Hugging Face's logo

cx-olquinjica
/
AngOFA

Fill-Mask
Transformers
PyTorch
xlm-roberta
Model card Files
xet
 Community
1
cx-olquinjica commited on
Commit
002e13a
·
verified ·
1 Parent(s): a29122a

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 +3070 -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 sha256:0da940ecc2f2b73a977b980baa212e16f98f0242050ea8e1149d2ce9732828c3
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 +1,3070 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 49.99630204866504,
5
+ "eval_steps": 500,
6
+ "global_step": 253500,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.1,
13
+ "learning_rate": 4.990295857988166e-05,
14
+ "loss": 6.3355,
15
+ "step": 500
16
+ },
17
+ {
18
+ "epoch": 0.2,
19
+ "learning_rate": 4.98043392504931e-05,
20
+ "loss": 4.6153,
21
+ "step": 1000
22
+ },
23
+ {
24
+ "epoch": 0.3,
25
+ "learning_rate": 4.9705719921104535e-05,
26
+ "loss": 4.0788,
27
+ "step": 1500
28
+ },
29
+ {
30
+ "epoch": 0.39,
31
+ "learning_rate": 4.9607100591715975e-05,
32
+ "loss": 3.7655,
33
+ "step": 2000
34
+ },
35
+ {
36
+ "epoch": 0.49,
37
+ "learning_rate": 4.9508481262327415e-05,
38
+ "loss": 3.5154,
39
+ "step": 2500
40
+ },
41
+ {
42
+ "epoch": 0.59,
43
+ "learning_rate": 4.9409861932938855e-05,
44
+ "loss": 3.3402,
45
+ "step": 3000
46
+ },
47
+ {
48
+ "epoch": 0.69,
49
+ "learning_rate": 4.93112426035503e-05,
50
+ "loss": 3.1954,
51
+ "step": 3500
52
+ },
53
+ {
54
+ "epoch": 0.79,
55
+ "learning_rate": 4.921262327416174e-05,
56
+ "loss": 3.0717,
57
+ "step": 4000
58
+ },
59
+ {
60
+ "epoch": 0.89,
61
+ "learning_rate": 4.9114003944773176e-05,
62
+ "loss": 2.9541,
63
+ "step": 4500
64
+ },
65
+ {
66
+ "epoch": 0.99,
67
+ "learning_rate": 4.9015384615384616e-05,
68
+ "loss": 2.8672,
69
+ "step": 5000
70
+ },
71
+ {
72
+ "epoch": 1.08,
73
+ "learning_rate": 4.8916765285996056e-05,
74
+ "loss": 2.7782,
75
+ "step": 5500
76
+ },
77
+ {
78
+ "epoch": 1.18,
79
+ "learning_rate": 4.88181459566075e-05,
80
+ "loss": 2.7074,
81
+ "step": 6000
82
+ },
83
+ {
84
+ "epoch": 1.28,
85
+ "learning_rate": 4.871952662721894e-05,
86
+ "loss": 2.6362,
87
+ "step": 6500
88
+ },
89
+ {
90
+ "epoch": 1.38,
91
+ "learning_rate": 4.862090729783038e-05,
92
+ "loss": 2.5867,
93
+ "step": 7000
94
+ },
95
+ {
96
+ "epoch": 1.48,
97
+ "learning_rate": 4.852228796844182e-05,
98
+ "loss": 2.5227,
99
+ "step": 7500
100
+ },
101
+ {
102
+ "epoch": 1.58,
103
+ "learning_rate": 4.842366863905326e-05,
104
+ "loss": 2.4813,
105
+ "step": 8000
106
+ },
107
+ {
108
+ "epoch": 1.68,
109
+ "learning_rate": 4.832504930966469e-05,
110
+ "loss": 2.4355,
111
+ "step": 8500
112
+ },
113
+ {
114
+ "epoch": 1.78,
115
+ "learning_rate": 4.822642998027614e-05,
116
+ "loss": 2.3717,
117
+ "step": 9000
118
+ },
119
+ {
120
+ "epoch": 1.87,
121
+ "learning_rate": 4.812781065088758e-05,
122
+ "loss": 2.345,
123
+ "step": 9500
124
+ },
125
+ {
126
+ "epoch": 1.97,
127
+ "learning_rate": 4.802919132149902e-05,
128
+ "loss": 2.3076,
129
+ "step": 10000
130
+ },
131
+ {
132
+ "epoch": 2.07,
133
+ "learning_rate": 4.793057199211046e-05,
134
+ "loss": 2.2789,
135
+ "step": 10500
136
+ },
137
+ {
138
+ "epoch": 2.17,
139
+ "learning_rate": 4.78319526627219e-05,
140
+ "loss": 2.2335,
141
+ "step": 11000
142
+ },
143
+ {
144
+ "epoch": 2.27,
145
+ "learning_rate": 4.773333333333333e-05,
146
+ "loss": 2.2165,
147
+ "step": 11500
148
+ },
149
+ {
150
+ "epoch": 2.37,
151
+ "learning_rate": 4.763471400394477e-05,
152
+ "loss": 2.185,
153
+ "step": 12000
154
+ },
155
+ {
156
+ "epoch": 2.47,
157
+ "learning_rate": 4.753609467455621e-05,
158
+ "loss": 2.1461,
159
+ "step": 12500
160
+ },
161
+ {
162
+ "epoch": 2.56,
163
+ "learning_rate": 4.743747534516765e-05,
164
+ "loss": 2.1114,
165
+ "step": 13000
166
+ },
167
+ {
168
+ "epoch": 2.66,
169
+ "learning_rate": 4.7338856015779094e-05,
170
+ "loss": 2.0961,
171
+ "step": 13500
172
+ },
173
+ {
174
+ "epoch": 2.76,
175
+ "learning_rate": 4.7240236686390534e-05,
176
+ "loss": 2.0726,
177
+ "step": 14000
178
+ },
179
+ {
180
+ "epoch": 2.86,
181
+ "learning_rate": 4.714181459566075e-05,
182
+ "loss": 2.0446,
183
+ "step": 14500
184
+ },
185
+ {
186
+ "epoch": 2.96,
187
+ "learning_rate": 4.70431952662722e-05,
188
+ "loss": 2.0268,
189
+ "step": 15000
190
+ },
191
+ {
192
+ "epoch": 3.06,
193
+ "learning_rate": 4.6944773175542406e-05,
194
+ "loss": 2.0076,
195
+ "step": 15500
196
+ },
197
+ {
198
+ "epoch": 3.16,
199
+ "learning_rate": 4.684615384615385e-05,
200
+ "loss": 1.9788,
201
+ "step": 16000
202
+ },
203
+ {
204
+ "epoch": 3.25,
205
+ "learning_rate": 4.674753451676529e-05,
206
+ "loss": 1.9637,
207
+ "step": 16500
208
+ },
209
+ {
210
+ "epoch": 3.35,
211
+ "learning_rate": 4.6648915187376726e-05,
212
+ "loss": 1.9425,
213
+ "step": 17000
214
+ },
215
+ {
216
+ "epoch": 3.45,
217
+ "learning_rate": 4.655029585798817e-05,
218
+ "loss": 1.92,
219
+ "step": 17500
220
+ },
221
+ {
222
+ "epoch": 3.55,
223
+ "learning_rate": 4.645167652859961e-05,
224
+ "loss": 1.9069,
225
+ "step": 18000
226
+ },
227
+ {
228
+ "epoch": 3.65,
229
+ "learning_rate": 4.635305719921105e-05,
230
+ "loss": 1.8903,
231
+ "step": 18500
232
+ },
233
+ {
234
+ "epoch": 3.75,
235
+ "learning_rate": 4.625443786982249e-05,
236
+ "loss": 1.8661,
237
+ "step": 19000
238
+ },
239
+ {
240
+ "epoch": 3.85,
241
+ "learning_rate": 4.61560157790927e-05,
242
+ "loss": 1.8637,
243
+ "step": 19500
244
+ },
245
+ {
246
+ "epoch": 3.94,
247
+ "learning_rate": 4.6057396449704143e-05,
248
+ "loss": 1.8444,
249
+ "step": 20000
250
+ },
251
+ {
252
+ "epoch": 4.04,
253
+ "learning_rate": 4.5958777120315584e-05,
254
+ "loss": 1.8321,
255
+ "step": 20500
256
+ },
257
+ {
258
+ "epoch": 4.14,
259
+ "learning_rate": 4.5860157790927024e-05,
260
+ "loss": 1.8032,
261
+ "step": 21000
262
+ },
263
+ {
264
+ "epoch": 4.24,
265
+ "learning_rate": 4.5761538461538464e-05,
266
+ "loss": 1.7901,
267
+ "step": 21500
268
+ },
269
+ {
270
+ "epoch": 4.34,
271
+ "learning_rate": 4.5662919132149904e-05,
272
+ "loss": 1.7781,
273
+ "step": 22000
274
+ },
275
+ {
276
+ "epoch": 4.44,
277
+ "learning_rate": 4.556449704142012e-05,
278
+ "loss": 1.7769,
279
+ "step": 22500
280
+ },
281
+ {
282
+ "epoch": 4.54,
283
+ "learning_rate": 4.546587771203156e-05,
284
+ "loss": 1.7731,
285
+ "step": 23000
286
+ },
287
+ {
288
+ "epoch": 4.63,
289
+ "learning_rate": 4.5367258382643e-05,
290
+ "loss": 1.7503,
291
+ "step": 23500
292
+ },
293
+ {
294
+ "epoch": 4.73,
295
+ "learning_rate": 4.526863905325444e-05,
296
+ "loss": 1.7306,
297
+ "step": 24000
298
+ },
299
+ {
300
+ "epoch": 4.83,
301
+ "learning_rate": 4.517021696252466e-05,
302
+ "loss": 1.729,
303
+ "step": 24500
304
+ },
305
+ {
306
+ "epoch": 4.93,
307
+ "learning_rate": 4.50715976331361e-05,
308
+ "loss": 1.7238,
309
+ "step": 25000
310
+ },
311
+ {
312
+ "epoch": 5.03,
313
+ "learning_rate": 4.497297830374754e-05,
314
+ "loss": 1.7157,
315
+ "step": 25500
316
+ },
317
+ {
318
+ "epoch": 5.13,
319
+ "learning_rate": 4.487435897435898e-05,
320
+ "loss": 1.6753,
321
+ "step": 26000
322
+ },
323
+ {
324
+ "epoch": 5.23,
325
+ "learning_rate": 4.477573964497042e-05,
326
+ "loss": 1.6714,
327
+ "step": 26500
328
+ },
329
+ {
330
+ "epoch": 5.33,
331
+ "learning_rate": 4.4677317554240634e-05,
332
+ "loss": 1.6588,
333
+ "step": 27000
334
+ },
335
+ {
336
+ "epoch": 5.42,
337
+ "learning_rate": 4.4578698224852074e-05,
338
+ "loss": 1.6559,
339
+ "step": 27500
340
+ },
341
+ {
342
+ "epoch": 5.52,
343
+ "learning_rate": 4.4480078895463514e-05,
344
+ "loss": 1.6571,
345
+ "step": 28000
346
+ },
347
+ {
348
+ "epoch": 5.62,
349
+ "learning_rate": 4.4381459566074954e-05,
350
+ "loss": 1.6449,
351
+ "step": 28500
352
+ },
353
+ {
354
+ "epoch": 5.72,
355
+ "learning_rate": 4.428284023668639e-05,
356
+ "loss": 1.6439,
357
+ "step": 29000
358
+ },
359
+ {
360
+ "epoch": 5.82,
361
+ "learning_rate": 4.4184220907297835e-05,
362
+ "loss": 1.6303,
363
+ "step": 29500
364
+ },
365
+ {
366
+ "epoch": 5.92,
367
+ "learning_rate": 4.4085798816568044e-05,
368
+ "loss": 1.6159,
369
+ "step": 30000
370
+ },
371
+ {
372
+ "epoch": 6.02,
373
+ "learning_rate": 4.398717948717949e-05,
374
+ "loss": 1.6218,
375
+ "step": 30500
376
+ },
377
+ {
378
+ "epoch": 6.11,
379
+ "learning_rate": 4.3888757396449707e-05,
380
+ "loss": 1.5967,
381
+ "step": 31000
382
+ },
383
+ {
384
+ "epoch": 6.21,
385
+ "learning_rate": 4.379013806706115e-05,
386
+ "loss": 1.5843,
387
+ "step": 31500
388
+ },
389
+ {
390
+ "epoch": 6.31,
391
+ "learning_rate": 4.369151873767259e-05,
392
+ "loss": 1.581,
393
+ "step": 32000
394
+ },
395
+ {
396
+ "epoch": 6.41,
397
+ "learning_rate": 4.359289940828403e-05,
398
+ "loss": 1.5737,
399
+ "step": 32500
400
+ },
401
+ {
402
+ "epoch": 6.51,
403
+ "learning_rate": 4.349428007889547e-05,
404
+ "loss": 1.5591,
405
+ "step": 33000
406
+ },
407
+ {
408
+ "epoch": 6.61,
409
+ "learning_rate": 4.339566074950691e-05,
410
+ "loss": 1.5654,
411
+ "step": 33500
412
+ },
413
+ {
414
+ "epoch": 6.71,
415
+ "learning_rate": 4.3297238658777124e-05,
416
+ "loss": 1.5536,
417
+ "step": 34000
418
+ },
419
+ {
420
+ "epoch": 6.8,
421
+ "learning_rate": 4.3198619329388564e-05,
422
+ "loss": 1.5458,
423
+ "step": 34500
424
+ },
425
+ {
426
+ "epoch": 6.9,
427
+ "learning_rate": 4.3100000000000004e-05,
428
+ "loss": 1.5353,
429
+ "step": 35000
430
+ },
431
+ {
432
+ "epoch": 7.0,
433
+ "learning_rate": 4.300138067061144e-05,
434
+ "loss": 1.5321,
435
+ "step": 35500
436
+ },
437
+ {
438
+ "epoch": 7.1,
439
+ "learning_rate": 4.290276134122288e-05,
440
+ "loss": 1.5217,
441
+ "step": 36000
442
+ },
443
+ {
444
+ "epoch": 7.2,
445
+ "learning_rate": 4.280414201183432e-05,
446
+ "loss": 1.5073,
447
+ "step": 36500
448
+ },
449
+ {
450
+ "epoch": 7.3,
451
+ "learning_rate": 4.270552268244576e-05,
452
+ "loss": 1.5073,
453
+ "step": 37000
454
+ },
455
+ {
456
+ "epoch": 7.4,
457
+ "learning_rate": 4.2606903353057205e-05,
458
+ "loss": 1.4939,
459
+ "step": 37500
460
+ },
461
+ {
462
+ "epoch": 7.49,
463
+ "learning_rate": 4.2508284023668646e-05,
464
+ "loss": 1.4914,
465
+ "step": 38000
466
+ },
467
+ {
468
+ "epoch": 7.59,
469
+ "learning_rate": 4.2409664694280086e-05,
470
+ "loss": 1.489,
471
+ "step": 38500
472
+ },
473
+ {
474
+ "epoch": 7.69,
475
+ "learning_rate": 4.23112426035503e-05,
476
+ "loss": 1.4797,
477
+ "step": 39000
478
+ },
479
+ {
480
+ "epoch": 7.79,
481
+ "learning_rate": 4.221262327416174e-05,
482
+ "loss": 1.4836,
483
+ "step": 39500
484
+ },
485
+ {
486
+ "epoch": 7.89,
487
+ "learning_rate": 4.2114003944773175e-05,
488
+ "loss": 1.4689,
489
+ "step": 40000
490
+ },
491
+ {
492
+ "epoch": 7.99,
493
+ "learning_rate": 4.20155818540434e-05,
494
+ "loss": 1.4639,
495
+ "step": 40500
496
+ },
497
+ {
498
+ "epoch": 8.09,
499
+ "learning_rate": 4.191696252465484e-05,
500
+ "loss": 1.4526,
501
+ "step": 41000
502
+ },
503
+ {
504
+ "epoch": 8.18,
505
+ "learning_rate": 4.181834319526627e-05,
506
+ "loss": 1.4486,
507
+ "step": 41500
508
+ },
509
+ {
510
+ "epoch": 8.28,
511
+ "learning_rate": 4.171972386587771e-05,
512
+ "loss": 1.4381,
513
+ "step": 42000
514
+ },
515
+ {
516
+ "epoch": 8.38,
517
+ "learning_rate": 4.162110453648915e-05,
518
+ "loss": 1.4321,
519
+ "step": 42500
520
+ },
521
+ {
522
+ "epoch": 8.48,
523
+ "learning_rate": 4.152248520710059e-05,
524
+ "loss": 1.4366,
525
+ "step": 43000
526
+ },
527
+ {
528
+ "epoch": 8.58,
529
+ "learning_rate": 4.142386587771203e-05,
530
+ "loss": 1.4324,
531
+ "step": 43500
532
+ },
533
+ {
534
+ "epoch": 8.68,
535
+ "learning_rate": 4.132524654832347e-05,
536
+ "loss": 1.4236,
537
+ "step": 44000
538
+ },
539
+ {
540
+ "epoch": 8.78,
541
+ "learning_rate": 4.122662721893491e-05,
542
+ "loss": 1.4231,
543
+ "step": 44500
544
+ },
545
+ {
546
+ "epoch": 8.88,
547
+ "learning_rate": 4.112800788954635e-05,
548
+ "loss": 1.416,
549
+ "step": 45000
550
+ },
551
+ {
552
+ "epoch": 8.97,
553
+ "learning_rate": 4.102958579881657e-05,
554
+ "loss": 1.4205,
555
+ "step": 45500
556
+ },
557
+ {
558
+ "epoch": 9.07,
559
+ "learning_rate": 4.093096646942801e-05,
560
+ "loss": 1.3942,
561
+ "step": 46000
562
+ },
563
+ {
564
+ "epoch": 9.17,
565
+ "learning_rate": 4.083234714003945e-05,
566
+ "loss": 1.402,
567
+ "step": 46500
568
+ },
569
+ {
570
+ "epoch": 9.27,
571
+ "learning_rate": 4.073372781065089e-05,
572
+ "loss": 1.3927,
573
+ "step": 47000
574
+ },
575
+ {
576
+ "epoch": 9.37,
577
+ "learning_rate": 4.063510848126233e-05,
578
+ "loss": 1.3817,
579
+ "step": 47500
580
+ },
581
+ {
582
+ "epoch": 9.47,
583
+ "learning_rate": 4.053648915187377e-05,
584
+ "loss": 1.3762,
585
+ "step": 48000
586
+ },
587
+ {
588
+ "epoch": 9.57,
589
+ "learning_rate": 4.043786982248521e-05,
590
+ "loss": 1.3781,
591
+ "step": 48500
592
+ },
593
+ {
594
+ "epoch": 9.66,
595
+ "learning_rate": 4.0339447731755426e-05,
596
+ "loss": 1.3754,
597
+ "step": 49000
598
+ },
599
+ {
600
+ "epoch": 9.76,
601
+ "learning_rate": 4.0240828402366867e-05,
602
+ "loss": 1.3764,
603
+ "step": 49500
604
+ },
605
+ {
606
+ "epoch": 9.86,
607
+ "learning_rate": 4.014240631163708e-05,
608
+ "loss": 1.3621,
609
+ "step": 50000
610
+ },
611
+ {
612
+ "epoch": 9.96,
613
+ "learning_rate": 4.004378698224852e-05,
614
+ "loss": 1.3692,
615
+ "step": 50500
616
+ },
617
+ {
618
+ "epoch": 10.06,
619
+ "learning_rate": 3.994516765285996e-05,
620
+ "loss": 1.3485,
621
+ "step": 51000
622
+ },
623
+ {
624
+ "epoch": 10.16,
625
+ "learning_rate": 3.98465483234714e-05,
626
+ "loss": 1.3479,
627
+ "step": 51500
628
+ },
629
+ {
630
+ "epoch": 10.26,
631
+ "learning_rate": 3.974792899408284e-05,
632
+ "loss": 1.343,
633
+ "step": 52000
634
+ },
635
+ {
636
+ "epoch": 10.35,
637
+ "learning_rate": 3.9649309664694284e-05,
638
+ "loss": 1.3521,
639
+ "step": 52500
640
+ },
641
+ {
642
+ "epoch": 10.45,
643
+ "learning_rate": 3.9550690335305724e-05,
644
+ "loss": 1.336,
645
+ "step": 53000
646
+ },
647
+ {
648
+ "epoch": 10.55,
649
+ "learning_rate": 3.9452071005917164e-05,
650
+ "loss": 1.3361,
651
+ "step": 53500
652
+ },
653
+ {
654
+ "epoch": 10.65,
655
+ "learning_rate": 3.9353451676528604e-05,
656
+ "loss": 1.3395,
657
+ "step": 54000
658
+ },
659
+ {
660
+ "epoch": 10.75,
661
+ "learning_rate": 3.9254832347140045e-05,
662
+ "loss": 1.3308,
663
+ "step": 54500
664
+ },
665
+ {
666
+ "epoch": 10.85,
667
+ "learning_rate": 3.915621301775148e-05,
668
+ "loss": 1.3235,
669
+ "step": 55000
670
+ },
671
+ {
672
+ "epoch": 10.95,
673
+ "learning_rate": 3.90577909270217e-05,
674
+ "loss": 1.3222,
675
+ "step": 55500
676
+ },
677
+ {
678
+ "epoch": 11.04,
679
+ "learning_rate": 3.8959171597633134e-05,
680
+ "loss": 1.322,
681
+ "step": 56000
682
+ },
683
+ {
684
+ "epoch": 11.14,
685
+ "learning_rate": 3.8860552268244574e-05,
686
+ "loss": 1.3015,
687
+ "step": 56500
688
+ },
689
+ {
690
+ "epoch": 11.24,
691
+ "learning_rate": 3.8761932938856015e-05,
692
+ "loss": 1.3081,
693
+ "step": 57000
694
+ },
695
+ {
696
+ "epoch": 11.34,
697
+ "learning_rate": 3.8663313609467455e-05,
698
+ "loss": 1.3034,
699
+ "step": 57500
700
+ },
701
+ {
702
+ "epoch": 11.44,
703
+ "learning_rate": 3.8564694280078895e-05,
704
+ "loss": 1.2988,
705
+ "step": 58000
706
+ },
707
+ {
708
+ "epoch": 11.54,
709
+ "learning_rate": 3.846607495069034e-05,
710
+ "loss": 1.287,
711
+ "step": 58500
712
+ },
713
+ {
714
+ "epoch": 11.64,
715
+ "learning_rate": 3.836765285996056e-05,
716
+ "loss": 1.2945,
717
+ "step": 59000
718
+ },
719
+ {
720
+ "epoch": 11.73,
721
+ "learning_rate": 3.8269033530572e-05,
722
+ "loss": 1.2901,
723
+ "step": 59500
724
+ },
725
+ {
726
+ "epoch": 11.83,
727
+ "learning_rate": 3.817041420118344e-05,
728
+ "loss": 1.2921,
729
+ "step": 60000
730
+ },
731
+ {
732
+ "epoch": 11.93,
733
+ "learning_rate": 3.8071992110453654e-05,
734
+ "loss": 1.2868,
735
+ "step": 60500
736
+ },
737
+ {
738
+ "epoch": 12.03,
739
+ "learning_rate": 3.7973372781065094e-05,
740
+ "loss": 1.2843,
741
+ "step": 61000
742
+ },
743
+ {
744
+ "epoch": 12.13,
745
+ "learning_rate": 3.787475345167653e-05,
746
+ "loss": 1.2763,
747
+ "step": 61500
748
+ },
749
+ {
750
+ "epoch": 12.23,
751
+ "learning_rate": 3.777613412228797e-05,
752
+ "loss": 1.2681,
753
+ "step": 62000
754
+ },
755
+ {
756
+ "epoch": 12.33,
757
+ "learning_rate": 3.767751479289941e-05,
758
+ "loss": 1.2677,
759
+ "step": 62500
760
+ },
761
+ {
762
+ "epoch": 12.43,
763
+ "learning_rate": 3.757889546351085e-05,
764
+ "loss": 1.2675,
765
+ "step": 63000
766
+ },
767
+ {
768
+ "epoch": 12.52,
769
+ "learning_rate": 3.7480473372781064e-05,
770
+ "loss": 1.2605,
771
+ "step": 63500
772
+ },
773
+ {
774
+ "epoch": 12.62,
775
+ "learning_rate": 3.7381854043392505e-05,
776
+ "loss": 1.262,
777
+ "step": 64000
778
+ },
779
+ {
780
+ "epoch": 12.72,
781
+ "learning_rate": 3.7283234714003945e-05,
782
+ "loss": 1.2483,
783
+ "step": 64500
784
+ },
785
+ {
786
+ "epoch": 12.82,
787
+ "learning_rate": 3.7184615384615385e-05,
788
+ "loss": 1.2615,
789
+ "step": 65000
790
+ },
791
+ {
792
+ "epoch": 12.92,
793
+ "learning_rate": 3.7085996055226825e-05,
794
+ "loss": 1.2499,
795
+ "step": 65500
796
+ },
797
+ {
798
+ "epoch": 13.02,
799
+ "learning_rate": 3.6987376725838266e-05,
800
+ "loss": 1.2475,
801
+ "step": 66000
802
+ },
803
+ {
804
+ "epoch": 13.12,
805
+ "learning_rate": 3.6888757396449706e-05,
806
+ "loss": 1.2385,
807
+ "step": 66500
808
+ },
809
+ {
810
+ "epoch": 13.21,
811
+ "learning_rate": 3.6790138067061146e-05,
812
+ "loss": 1.2381,
813
+ "step": 67000
814
+ },
815
+ {
816
+ "epoch": 13.31,
817
+ "learning_rate": 3.6691518737672586e-05,
818
+ "loss": 1.2394,
819
+ "step": 67500
820
+ },
821
+ {
822
+ "epoch": 13.41,
823
+ "learning_rate": 3.65930966469428e-05,
824
+ "loss": 1.2345,
825
+ "step": 68000
826
+ },
827
+ {
828
+ "epoch": 13.51,
829
+ "learning_rate": 3.649447731755424e-05,
830
+ "loss": 1.2348,
831
+ "step": 68500
832
+ },
833
+ {
834
+ "epoch": 13.61,
835
+ "learning_rate": 3.639585798816568e-05,
836
+ "loss": 1.2237,
837
+ "step": 69000
838
+ },
839
+ {
840
+ "epoch": 13.71,
841
+ "learning_rate": 3.62974358974359e-05,
842
+ "loss": 1.2346,
843
+ "step": 69500
844
+ },
845
+ {
846
+ "epoch": 13.81,
847
+ "learning_rate": 3.619881656804734e-05,
848
+ "loss": 1.2205,
849
+ "step": 70000
850
+ },
851
+ {
852
+ "epoch": 13.9,
853
+ "learning_rate": 3.610019723865878e-05,
854
+ "loss": 1.2187,
855
+ "step": 70500
856
+ },
857
+ {
858
+ "epoch": 14.0,
859
+ "learning_rate": 3.600157790927022e-05,
860
+ "loss": 1.222,
861
+ "step": 71000
862
+ },
863
+ {
864
+ "epoch": 14.1,
865
+ "learning_rate": 3.590295857988166e-05,
866
+ "loss": 1.2185,
867
+ "step": 71500
868
+ },
869
+ {
870
+ "epoch": 14.2,
871
+ "learning_rate": 3.58043392504931e-05,
872
+ "loss": 1.2069,
873
+ "step": 72000
874
+ },
875
+ {
876
+ "epoch": 14.3,
877
+ "learning_rate": 3.570571992110453e-05,
878
+ "loss": 1.2022,
879
+ "step": 72500
880
+ },
881
+ {
882
+ "epoch": 14.4,
883
+ "learning_rate": 3.5607297830374756e-05,
884
+ "loss": 1.2077,
885
+ "step": 73000
886
+ },
887
+ {
888
+ "epoch": 14.5,
889
+ "learning_rate": 3.5508678500986196e-05,
890
+ "loss": 1.2086,
891
+ "step": 73500
892
+ },
893
+ {
894
+ "epoch": 14.59,
895
+ "learning_rate": 3.5410059171597636e-05,
896
+ "loss": 1.2018,
897
+ "step": 74000
898
+ },
899
+ {
900
+ "epoch": 14.69,
901
+ "learning_rate": 3.5311439842209076e-05,
902
+ "loss": 1.1921,
903
+ "step": 74500
904
+ },
905
+ {
906
+ "epoch": 14.79,
907
+ "learning_rate": 3.521282051282052e-05,
908
+ "loss": 1.1921,
909
+ "step": 75000
910
+ },
911
+ {
912
+ "epoch": 14.89,
913
+ "learning_rate": 3.511420118343196e-05,
914
+ "loss": 1.195,
915
+ "step": 75500
916
+ },
917
+ {
918
+ "epoch": 14.99,
919
+ "learning_rate": 3.50155818540434e-05,
920
+ "loss": 1.1917,
921
+ "step": 76000
922
+ },
923
+ {
924
+ "epoch": 15.09,
925
+ "learning_rate": 3.491715976331361e-05,
926
+ "loss": 1.1868,
927
+ "step": 76500
928
+ },
929
+ {
930
+ "epoch": 15.19,
931
+ "learning_rate": 3.481854043392505e-05,
932
+ "loss": 1.188,
933
+ "step": 77000
934
+ },
935
+ {
936
+ "epoch": 15.28,
937
+ "learning_rate": 3.4719921104536493e-05,
938
+ "loss": 1.1823,
939
+ "step": 77500
940
+ },
941
+ {
942
+ "epoch": 15.38,
943
+ "learning_rate": 3.462130177514793e-05,
944
+ "loss": 1.1842,
945
+ "step": 78000
946
+ },
947
+ {
948
+ "epoch": 15.48,
949
+ "learning_rate": 3.452268244575937e-05,
950
+ "loss": 1.1753,
951
+ "step": 78500
952
+ },
953
+ {
954
+ "epoch": 15.58,
955
+ "learning_rate": 3.442406311637081e-05,
956
+ "loss": 1.1706,
957
+ "step": 79000
958
+ },
959
+ {
960
+ "epoch": 15.68,
961
+ "learning_rate": 3.432544378698225e-05,
962
+ "loss": 1.176,
963
+ "step": 79500
964
+ },
965
+ {
966
+ "epoch": 15.78,
967
+ "learning_rate": 3.4226824457593695e-05,
968
+ "loss": 1.1707,
969
+ "step": 80000
970
+ },
971
+ {
972
+ "epoch": 15.88,
973
+ "learning_rate": 3.4128205128205135e-05,
974
+ "loss": 1.1627,
975
+ "step": 80500
976
+ },
977
+ {
978
+ "epoch": 15.98,
979
+ "learning_rate": 3.402978303747535e-05,
980
+ "loss": 1.166,
981
+ "step": 81000
982
+ },
983
+ {
984
+ "epoch": 16.07,
985
+ "learning_rate": 3.393116370808679e-05,
986
+ "loss": 1.1525,
987
+ "step": 81500
988
+ },
989
+ {
990
+ "epoch": 16.17,
991
+ "learning_rate": 3.3832544378698224e-05,
992
+ "loss": 1.1591,
993
+ "step": 82000
994
+ },
995
+ {
996
+ "epoch": 16.27,
997
+ "learning_rate": 3.3733925049309665e-05,
998
+ "loss": 1.159,
999
+ "step": 82500
1000
+ },
1001
+ {
1002
+ "epoch": 16.37,
1003
+ "learning_rate": 3.3635305719921105e-05,
1004
+ "loss": 1.1587,
1005
+ "step": 83000
1006
+ },
1007
+ {
1008
+ "epoch": 16.47,
1009
+ "learning_rate": 3.353688362919132e-05,
1010
+ "loss": 1.156,
1011
+ "step": 83500
1012
+ },
1013
+ {
1014
+ "epoch": 16.57,
1015
+ "learning_rate": 3.343826429980276e-05,
1016
+ "loss": 1.1462,
1017
+ "step": 84000
1018
+ },
1019
+ {
1020
+ "epoch": 16.67,
1021
+ "learning_rate": 3.33396449704142e-05,
1022
+ "loss": 1.1573,
1023
+ "step": 84500
1024
+ },
1025
+ {
1026
+ "epoch": 16.76,
1027
+ "learning_rate": 3.324102564102564e-05,
1028
+ "loss": 1.1403,
1029
+ "step": 85000
1030
+ },
1031
+ {
1032
+ "epoch": 16.86,
1033
+ "learning_rate": 3.314260355029586e-05,
1034
+ "loss": 1.1419,
1035
+ "step": 85500
1036
+ },
1037
+ {
1038
+ "epoch": 16.96,
1039
+ "learning_rate": 3.30439842209073e-05,
1040
+ "loss": 1.1434,
1041
+ "step": 86000
1042
+ },
1043
+ {
1044
+ "epoch": 17.06,
1045
+ "learning_rate": 3.294536489151874e-05,
1046
+ "loss": 1.1452,
1047
+ "step": 86500
1048
+ },
1049
+ {
1050
+ "epoch": 17.16,
1051
+ "learning_rate": 3.2846942800788954e-05,
1052
+ "loss": 1.128,
1053
+ "step": 87000
1054
+ },
1055
+ {
1056
+ "epoch": 17.26,
1057
+ "learning_rate": 3.2748323471400394e-05,
1058
+ "loss": 1.1396,
1059
+ "step": 87500
1060
+ },
1061
+ {
1062
+ "epoch": 17.36,
1063
+ "learning_rate": 3.2649704142011834e-05,
1064
+ "loss": 1.13,
1065
+ "step": 88000
1066
+ },
1067
+ {
1068
+ "epoch": 17.45,
1069
+ "learning_rate": 3.255108481262328e-05,
1070
+ "loss": 1.1303,
1071
+ "step": 88500
1072
+ },
1073
+ {
1074
+ "epoch": 17.55,
1075
+ "learning_rate": 3.2452465483234715e-05,
1076
+ "loss": 1.1292,
1077
+ "step": 89000
1078
+ },
1079
+ {
1080
+ "epoch": 17.65,
1081
+ "learning_rate": 3.2353846153846155e-05,
1082
+ "loss": 1.1313,
1083
+ "step": 89500
1084
+ },
1085
+ {
1086
+ "epoch": 17.75,
1087
+ "learning_rate": 3.2255226824457595e-05,
1088
+ "loss": 1.1294,
1089
+ "step": 90000
1090
+ },
1091
+ {
1092
+ "epoch": 17.85,
1093
+ "learning_rate": 3.2156607495069035e-05,
1094
+ "loss": 1.1237,
1095
+ "step": 90500
1096
+ },
1097
+ {
1098
+ "epoch": 17.95,
1099
+ "learning_rate": 3.2057988165680475e-05,
1100
+ "loss": 1.1252,
1101
+ "step": 91000
1102
+ },
1103
+ {
1104
+ "epoch": 18.05,
1105
+ "learning_rate": 3.1959368836291916e-05,
1106
+ "loss": 1.1231,
1107
+ "step": 91500
1108
+ },
1109
+ {
1110
+ "epoch": 18.14,
1111
+ "learning_rate": 3.1860749506903356e-05,
1112
+ "loss": 1.1169,
1113
+ "step": 92000
1114
+ },
1115
+ {
1116
+ "epoch": 18.24,
1117
+ "learning_rate": 3.176232741617357e-05,
1118
+ "loss": 1.1104,
1119
+ "step": 92500
1120
+ },
1121
+ {
1122
+ "epoch": 18.34,
1123
+ "learning_rate": 3.166370808678501e-05,
1124
+ "loss": 1.1108,
1125
+ "step": 93000
1126
+ },
1127
+ {
1128
+ "epoch": 18.44,
1129
+ "learning_rate": 3.156508875739645e-05,
1130
+ "loss": 1.1034,
1131
+ "step": 93500
1132
+ },
1133
+ {
1134
+ "epoch": 18.54,
1135
+ "learning_rate": 3.1466469428007886e-05,
1136
+ "loss": 1.1067,
1137
+ "step": 94000
1138
+ },
1139
+ {
1140
+ "epoch": 18.64,
1141
+ "learning_rate": 3.136785009861933e-05,
1142
+ "loss": 1.1096,
1143
+ "step": 94500
1144
+ },
1145
+ {
1146
+ "epoch": 18.74,
1147
+ "learning_rate": 3.126923076923077e-05,
1148
+ "loss": 1.103,
1149
+ "step": 95000
1150
+ },
1151
+ {
1152
+ "epoch": 18.83,
1153
+ "learning_rate": 3.117080867850099e-05,
1154
+ "loss": 1.1128,
1155
+ "step": 95500
1156
+ },
1157
+ {
1158
+ "epoch": 18.93,
1159
+ "learning_rate": 3.107218934911243e-05,
1160
+ "loss": 1.0998,
1161
+ "step": 96000
1162
+ },
1163
+ {
1164
+ "epoch": 19.03,
1165
+ "learning_rate": 3.097357001972387e-05,
1166
+ "loss": 1.0949,
1167
+ "step": 96500
1168
+ },
1169
+ {
1170
+ "epoch": 19.13,
1171
+ "learning_rate": 3.087495069033531e-05,
1172
+ "loss": 1.0935,
1173
+ "step": 97000
1174
+ },
1175
+ {
1176
+ "epoch": 19.23,
1177
+ "learning_rate": 3.077633136094675e-05,
1178
+ "loss": 1.1058,
1179
+ "step": 97500
1180
+ },
1181
+ {
1182
+ "epoch": 19.33,
1183
+ "learning_rate": 3.067771203155819e-05,
1184
+ "loss": 1.0914,
1185
+ "step": 98000
1186
+ },
1187
+ {
1188
+ "epoch": 19.43,
1189
+ "learning_rate": 3.0579092702169623e-05,
1190
+ "loss": 1.0907,
1191
+ "step": 98500
1192
+ },
1193
+ {
1194
+ "epoch": 19.53,
1195
+ "learning_rate": 3.0480473372781067e-05,
1196
+ "loss": 1.0849,
1197
+ "step": 99000
1198
+ },
1199
+ {
1200
+ "epoch": 19.62,
1201
+ "learning_rate": 3.0382248520710062e-05,
1202
+ "loss": 1.0866,
1203
+ "step": 99500
1204
+ },
1205
+ {
1206
+ "epoch": 19.72,
1207
+ "learning_rate": 3.02836291913215e-05,
1208
+ "loss": 1.0957,
1209
+ "step": 100000
1210
+ },
1211
+ {
1212
+ "epoch": 19.82,
1213
+ "learning_rate": 3.018500986193294e-05,
1214
+ "loss": 1.0817,
1215
+ "step": 100500
1216
+ },
1217
+ {
1218
+ "epoch": 19.92,
1219
+ "learning_rate": 3.008639053254438e-05,
1220
+ "loss": 1.0924,
1221
+ "step": 101000
1222
+ },
1223
+ {
1224
+ "epoch": 20.02,
1225
+ "learning_rate": 2.998777120315582e-05,
1226
+ "loss": 1.0742,
1227
+ "step": 101500
1228
+ },
1229
+ {
1230
+ "epoch": 20.12,
1231
+ "learning_rate": 2.9889151873767256e-05,
1232
+ "loss": 1.0749,
1233
+ "step": 102000
1234
+ },
1235
+ {
1236
+ "epoch": 20.22,
1237
+ "learning_rate": 2.9790729783037475e-05,
1238
+ "loss": 1.0764,
1239
+ "step": 102500
1240
+ },
1241
+ {
1242
+ "epoch": 20.31,
1243
+ "learning_rate": 2.969211045364892e-05,
1244
+ "loss": 1.0746,
1245
+ "step": 103000
1246
+ },
1247
+ {
1248
+ "epoch": 20.41,
1249
+ "learning_rate": 2.959349112426036e-05,
1250
+ "loss": 1.073,
1251
+ "step": 103500
1252
+ },
1253
+ {
1254
+ "epoch": 20.51,
1255
+ "learning_rate": 2.9494871794871796e-05,
1256
+ "loss": 1.0738,
1257
+ "step": 104000
1258
+ },
1259
+ {
1260
+ "epoch": 20.61,
1261
+ "learning_rate": 2.9396252465483236e-05,
1262
+ "loss": 1.0642,
1263
+ "step": 104500
1264
+ },
1265
+ {
1266
+ "epoch": 20.71,
1267
+ "learning_rate": 2.9297633136094677e-05,
1268
+ "loss": 1.0707,
1269
+ "step": 105000
1270
+ },
1271
+ {
1272
+ "epoch": 20.81,
1273
+ "learning_rate": 2.9199013806706117e-05,
1274
+ "loss": 1.0765,
1275
+ "step": 105500
1276
+ },
1277
+ {
1278
+ "epoch": 20.91,
1279
+ "learning_rate": 2.9100394477317554e-05,
1280
+ "loss": 1.0729,
1281
+ "step": 106000
1282
+ },
1283
+ {
1284
+ "epoch": 21.0,
1285
+ "learning_rate": 2.9001775147928994e-05,
1286
+ "loss": 1.0632,
1287
+ "step": 106500
1288
+ },
1289
+ {
1290
+ "epoch": 21.1,
1291
+ "learning_rate": 2.890355029585799e-05,
1292
+ "loss": 1.0581,
1293
+ "step": 107000
1294
+ },
1295
+ {
1296
+ "epoch": 21.2,
1297
+ "learning_rate": 2.880493096646943e-05,
1298
+ "loss": 1.058,
1299
+ "step": 107500
1300
+ },
1301
+ {
1302
+ "epoch": 21.3,
1303
+ "learning_rate": 2.870631163708087e-05,
1304
+ "loss": 1.0627,
1305
+ "step": 108000
1306
+ },
1307
+ {
1308
+ "epoch": 21.4,
1309
+ "learning_rate": 2.8607692307692306e-05,
1310
+ "loss": 1.0581,
1311
+ "step": 108500
1312
+ },
1313
+ {
1314
+ "epoch": 21.5,
1315
+ "learning_rate": 2.8509270216962525e-05,
1316
+ "loss": 1.0568,
1317
+ "step": 109000
1318
+ },
1319
+ {
1320
+ "epoch": 21.6,
1321
+ "learning_rate": 2.8410650887573966e-05,
1322
+ "loss": 1.05,
1323
+ "step": 109500
1324
+ },
1325
+ {
1326
+ "epoch": 21.69,
1327
+ "learning_rate": 2.8312031558185402e-05,
1328
+ "loss": 1.0602,
1329
+ "step": 110000
1330
+ },
1331
+ {
1332
+ "epoch": 21.79,
1333
+ "learning_rate": 2.8213412228796843e-05,
1334
+ "loss": 1.0533,
1335
+ "step": 110500
1336
+ },
1337
+ {
1338
+ "epoch": 21.89,
1339
+ "learning_rate": 2.8114792899408286e-05,
1340
+ "loss": 1.0573,
1341
+ "step": 111000
1342
+ },
1343
+ {
1344
+ "epoch": 21.99,
1345
+ "learning_rate": 2.8016370808678506e-05,
1346
+ "loss": 1.0523,
1347
+ "step": 111500
1348
+ },
1349
+ {
1350
+ "epoch": 22.09,
1351
+ "learning_rate": 2.7917751479289946e-05,
1352
+ "loss": 1.0414,
1353
+ "step": 112000
1354
+ },
1355
+ {
1356
+ "epoch": 22.19,
1357
+ "learning_rate": 2.7819132149901383e-05,
1358
+ "loss": 1.0416,
1359
+ "step": 112500
1360
+ },
1361
+ {
1362
+ "epoch": 22.29,
1363
+ "learning_rate": 2.7720512820512823e-05,
1364
+ "loss": 1.042,
1365
+ "step": 113000
1366
+ },
1367
+ {
1368
+ "epoch": 22.38,
1369
+ "learning_rate": 2.7621893491124263e-05,
1370
+ "loss": 1.0446,
1371
+ "step": 113500
1372
+ },
1373
+ {
1374
+ "epoch": 22.48,
1375
+ "learning_rate": 2.752347140039448e-05,
1376
+ "loss": 1.0371,
1377
+ "step": 114000
1378
+ },
1379
+ {
1380
+ "epoch": 22.58,
1381
+ "learning_rate": 2.742485207100592e-05,
1382
+ "loss": 1.0334,
1383
+ "step": 114500
1384
+ },
1385
+ {
1386
+ "epoch": 22.68,
1387
+ "learning_rate": 2.732623274161736e-05,
1388
+ "loss": 1.0409,
1389
+ "step": 115000
1390
+ },
1391
+ {
1392
+ "epoch": 22.78,
1393
+ "learning_rate": 2.7227810650887575e-05,
1394
+ "loss": 1.043,
1395
+ "step": 115500
1396
+ },
1397
+ {
1398
+ "epoch": 22.88,
1399
+ "learning_rate": 2.7129191321499015e-05,
1400
+ "loss": 1.0459,
1401
+ "step": 116000
1402
+ },
1403
+ {
1404
+ "epoch": 22.98,
1405
+ "learning_rate": 2.7030571992110452e-05,
1406
+ "loss": 1.0381,
1407
+ "step": 116500
1408
+ },
1409
+ {
1410
+ "epoch": 23.08,
1411
+ "learning_rate": 2.6931952662721893e-05,
1412
+ "loss": 1.0289,
1413
+ "step": 117000
1414
+ },
1415
+ {
1416
+ "epoch": 23.17,
1417
+ "learning_rate": 2.6833333333333333e-05,
1418
+ "loss": 1.0328,
1419
+ "step": 117500
1420
+ },
1421
+ {
1422
+ "epoch": 23.27,
1423
+ "learning_rate": 2.6734714003944773e-05,
1424
+ "loss": 1.0265,
1425
+ "step": 118000
1426
+ },
1427
+ {
1428
+ "epoch": 23.37,
1429
+ "learning_rate": 2.6636094674556217e-05,
1430
+ "loss": 1.0294,
1431
+ "step": 118500
1432
+ },
1433
+ {
1434
+ "epoch": 23.47,
1435
+ "learning_rate": 2.6537475345167657e-05,
1436
+ "loss": 1.0243,
1437
+ "step": 119000
1438
+ },
1439
+ {
1440
+ "epoch": 23.57,
1441
+ "learning_rate": 2.6438856015779094e-05,
1442
+ "loss": 1.0317,
1443
+ "step": 119500
1444
+ },
1445
+ {
1446
+ "epoch": 23.67,
1447
+ "learning_rate": 2.6340236686390534e-05,
1448
+ "loss": 1.0321,
1449
+ "step": 120000
1450
+ },
1451
+ {
1452
+ "epoch": 23.77,
1453
+ "learning_rate": 2.6241617357001974e-05,
1454
+ "loss": 1.0258,
1455
+ "step": 120500
1456
+ },
1457
+ {
1458
+ "epoch": 23.86,
1459
+ "learning_rate": 2.614319526627219e-05,
1460
+ "loss": 1.0245,
1461
+ "step": 121000
1462
+ },
1463
+ {
1464
+ "epoch": 23.96,
1465
+ "learning_rate": 2.604457593688363e-05,
1466
+ "loss": 1.0255,
1467
+ "step": 121500
1468
+ },
1469
+ {
1470
+ "epoch": 24.06,
1471
+ "learning_rate": 2.594615384615385e-05,
1472
+ "loss": 1.0241,
1473
+ "step": 122000
1474
+ },
1475
+ {
1476
+ "epoch": 24.16,
1477
+ "learning_rate": 2.5847534516765286e-05,
1478
+ "loss": 1.0119,
1479
+ "step": 122500
1480
+ },
1481
+ {
1482
+ "epoch": 24.26,
1483
+ "learning_rate": 2.5748915187376727e-05,
1484
+ "loss": 1.0134,
1485
+ "step": 123000
1486
+ },
1487
+ {
1488
+ "epoch": 24.36,
1489
+ "learning_rate": 2.5650295857988167e-05,
1490
+ "loss": 1.0086,
1491
+ "step": 123500
1492
+ },
1493
+ {
1494
+ "epoch": 24.46,
1495
+ "learning_rate": 2.5551676528599604e-05,
1496
+ "loss": 1.0143,
1497
+ "step": 124000
1498
+ },
1499
+ {
1500
+ "epoch": 24.55,
1501
+ "learning_rate": 2.5453057199211044e-05,
1502
+ "loss": 1.018,
1503
+ "step": 124500
1504
+ },
1505
+ {
1506
+ "epoch": 24.65,
1507
+ "learning_rate": 2.5354437869822484e-05,
1508
+ "loss": 1.0141,
1509
+ "step": 125000
1510
+ },
1511
+ {
1512
+ "epoch": 24.75,
1513
+ "learning_rate": 2.5255818540433928e-05,
1514
+ "loss": 1.0027,
1515
+ "step": 125500
1516
+ },
1517
+ {
1518
+ "epoch": 24.85,
1519
+ "learning_rate": 2.515759368836292e-05,
1520
+ "loss": 1.0171,
1521
+ "step": 126000
1522
+ },
1523
+ {
1524
+ "epoch": 24.95,
1525
+ "learning_rate": 2.5058974358974356e-05,
1526
+ "loss": 1.0162,
1527
+ "step": 126500
1528
+ },
1529
+ {
1530
+ "epoch": 25.05,
1531
+ "learning_rate": 2.49603550295858e-05,
1532
+ "loss": 1.0012,
1533
+ "step": 127000
1534
+ },
1535
+ {
1536
+ "epoch": 25.15,
1537
+ "learning_rate": 2.486173570019724e-05,
1538
+ "loss": 0.9968,
1539
+ "step": 127500
1540
+ },
1541
+ {
1542
+ "epoch": 25.24,
1543
+ "learning_rate": 2.4763116370808677e-05,
1544
+ "loss": 1.0019,
1545
+ "step": 128000
1546
+ },
1547
+ {
1548
+ "epoch": 25.34,
1549
+ "learning_rate": 2.466449704142012e-05,
1550
+ "loss": 0.9994,
1551
+ "step": 128500
1552
+ },
1553
+ {
1554
+ "epoch": 25.44,
1555
+ "learning_rate": 2.456587771203156e-05,
1556
+ "loss": 1.0005,
1557
+ "step": 129000
1558
+ },
1559
+ {
1560
+ "epoch": 25.54,
1561
+ "learning_rate": 2.4467258382642997e-05,
1562
+ "loss": 1.0004,
1563
+ "step": 129500
1564
+ },
1565
+ {
1566
+ "epoch": 25.64,
1567
+ "learning_rate": 2.4368639053254438e-05,
1568
+ "loss": 1.0023,
1569
+ "step": 130000
1570
+ },
1571
+ {
1572
+ "epoch": 25.74,
1573
+ "learning_rate": 2.4270019723865878e-05,
1574
+ "loss": 0.9932,
1575
+ "step": 130500
1576
+ },
1577
+ {
1578
+ "epoch": 25.84,
1579
+ "learning_rate": 2.4171400394477318e-05,
1580
+ "loss": 0.9891,
1581
+ "step": 131000
1582
+ },
1583
+ {
1584
+ "epoch": 25.93,
1585
+ "learning_rate": 2.4072978303747537e-05,
1586
+ "loss": 1.0018,
1587
+ "step": 131500
1588
+ },
1589
+ {
1590
+ "epoch": 26.03,
1591
+ "learning_rate": 2.3974358974358978e-05,
1592
+ "loss": 0.9974,
1593
+ "step": 132000
1594
+ },
1595
+ {
1596
+ "epoch": 26.13,
1597
+ "learning_rate": 2.3875739644970414e-05,
1598
+ "loss": 0.9823,
1599
+ "step": 132500
1600
+ },
1601
+ {
1602
+ "epoch": 26.23,
1603
+ "learning_rate": 2.3777317554240634e-05,
1604
+ "loss": 0.989,
1605
+ "step": 133000
1606
+ },
1607
+ {
1608
+ "epoch": 26.33,
1609
+ "learning_rate": 2.367889546351085e-05,
1610
+ "loss": 0.9932,
1611
+ "step": 133500
1612
+ },
1613
+ {
1614
+ "epoch": 26.43,
1615
+ "learning_rate": 2.358027613412229e-05,
1616
+ "loss": 0.9848,
1617
+ "step": 134000
1618
+ },
1619
+ {
1620
+ "epoch": 26.53,
1621
+ "learning_rate": 2.348165680473373e-05,
1622
+ "loss": 0.9874,
1623
+ "step": 134500
1624
+ },
1625
+ {
1626
+ "epoch": 26.63,
1627
+ "learning_rate": 2.3383037475345167e-05,
1628
+ "loss": 0.9876,
1629
+ "step": 135000
1630
+ },
1631
+ {
1632
+ "epoch": 26.72,
1633
+ "learning_rate": 2.3284418145956607e-05,
1634
+ "loss": 0.9832,
1635
+ "step": 135500
1636
+ },
1637
+ {
1638
+ "epoch": 26.82,
1639
+ "learning_rate": 2.318579881656805e-05,
1640
+ "loss": 0.9902,
1641
+ "step": 136000
1642
+ },
1643
+ {
1644
+ "epoch": 26.92,
1645
+ "learning_rate": 2.3087179487179488e-05,
1646
+ "loss": 0.9808,
1647
+ "step": 136500
1648
+ },
1649
+ {
1650
+ "epoch": 27.02,
1651
+ "learning_rate": 2.2988560157790928e-05,
1652
+ "loss": 0.9789,
1653
+ "step": 137000
1654
+ },
1655
+ {
1656
+ "epoch": 27.12,
1657
+ "learning_rate": 2.2890138067061147e-05,
1658
+ "loss": 0.9742,
1659
+ "step": 137500
1660
+ },
1661
+ {
1662
+ "epoch": 27.22,
1663
+ "learning_rate": 2.2791518737672584e-05,
1664
+ "loss": 0.9733,
1665
+ "step": 138000
1666
+ },
1667
+ {
1668
+ "epoch": 27.32,
1669
+ "learning_rate": 2.2692899408284024e-05,
1670
+ "loss": 0.9864,
1671
+ "step": 138500
1672
+ },
1673
+ {
1674
+ "epoch": 27.41,
1675
+ "learning_rate": 2.2594280078895464e-05,
1676
+ "loss": 0.9769,
1677
+ "step": 139000
1678
+ },
1679
+ {
1680
+ "epoch": 27.51,
1681
+ "learning_rate": 2.2495660749506905e-05,
1682
+ "loss": 0.9751,
1683
+ "step": 139500
1684
+ },
1685
+ {
1686
+ "epoch": 27.61,
1687
+ "learning_rate": 2.2397041420118345e-05,
1688
+ "loss": 0.9776,
1689
+ "step": 140000
1690
+ },
1691
+ {
1692
+ "epoch": 27.71,
1693
+ "learning_rate": 2.2298422090729785e-05,
1694
+ "loss": 0.9781,
1695
+ "step": 140500
1696
+ },
1697
+ {
1698
+ "epoch": 27.81,
1699
+ "learning_rate": 2.2199802761341222e-05,
1700
+ "loss": 0.9687,
1701
+ "step": 141000
1702
+ },
1703
+ {
1704
+ "epoch": 27.91,
1705
+ "learning_rate": 2.210138067061144e-05,
1706
+ "loss": 0.9738,
1707
+ "step": 141500
1708
+ },
1709
+ {
1710
+ "epoch": 28.01,
1711
+ "learning_rate": 2.200276134122288e-05,
1712
+ "loss": 0.9686,
1713
+ "step": 142000
1714
+ },
1715
+ {
1716
+ "epoch": 28.1,
1717
+ "learning_rate": 2.1904339250493097e-05,
1718
+ "loss": 0.9642,
1719
+ "step": 142500
1720
+ },
1721
+ {
1722
+ "epoch": 28.2,
1723
+ "learning_rate": 2.1805719921104537e-05,
1724
+ "loss": 0.9571,
1725
+ "step": 143000
1726
+ },
1727
+ {
1728
+ "epoch": 28.3,
1729
+ "learning_rate": 2.1707100591715978e-05,
1730
+ "loss": 0.9686,
1731
+ "step": 143500
1732
+ },
1733
+ {
1734
+ "epoch": 28.4,
1735
+ "learning_rate": 2.1608481262327418e-05,
1736
+ "loss": 0.9633,
1737
+ "step": 144000
1738
+ },
1739
+ {
1740
+ "epoch": 28.5,
1741
+ "learning_rate": 2.1509861932938858e-05,
1742
+ "loss": 0.9627,
1743
+ "step": 144500
1744
+ },
1745
+ {
1746
+ "epoch": 28.6,
1747
+ "learning_rate": 2.1411242603550295e-05,
1748
+ "loss": 0.9615,
1749
+ "step": 145000
1750
+ },
1751
+ {
1752
+ "epoch": 28.7,
1753
+ "learning_rate": 2.1312623274161735e-05,
1754
+ "loss": 0.9633,
1755
+ "step": 145500
1756
+ },
1757
+ {
1758
+ "epoch": 28.79,
1759
+ "learning_rate": 2.1214003944773175e-05,
1760
+ "loss": 0.9609,
1761
+ "step": 146000
1762
+ },
1763
+ {
1764
+ "epoch": 28.89,
1765
+ "learning_rate": 2.111538461538462e-05,
1766
+ "loss": 0.962,
1767
+ "step": 146500
1768
+ },
1769
+ {
1770
+ "epoch": 28.99,
1771
+ "learning_rate": 2.1016765285996056e-05,
1772
+ "loss": 0.9561,
1773
+ "step": 147000
1774
+ },
1775
+ {
1776
+ "epoch": 29.09,
1777
+ "learning_rate": 2.0918145956607496e-05,
1778
+ "loss": 0.9611,
1779
+ "step": 147500
1780
+ },
1781
+ {
1782
+ "epoch": 29.19,
1783
+ "learning_rate": 2.0819526627218936e-05,
1784
+ "loss": 0.9502,
1785
+ "step": 148000
1786
+ },
1787
+ {
1788
+ "epoch": 29.29,
1789
+ "learning_rate": 2.0720907297830373e-05,
1790
+ "loss": 0.9651,
1791
+ "step": 148500
1792
+ },
1793
+ {
1794
+ "epoch": 29.39,
1795
+ "learning_rate": 2.062268244575937e-05,
1796
+ "loss": 0.9502,
1797
+ "step": 149000
1798
+ },
1799
+ {
1800
+ "epoch": 29.48,
1801
+ "learning_rate": 2.052406311637081e-05,
1802
+ "loss": 0.9558,
1803
+ "step": 149500
1804
+ },
1805
+ {
1806
+ "epoch": 29.58,
1807
+ "learning_rate": 2.042544378698225e-05,
1808
+ "loss": 0.9614,
1809
+ "step": 150000
1810
+ },
1811
+ {
1812
+ "epoch": 29.68,
1813
+ "learning_rate": 2.0326824457593692e-05,
1814
+ "loss": 0.9507,
1815
+ "step": 150500
1816
+ },
1817
+ {
1818
+ "epoch": 29.78,
1819
+ "learning_rate": 2.0228402366863905e-05,
1820
+ "loss": 0.9465,
1821
+ "step": 151000
1822
+ },
1823
+ {
1824
+ "epoch": 29.88,
1825
+ "learning_rate": 2.0129783037475348e-05,
1826
+ "loss": 0.9483,
1827
+ "step": 151500
1828
+ },
1829
+ {
1830
+ "epoch": 29.98,
1831
+ "learning_rate": 2.0031163708086785e-05,
1832
+ "loss": 0.9488,
1833
+ "step": 152000
1834
+ },
1835
+ {
1836
+ "epoch": 30.08,
1837
+ "learning_rate": 1.9932544378698225e-05,
1838
+ "loss": 0.9501,
1839
+ "step": 152500
1840
+ },
1841
+ {
1842
+ "epoch": 30.18,
1843
+ "learning_rate": 1.9833925049309666e-05,
1844
+ "loss": 0.9477,
1845
+ "step": 153000
1846
+ },
1847
+ {
1848
+ "epoch": 30.27,
1849
+ "learning_rate": 1.9735305719921106e-05,
1850
+ "loss": 0.9401,
1851
+ "step": 153500
1852
+ },
1853
+ {
1854
+ "epoch": 30.37,
1855
+ "learning_rate": 1.9636686390532546e-05,
1856
+ "loss": 0.9528,
1857
+ "step": 154000
1858
+ },
1859
+ {
1860
+ "epoch": 30.47,
1861
+ "learning_rate": 1.9538067061143986e-05,
1862
+ "loss": 0.9412,
1863
+ "step": 154500
1864
+ },
1865
+ {
1866
+ "epoch": 30.57,
1867
+ "learning_rate": 1.9439644970414202e-05,
1868
+ "loss": 0.9421,
1869
+ "step": 155000
1870
+ },
1871
+ {
1872
+ "epoch": 30.67,
1873
+ "learning_rate": 1.934122287968442e-05,
1874
+ "loss": 0.9455,
1875
+ "step": 155500
1876
+ },
1877
+ {
1878
+ "epoch": 30.77,
1879
+ "learning_rate": 1.9242603550295858e-05,
1880
+ "loss": 0.9401,
1881
+ "step": 156000
1882
+ },
1883
+ {
1884
+ "epoch": 30.87,
1885
+ "learning_rate": 1.91439842209073e-05,
1886
+ "loss": 0.9396,
1887
+ "step": 156500
1888
+ },
1889
+ {
1890
+ "epoch": 30.96,
1891
+ "learning_rate": 1.904536489151874e-05,
1892
+ "loss": 0.937,
1893
+ "step": 157000
1894
+ },
1895
+ {
1896
+ "epoch": 31.06,
1897
+ "learning_rate": 1.894674556213018e-05,
1898
+ "loss": 0.9382,
1899
+ "step": 157500
1900
+ },
1901
+ {
1902
+ "epoch": 31.16,
1903
+ "learning_rate": 1.8848126232741616e-05,
1904
+ "loss": 0.9368,
1905
+ "step": 158000
1906
+ },
1907
+ {
1908
+ "epoch": 31.26,
1909
+ "learning_rate": 1.874950690335306e-05,
1910
+ "loss": 0.9322,
1911
+ "step": 158500
1912
+ },
1913
+ {
1914
+ "epoch": 31.36,
1915
+ "learning_rate": 1.86508875739645e-05,
1916
+ "loss": 0.938,
1917
+ "step": 159000
1918
+ },
1919
+ {
1920
+ "epoch": 31.46,
1921
+ "learning_rate": 1.8552268244575936e-05,
1922
+ "loss": 0.9279,
1923
+ "step": 159500
1924
+ },
1925
+ {
1926
+ "epoch": 31.56,
1927
+ "learning_rate": 1.8453648915187377e-05,
1928
+ "loss": 0.9364,
1929
+ "step": 160000
1930
+ },
1931
+ {
1932
+ "epoch": 31.65,
1933
+ "learning_rate": 1.8355226824457593e-05,
1934
+ "loss": 0.9404,
1935
+ "step": 160500
1936
+ },
1937
+ {
1938
+ "epoch": 31.75,
1939
+ "learning_rate": 1.8256607495069033e-05,
1940
+ "loss": 0.9264,
1941
+ "step": 161000
1942
+ },
1943
+ {
1944
+ "epoch": 31.85,
1945
+ "learning_rate": 1.8158185404339252e-05,
1946
+ "loss": 0.9406,
1947
+ "step": 161500
1948
+ },
1949
+ {
1950
+ "epoch": 31.95,
1951
+ "learning_rate": 1.805956607495069e-05,
1952
+ "loss": 0.9316,
1953
+ "step": 162000
1954
+ },
1955
+ {
1956
+ "epoch": 32.05,
1957
+ "learning_rate": 1.7961143984220908e-05,
1958
+ "loss": 0.9257,
1959
+ "step": 162500
1960
+ },
1961
+ {
1962
+ "epoch": 32.15,
1963
+ "learning_rate": 1.7862524654832348e-05,
1964
+ "loss": 0.9292,
1965
+ "step": 163000
1966
+ },
1967
+ {
1968
+ "epoch": 32.25,
1969
+ "learning_rate": 1.776390532544379e-05,
1970
+ "loss": 0.9198,
1971
+ "step": 163500
1972
+ },
1973
+ {
1974
+ "epoch": 32.34,
1975
+ "learning_rate": 1.766528599605523e-05,
1976
+ "loss": 0.9245,
1977
+ "step": 164000
1978
+ },
1979
+ {
1980
+ "epoch": 32.44,
1981
+ "learning_rate": 1.756666666666667e-05,
1982
+ "loss": 0.9244,
1983
+ "step": 164500
1984
+ },
1985
+ {
1986
+ "epoch": 32.54,
1987
+ "learning_rate": 1.7468047337278106e-05,
1988
+ "loss": 0.9294,
1989
+ "step": 165000
1990
+ },
1991
+ {
1992
+ "epoch": 32.64,
1993
+ "learning_rate": 1.7369428007889546e-05,
1994
+ "loss": 0.9284,
1995
+ "step": 165500
1996
+ },
1997
+ {
1998
+ "epoch": 32.74,
1999
+ "learning_rate": 1.727080867850099e-05,
2000
+ "loss": 0.9224,
2001
+ "step": 166000
2002
+ },
2003
+ {
2004
+ "epoch": 32.84,
2005
+ "learning_rate": 1.7172189349112427e-05,
2006
+ "loss": 0.9269,
2007
+ "step": 166500
2008
+ },
2009
+ {
2010
+ "epoch": 32.94,
2011
+ "learning_rate": 1.7073570019723867e-05,
2012
+ "loss": 0.9204,
2013
+ "step": 167000
2014
+ },
2015
+ {
2016
+ "epoch": 33.04,
2017
+ "learning_rate": 1.6974950690335307e-05,
2018
+ "loss": 0.9203,
2019
+ "step": 167500
2020
+ },
2021
+ {
2022
+ "epoch": 33.13,
2023
+ "learning_rate": 1.6876331360946744e-05,
2024
+ "loss": 0.9213,
2025
+ "step": 168000
2026
+ },
2027
+ {
2028
+ "epoch": 33.23,
2029
+ "learning_rate": 1.6777712031558184e-05,
2030
+ "loss": 0.9177,
2031
+ "step": 168500
2032
+ },
2033
+ {
2034
+ "epoch": 33.33,
2035
+ "learning_rate": 1.6679289940828403e-05,
2036
+ "loss": 0.9117,
2037
+ "step": 169000
2038
+ },
2039
+ {
2040
+ "epoch": 33.43,
2041
+ "learning_rate": 1.6580670611439844e-05,
2042
+ "loss": 0.9109,
2043
+ "step": 169500
2044
+ },
2045
+ {
2046
+ "epoch": 33.53,
2047
+ "learning_rate": 1.6482051282051284e-05,
2048
+ "loss": 0.919,
2049
+ "step": 170000
2050
+ },
2051
+ {
2052
+ "epoch": 33.63,
2053
+ "learning_rate": 1.6383431952662724e-05,
2054
+ "loss": 0.917,
2055
+ "step": 170500
2056
+ },
2057
+ {
2058
+ "epoch": 33.73,
2059
+ "learning_rate": 1.628500986193294e-05,
2060
+ "loss": 0.9071,
2061
+ "step": 171000
2062
+ },
2063
+ {
2064
+ "epoch": 33.82,
2065
+ "learning_rate": 1.618639053254438e-05,
2066
+ "loss": 0.9128,
2067
+ "step": 171500
2068
+ },
2069
+ {
2070
+ "epoch": 33.92,
2071
+ "learning_rate": 1.6087771203155817e-05,
2072
+ "loss": 0.9237,
2073
+ "step": 172000
2074
+ },
2075
+ {
2076
+ "epoch": 34.02,
2077
+ "learning_rate": 1.5989151873767257e-05,
2078
+ "loss": 0.9179,
2079
+ "step": 172500
2080
+ },
2081
+ {
2082
+ "epoch": 34.12,
2083
+ "learning_rate": 1.58905325443787e-05,
2084
+ "loss": 0.9095,
2085
+ "step": 173000
2086
+ },
2087
+ {
2088
+ "epoch": 34.22,
2089
+ "learning_rate": 1.579191321499014e-05,
2090
+ "loss": 0.9104,
2091
+ "step": 173500
2092
+ },
2093
+ {
2094
+ "epoch": 34.32,
2095
+ "learning_rate": 1.5693491124260357e-05,
2096
+ "loss": 0.9134,
2097
+ "step": 174000
2098
+ },
2099
+ {
2100
+ "epoch": 34.42,
2101
+ "learning_rate": 1.5594871794871797e-05,
2102
+ "loss": 0.9078,
2103
+ "step": 174500
2104
+ },
2105
+ {
2106
+ "epoch": 34.51,
2107
+ "learning_rate": 1.5496449704142013e-05,
2108
+ "loss": 0.9037,
2109
+ "step": 175000
2110
+ },
2111
+ {
2112
+ "epoch": 34.61,
2113
+ "learning_rate": 1.5397830374753453e-05,
2114
+ "loss": 0.9107,
2115
+ "step": 175500
2116
+ },
2117
+ {
2118
+ "epoch": 34.71,
2119
+ "learning_rate": 1.5299211045364893e-05,
2120
+ "loss": 0.9082,
2121
+ "step": 176000
2122
+ },
2123
+ {
2124
+ "epoch": 34.81,
2125
+ "learning_rate": 1.520059171597633e-05,
2126
+ "loss": 0.914,
2127
+ "step": 176500
2128
+ },
2129
+ {
2130
+ "epoch": 34.91,
2131
+ "learning_rate": 1.5101972386587774e-05,
2132
+ "loss": 0.9087,
2133
+ "step": 177000
2134
+ },
2135
+ {
2136
+ "epoch": 35.01,
2137
+ "learning_rate": 1.5003353057199212e-05,
2138
+ "loss": 0.9059,
2139
+ "step": 177500
2140
+ },
2141
+ {
2142
+ "epoch": 35.11,
2143
+ "learning_rate": 1.490493096646943e-05,
2144
+ "loss": 0.8985,
2145
+ "step": 178000
2146
+ },
2147
+ {
2148
+ "epoch": 35.2,
2149
+ "learning_rate": 1.4806311637080869e-05,
2150
+ "loss": 0.9007,
2151
+ "step": 178500
2152
+ },
2153
+ {
2154
+ "epoch": 35.3,
2155
+ "learning_rate": 1.4707692307692309e-05,
2156
+ "loss": 0.9071,
2157
+ "step": 179000
2158
+ },
2159
+ {
2160
+ "epoch": 35.4,
2161
+ "learning_rate": 1.4609072978303747e-05,
2162
+ "loss": 0.9013,
2163
+ "step": 179500
2164
+ },
2165
+ {
2166
+ "epoch": 35.5,
2167
+ "learning_rate": 1.4510453648915188e-05,
2168
+ "loss": 0.9006,
2169
+ "step": 180000
2170
+ },
2171
+ {
2172
+ "epoch": 35.6,
2173
+ "learning_rate": 1.4412031558185405e-05,
2174
+ "loss": 0.9026,
2175
+ "step": 180500
2176
+ },
2177
+ {
2178
+ "epoch": 35.7,
2179
+ "learning_rate": 1.4313412228796844e-05,
2180
+ "loss": 0.9051,
2181
+ "step": 181000
2182
+ },
2183
+ {
2184
+ "epoch": 35.8,
2185
+ "learning_rate": 1.4214792899408286e-05,
2186
+ "loss": 0.9061,
2187
+ "step": 181500
2188
+ },
2189
+ {
2190
+ "epoch": 35.89,
2191
+ "learning_rate": 1.4116173570019726e-05,
2192
+ "loss": 0.9025,
2193
+ "step": 182000
2194
+ },
2195
+ {
2196
+ "epoch": 35.99,
2197
+ "learning_rate": 1.4017554240631164e-05,
2198
+ "loss": 0.8981,
2199
+ "step": 182500
2200
+ },
2201
+ {
2202
+ "epoch": 36.09,
2203
+ "learning_rate": 1.3918934911242603e-05,
2204
+ "loss": 0.8951,
2205
+ "step": 183000
2206
+ },
2207
+ {
2208
+ "epoch": 36.19,
2209
+ "learning_rate": 1.3820315581854043e-05,
2210
+ "loss": 0.8995,
2211
+ "step": 183500
2212
+ },
2213
+ {
2214
+ "epoch": 36.29,
2215
+ "learning_rate": 1.3721696252465485e-05,
2216
+ "loss": 0.8995,
2217
+ "step": 184000
2218
+ },
2219
+ {
2220
+ "epoch": 36.39,
2221
+ "learning_rate": 1.36232741617357e-05,
2222
+ "loss": 0.8954,
2223
+ "step": 184500
2224
+ },
2225
+ {
2226
+ "epoch": 36.49,
2227
+ "learning_rate": 1.3524654832347141e-05,
2228
+ "loss": 0.8927,
2229
+ "step": 185000
2230
+ },
2231
+ {
2232
+ "epoch": 36.59,
2233
+ "learning_rate": 1.3426035502958581e-05,
2234
+ "loss": 0.8843,
2235
+ "step": 185500
2236
+ },
2237
+ {
2238
+ "epoch": 36.68,
2239
+ "learning_rate": 1.332741617357002e-05,
2240
+ "loss": 0.891,
2241
+ "step": 186000
2242
+ },
2243
+ {
2244
+ "epoch": 36.78,
2245
+ "learning_rate": 1.322879684418146e-05,
2246
+ "loss": 0.9012,
2247
+ "step": 186500
2248
+ },
2249
+ {
2250
+ "epoch": 36.88,
2251
+ "learning_rate": 1.3130177514792899e-05,
2252
+ "loss": 0.8937,
2253
+ "step": 187000
2254
+ },
2255
+ {
2256
+ "epoch": 36.98,
2257
+ "learning_rate": 1.303155818540434e-05,
2258
+ "loss": 0.8982,
2259
+ "step": 187500
2260
+ },
2261
+ {
2262
+ "epoch": 37.08,
2263
+ "learning_rate": 1.2932938856015781e-05,
2264
+ "loss": 0.8958,
2265
+ "step": 188000
2266
+ },
2267
+ {
2268
+ "epoch": 37.18,
2269
+ "learning_rate": 1.283431952662722e-05,
2270
+ "loss": 0.893,
2271
+ "step": 188500
2272
+ },
2273
+ {
2274
+ "epoch": 37.28,
2275
+ "learning_rate": 1.273570019723866e-05,
2276
+ "loss": 0.8875,
2277
+ "step": 189000
2278
+ },
2279
+ {
2280
+ "epoch": 37.37,
2281
+ "learning_rate": 1.2637278106508877e-05,
2282
+ "loss": 0.8863,
2283
+ "step": 189500
2284
+ },
2285
+ {
2286
+ "epoch": 37.47,
2287
+ "learning_rate": 1.2538658777120316e-05,
2288
+ "loss": 0.881,
2289
+ "step": 190000
2290
+ },
2291
+ {
2292
+ "epoch": 37.57,
2293
+ "learning_rate": 1.2440039447731756e-05,
2294
+ "loss": 0.8837,
2295
+ "step": 190500
2296
+ },
2297
+ {
2298
+ "epoch": 37.67,
2299
+ "learning_rate": 1.2341420118343196e-05,
2300
+ "loss": 0.8899,
2301
+ "step": 191000
2302
+ },
2303
+ {
2304
+ "epoch": 37.77,
2305
+ "learning_rate": 1.2242800788954635e-05,
2306
+ "loss": 0.8834,
2307
+ "step": 191500
2308
+ },
2309
+ {
2310
+ "epoch": 37.87,
2311
+ "learning_rate": 1.2144181459566075e-05,
2312
+ "loss": 0.889,
2313
+ "step": 192000
2314
+ },
2315
+ {
2316
+ "epoch": 37.97,
2317
+ "learning_rate": 1.2045562130177515e-05,
2318
+ "loss": 0.8911,
2319
+ "step": 192500
2320
+ },
2321
+ {
2322
+ "epoch": 38.06,
2323
+ "learning_rate": 1.1947140039447733e-05,
2324
+ "loss": 0.887,
2325
+ "step": 193000
2326
+ },
2327
+ {
2328
+ "epoch": 38.16,
2329
+ "learning_rate": 1.1848520710059171e-05,
2330
+ "loss": 0.8768,
2331
+ "step": 193500
2332
+ },
2333
+ {
2334
+ "epoch": 38.26,
2335
+ "learning_rate": 1.1749901380670612e-05,
2336
+ "loss": 0.8869,
2337
+ "step": 194000
2338
+ },
2339
+ {
2340
+ "epoch": 38.36,
2341
+ "learning_rate": 1.1651282051282052e-05,
2342
+ "loss": 0.8802,
2343
+ "step": 194500
2344
+ },
2345
+ {
2346
+ "epoch": 38.46,
2347
+ "learning_rate": 1.155266272189349e-05,
2348
+ "loss": 0.8752,
2349
+ "step": 195000
2350
+ },
2351
+ {
2352
+ "epoch": 38.56,
2353
+ "learning_rate": 1.1454240631163708e-05,
2354
+ "loss": 0.8792,
2355
+ "step": 195500
2356
+ },
2357
+ {
2358
+ "epoch": 38.66,
2359
+ "learning_rate": 1.135562130177515e-05,
2360
+ "loss": 0.8815,
2361
+ "step": 196000
2362
+ },
2363
+ {
2364
+ "epoch": 38.75,
2365
+ "learning_rate": 1.1257199211045366e-05,
2366
+ "loss": 0.8719,
2367
+ "step": 196500
2368
+ },
2369
+ {
2370
+ "epoch": 38.85,
2371
+ "learning_rate": 1.1158579881656806e-05,
2372
+ "loss": 0.8821,
2373
+ "step": 197000
2374
+ },
2375
+ {
2376
+ "epoch": 38.95,
2377
+ "learning_rate": 1.1059960552268244e-05,
2378
+ "loss": 0.88,
2379
+ "step": 197500
2380
+ },
2381
+ {
2382
+ "epoch": 39.05,
2383
+ "learning_rate": 1.0961341222879686e-05,
2384
+ "loss": 0.878,
2385
+ "step": 198000
2386
+ },
2387
+ {
2388
+ "epoch": 39.15,
2389
+ "learning_rate": 1.0862919132149902e-05,
2390
+ "loss": 0.8813,
2391
+ "step": 198500
2392
+ },
2393
+ {
2394
+ "epoch": 39.25,
2395
+ "learning_rate": 1.0764299802761342e-05,
2396
+ "loss": 0.8756,
2397
+ "step": 199000
2398
+ },
2399
+ {
2400
+ "epoch": 39.35,
2401
+ "learning_rate": 1.0665680473372781e-05,
2402
+ "loss": 0.8748,
2403
+ "step": 199500
2404
+ },
2405
+ {
2406
+ "epoch": 39.44,
2407
+ "learning_rate": 1.0567061143984223e-05,
2408
+ "loss": 0.8762,
2409
+ "step": 200000
2410
+ },
2411
+ {
2412
+ "epoch": 39.54,
2413
+ "learning_rate": 1.0468441814595661e-05,
2414
+ "loss": 0.876,
2415
+ "step": 200500
2416
+ },
2417
+ {
2418
+ "epoch": 39.64,
2419
+ "learning_rate": 1.03698224852071e-05,
2420
+ "loss": 0.8688,
2421
+ "step": 201000
2422
+ },
2423
+ {
2424
+ "epoch": 39.74,
2425
+ "learning_rate": 1.0271203155818542e-05,
2426
+ "loss": 0.8772,
2427
+ "step": 201500
2428
+ },
2429
+ {
2430
+ "epoch": 39.84,
2431
+ "learning_rate": 1.017258382642998e-05,
2432
+ "loss": 0.8721,
2433
+ "step": 202000
2434
+ },
2435
+ {
2436
+ "epoch": 39.94,
2437
+ "learning_rate": 1.007396449704142e-05,
2438
+ "loss": 0.8711,
2439
+ "step": 202500
2440
+ },
2441
+ {
2442
+ "epoch": 40.04,
2443
+ "learning_rate": 9.975542406311638e-06,
2444
+ "loss": 0.8737,
2445
+ "step": 203000
2446
+ },
2447
+ {
2448
+ "epoch": 40.14,
2449
+ "learning_rate": 9.876923076923078e-06,
2450
+ "loss": 0.8682,
2451
+ "step": 203500
2452
+ },
2453
+ {
2454
+ "epoch": 40.23,
2455
+ "learning_rate": 9.778303747534517e-06,
2456
+ "loss": 0.8691,
2457
+ "step": 204000
2458
+ },
2459
+ {
2460
+ "epoch": 40.33,
2461
+ "learning_rate": 9.679684418145957e-06,
2462
+ "loss": 0.8727,
2463
+ "step": 204500
2464
+ },
2465
+ {
2466
+ "epoch": 40.43,
2467
+ "learning_rate": 9.581065088757397e-06,
2468
+ "loss": 0.8725,
2469
+ "step": 205000
2470
+ },
2471
+ {
2472
+ "epoch": 40.53,
2473
+ "learning_rate": 9.482642998027613e-06,
2474
+ "loss": 0.8707,
2475
+ "step": 205500
2476
+ },
2477
+ {
2478
+ "epoch": 40.63,
2479
+ "learning_rate": 9.384023668639053e-06,
2480
+ "loss": 0.8761,
2481
+ "step": 206000
2482
+ },
2483
+ {
2484
+ "epoch": 40.73,
2485
+ "learning_rate": 9.285404339250494e-06,
2486
+ "loss": 0.8674,
2487
+ "step": 206500
2488
+ },
2489
+ {
2490
+ "epoch": 40.83,
2491
+ "learning_rate": 9.186785009861934e-06,
2492
+ "loss": 0.8629,
2493
+ "step": 207000
2494
+ },
2495
+ {
2496
+ "epoch": 40.92,
2497
+ "learning_rate": 9.088165680473374e-06,
2498
+ "loss": 0.8733,
2499
+ "step": 207500
2500
+ },
2501
+ {
2502
+ "epoch": 41.02,
2503
+ "learning_rate": 8.989546351084813e-06,
2504
+ "loss": 0.8634,
2505
+ "step": 208000
2506
+ },
2507
+ {
2508
+ "epoch": 41.12,
2509
+ "learning_rate": 8.89112426035503e-06,
2510
+ "loss": 0.8653,
2511
+ "step": 208500
2512
+ },
2513
+ {
2514
+ "epoch": 41.22,
2515
+ "learning_rate": 8.792504930966469e-06,
2516
+ "loss": 0.8645,
2517
+ "step": 209000
2518
+ },
2519
+ {
2520
+ "epoch": 41.32,
2521
+ "learning_rate": 8.69388560157791e-06,
2522
+ "loss": 0.8623,
2523
+ "step": 209500
2524
+ },
2525
+ {
2526
+ "epoch": 41.42,
2527
+ "learning_rate": 8.59526627218935e-06,
2528
+ "loss": 0.8655,
2529
+ "step": 210000
2530
+ },
2531
+ {
2532
+ "epoch": 41.52,
2533
+ "learning_rate": 8.496844181459567e-06,
2534
+ "loss": 0.8619,
2535
+ "step": 210500
2536
+ },
2537
+ {
2538
+ "epoch": 41.61,
2539
+ "learning_rate": 8.398224852071005e-06,
2540
+ "loss": 0.8728,
2541
+ "step": 211000
2542
+ },
2543
+ {
2544
+ "epoch": 41.71,
2545
+ "learning_rate": 8.299605522682447e-06,
2546
+ "loss": 0.8671,
2547
+ "step": 211500
2548
+ },
2549
+ {
2550
+ "epoch": 41.81,
2551
+ "learning_rate": 8.200986193293886e-06,
2552
+ "loss": 0.8606,
2553
+ "step": 212000
2554
+ },
2555
+ {
2556
+ "epoch": 41.91,
2557
+ "learning_rate": 8.102366863905324e-06,
2558
+ "loss": 0.8578,
2559
+ "step": 212500
2560
+ },
2561
+ {
2562
+ "epoch": 42.01,
2563
+ "learning_rate": 8.003747534516766e-06,
2564
+ "loss": 0.8644,
2565
+ "step": 213000
2566
+ },
2567
+ {
2568
+ "epoch": 42.11,
2569
+ "learning_rate": 7.905128205128205e-06,
2570
+ "loss": 0.8576,
2571
+ "step": 213500
2572
+ },
2573
+ {
2574
+ "epoch": 42.21,
2575
+ "learning_rate": 7.806508875739647e-06,
2576
+ "loss": 0.855,
2577
+ "step": 214000
2578
+ },
2579
+ {
2580
+ "epoch": 42.3,
2581
+ "learning_rate": 7.707889546351085e-06,
2582
+ "loss": 0.85,
2583
+ "step": 214500
2584
+ },
2585
+ {
2586
+ "epoch": 42.4,
2587
+ "learning_rate": 7.609270216962525e-06,
2588
+ "loss": 0.8654,
2589
+ "step": 215000
2590
+ },
2591
+ {
2592
+ "epoch": 42.5,
2593
+ "learning_rate": 7.510650887573965e-06,
2594
+ "loss": 0.8571,
2595
+ "step": 215500
2596
+ },
2597
+ {
2598
+ "epoch": 42.6,
2599
+ "learning_rate": 7.412031558185404e-06,
2600
+ "loss": 0.8673,
2601
+ "step": 216000
2602
+ },
2603
+ {
2604
+ "epoch": 42.7,
2605
+ "learning_rate": 7.313609467455622e-06,
2606
+ "loss": 0.8577,
2607
+ "step": 216500
2608
+ },
2609
+ {
2610
+ "epoch": 42.8,
2611
+ "learning_rate": 7.214990138067061e-06,
2612
+ "loss": 0.8535,
2613
+ "step": 217000
2614
+ },
2615
+ {
2616
+ "epoch": 42.9,
2617
+ "learning_rate": 7.116370808678502e-06,
2618
+ "loss": 0.8613,
2619
+ "step": 217500
2620
+ },
2621
+ {
2622
+ "epoch": 42.99,
2623
+ "learning_rate": 7.017751479289941e-06,
2624
+ "loss": 0.8586,
2625
+ "step": 218000
2626
+ },
2627
+ {
2628
+ "epoch": 43.09,
2629
+ "learning_rate": 6.919329388560158e-06,
2630
+ "loss": 0.8546,
2631
+ "step": 218500
2632
+ },
2633
+ {
2634
+ "epoch": 43.19,
2635
+ "learning_rate": 6.820710059171598e-06,
2636
+ "loss": 0.8521,
2637
+ "step": 219000
2638
+ },
2639
+ {
2640
+ "epoch": 43.29,
2641
+ "learning_rate": 6.722090729783037e-06,
2642
+ "loss": 0.8598,
2643
+ "step": 219500
2644
+ },
2645
+ {
2646
+ "epoch": 43.39,
2647
+ "learning_rate": 6.623668639053255e-06,
2648
+ "loss": 0.8527,
2649
+ "step": 220000
2650
+ },
2651
+ {
2652
+ "epoch": 43.49,
2653
+ "learning_rate": 6.525049309664695e-06,
2654
+ "loss": 0.8512,
2655
+ "step": 220500
2656
+ },
2657
+ {
2658
+ "epoch": 43.59,
2659
+ "learning_rate": 6.426429980276134e-06,
2660
+ "loss": 0.8522,
2661
+ "step": 221000
2662
+ },
2663
+ {
2664
+ "epoch": 43.69,
2665
+ "learning_rate": 6.327810650887574e-06,
2666
+ "loss": 0.855,
2667
+ "step": 221500
2668
+ },
2669
+ {
2670
+ "epoch": 43.78,
2671
+ "learning_rate": 6.229191321499015e-06,
2672
+ "loss": 0.854,
2673
+ "step": 222000
2674
+ },
2675
+ {
2676
+ "epoch": 43.88,
2677
+ "learning_rate": 6.130769230769231e-06,
2678
+ "loss": 0.8585,
2679
+ "step": 222500
2680
+ },
2681
+ {
2682
+ "epoch": 43.98,
2683
+ "learning_rate": 6.032149901380671e-06,
2684
+ "loss": 0.8577,
2685
+ "step": 223000
2686
+ },
2687
+ {
2688
+ "epoch": 44.08,
2689
+ "learning_rate": 5.933530571992111e-06,
2690
+ "loss": 0.8501,
2691
+ "step": 223500
2692
+ },
2693
+ {
2694
+ "epoch": 44.18,
2695
+ "learning_rate": 5.8349112426035505e-06,
2696
+ "loss": 0.8473,
2697
+ "step": 224000
2698
+ },
2699
+ {
2700
+ "epoch": 44.28,
2701
+ "learning_rate": 5.73629191321499e-06,
2702
+ "loss": 0.8528,
2703
+ "step": 224500
2704
+ },
2705
+ {
2706
+ "epoch": 44.38,
2707
+ "learning_rate": 5.6378698224852074e-06,
2708
+ "loss": 0.8459,
2709
+ "step": 225000
2710
+ },
2711
+ {
2712
+ "epoch": 44.47,
2713
+ "learning_rate": 5.539250493096648e-06,
2714
+ "loss": 0.8507,
2715
+ "step": 225500
2716
+ },
2717
+ {
2718
+ "epoch": 44.57,
2719
+ "learning_rate": 5.440631163708087e-06,
2720
+ "loss": 0.844,
2721
+ "step": 226000
2722
+ },
2723
+ {
2724
+ "epoch": 44.67,
2725
+ "learning_rate": 5.342011834319527e-06,
2726
+ "loss": 0.8505,
2727
+ "step": 226500
2728
+ },
2729
+ {
2730
+ "epoch": 44.77,
2731
+ "learning_rate": 5.243392504930967e-06,
2732
+ "loss": 0.8497,
2733
+ "step": 227000
2734
+ },
2735
+ {
2736
+ "epoch": 44.87,
2737
+ "learning_rate": 5.144773175542406e-06,
2738
+ "loss": 0.8534,
2739
+ "step": 227500
2740
+ },
2741
+ {
2742
+ "epoch": 44.97,
2743
+ "learning_rate": 5.046153846153846e-06,
2744
+ "loss": 0.8451,
2745
+ "step": 228000
2746
+ },
2747
+ {
2748
+ "epoch": 45.07,
2749
+ "learning_rate": 4.9475345167652866e-06,
2750
+ "loss": 0.8426,
2751
+ "step": 228500
2752
+ },
2753
+ {
2754
+ "epoch": 45.16,
2755
+ "learning_rate": 4.848915187376726e-06,
2756
+ "loss": 0.8503,
2757
+ "step": 229000
2758
+ },
2759
+ {
2760
+ "epoch": 45.26,
2761
+ "learning_rate": 4.750493096646943e-06,
2762
+ "loss": 0.8476,
2763
+ "step": 229500
2764
+ },
2765
+ {
2766
+ "epoch": 45.36,
2767
+ "learning_rate": 4.65207100591716e-06,
2768
+ "loss": 0.8432,
2769
+ "step": 230000
2770
+ },
2771
+ {
2772
+ "epoch": 45.46,
2773
+ "learning_rate": 4.5534516765286e-06,
2774
+ "loss": 0.8452,
2775
+ "step": 230500
2776
+ },
2777
+ {
2778
+ "epoch": 45.56,
2779
+ "learning_rate": 4.45483234714004e-06,
2780
+ "loss": 0.8435,
2781
+ "step": 231000
2782
+ },
2783
+ {
2784
+ "epoch": 45.66,
2785
+ "learning_rate": 4.356213017751479e-06,
2786
+ "loss": 0.8441,
2787
+ "step": 231500
2788
+ },
2789
+ {
2790
+ "epoch": 45.76,
2791
+ "learning_rate": 4.257593688362919e-06,
2792
+ "loss": 0.8435,
2793
+ "step": 232000
2794
+ },
2795
+ {
2796
+ "epoch": 45.85,
2797
+ "learning_rate": 4.15897435897436e-06,
2798
+ "loss": 0.8458,
2799
+ "step": 232500
2800
+ },
2801
+ {
2802
+ "epoch": 45.95,
2803
+ "learning_rate": 4.060355029585799e-06,
2804
+ "loss": 0.8538,
2805
+ "step": 233000
2806
+ },
2807
+ {
2808
+ "epoch": 46.05,
2809
+ "learning_rate": 3.961932938856016e-06,
2810
+ "loss": 0.8452,
2811
+ "step": 233500
2812
+ },
2813
+ {
2814
+ "epoch": 46.15,
2815
+ "learning_rate": 3.863510848126233e-06,
2816
+ "loss": 0.8372,
2817
+ "step": 234000
2818
+ },
2819
+ {
2820
+ "epoch": 46.25,
2821
+ "learning_rate": 3.7648915187376726e-06,
2822
+ "loss": 0.8456,
2823
+ "step": 234500
2824
+ },
2825
+ {
2826
+ "epoch": 46.35,
2827
+ "learning_rate": 3.6662721893491124e-06,
2828
+ "loss": 0.8453,
2829
+ "step": 235000
2830
+ },
2831
+ {
2832
+ "epoch": 46.45,
2833
+ "learning_rate": 3.5676528599605527e-06,
2834
+ "loss": 0.8438,
2835
+ "step": 235500
2836
+ },
2837
+ {
2838
+ "epoch": 46.54,
2839
+ "learning_rate": 3.4690335305719925e-06,
2840
+ "loss": 0.8473,
2841
+ "step": 236000
2842
+ },
2843
+ {
2844
+ "epoch": 46.64,
2845
+ "learning_rate": 3.370414201183432e-06,
2846
+ "loss": 0.8404,
2847
+ "step": 236500
2848
+ },
2849
+ {
2850
+ "epoch": 46.74,
2851
+ "learning_rate": 3.271794871794872e-06,
2852
+ "loss": 0.843,
2853
+ "step": 237000
2854
+ },
2855
+ {
2856
+ "epoch": 46.84,
2857
+ "learning_rate": 3.173175542406312e-06,
2858
+ "loss": 0.8439,
2859
+ "step": 237500
2860
+ },
2861
+ {
2862
+ "epoch": 46.94,
2863
+ "learning_rate": 3.0745562130177517e-06,
2864
+ "loss": 0.8492,
2865
+ "step": 238000
2866
+ },
2867
+ {
2868
+ "epoch": 47.04,
2869
+ "learning_rate": 2.975936883629191e-06,
2870
+ "loss": 0.8425,
2871
+ "step": 238500
2872
+ },
2873
+ {
2874
+ "epoch": 47.14,
2875
+ "learning_rate": 2.8775147928994087e-06,
2876
+ "loss": 0.8401,
2877
+ "step": 239000
2878
+ },
2879
+ {
2880
+ "epoch": 47.24,
2881
+ "learning_rate": 2.778895463510848e-06,
2882
+ "loss": 0.8437,
2883
+ "step": 239500
2884
+ },
2885
+ {
2886
+ "epoch": 47.33,
2887
+ "learning_rate": 2.6802761341222883e-06,
2888
+ "loss": 0.8375,
2889
+ "step": 240000
2890
+ },
2891
+ {
2892
+ "epoch": 47.43,
2893
+ "learning_rate": 2.581656804733728e-06,
2894
+ "loss": 0.8448,
2895
+ "step": 240500
2896
+ },
2897
+ {
2898
+ "epoch": 47.53,
2899
+ "learning_rate": 2.4830374753451675e-06,
2900
+ "loss": 0.8421,
2901
+ "step": 241000
2902
+ },
2903
+ {
2904
+ "epoch": 47.63,
2905
+ "learning_rate": 2.3844181459566077e-06,
2906
+ "loss": 0.8446,
2907
+ "step": 241500
2908
+ },
2909
+ {
2910
+ "epoch": 47.73,
2911
+ "learning_rate": 2.2859960552268244e-06,
2912
+ "loss": 0.8403,
2913
+ "step": 242000
2914
+ },
2915
+ {
2916
+ "epoch": 47.83,
2917
+ "learning_rate": 2.1873767258382646e-06,
2918
+ "loss": 0.8343,
2919
+ "step": 242500
2920
+ },
2921
+ {
2922
+ "epoch": 47.93,
2923
+ "learning_rate": 2.088757396449704e-06,
2924
+ "loss": 0.8386,
2925
+ "step": 243000
2926
+ },
2927
+ {
2928
+ "epoch": 48.02,
2929
+ "learning_rate": 1.9901380670611443e-06,
2930
+ "loss": 0.8284,
2931
+ "step": 243500
2932
+ },
2933
+ {
2934
+ "epoch": 48.12,
2935
+ "learning_rate": 1.8915187376725839e-06,
2936
+ "loss": 0.8372,
2937
+ "step": 244000
2938
+ },
2939
+ {
2940
+ "epoch": 48.22,
2941
+ "learning_rate": 1.7930966469428008e-06,
2942
+ "loss": 0.8392,
2943
+ "step": 244500
2944
+ },
2945
+ {
2946
+ "epoch": 48.32,
2947
+ "learning_rate": 1.6944773175542408e-06,
2948
+ "loss": 0.8418,
2949
+ "step": 245000
2950
+ },
2951
+ {
2952
+ "epoch": 48.42,
2953
+ "learning_rate": 1.5958579881656804e-06,
2954
+ "loss": 0.8393,
2955
+ "step": 245500
2956
+ },
2957
+ {
2958
+ "epoch": 48.52,
2959
+ "learning_rate": 1.4972386587771204e-06,
2960
+ "loss": 0.838,
2961
+ "step": 246000
2962
+ },
2963
+ {
2964
+ "epoch": 48.62,
2965
+ "learning_rate": 1.3986193293885602e-06,
2966
+ "loss": 0.8403,
2967
+ "step": 246500
2968
+ },
2969
+ {
2970
+ "epoch": 48.71,
2971
+ "learning_rate": 1.3001972386587771e-06,
2972
+ "loss": 0.8346,
2973
+ "step": 247000
2974
+ },
2975
+ {
2976
+ "epoch": 48.81,
2977
+ "learning_rate": 1.201577909270217e-06,
2978
+ "loss": 0.8348,
2979
+ "step": 247500
2980
+ },
2981
+ {
2982
+ "epoch": 48.91,
2983
+ "learning_rate": 1.102958579881657e-06,
2984
+ "loss": 0.8375,
2985
+ "step": 248000
2986
+ },
2987
+ {
2988
+ "epoch": 49.01,
2989
+ "learning_rate": 1.0043392504930968e-06,
2990
+ "loss": 0.838,
2991
+ "step": 248500
2992
+ },
2993
+ {
2994
+ "epoch": 49.11,
2995
+ "learning_rate": 9.057199211045366e-07,
2996
+ "loss": 0.831,
2997
+ "step": 249000
2998
+ },
2999
+ {
3000
+ "epoch": 49.21,
3001
+ "learning_rate": 8.072978303747534e-07,
3002
+ "loss": 0.8363,
3003
+ "step": 249500
3004
+ },
3005
+ {
3006
+ "epoch": 49.31,
3007
+ "learning_rate": 7.086785009861934e-07,
3008
+ "loss": 0.8378,
3009
+ "step": 250000
3010
+ },
3011
+ {
3012
+ "epoch": 49.4,
3013
+ "learning_rate": 6.102564102564103e-07,
3014
+ "loss": 0.8329,
3015
+ "step": 250500
3016
+ },
3017
+ {
3018
+ "epoch": 49.5,
3019
+ "learning_rate": 5.116370808678501e-07,
3020
+ "loss": 0.8419,
3021
+ "step": 251000
3022
+ },
3023
+ {
3024
+ "epoch": 49.6,
3025
+ "learning_rate": 4.1301775147929e-07,
3026
+ "loss": 0.8304,
3027
+ "step": 251500
3028
+ },
3029
+ {
3030
+ "epoch": 49.7,
3031
+ "learning_rate": 3.1439842209072983e-07,
3032
+ "loss": 0.8399,
3033
+ "step": 252000
3034
+ },
3035
+ {
3036
+ "epoch": 49.8,
3037
+ "learning_rate": 2.1577909270216962e-07,
3038
+ "loss": 0.834,
3039
+ "step": 252500
3040
+ },
3041
+ {
3042
+ "epoch": 49.9,
3043
+ "learning_rate": 1.1715976331360947e-07,
3044
+ "loss": 0.838,
3045
+ "step": 253000
3046
+ },
3047
+ {
3048
+ "epoch": 50.0,
3049
+ "learning_rate": 1.854043392504931e-08,
3050
+ "loss": 0.8346,
3051
+ "step": 253500
3052
+ },
3053
+ {
3054
+ "epoch": 50.0,
3055
+ "step": 253500,
3056
+ "total_flos": 2.1443533265845617e+18,
3057
+ "train_loss": 1.184410398825621,
3058
+ "train_runtime": 354987.1489,
3059
+ "train_samples_per_second": 22.853,
3060
+ "train_steps_per_second": 0.714
3061
+ }
3062
+ ],
3063
+ "logging_steps": 500,
3064
+ "max_steps": 253500,
3065
+ "num_train_epochs": 50,
3066
+ "save_steps": 10000,
3067
+ "total_flos": 2.1443533265845617e+18,
3068
+ "trial_name": null,
3069
+ "trial_params": null
3070
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b8c5d5d0a4e0249ebde874988b00f81533dbf641c765c56b012e5ebd93ee9843
3
+ size 4472