1895
+ "epoch": 1.73,
1896
+ "learning_rate": 2.1105901000561373e-05,
1897
+ "loss": 0.7113,
1898
+ "step": 157500
1899
+ },
1900
+ {
1901
+ "epoch": 1.74,
1902
+ "learning_rate": 2.1014173702150454e-05,
1903
+ "loss": 0.7088,
1904
+ "step": 158000
1905
+ },
1906
+ {
1907
+ "epoch": 1.74,
1908
+ "learning_rate": 2.092244640373954e-05,
1909
+ "loss": 0.7178,
1910
+ "step": 158500
1911
+ },
1912
+ {
1913
+ "epoch": 1.75,
1914
+ "learning_rate": 2.083071910532862e-05,
1915
+ "loss": 0.7199,
1916
+ "step": 159000
1917
+ },
1918
+ {
1919
+ "epoch": 1.76,
1920
+ "learning_rate": 2.0738991806917705e-05,
1921
+ "loss": 0.7073,
1922
+ "step": 159500
1923
+ },
1924
+ {
1925
+ "epoch": 1.76,
1926
+ "learning_rate": 2.064726450850679e-05,
1927
+ "loss": 0.7113,
1928
+ "step": 160000
1929
+ },
1930
+ {
1931
+ "epoch": 1.77,
1932
+ "learning_rate": 2.0555537210095875e-05,
1933
+ "loss": 0.7121,
1934
+ "step": 160500
1935
+ },
1936
+ {
1937
+ "epoch": 1.77,
1938
+ "learning_rate": 2.046380991168496e-05,
1939
+ "loss": 0.714,
1940
+ "step": 161000
1941
+ },
1942
+ {
1943
+ "epoch": 1.78,
1944
+ "learning_rate": 2.037208261327404e-05,
1945
+ "loss": 0.7151,
1946
+ "step": 161500
1947
+ },
1948
+ {
1949
+ "epoch": 1.78,
1950
+ "learning_rate": 2.0280355314863126e-05,
1951
+ "loss": 0.7062,
1952
+ "step": 162000
1953
+ },
1954
+ {
1955
+ "epoch": 1.79,
1956
+ "learning_rate": 2.018862801645221e-05,
1957
+ "loss": 0.7053,
1958
+ "step": 162500
1959
+ },
1960
+ {
1961
+ "epoch": 1.79,
1962
+ "learning_rate": 2.0096900718041292e-05,
1963
+ "loss": 0.7097,
1964
+ "step": 163000
1965
+ },
1966
+ {
1967
+ "epoch": 1.8,
1968
+ "learning_rate": 2.0005173419630377e-05,
1969
+ "loss": 0.7139,
1970
+ "step": 163500
1971
+ },
1972
+ {
1973
+ "epoch": 1.81,
1974
+ "learning_rate": 1.9913446121219462e-05,
1975
+ "loss": 0.7163,
1976
+ "step": 164000
1977
+ },
1978
+ {
1979
+ "epoch": 1.81,
1980
+ "learning_rate": 1.9821718822808543e-05,
1981
+ "loss": 0.7031,
1982
+ "step": 164500
1983
+ },
1984
+ {
1985
+ "epoch": 1.82,
1986
+ "learning_rate": 1.9729991524397628e-05,
1987
+ "loss": 0.7035,
1988
+ "step": 165000
1989
+ },
1990
+ {
1991
+ "epoch": 1.82,
1992
+ "learning_rate": 1.963826422598671e-05,
1993
+ "loss": 0.7138,
1994
+ "step": 165500
1995
+ },
1996
+ {
1997
+ "epoch": 1.83,
1998
+ "learning_rate": 1.9546536927575794e-05,
1999
+ "loss": 0.72,
2000
+ "step": 166000
2001
+ },
2002
+ {
2003
+ "epoch": 1.83,
2004
+ "learning_rate": 1.945480962916488e-05,
2005
+ "loss": 0.7167,
2006
+ "step": 166500
2007
+ },
2008
+ {
2009
+ "epoch": 1.84,
2010
+ "learning_rate": 1.9363082330753964e-05,
2011
+ "loss": 0.7039,
2012
+ "step": 167000
2013
+ },
2014
+ {
2015
+ "epoch": 1.84,
2016
+ "learning_rate": 1.927135503234305e-05,
2017
+ "loss": 0.701,
2018
+ "step": 167500
2019
+ },
2020
+ {
2021
+ "epoch": 1.85,
2022
+ "learning_rate": 1.917962773393213e-05,
2023
+ "loss": 0.7144,
2024
+ "step": 168000
2025
+ },
2026
+ {
2027
+ "epoch": 1.85,
2028
+ "learning_rate": 1.9087900435521215e-05,
2029
+ "loss": 0.7003,
2030
+ "step": 168500
2031
+ },
2032
+ {
2033
+ "epoch": 1.86,
2034
+ "learning_rate": 1.89961731371103e-05,
2035
+ "loss": 0.7058,
2036
+ "step": 169000
2037
+ },
2038
+ {
2039
+ "epoch": 1.87,
2040
+ "learning_rate": 1.890444583869938e-05,
2041
+ "loss": 0.7048,
2042
+ "step": 169500
2043
+ },
2044
+ {
2045
+ "epoch": 1.87,
2046
+ "learning_rate": 1.8812718540288466e-05,
2047
+ "loss": 0.7109,
2048
+ "step": 170000
2049
+ },
2050
+ {
2051
+ "epoch": 1.88,
2052
+ "learning_rate": 1.8720991241877547e-05,
2053
+ "loss": 0.7145,
2054
+ "step": 170500
2055
+ },
2056
+ {
2057
+ "epoch": 1.88,
2058
+ "learning_rate": 1.8629263943466632e-05,
2059
+ "loss": 0.7051,
2060
+ "step": 171000
2061
+ },
2062
+ {
2063
+ "epoch": 1.89,
2064
+ "learning_rate": 1.8537536645055717e-05,
2065
+ "loss": 0.7063,
2066
+ "step": 171500
2067
+ },
2068
+ {
2069
+ "epoch": 1.89,
2070
+ "learning_rate": 1.8445809346644798e-05,
2071
+ "loss": 0.7145,
2072
+ "step": 172000
2073
+ },
2074
+ {
2075
+ "epoch": 1.9,
2076
+ "learning_rate": 1.8354082048233883e-05,
2077
+ "loss": 0.7064,
2078
+ "step": 172500
2079
+ },
2080
+ {
2081
+ "epoch": 1.9,
2082
+ "learning_rate": 1.8262354749822968e-05,
2083
+ "loss": 0.7164,
2084
+ "step": 173000
2085
+ },
2086
+ {
2087
+ "epoch": 1.91,
2088
+ "learning_rate": 1.817062745141205e-05,
2089
+ "loss": 0.7076,
2090
+ "step": 173500
2091
+ },
2092
+ {
2093
+ "epoch": 1.92,
2094
+ "learning_rate": 1.8078900153001134e-05,
2095
+ "loss": 0.7097,
2096
+ "step": 174000
2097
+ },
2098
+ {
2099
+ "epoch": 1.92,
2100
+ "learning_rate": 1.798717285459022e-05,
2101
+ "loss": 0.6915,
2102
+ "step": 174500
2103
+ },
2104
+ {
2105
+ "epoch": 1.93,
2106
+ "learning_rate": 1.7895445556179304e-05,
2107
+ "loss": 0.7079,
2108
+ "step": 175000
2109
+ },
2110
+ {
2111
+ "epoch": 1.93,
2112
+ "learning_rate": 1.780371825776839e-05,
2113
+ "loss": 0.7072,
2114
+ "step": 175500
2115
+ },
2116
+ {
2117
+ "epoch": 1.94,
2118
+ "learning_rate": 1.771199095935747e-05,
2119
+ "loss": 0.7064,
2120
+ "step": 176000
2121
+ },
2122
+ {
2123
+ "epoch": 1.94,
2124
+ "learning_rate": 1.7620263660946555e-05,
2125
+ "loss": 0.7004,
2126
+ "step": 176500
2127
+ },
2128
+ {
2129
+ "epoch": 1.95,
2130
+ "learning_rate": 1.7528536362535636e-05,
2131
+ "loss": 0.6992,
2132
+ "step": 177000
2133
+ },
2134
+ {
2135
+ "epoch": 1.95,
2136
+ "learning_rate": 1.743680906412472e-05,
2137
+ "loss": 0.7017,
2138
+ "step": 177500
2139
+ },
2140
+ {
2141
+ "epoch": 1.96,
2142
+ "learning_rate": 1.7345081765713806e-05,
2143
+ "loss": 0.6989,
2144
+ "step": 178000
2145
+ },
2146
+ {
2147
+ "epoch": 1.96,
2148
+ "learning_rate": 1.7253354467302887e-05,
2149
+ "loss": 0.6957,
2150
+ "step": 178500
2151
+ },
2152
+ {
2153
+ "epoch": 1.97,
2154
+ "learning_rate": 1.7161627168891972e-05,
2155
+ "loss": 0.6949,
2156
+ "step": 179000
2157
+ },
2158
+ {
2159
+ "epoch": 1.98,
2160
+ "learning_rate": 1.7069899870481053e-05,
2161
+ "loss": 0.6979,
2162
+ "step": 179500
2163
+ },
2164
+ {
2165
+ "epoch": 1.98,
2166
+ "learning_rate": 1.6978172572070138e-05,
2167
+ "loss": 0.7044,
2168
+ "step": 180000
2169
+ },
2170
+ {
2171
+ "epoch": 1.99,
2172
+ "learning_rate": 1.6886445273659223e-05,
2173
+ "loss": 0.6996,
2174
+ "step": 180500
2175
+ },
2176
+ {
2177
+ "epoch": 1.99,
2178
+ "learning_rate": 1.6794717975248304e-05,
2179
+ "loss": 0.6996,
2180
+ "step": 181000
2181
+ },
2182
+ {
2183
+ "epoch": 2.0,
2184
+ "learning_rate": 1.670299067683739e-05,
2185
+ "loss": 0.7085,
2186
+ "step": 181500
2187
+ },
2188
+ {
2189
+ "epoch": 2.0,
2190
+ "learning_rate": 1.6611263378426474e-05,
2191
+ "loss": 0.6935,
2192
+ "step": 182000
2193
+ },
2194
+ {
2195
+ "epoch": 2.01,
2196
+ "learning_rate": 1.651953608001556e-05,
2197
+ "loss": 0.6947,
2198
+ "step": 182500
2199
+ },
2200
+ {
2201
+ "epoch": 2.01,
2202
+ "learning_rate": 1.6427808781604644e-05,
2203
+ "loss": 0.6891,
2204
+ "step": 183000
2205
+ },
2206
+ {
2207
+ "epoch": 2.02,
2208
+ "learning_rate": 1.6336081483193725e-05,
2209
+ "loss": 0.6922,
2210
+ "step": 183500
2211
+ },
2212
+ {
2213
+ "epoch": 2.03,
2214
+ "learning_rate": 1.624435418478281e-05,
2215
+ "loss": 0.6943,
2216
+ "step": 184000
2217
+ },
2218
+ {
2219
+ "epoch": 2.03,
2220
+ "learning_rate": 1.6152626886371895e-05,
2221
+ "loss": 0.6984,
2222
+ "step": 184500
2223
+ },
2224
+ {
2225
+ "epoch": 2.04,
2226
+ "learning_rate": 1.6060899587960976e-05,
2227
+ "loss": 0.6979,
2228
+ "step": 185000
2229
+ },
2230
+ {
2231
+ "epoch": 2.04,
2232
+ "learning_rate": 1.596917228955006e-05,
2233
+ "loss": 0.6936,
2234
+ "step": 185500
2235
+ },
2236
+ {
2237
+ "epoch": 2.05,
2238
+ "learning_rate": 1.5877444991139142e-05,
2239
+ "loss": 0.6989,
2240
+ "step": 186000
2241
+ },
2242
+ {
2243
+ "epoch": 2.05,
2244
+ "learning_rate": 1.5785717692728227e-05,
2245
+ "loss": 0.6923,
2246
+ "step": 186500
2247
+ },
2248
+ {
2249
+ "epoch": 2.06,
2250
+ "learning_rate": 1.5693990394317312e-05,
2251
+ "loss": 0.6899,
2252
+ "step": 187000
2253
+ },
2254
+ {
2255
+ "epoch": 2.06,
2256
+ "learning_rate": 1.5602263095906393e-05,
2257
+ "loss": 0.6926,
2258
+ "step": 187500
2259
+ },
2260
+ {
2261
+ "epoch": 2.07,
2262
+ "learning_rate": 1.5510535797495478e-05,
2263
+ "loss": 0.6898,
2264
+ "step": 188000
2265
+ },
2266
+ {
2267
+ "epoch": 2.07,
2268
+ "learning_rate": 1.541880849908456e-05,
2269
+ "loss": 0.6887,
2270
+ "step": 188500
2271
+ },
2272
+ {
2273
+ "epoch": 2.08,
2274
+ "learning_rate": 1.5327081200673644e-05,
2275
+ "loss": 0.6907,
2276
+ "step": 189000
2277
+ },
2278
+ {
2279
+ "epoch": 2.09,
2280
+ "learning_rate": 1.523535390226273e-05,
2281
+ "loss": 0.6941,
2282
+ "step": 189500
2283
+ },
2284
+ {
2285
+ "epoch": 2.09,
2286
+ "learning_rate": 1.5143626603851812e-05,
2287
+ "loss": 0.6956,
2288
+ "step": 190000
2289
+ },
2290
+ {
2291
+ "epoch": 2.1,
2292
+ "learning_rate": 1.5051899305440897e-05,
2293
+ "loss": 0.6927,
2294
+ "step": 190500
2295
+ },
2296
+ {
2297
+ "epoch": 2.1,
2298
+ "learning_rate": 1.4960172007029982e-05,
2299
+ "loss": 0.6887,
2300
+ "step": 191000
2301
+ },
2302
+ {
2303
+ "epoch": 2.11,
2304
+ "learning_rate": 1.4868444708619065e-05,
2305
+ "loss": 0.6817,
2306
+ "step": 191500
2307
+ },
2308
+ {
2309
+ "epoch": 2.11,
2310
+ "learning_rate": 1.477671741020815e-05,
2311
+ "loss": 0.6842,
2312
+ "step": 192000
2313
+ },
2314
+ {
2315
+ "epoch": 2.12,
2316
+ "learning_rate": 1.4684990111797231e-05,
2317
+ "loss": 0.6888,
2318
+ "step": 192500
2319
+ },
2320
+ {
2321
+ "epoch": 2.12,
2322
+ "learning_rate": 1.4593262813386316e-05,
2323
+ "loss": 0.6862,
2324
+ "step": 193000
2325
+ },
2326
+ {
2327
+ "epoch": 2.13,
2328
+ "learning_rate": 1.45015355149754e-05,
2329
+ "loss": 0.6882,
2330
+ "step": 193500
2331
+ },
2332
+ {
2333
+ "epoch": 2.14,
2334
+ "learning_rate": 1.4409808216564482e-05,
2335
+ "loss": 0.6889,
2336
+ "step": 194000
2337
+ },
2338
+ {
2339
+ "epoch": 2.14,
2340
+ "learning_rate": 1.4318080918153567e-05,
2341
+ "loss": 0.6904,
2342
+ "step": 194500
2343
+ },
2344
+ {
2345
+ "epoch": 2.15,
2346
+ "learning_rate": 1.422635361974265e-05,
2347
+ "loss": 0.6932,
2348
+ "step": 195000
2349
+ },
2350
+ {
2351
+ "epoch": 2.15,
2352
+ "learning_rate": 1.4134626321331735e-05,
2353
+ "loss": 0.687,
2354
+ "step": 195500
2355
+ },
2356
+ {
2357
+ "epoch": 2.16,
2358
+ "learning_rate": 1.404289902292082e-05,
2359
+ "loss": 0.6928,
2360
+ "step": 196000
2361
+ },
2362
+ {
2363
+ "epoch": 2.16,
2364
+ "learning_rate": 1.3951171724509901e-05,
2365
+ "loss": 0.6837,
2366
+ "step": 196500
2367
+ },
2368
+ {
2369
+ "epoch": 2.17,
2370
+ "learning_rate": 1.3859444426098986e-05,
2371
+ "loss": 0.6757,
2372
+ "step": 197000
2373
+ },
2374
+ {
2375
+ "epoch": 2.17,
2376
+ "learning_rate": 1.3767717127688067e-05,
2377
+ "loss": 0.6889,
2378
+ "step": 197500
2379
+ },
2380
+ {
2381
+ "epoch": 2.18,
2382
+ "learning_rate": 1.3675989829277152e-05,
2383
+ "loss": 0.6837,
2384
+ "step": 198000
2385
+ },
2386
+ {
2387
+ "epoch": 2.18,
2388
+ "learning_rate": 1.3584262530866237e-05,
2389
+ "loss": 0.6903,
2390
+ "step": 198500
2391
+ },
2392
+ {
2393
+ "epoch": 2.19,
2394
+ "learning_rate": 1.349253523245532e-05,
2395
+ "loss": 0.6916,
2396
+ "step": 199000
2397
+ },
2398
+ {
2399
+ "epoch": 2.2,
2400
+ "learning_rate": 1.3400807934044405e-05,
2401
+ "loss": 0.6816,
2402
+ "step": 199500
2403
+ },
2404
+ {
2405
+ "epoch": 2.2,
2406
+ "learning_rate": 1.330908063563349e-05,
2407
+ "loss": 0.6796,
2408
+ "step": 200000
2409
+ },
2410
+ {
2411
+ "epoch": 2.21,
2412
+ "learning_rate": 1.3217353337222571e-05,
2413
+ "loss": 0.6864,
2414
+ "step": 200500
2415
+ },
2416
+ {
2417
+ "epoch": 2.21,
2418
+ "learning_rate": 1.3125626038811656e-05,
2419
+ "loss": 0.6848,
2420
+ "step": 201000
2421
+ },
2422
+ {
2423
+ "epoch": 2.22,
2424
+ "learning_rate": 1.3033898740400737e-05,
2425
+ "loss": 0.6887,
2426
+ "step": 201500
2427
+ },
2428
+ {
2429
+ "epoch": 2.22,
2430
+ "learning_rate": 1.2942171441989822e-05,
2431
+ "loss": 0.687,
2432
+ "step": 202000
2433
+ },
2434
+ {
2435
+ "epoch": 2.23,
2436
+ "learning_rate": 1.2850444143578907e-05,
2437
+ "loss": 0.6797,
2438
+ "step": 202500
2439
+ },
2440
+ {
2441
+ "epoch": 2.23,
2442
+ "learning_rate": 1.275871684516799e-05,
2443
+ "loss": 0.6891,
2444
+ "step": 203000
2445
+ },
2446
+ {
2447
+ "epoch": 2.24,
2448
+ "learning_rate": 1.2666989546757075e-05,
2449
+ "loss": 0.6847,
2450
+ "step": 203500
2451
+ },
2452
+ {
2453
+ "epoch": 2.25,
2454
+ "learning_rate": 1.2575262248346156e-05,
2455
+ "loss": 0.6854,
2456
+ "step": 204000
2457
+ },
2458
+ {
2459
+ "epoch": 2.25,
2460
+ "learning_rate": 1.2483534949935241e-05,
2461
+ "loss": 0.6755,
2462
+ "step": 204500
2463
+ },
2464
+ {
2465
+ "epoch": 2.26,
2466
+ "learning_rate": 1.2391807651524324e-05,
2467
+ "loss": 0.682,
2468
+ "step": 205000
2469
+ },
2470
+ {
2471
+ "epoch": 2.26,
2472
+ "learning_rate": 1.2300080353113407e-05,
2473
+ "loss": 0.6939,
2474
+ "step": 205500
2475
+ },
2476
+ {
2477
+ "epoch": 2.27,
2478
+ "learning_rate": 1.2208353054702492e-05,
2479
+ "loss": 0.6786,
2480
+ "step": 206000
2481
+ },
2482
+ {
2483
+ "epoch": 2.27,
2484
+ "learning_rate": 1.2116625756291577e-05,
2485
+ "loss": 0.6778,
2486
+ "step": 206500
2487
+ },
2488
+ {
2489
+ "epoch": 2.28,
2490
+ "learning_rate": 1.202489845788066e-05,
2491
+ "loss": 0.6784,
2492
+ "step": 207000
2493
+ },
2494
+ {
2495
+ "epoch": 2.28,
2496
+ "learning_rate": 1.1933171159469743e-05,
2497
+ "loss": 0.6819,
2498
+ "step": 207500
2499
+ },
2500
+ {
2501
+ "epoch": 2.29,
2502
+ "learning_rate": 1.1841443861058828e-05,
2503
+ "loss": 0.6806,
2504
+ "step": 208000
2505
+ },
2506
+ {
2507
+ "epoch": 2.3,
2508
+ "learning_rate": 1.174971656264791e-05,
2509
+ "loss": 0.6775,
2510
+ "step": 208500
2511
+ },
2512
+ {
2513
+ "epoch": 2.3,
2514
+ "learning_rate": 1.1657989264236994e-05,
2515
+ "loss": 0.6793,
2516
+ "step": 209000
2517
+ },
2518
+ {
2519
+ "epoch": 2.31,
2520
+ "learning_rate": 1.1566261965826079e-05,
2521
+ "loss": 0.6723,
2522
+ "step": 209500
2523
+ },
2524
+ {
2525
+ "epoch": 2.31,
2526
+ "learning_rate": 1.1474534667415162e-05,
2527
+ "loss": 0.6744,
2528
+ "step": 210000
2529
+ },
2530
+ {
2531
+ "epoch": 2.32,
2532
+ "learning_rate": 1.1382807369004247e-05,
2533
+ "loss": 0.6802,
2534
+ "step": 210500
2535
+ },
2536
+ {
2537
+ "epoch": 2.32,
2538
+ "learning_rate": 1.129108007059333e-05,
2539
+ "loss": 0.6739,
2540
+ "step": 211000
2541
+ },
2542
+ {
2543
+ "epoch": 2.33,
2544
+ "learning_rate": 1.1199352772182413e-05,
2545
+ "loss": 0.6806,
2546
+ "step": 211500
2547
+ },
2548
+ {
2549
+ "epoch": 2.33,
2550
+ "learning_rate": 1.1107625473771496e-05,
2551
+ "loss": 0.6803,
2552
+ "step": 212000
2553
+ },
2554
+ {
2555
+ "epoch": 2.34,
2556
+ "learning_rate": 1.101589817536058e-05,
2557
+ "loss": 0.6821,
2558
+ "step": 212500
2559
+ },
2560
+ {
2561
+ "epoch": 2.34,
2562
+ "learning_rate": 1.0924170876949664e-05,
2563
+ "loss": 0.6778,
2564
+ "step": 213000
2565
+ },
2566
+ {
2567
+ "epoch": 2.35,
2568
+ "learning_rate": 1.0832443578538749e-05,
2569
+ "loss": 0.6777,
2570
+ "step": 213500
2571
+ },
2572
+ {
2573
+ "epoch": 2.36,
2574
+ "learning_rate": 1.0740716280127832e-05,
2575
+ "loss": 0.6766,
2576
+ "step": 214000
2577
+ },
2578
+ {
2579
+ "epoch": 2.36,
2580
+ "learning_rate": 1.0648988981716915e-05,
2581
+ "loss": 0.6837,
2582
+ "step": 214500
2583
+ },
2584
+ {
2585
+ "epoch": 2.37,
2586
+ "learning_rate": 1.0557261683306e-05,
2587
+ "loss": 0.6731,
2588
+ "step": 215000
2589
+ },
2590
+ {
2591
+ "epoch": 2.37,
2592
+ "learning_rate": 1.0465534384895083e-05,
2593
+ "loss": 0.6738,
2594
+ "step": 215500
2595
+ },
2596
+ {
2597
+ "epoch": 2.38,
2598
+ "learning_rate": 1.0373807086484166e-05,
2599
+ "loss": 0.673,
2600
+ "step": 216000
2601
+ },
2602
+ {
2603
+ "epoch": 2.38,
2604
+ "learning_rate": 1.0282079788073249e-05,
2605
+ "loss": 0.6807,
2606
+ "step": 216500
2607
+ },
2608
+ {
2609
+ "epoch": 2.39,
2610
+ "learning_rate": 1.0190352489662334e-05,
2611
+ "loss": 0.6743,
2612
+ "step": 217000
2613
+ },
2614
+ {
2615
+ "epoch": 2.39,
2616
+ "learning_rate": 1.0098625191251419e-05,
2617
+ "loss": 0.677,
2618
+ "step": 217500
2619
+ },
2620
+ {
2621
+ "epoch": 2.4,
2622
+ "learning_rate": 1.0006897892840502e-05,
2623
+ "loss": 0.6549,
2624
+ "step": 218000
2625
+ },
2626
+ {
2627
+ "epoch": 2.41,
2628
+ "learning_rate": 9.915170594429585e-06,
2629
+ "loss": 0.6653,
2630
+ "step": 218500
2631
+ },
2632
+ {
2633
+ "epoch": 2.41,
2634
+ "learning_rate": 9.823443296018668e-06,
2635
+ "loss": 0.6823,
2636
+ "step": 219000
2637
+ },
2638
+ {
2639
+ "epoch": 2.42,
2640
+ "learning_rate": 9.731715997607753e-06,
2641
+ "loss": 0.6695,
2642
+ "step": 219500
2643
+ },
2644
+ {
2645
+ "epoch": 2.42,
2646
+ "learning_rate": 9.639988699196836e-06,
2647
+ "loss": 0.6694,
2648
+ "step": 220000
2649
+ },
2650
+ {
2651
+ "epoch": 2.43,
2652
+ "learning_rate": 9.548261400785919e-06,
2653
+ "loss": 0.6707,
2654
+ "step": 220500
2655
+ },
2656
+ {
2657
+ "epoch": 2.43,
2658
+ "learning_rate": 9.456534102375004e-06,
2659
+ "loss": 0.6751,
2660
+ "step": 221000
2661
+ },
2662
+ {
2663
+ "epoch": 2.44,
2664
+ "learning_rate": 9.364806803964089e-06,
2665
+ "loss": 0.6709,
2666
+ "step": 221500
2667
+ },
2668
+ {
2669
+ "epoch": 2.44,
2670
+ "learning_rate": 9.273079505553172e-06,
2671
+ "loss": 0.6726,
2672
+ "step": 222000
2673
+ },
2674
+ {
2675
+ "epoch": 2.45,
2676
+ "learning_rate": 9.181352207142255e-06,
2677
+ "loss": 0.6773,
2678
+ "step": 222500
2679
+ },
2680
+ {
2681
+ "epoch": 2.45,
2682
+ "learning_rate": 9.089624908731338e-06,
2683
+ "loss": 0.6724,
2684
+ "step": 223000
2685
+ },
2686
+ {
2687
+ "epoch": 2.46,
2688
+ "learning_rate": 8.997897610320421e-06,
2689
+ "loss": 0.6686,
2690
+ "step": 223500
2691
+ },
2692
+ {
2693
+ "epoch": 2.47,
2694
+ "learning_rate": 8.906170311909506e-06,
2695
+ "loss": 0.6654,
2696
+ "step": 224000
2697
+ },
2698
+ {
2699
+ "epoch": 2.47,
2700
+ "learning_rate": 8.814443013498589e-06,
2701
+ "loss": 0.6686,
2702
+ "step": 224500
2703
+ },
2704
+ {
2705
+ "epoch": 2.48,
2706
+ "learning_rate": 8.722715715087674e-06,
2707
+ "loss": 0.6622,
2708
+ "step": 225000
2709
+ },
2710
+ {
2711
+ "epoch": 2.48,
2712
+ "learning_rate": 8.630988416676757e-06,
2713
+ "loss": 0.6732,
2714
+ "step": 225500
2715
+ },
2716
+ {
2717
+ "epoch": 2.49,
2718
+ "learning_rate": 8.539261118265842e-06,
2719
+ "loss": 0.6694,
2720
+ "step": 226000
2721
+ },
2722
+ {
2723
+ "epoch": 2.49,
2724
+ "learning_rate": 8.447533819854925e-06,
2725
+ "loss": 0.6663,
2726
+ "step": 226500
2727
+ },
2728
+ {
2729
+ "epoch": 2.5,
2730
+ "learning_rate": 8.355806521444008e-06,
2731
+ "loss": 0.6632,
2732
+ "step": 227000
2733
+ },
2734
+ {
2735
+ "epoch": 2.5,
2736
+ "learning_rate": 8.264079223033091e-06,
2737
+ "loss": 0.6753,
2738
+ "step": 227500
2739
+ },
2740
+ {
2741
+ "epoch": 2.51,
2742
+ "learning_rate": 8.172351924622176e-06,
2743
+ "loss": 0.6691,
2744
+ "step": 228000
2745
+ },
2746
+ {
2747
+ "epoch": 2.52,
2748
+ "learning_rate": 8.08062462621126e-06,
2749
+ "loss": 0.6682,
2750
+ "step": 228500
2751
+ },
2752
+ {
2753
+ "epoch": 2.52,
2754
+ "learning_rate": 7.988897327800344e-06,
2755
+ "loss": 0.6718,
2756
+ "step": 229000
2757
+ },
2758
+ {
2759
+ "epoch": 2.53,
2760
+ "learning_rate": 7.897170029389427e-06,
2761
+ "loss": 0.6662,
2762
+ "step": 229500
2763
+ },
2764
+ {
2765
+ "epoch": 2.53,
2766
+ "learning_rate": 7.80544273097851e-06,
2767
+ "loss": 0.6651,
2768
+ "step": 230000
2769
+ },
2770
+ {
2771
+ "epoch": 2.54,
2772
+ "learning_rate": 7.713715432567595e-06,
2773
+ "loss": 0.6678,
2774
+ "step": 230500
2775
+ },
2776
+ {
2777
+ "epoch": 2.54,
2778
+ "learning_rate": 7.621988134156678e-06,
2779
+ "loss": 0.6716,
2780
+ "step": 231000
2781
+ },
2782
+ {
2783
+ "epoch": 2.55,
2784
+ "learning_rate": 7.530260835745762e-06,
2785
+ "loss": 0.6742,
2786
+ "step": 231500
2787
+ },
2788
+ {
2789
+ "epoch": 2.55,
2790
+ "learning_rate": 7.438533537334845e-06,
2791
+ "loss": 0.6729,
2792
+ "step": 232000
2793
+ },
2794
+ {
2795
+ "epoch": 2.56,
2796
+ "learning_rate": 7.346806238923929e-06,
2797
+ "loss": 0.6708,
2798
+ "step": 232500
2799
+ },
2800
+ {
2801
+ "epoch": 2.56,
2802
+ "learning_rate": 7.255078940513013e-06,
2803
+ "loss": 0.6605,
2804
+ "step": 233000
2805
+ },
2806
+ {
2807
+ "epoch": 2.57,
2808
+ "learning_rate": 7.163351642102097e-06,
2809
+ "loss": 0.6619,
2810
+ "step": 233500
2811
+ },
2812
+ {
2813
+ "epoch": 2.58,
2814
+ "learning_rate": 7.07162434369118e-06,
2815
+ "loss": 0.6688,
2816
+ "step": 234000
2817
+ },
2818
+ {
2819
+ "epoch": 2.58,
2820
+ "learning_rate": 6.979897045280264e-06,
2821
+ "loss": 0.6693,
2822
+ "step": 234500
2823
+ },
2824
+ {
2825
+ "epoch": 2.59,
2826
+ "learning_rate": 6.8881697468693486e-06,
2827
+ "loss": 0.6598,
2828
+ "step": 235000
2829
+ },
2830
+ {
2831
+ "epoch": 2.59,
2832
+ "learning_rate": 6.796442448458432e-06,
2833
+ "loss": 0.6647,
2834
+ "step": 235500
2835
+ },
2836
+ {
2837
+ "epoch": 2.6,
2838
+ "learning_rate": 6.704715150047515e-06,
2839
+ "loss": 0.6691,
2840
+ "step": 236000
2841
+ },
2842
+ {
2843
+ "epoch": 2.6,
2844
+ "learning_rate": 6.612987851636599e-06,
2845
+ "loss": 0.661,
2846
+ "step": 236500
2847
+ },
2848
+ {
2849
+ "epoch": 2.61,
2850
+ "learning_rate": 6.521260553225682e-06,
2851
+ "loss": 0.6545,
2852
+ "step": 237000
2853
+ },
2854
+ {
2855
+ "epoch": 2.61,
2856
+ "learning_rate": 6.429533254814767e-06,
2857
+ "loss": 0.6663,
2858
+ "step": 237500
2859
+ },
2860
+ {
2861
+ "epoch": 2.62,
2862
+ "learning_rate": 6.33780595640385e-06,
2863
+ "loss": 0.6671,
2864
+ "step": 238000
2865
+ },
2866
+ {
2867
+ "epoch": 2.63,
2868
+ "learning_rate": 6.246078657992934e-06,
2869
+ "loss": 0.6586,
2870
+ "step": 238500
2871
+ },
2872
+ {
2873
+ "epoch": 2.63,
2874
+ "learning_rate": 6.154351359582018e-06,
2875
+ "loss": 0.6642,
2876
+ "step": 239000
2877
+ },
2878
+ {
2879
+ "epoch": 2.64,
2880
+ "learning_rate": 6.062624061171101e-06,
2881
+ "loss": 0.6614,
2882
+ "step": 239500
2883
+ },
2884
+ {
2885
+ "epoch": 2.64,
2886
+ "learning_rate": 5.970896762760185e-06,
2887
+ "loss": 0.6537,
2888
+ "step": 240000
2889
+ },
2890
+ {
2891
+ "epoch": 2.65,
2892
+ "learning_rate": 5.879169464349269e-06,
2893
+ "loss": 0.6559,
2894
+ "step": 240500
2895
+ },
2896
+ {
2897
+ "epoch": 2.65,
2898
+ "learning_rate": 5.787442165938353e-06,
2899
+ "loss": 0.6559,
2900
+ "step": 241000
2901
+ },
2902
+ {
2903
+ "epoch": 2.66,
2904
+ "learning_rate": 5.695714867527436e-06,
2905
+ "loss": 0.6585,
2906
+ "step": 241500
2907
+ },
2908
+ {
2909
+ "epoch": 2.66,
2910
+ "learning_rate": 5.60398756911652e-06,
2911
+ "loss": 0.6646,
2912
+ "step": 242000
2913
+ },
2914
+ {
2915
+ "epoch": 2.67,
2916
+ "learning_rate": 5.512260270705604e-06,
2917
+ "loss": 0.6567,
2918
+ "step": 242500
2919
+ },
2920
+ {
2921
+ "epoch": 2.67,
2922
+ "learning_rate": 5.420532972294687e-06,
2923
+ "loss": 0.6684,
2924
+ "step": 243000
2925
+ },
2926
+ {
2927
+ "epoch": 2.68,
2928
+ "learning_rate": 5.328805673883771e-06,
2929
+ "loss": 0.6617,
2930
+ "step": 243500
2931
+ },
2932
+ {
2933
+ "epoch": 2.69,
2934
+ "learning_rate": 5.237078375472855e-06,
2935
+ "loss": 0.6658,
2936
+ "step": 244000
2937
+ },
2938
+ {
2939
+ "epoch": 2.69,
2940
+ "learning_rate": 5.145351077061939e-06,
2941
+ "loss": 0.6589,
2942
+ "step": 244500
2943
+ },
2944
+ {
2945
+ "epoch": 2.7,
2946
+ "learning_rate": 5.053623778651022e-06,
2947
+ "loss": 0.6683,
2948
+ "step": 245000
2949
+ },
2950
+ {
2951
+ "epoch": 2.7,
2952
+ "learning_rate": 4.961896480240106e-06,
2953
+ "loss": 0.6542,
2954
+ "step": 245500
2955
+ },
2956
+ {
2957
+ "epoch": 2.71,
2958
+ "learning_rate": 4.87016918182919e-06,
2959
+ "loss": 0.6531,
2960
+ "step": 246000
2961
+ },
2962
+ {
2963
+ "epoch": 2.71,
2964
+ "learning_rate": 4.778441883418273e-06,
2965
+ "loss": 0.662,
2966
+ "step": 246500
2967
+ },
2968
+ {
2969
+ "epoch": 2.72,
2970
+ "learning_rate": 4.686714585007357e-06,
2971
+ "loss": 0.6601,
2972
+ "step": 247000
2973
+ },
2974
+ {
2975
+ "epoch": 2.72,
2976
+ "learning_rate": 4.59498728659644e-06,
2977
+ "loss": 0.6576,
2978
+ "step": 247500
2979
+ },
2980
+ {
2981
+ "epoch": 2.73,
2982
+ "learning_rate": 4.503259988185525e-06,
2983
+ "loss": 0.6587,
2984
+ "step": 248000
2985
+ },
2986
+ {
2987
+ "epoch": 2.74,
2988
+ "learning_rate": 4.411532689774608e-06,
2989
+ "loss": 0.6617,
2990
+ "step": 248500
2991
+ },
2992
+ {
2993
+ "epoch": 2.74,
2994
+ "learning_rate": 4.319805391363692e-06,
2995
+ "loss": 0.6674,
2996
+ "step": 249000
2997
+ },
2998
+ {
2999
+ "epoch": 2.75,
3000
+ "learning_rate": 4.228078092952776e-06,
3001
+ "loss": 0.6614,
3002
+ "step": 249500
3003
+ },
3004
+ {
3005
+ "epoch": 2.75,
3006
+ "learning_rate": 4.13635079454186e-06,
3007
+ "loss": 0.6621,
3008
+ "step": 250000
3009
+ },
3010
+ {
3011
+ "epoch": 2.76,
3012
+ "learning_rate": 4.044623496130943e-06,
3013
+ "loss": 0.6566,
3014
+ "step": 250500
3015
+ },
3016
+ {
3017
+ "epoch": 2.76,
3018
+ "learning_rate": 3.952896197720026e-06,
3019
+ "loss": 0.6596,
3020
+ "step": 251000
3021
+ },
3022
+ {
3023
+ "epoch": 2.77,
3024
+ "learning_rate": 3.861168899309111e-06,
3025
+ "loss": 0.6507,
3026
+ "step": 251500
3027
+ },
3028
+ {
3029
+ "epoch": 2.77,
3030
+ "learning_rate": 3.7694416008981937e-06,
3031
+ "loss": 0.6612,
3032
+ "step": 252000
3033
+ },
3034
+ {
3035
+ "epoch": 2.78,
3036
+ "learning_rate": 3.6777143024872777e-06,
3037
+ "loss": 0.6589,
3038
+ "step": 252500
3039
+ },
3040
+ {
3041
+ "epoch": 2.78,
3042
+ "learning_rate": 3.5859870040763612e-06,
3043
+ "loss": 0.6532,
3044
+ "step": 253000
3045
+ },
3046
+ {
3047
+ "epoch": 2.79,
3048
+ "learning_rate": 3.494259705665445e-06,
3049
+ "loss": 0.6619,
3050
+ "step": 253500
3051
+ },
3052
+ {
3053
+ "epoch": 2.8,
3054
+ "learning_rate": 3.4025324072545287e-06,
3055
+ "loss": 0.6534,
3056
+ "step": 254000
3057
+ },
3058
+ {
3059
+ "epoch": 2.8,
3060
+ "learning_rate": 3.3108051088436127e-06,
3061
+ "loss": 0.651,
3062
+ "step": 254500
3063
+ },
3064
+ {
3065
+ "epoch": 2.81,
3066
+ "learning_rate": 3.219077810432696e-06,
3067
+ "loss": 0.6567,
3068
+ "step": 255000
3069
+ },
3070
+ {
3071
+ "epoch": 2.81,
3072
+ "learning_rate": 3.1273505120217797e-06,
3073
+ "loss": 0.6552,
3074
+ "step": 255500
3075
+ },
3076
+ {
3077
+ "epoch": 2.82,
3078
+ "learning_rate": 3.0356232136108637e-06,
3079
+ "loss": 0.6577,
3080
+ "step": 256000
3081
+ },
3082
+ {
3083
+ "epoch": 2.82,
3084
+ "learning_rate": 2.943895915199947e-06,
3085
+ "loss": 0.6568,
3086
+ "step": 256500
3087
+ },
3088
+ {
3089
+ "epoch": 2.83,
3090
+ "learning_rate": 2.852168616789031e-06,
3091
+ "loss": 0.6576,
3092
+ "step": 257000
3093
+ },
3094
+ {
3095
+ "epoch": 2.83,
3096
+ "learning_rate": 2.7604413183781147e-06,
3097
+ "loss": 0.6599,
3098
+ "step": 257500
3099
+ },
3100
+ {
3101
+ "epoch": 2.84,
3102
+ "learning_rate": 2.6687140199671982e-06,
3103
+ "loss": 0.6568,
3104
+ "step": 258000
3105
+ },
3106
+ {
3107
+ "epoch": 2.85,
3108
+ "learning_rate": 2.576986721556282e-06,
3109
+ "loss": 0.6484,
3110
+ "step": 258500
3111
+ },
3112
+ {
3113
+ "epoch": 2.85,
3114
+ "learning_rate": 2.4852594231453657e-06,
3115
+ "loss": 0.6508,
3116
+ "step": 259000
3117
+ },
3118
+ {
3119
+ "epoch": 2.86,
3120
+ "learning_rate": 2.3935321247344497e-06,
3121
+ "loss": 0.6482,
3122
+ "step": 259500
3123
+ },
3124
+ {
3125
+ "epoch": 2.86,
3126
+ "learning_rate": 2.301804826323533e-06,
3127
+ "loss": 0.6533,
3128
+ "step": 260000
3129
+ },
3130
+ {
3131
+ "epoch": 2.87,
3132
+ "learning_rate": 2.210077527912617e-06,
3133
+ "loss": 0.6555,
3134
+ "step": 260500
3135
+ },
3136
+ {
3137
+ "epoch": 2.87,
3138
+ "learning_rate": 2.1183502295017007e-06,
3139
+ "loss": 0.6548,
3140
+ "step": 261000
3141
+ },
3142
+ {
3143
+ "epoch": 2.88,
3144
+ "learning_rate": 2.0266229310907846e-06,
3145
+ "loss": 0.6529,
3146
+ "step": 261500
3147
+ },
3148
+ {
3149
+ "epoch": 2.88,
3150
+ "learning_rate": 1.934895632679868e-06,
3151
+ "loss": 0.6423,
3152
+ "step": 262000
3153
+ },
3154
+ {
3155
+ "epoch": 2.89,
3156
+ "learning_rate": 1.8431683342689517e-06,
3157
+ "loss": 0.6485,
3158
+ "step": 262500
3159
+ },
3160
+ {
3161
+ "epoch": 2.89,
3162
+ "learning_rate": 1.7514410358580354e-06,
3163
+ "loss": 0.6594,
3164
+ "step": 263000
3165
+ },
3166
+ {
3167
+ "epoch": 2.9,
3168
+ "learning_rate": 1.6597137374471192e-06,
3169
+ "loss": 0.6487,
3170
+ "step": 263500
3171
+ },
3172
+ {
3173
+ "epoch": 2.91,
3174
+ "learning_rate": 1.567986439036203e-06,
3175
+ "loss": 0.6515,
3176
+ "step": 264000
3177
+ },
3178
+ {
3179
+ "epoch": 2.91,
3180
+ "learning_rate": 1.4762591406252867e-06,
3181
+ "loss": 0.6538,
3182
+ "step": 264500
3183
+ },
3184
+ {
3185
+ "epoch": 2.92,
3186
+ "learning_rate": 1.3845318422143704e-06,
3187
+ "loss": 0.6497,
3188
+ "step": 265000
3189
+ },
3190
+ {
3191
+ "epoch": 2.92,
3192
+ "learning_rate": 1.2928045438034542e-06,
3193
+ "loss": 0.6484,
3194
+ "step": 265500
3195
+ },
3196
+ {
3197
+ "epoch": 2.93,
3198
+ "learning_rate": 1.201077245392538e-06,
3199
+ "loss": 0.6508,
3200
+ "step": 266000
3201
+ },
3202
+ {
3203
+ "epoch": 2.93,
3204
+ "learning_rate": 1.1093499469816216e-06,
3205
+ "loss": 0.6572,
3206
+ "step": 266500
3207
+ },
3208
+ {
3209
+ "epoch": 2.94,
3210
+ "learning_rate": 1.0176226485707054e-06,
3211
+ "loss": 0.6607,
3212
+ "step": 267000
3213
+ },
3214
+ {
3215
+ "epoch": 2.94,
3216
+ "learning_rate": 9.25895350159789e-07,
3217
+ "loss": 0.6484,
3218
+ "step": 267500
3219
+ },
3220
+ {
3221
+ "epoch": 2.95,
3222
+ "learning_rate": 8.341680517488727e-07,
3223
+ "loss": 0.66,
3224
+ "step": 268000
3225
+ },
3226
+ {
3227
+ "epoch": 2.96,
3228
+ "learning_rate": 7.424407533379564e-07,
3229
+ "loss": 0.6494,
3230
+ "step": 268500
3231
+ },
3232
+ {
3233
+ "epoch": 2.96,
3234
+ "learning_rate": 6.507134549270401e-07,
3235
+ "loss": 0.6547,
3236
+ "step": 269000
3237
+ },
3238
+ {
3239
+ "epoch": 2.97,
3240
+ "learning_rate": 5.589861565161238e-07,
3241
+ "loss": 0.6514,
3242
+ "step": 269500
3243
+ },
3244
+ {
3245
+ "epoch": 2.97,
3246
+ "learning_rate": 4.672588581052076e-07,
3247
+ "loss": 0.6421,
3248
+ "step": 270000
3249
+ },
3250
+ {
3251
+ "epoch": 2.98,
3252
+ "learning_rate": 3.7553155969429127e-07,
3253
+ "loss": 0.6482,
3254
+ "step": 270500
3255
+ },
3256
+ {
3257
+ "epoch": 2.98,
3258
+ "learning_rate": 2.8380426128337496e-07,
3259
+ "loss": 0.6474,
3260
+ "step": 271000
3261
+ },
3262
+ {
3263
+ "epoch": 2.99,
3264
+ "learning_rate": 1.920769628724587e-07,
3265
+ "loss": 0.6526,
3266
+ "step": 271500
3267
+ },
3268
+ {
3269
+ "epoch": 2.99,
3270
+ "learning_rate": 1.0034966446154242e-07,
3271
+ "loss": 0.655,
3272
+ "step": 272000
3273
+ },
3274
+ {
3275
+ "epoch": 3.0,
3276
+ "learning_rate": 8.62236605062613e-09,
3277
+ "loss": 0.6447,
3278
+ "step": 272500
3279
+ },
3280
+ {
3281
+ "epoch": 3.0,
3282
+ "step": 272547,
3283
+ "total_flos": 1.1358886111610301e+18,
3284
+ "train_loss": 0.7392136628871132,
3285
+ "train_runtime": 175400.2766,
3286
+ "train_samples_per_second": 24.861,
3287
+ "train_steps_per_second": 1.554
3288
+ }
3289
+ ],
3290
+ "max_steps": 272547,
3291
+ "num_train_epochs": 3,
3292
+ "total_flos": 1.1358886111610301e+18,
3293
+ "trial_name": null,
3294
+ "trial_params": null
3295
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b36c04b232856d3cce394d15ec8f1433d8884bacac0696121992e5d03ea0455a
3
+ size 3055
vocab.txt ADDED
The diff for this file is too large to render. See raw diff