Hugging Face's logo

yesj1234
/
enko_xlsr_100p_run2

Automatic Speech Recognition
Transformers
Safetensors
wav2vec2
./train_dataset.py
Generated from Trainer
Model card Files
xet
Community
yesj1234 commited on
Commit
85115cb
·
1 Parent(s): 9a934a5

Upload folder using huggingface_hub

Browse files
Files changed (42) hide show
  1. README.md +60 -0
  2. added_tokens.json +4 -0
  3. all_results.json +8 -0
  4. checkpoint-1252379/added_tokens.json +4 -0
  5. checkpoint-1252379/config.json +117 -0
  6. checkpoint-1252379/model.safetensors +3 -0
  7. checkpoint-1252379/optimizer.pt +3 -0
  8. checkpoint-1252379/preprocessor_config.json +10 -0
  9. checkpoint-1252379/rng_state_0.pth +3 -0
  10. checkpoint-1252379/rng_state_1.pth +3 -0
  11. checkpoint-1252379/rng_state_2.pth +3 -0
  12. checkpoint-1252379/rng_state_3.pth +3 -0
  13. checkpoint-1252379/scheduler.pt +3 -0
  14. checkpoint-1252379/special_tokens_map.json +30 -0
  15. checkpoint-1252379/tokenizer_config.json +51 -0
  16. checkpoint-1252379/trainer_state.json +3775 -0
  17. checkpoint-1252379/training_args.bin +3 -0
  18. checkpoint-1252379/vocab.json +53 -0
  19. checkpoint-1295550/added_tokens.json +4 -0
  20. checkpoint-1295550/config.json +117 -0
  21. checkpoint-1295550/model.safetensors +3 -0
  22. checkpoint-1295550/optimizer.pt +3 -0
  23. checkpoint-1295550/preprocessor_config.json +10 -0
  24. checkpoint-1295550/rng_state_0.pth +3 -0
  25. checkpoint-1295550/rng_state_1.pth +3 -0
  26. checkpoint-1295550/rng_state_2.pth +3 -0
  27. checkpoint-1295550/rng_state_3.pth +3 -0
  28. checkpoint-1295550/scheduler.pt +3 -0
  29. checkpoint-1295550/special_tokens_map.json +30 -0
  30. checkpoint-1295550/tokenizer_config.json +51 -0
  31. checkpoint-1295550/trainer_state.json +3901 -0
  32. checkpoint-1295550/training_args.bin +3 -0
  33. checkpoint-1295550/vocab.json +53 -0
  34. config.json +117 -0
  35. model.safetensors +3 -0
  36. preprocessor_config.json +10 -0
  37. special_tokens_map.json +30 -0
  38. tokenizer_config.json +51 -0
  39. train_results.json +8 -0
  40. trainer_state.json +3910 -0
  41. training_args.bin +3 -0
  42. vocab.json +53 -0
README.md ADDED
@@ -0,0 +1,60 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ base_model: facebook/wav2vec2-large-xlsr-53
4
+ tags:
5
+ - automatic-speech-recognition
6
+ - ./train_dataset.py
7
+ - generated_from_trainer
8
+ model-index:
9
+ - name: en_xlsr_100p_run2
10
+ results: []
11
+ ---
12
+
13
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
14
+ should probably proofread and complete it, then remove this comment. -->
15
+
16
+ # en_xlsr_100p_run2
17
+
18
+ This model is a fine-tuned version of [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) on the ./TRAIN_DATASET.PY - NA dataset.
19
+
20
+ ## Model description
21
+
22
+ More information needed
23
+
24
+ ## Intended uses & limitations
25
+
26
+ More information needed
27
+
28
+ ## Training and evaluation data
29
+
30
+ More information needed
31
+
32
+ ## Training procedure
33
+
34
+ ### Training hyperparameters
35
+
36
+ The following hyperparameters were used during training:
37
+ - learning_rate: 0.0003
38
+ - train_batch_size: 2
39
+ - eval_batch_size: 8
40
+ - seed: 42
41
+ - distributed_type: multi-GPU
42
+ - num_devices: 4
43
+ - gradient_accumulation_steps: 2
44
+ - total_train_batch_size: 16
45
+ - total_eval_batch_size: 32
46
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
47
+ - lr_scheduler_type: linear
48
+ - lr_scheduler_warmup_ratio: 0.01
49
+ - num_epochs: 30
50
+
51
+ ### Training results
52
+
53
+
54
+
55
+ ### Framework versions
56
+
57
+ - Transformers 4.35.0
58
+ - Pytorch 2.1.0+cu121
59
+ - Datasets 2.14.6
60
+ - Tokenizers 0.14.1
added_tokens.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "</s>": 52,
3
+ "<s>": 51
4
+ }
all_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 30.0,
3
+ "train_loss": 0.7350748663054241,
4
+ "train_runtime": 658563.1153,
5
+ "train_samples": 690961,
6
+ "train_samples_per_second": 31.476,
7
+ "train_steps_per_second": 1.967
8
+ }
checkpoint-1252379/added_tokens.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "</s>": 52,
3
+ "<s>": 51
4
+ }
checkpoint-1252379/config.json ADDED
@@ -0,0 +1,117 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/wav2vec2-large-xlsr-53",
3
+ "activation_dropout": 0.0,
4
+ "adapter_attn_dim": null,
5
+ "adapter_kernel_size": 3,
6
+ "adapter_stride": 2,
7
+ "add_adapter": false,
8
+ "apply_spec_augment": true,
9
+ "architectures": [
10
+ "Wav2Vec2ForCTC"
11
+ ],
12
+ "attention_dropout": 0.05,
13
+ "bos_token_id": 1,
14
+ "classifier_proj_size": 256,
15
+ "codevector_dim": 768,
16
+ "contrastive_logits_temperature": 0.1,
17
+ "conv_bias": true,
18
+ "conv_dim": [
19
+ 512,
20
+ 512,
21
+ 512,
22
+ 512,
23
+ 512,
24
+ 512,
25
+ 512
26
+ ],
27
+ "conv_kernel": [
28
+ 10,
29
+ 3,
30
+ 3,
31
+ 3,
32
+ 3,
33
+ 2,
34
+ 2
35
+ ],
36
+ "conv_stride": [
37
+ 5,
38
+ 2,
39
+ 2,
40
+ 2,
41
+ 2,
42
+ 2,
43
+ 2
44
+ ],
45
+ "ctc_loss_reduction": "mean",
46
+ "ctc_zero_infinity": true,
47
+ "diversity_loss_weight": 0.1,
48
+ "do_stable_layer_norm": true,
49
+ "eos_token_id": 2,
50
+ "feat_extract_activation": "gelu",
51
+ "feat_extract_dropout": 0.0,
52
+ "feat_extract_norm": "layer",
53
+ "feat_proj_dropout": 0.05,
54
+ "feat_quantizer_dropout": 0.0,
55
+ "final_dropout": 0.0,
56
+ "gradient_checkpointing": false,
57
+ "hidden_act": "gelu",
58
+ "hidden_dropout": 0.05,
59
+ "hidden_size": 1024,
60
+ "initializer_range": 0.02,
61
+ "intermediate_size": 4096,
62
+ "layer_norm_eps": 1e-05,
63
+ "layerdrop": 0.05,
64
+ "mask_channel_length": 10,
65
+ "mask_channel_min_space": 1,
66
+ "mask_channel_other": 0.0,
67
+ "mask_channel_prob": 0.0,
68
+ "mask_channel_selection": "static",
69
+ "mask_feature_length": 10,
70
+ "mask_feature_min_masks": 0,
71
+ "mask_feature_prob": 0.0,
72
+ "mask_time_length": 10,
73
+ "mask_time_min_masks": 2,
74
+ "mask_time_min_space": 1,
75
+ "mask_time_other": 0.0,
76
+ "mask_time_prob": 0.05,
77
+ "mask_time_selection": "static",
78
+ "model_type": "wav2vec2",
79
+ "num_adapter_layers": 3,
80
+ "num_attention_heads": 16,
81
+ "num_codevector_groups": 2,
82
+ "num_codevectors_per_group": 320,
83
+ "num_conv_pos_embedding_groups": 16,
84
+ "num_conv_pos_embeddings": 128,
85
+ "num_feat_extract_layers": 7,
86
+ "num_hidden_layers": 24,
87
+ "num_negatives": 100,
88
+ "output_hidden_size": 1024,
89
+ "pad_token_id": 50,
90
+ "proj_codevector_dim": 768,
91
+ "tdnn_dilation": [
92
+ 1,
93
+ 2,
94
+ 3,
95
+ 1,
96
+ 1
97
+ ],
98
+ "tdnn_dim": [
99
+ 512,
100
+ 512,
101
+ 512,
102
+ 512,
103
+ 1500
104
+ ],
105
+ "tdnn_kernel": [
106
+ 5,
107
+ 3,
108
+ 3,
109
+ 1,
110
+ 1
111
+ ],
112
+ "torch_dtype": "float32",
113
+ "transformers_version": "4.35.0",
114
+ "use_weighted_layer_sum": false,
115
+ "vocab_size": 53,
116
+ "xvector_output_dim": 512
117
+ }
checkpoint-1252379/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c702e9df7d6016bff8739041000cca2d87c84032fec766def995c5aa3514cdd5
3
+ size 1262024780
checkpoint-1252379/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:12b957bc1d45934a2cce41e13a82dd21a021ff8d2e290d1ec14afd9cfe41e4f4
3
+ size 2490594358
checkpoint-1252379/preprocessor_config.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "feature_extractor_type": "Wav2Vec2FeatureExtractor",
4
+ "feature_size": 1,
5
+ "padding_side": "right",
6
+ "padding_value": 0,
7
+ "processor_class": "Wav2Vec2Processor",
8
+ "return_attention_mask": true,
9
+ "sampling_rate": 16000
10
+ }
checkpoint-1252379/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:547afd52601eb579b6c531b2ed8b775e0d9d926a8b450d45fb54bafa7187c8b7
3
+ size 15024
checkpoint-1252379/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c7534fe5568e7b5f3ab4fe0ddb954eaac1451fa3a2ad43e436b7ba380fd5ba6c
3
+ size 15024
checkpoint-1252379/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:556b7581e33217308bb7e2d60c6c3dbf13b8bbc968aa9537b788741776352c9b
3
+ size 15024
checkpoint-1252379/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:36c26fca8bfcfc8ce0a3daa0e3516294ae90eda3da9f06d290d940de84b84d78
3
+ size 15024
checkpoint-1252379/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cdb1737aa255b8f35915d8a1645fa8ba16910eb72156c335c84122cecd94d381
3
+ size 1064
checkpoint-1252379/special_tokens_map.json ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": true,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "</s>",
11
+ "lstrip": false,
12
+ "normalized": true,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": {
17
+ "content": "[PAD]",
18
+ "lstrip": true,
19
+ "normalized": false,
20
+ "rstrip": true,
21
+ "single_word": false
22
+ },
23
+ "unk_token": {
24
+ "content": "[UNK]",
25
+ "lstrip": true,
26
+ "normalized": false,
27
+ "rstrip": true,
28
+ "single_word": false
29
+ }
30
+ }
checkpoint-1252379/tokenizer_config.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "49": {
4
+ "content": "[UNK]",
5
+ "lstrip": true,
6
+ "normalized": false,
7
+ "rstrip": true,
8
+ "single_word": false,
9
+ "special": false
10
+ },
11
+ "50": {
12
+ "content": "[PAD]",
13
+ "lstrip": true,
14
+ "normalized": false,
15
+ "rstrip": true,
16
+ "single_word": false,
17
+ "special": false
18
+ },
19
+ "51": {
20
+ "content": "<s>",
21
+ "lstrip": false,
22
+ "normalized": true,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "52": {
28
+ "content": "</s>",
29
+ "lstrip": false,
30
+ "normalized": true,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ }
35
+ },
36
+ "bos_token": "<s>",
37
+ "clean_up_tokenization_spaces": true,
38
+ "config": null,
39
+ "do_lower_case": false,
40
+ "eos_token": "</s>",
41
+ "model_max_length": 1000000000000000019884624838656,
42
+ "pad_token": "[PAD]",
43
+ "processor_class": "Wav2Vec2Processor",
44
+ "replace_word_delimiter_char": " ",
45
+ "target_lang": null,
46
+ "tokenizer_class": "Wav2Vec2CTCTokenizer",
47
+ "tokenizer_type": "wav2vec2",
48
+ "trust_remote_code": false,
49
+ "unk_token": "[UNK]",
50
+ "word_delimiter_token": "|"
51
+ }
checkpoint-1252379/trainer_state.json ADDED
@@ -0,0 +1,3775 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 28.999988422039806,
5
+ "eval_steps": 500,
6
+ "global_step": 1252379,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.05,
13
+ "learning_rate": 4.6310589688175354e-05,
14
+ "loss": 4.5055,
15
+ "step": 2000
16
+ },
17
+ {
18
+ "epoch": 0.09,
19
+ "learning_rate": 9.262117937635071e-05,
20
+ "loss": 1.4851,
21
+ "step": 4000
22
+ },
23
+ {
24
+ "epoch": 0.14,
25
+ "learning_rate": 0.00013893176906452608,
26
+ "loss": 1.1808,
27
+ "step": 6000
28
+ },
29
+ {
30
+ "epoch": 0.19,
31
+ "learning_rate": 0.00018524235875270141,
32
+ "loss": 1.132,
33
+ "step": 8000
34
+ },
35
+ {
36
+ "epoch": 0.23,
37
+ "learning_rate": 0.0002315529484408768,
38
+ "loss": 1.1349,
39
+ "step": 10000
40
+ },
41
+ {
42
+ "epoch": 0.28,
43
+ "learning_rate": 0.00027786353812905216,
44
+ "loss": 1.1422,
45
+ "step": 12000
46
+ },
47
+ {
48
+ "epoch": 0.32,
49
+ "learning_rate": 0.0002997558073716234,
50
+ "loss": 1.1513,
51
+ "step": 14000
52
+ },
53
+ {
54
+ "epoch": 0.37,
55
+ "learning_rate": 0.0002992880054015534,
56
+ "loss": 1.142,
57
+ "step": 16000
58
+ },
59
+ {
60
+ "epoch": 0.42,
61
+ "learning_rate": 0.00029882020343148334,
62
+ "loss": 1.1303,
63
+ "step": 18000
64
+ },
65
+ {
66
+ "epoch": 0.46,
67
+ "learning_rate": 0.0002983524014614133,
68
+ "loss": 1.1215,
69
+ "step": 20000
70
+ },
71
+ {
72
+ "epoch": 0.51,
73
+ "learning_rate": 0.0002978845994913433,
74
+ "loss": 1.109,
75
+ "step": 22000
76
+ },
77
+ {
78
+ "epoch": 0.56,
79
+ "learning_rate": 0.00029741679752127325,
80
+ "loss": 1.1034,
81
+ "step": 24000
82
+ },
83
+ {
84
+ "epoch": 0.6,
85
+ "learning_rate": 0.0002969489955512032,
86
+ "loss": 1.0998,
87
+ "step": 26000
88
+ },
89
+ {
90
+ "epoch": 0.65,
91
+ "learning_rate": 0.00029648119358113324,
92
+ "loss": 1.0979,
93
+ "step": 28000
94
+ },
95
+ {
96
+ "epoch": 0.69,
97
+ "learning_rate": 0.00029601339161106316,
98
+ "loss": 1.083,
99
+ "step": 30000
100
+ },
101
+ {
102
+ "epoch": 0.74,
103
+ "learning_rate": 0.0002955455896409932,
104
+ "loss": 1.0741,
105
+ "step": 32000
106
+ },
107
+ {
108
+ "epoch": 0.79,
109
+ "learning_rate": 0.0002950777876709231,
110
+ "loss": 1.0741,
111
+ "step": 34000
112
+ },
113
+ {
114
+ "epoch": 0.83,
115
+ "learning_rate": 0.0002946099857008531,
116
+ "loss": 1.0707,
117
+ "step": 36000
118
+ },
119
+ {
120
+ "epoch": 0.88,
121
+ "learning_rate": 0.00029414218373078303,
122
+ "loss": 1.0689,
123
+ "step": 38000
124
+ },
125
+ {
126
+ "epoch": 0.93,
127
+ "learning_rate": 0.00029367438176071306,
128
+ "loss": 1.0536,
129
+ "step": 40000
130
+ },
131
+ {
132
+ "epoch": 0.97,
133
+ "learning_rate": 0.00029320657979064297,
134
+ "loss": 1.0556,
135
+ "step": 42000
136
+ },
137
+ {
138
+ "epoch": 1.02,
139
+ "learning_rate": 0.000292738777820573,
140
+ "loss": 1.0358,
141
+ "step": 44000
142
+ },
143
+ {
144
+ "epoch": 1.07,
145
+ "learning_rate": 0.0002922709758505029,
146
+ "loss": 1.018,
147
+ "step": 46000
148
+ },
149
+ {
150
+ "epoch": 1.11,
151
+ "learning_rate": 0.00029180317388043293,
152
+ "loss": 1.0162,
153
+ "step": 48000
154
+ },
155
+ {
156
+ "epoch": 1.16,
157
+ "learning_rate": 0.00029133537191036285,
158
+ "loss": 1.017,
159
+ "step": 50000
160
+ },
161
+ {
162
+ "epoch": 1.2,
163
+ "learning_rate": 0.00029086756994029287,
164
+ "loss": 1.0145,
165
+ "step": 52000
166
+ },
167
+ {
168
+ "epoch": 1.25,
169
+ "learning_rate": 0.0002903997679702228,
170
+ "loss": 1.0103,
171
+ "step": 54000
172
+ },
173
+ {
174
+ "epoch": 1.3,
175
+ "learning_rate": 0.0002899319660001528,
176
+ "loss": 1.0046,
177
+ "step": 56000
178
+ },
179
+ {
180
+ "epoch": 1.34,
181
+ "learning_rate": 0.0002894641640300827,
182
+ "loss": 1.01,
183
+ "step": 58000
184
+ },
185
+ {
186
+ "epoch": 1.39,
187
+ "learning_rate": 0.00028899636206001274,
188
+ "loss": 1.0146,
189
+ "step": 60000
190
+ },
191
+ {
192
+ "epoch": 1.44,
193
+ "learning_rate": 0.00028852856008994266,
194
+ "loss": 1.0122,
195
+ "step": 62000
196
+ },
197
+ {
198
+ "epoch": 1.48,
199
+ "learning_rate": 0.0002880607581198727,
200
+ "loss": 0.9961,
201
+ "step": 64000
202
+ },
203
+ {
204
+ "epoch": 1.53,
205
+ "learning_rate": 0.00028759295614980265,
206
+ "loss": 0.9914,
207
+ "step": 66000
208
+ },
209
+ {
210
+ "epoch": 1.57,
211
+ "learning_rate": 0.0002871251541797326,
212
+ "loss": 0.9936,
213
+ "step": 68000
214
+ },
215
+ {
216
+ "epoch": 1.62,
217
+ "learning_rate": 0.0002866573522096626,
218
+ "loss": 0.9998,
219
+ "step": 70000
220
+ },
221
+ {
222
+ "epoch": 1.67,
223
+ "learning_rate": 0.00028618955023959256,
224
+ "loss": 0.9833,
225
+ "step": 72000
226
+ },
227
+ {
228
+ "epoch": 1.71,
229
+ "learning_rate": 0.0002857217482695225,
230
+ "loss": 0.9901,
231
+ "step": 74000
232
+ },
233
+ {
234
+ "epoch": 1.76,
235
+ "learning_rate": 0.0002852539462994525,
236
+ "loss": 0.9905,
237
+ "step": 76000
238
+ },
239
+ {
240
+ "epoch": 1.81,
241
+ "learning_rate": 0.00028478614432938246,
242
+ "loss": 0.9845,
243
+ "step": 78000
244
+ },
245
+ {
246
+ "epoch": 1.85,
247
+ "learning_rate": 0.00028431834235931243,
248
+ "loss": 0.9886,
249
+ "step": 80000
250
+ },
251
+ {
252
+ "epoch": 1.9,
253
+ "learning_rate": 0.0002838505403892424,
254
+ "loss": 0.9848,
255
+ "step": 82000
256
+ },
257
+ {
258
+ "epoch": 1.95,
259
+ "learning_rate": 0.00028338273841917237,
260
+ "loss": 0.9827,
261
+ "step": 84000
262
+ },
263
+ {
264
+ "epoch": 1.99,
265
+ "learning_rate": 0.00028291493644910234,
266
+ "loss": 0.9764,
267
+ "step": 86000
268
+ },
269
+ {
270
+ "epoch": 2.04,
271
+ "learning_rate": 0.0002824471344790323,
272
+ "loss": 0.9546,
273
+ "step": 88000
274
+ },
275
+ {
276
+ "epoch": 2.08,
277
+ "learning_rate": 0.0002819793325089623,
278
+ "loss": 0.9388,
279
+ "step": 90000
280
+ },
281
+ {
282
+ "epoch": 2.13,
283
+ "learning_rate": 0.00028151153053889225,
284
+ "loss": 0.9432,
285
+ "step": 92000
286
+ },
287
+ {
288
+ "epoch": 2.18,
289
+ "learning_rate": 0.0002810437285688222,
290
+ "loss": 0.9591,
291
+ "step": 94000
292
+ },
293
+ {
294
+ "epoch": 2.22,
295
+ "learning_rate": 0.0002805759265987522,
296
+ "loss": 0.9431,
297
+ "step": 96000
298
+ },
299
+ {
300
+ "epoch": 2.27,
301
+ "learning_rate": 0.00028010812462868215,
302
+ "loss": 0.9388,
303
+ "step": 98000
304
+ },
305
+ {
306
+ "epoch": 2.32,
307
+ "learning_rate": 0.0002796403226586121,
308
+ "loss": 0.9421,
309
+ "step": 100000
310
+ },
311
+ {
312
+ "epoch": 2.36,
313
+ "learning_rate": 0.0002791725206885421,
314
+ "loss": 0.9452,
315
+ "step": 102000
316
+ },
317
+ {
318
+ "epoch": 2.41,
319
+ "learning_rate": 0.0002787047187184721,
320
+ "loss": 0.9535,
321
+ "step": 104000
322
+ },
323
+ {
324
+ "epoch": 2.45,
325
+ "learning_rate": 0.00027823691674840203,
326
+ "loss": 0.9373,
327
+ "step": 106000
328
+ },
329
+ {
330
+ "epoch": 2.5,
331
+ "learning_rate": 0.00027776911477833205,
332
+ "loss": 0.9345,
333
+ "step": 108000
334
+ },
335
+ {
336
+ "epoch": 2.55,
337
+ "learning_rate": 0.00027730131280826197,
338
+ "loss": 0.9478,
339
+ "step": 110000
340
+ },
341
+ {
342
+ "epoch": 2.59,
343
+ "learning_rate": 0.000276833510838192,
344
+ "loss": 0.9445,
345
+ "step": 112000
346
+ },
347
+ {
348
+ "epoch": 2.64,
349
+ "learning_rate": 0.0002763657088681219,
350
+ "loss": 0.9297,
351
+ "step": 114000
352
+ },
353
+ {
354
+ "epoch": 2.69,
355
+ "learning_rate": 0.0002758979068980519,
356
+ "loss": 0.9294,
357
+ "step": 116000
358
+ },
359
+ {
360
+ "epoch": 2.73,
361
+ "learning_rate": 0.00027543010492798184,
362
+ "loss": 0.9273,
363
+ "step": 118000
364
+ },
365
+ {
366
+ "epoch": 2.78,
367
+ "learning_rate": 0.00027496230295791186,
368
+ "loss": 0.9242,
369
+ "step": 120000
370
+ },
371
+ {
372
+ "epoch": 2.83,
373
+ "learning_rate": 0.0002744945009878418,
374
+ "loss": 0.9228,
375
+ "step": 122000
376
+ },
377
+ {
378
+ "epoch": 2.87,
379
+ "learning_rate": 0.0002740266990177718,
380
+ "loss": 0.9316,
381
+ "step": 124000
382
+ },
383
+ {
384
+ "epoch": 2.92,
385
+ "learning_rate": 0.0002735588970477017,
386
+ "loss": 0.9448,
387
+ "step": 126000
388
+ },
389
+ {
390
+ "epoch": 2.96,
391
+ "learning_rate": 0.00027309109507763174,
392
+ "loss": 0.9253,
393
+ "step": 128000
394
+ },
395
+ {
396
+ "epoch": 3.01,
397
+ "learning_rate": 0.00027262329310756165,
398
+ "loss": 0.9197,
399
+ "step": 130000
400
+ },
401
+ {
402
+ "epoch": 3.06,
403
+ "learning_rate": 0.0002721554911374917,
404
+ "loss": 0.9025,
405
+ "step": 132000
406
+ },
407
+ {
408
+ "epoch": 3.1,
409
+ "learning_rate": 0.0002716876891674216,
410
+ "loss": 0.8983,
411
+ "step": 134000
412
+ },
413
+ {
414
+ "epoch": 3.15,
415
+ "learning_rate": 0.0002712198871973516,
416
+ "loss": 0.9013,
417
+ "step": 136000
418
+ },
419
+ {
420
+ "epoch": 3.2,
421
+ "learning_rate": 0.0002707520852272816,
422
+ "loss": 0.9019,
423
+ "step": 138000
424
+ },
425
+ {
426
+ "epoch": 3.24,
427
+ "learning_rate": 0.00027028428325721155,
428
+ "loss": 0.9016,
429
+ "step": 140000
430
+ },
431
+ {
432
+ "epoch": 3.29,
433
+ "learning_rate": 0.0002698164812871415,
434
+ "loss": 0.8979,
435
+ "step": 142000
436
+ },
437
+ {
438
+ "epoch": 3.33,
439
+ "learning_rate": 0.0002693486793170715,
440
+ "loss": 0.8961,
441
+ "step": 144000
442
+ },
443
+ {
444
+ "epoch": 3.38,
445
+ "learning_rate": 0.00026888087734700146,
446
+ "loss": 0.9007,
447
+ "step": 146000
448
+ },
449
+ {
450
+ "epoch": 3.43,
451
+ "learning_rate": 0.00026841307537693143,
452
+ "loss": 0.8951,
453
+ "step": 148000
454
+ },
455
+ {
456
+ "epoch": 3.47,
457
+ "learning_rate": 0.0002679452734068614,
458
+ "loss": 0.8926,
459
+ "step": 150000
460
+ },
461
+ {
462
+ "epoch": 3.52,
463
+ "learning_rate": 0.00026747747143679137,
464
+ "loss": 0.8924,
465
+ "step": 152000
466
+ },
467
+ {
468
+ "epoch": 3.57,
469
+ "learning_rate": 0.00026700966946672134,
470
+ "loss": 0.9044,
471
+ "step": 154000
472
+ },
473
+ {
474
+ "epoch": 3.61,
475
+ "learning_rate": 0.0002665418674966513,
476
+ "loss": 0.8952,
477
+ "step": 156000
478
+ },
479
+ {
480
+ "epoch": 3.66,
481
+ "learning_rate": 0.00026607406552658127,
482
+ "loss": 0.9001,
483
+ "step": 158000
484
+ },
485
+ {
486
+ "epoch": 3.7,
487
+ "learning_rate": 0.00026560626355651124,
488
+ "loss": 0.8898,
489
+ "step": 160000
490
+ },
491
+ {
492
+ "epoch": 3.75,
493
+ "learning_rate": 0.0002651384615864412,
494
+ "loss": 0.895,
495
+ "step": 162000
496
+ },
497
+ {
498
+ "epoch": 3.8,
499
+ "learning_rate": 0.0002646706596163712,
500
+ "loss": 0.9015,
501
+ "step": 164000
502
+ },
503
+ {
504
+ "epoch": 3.84,
505
+ "learning_rate": 0.00026420285764630115,
506
+ "loss": 0.892,
507
+ "step": 166000
508
+ },
509
+ {
510
+ "epoch": 3.89,
511
+ "learning_rate": 0.0002637350556762311,
512
+ "loss": 0.8903,
513
+ "step": 168000
514
+ },
515
+ {
516
+ "epoch": 3.94,
517
+ "learning_rate": 0.0002632672537061611,
518
+ "loss": 0.8916,
519
+ "step": 170000
520
+ },
521
+ {
522
+ "epoch": 3.98,
523
+ "learning_rate": 0.00026279945173609105,
524
+ "loss": 0.8941,
525
+ "step": 172000
526
+ },
527
+ {
528
+ "epoch": 4.03,
529
+ "learning_rate": 0.000262331649766021,
530
+ "loss": 0.8771,
531
+ "step": 174000
532
+ },
533
+ {
534
+ "epoch": 4.08,
535
+ "learning_rate": 0.000261863847795951,
536
+ "loss": 0.8716,
537
+ "step": 176000
538
+ },
539
+ {
540
+ "epoch": 4.12,
541
+ "learning_rate": 0.00026139604582588096,
542
+ "loss": 0.8632,
543
+ "step": 178000
544
+ },
545
+ {
546
+ "epoch": 4.17,
547
+ "learning_rate": 0.00026092824385581093,
548
+ "loss": 0.8573,
549
+ "step": 180000
550
+ },
551
+ {
552
+ "epoch": 4.21,
553
+ "learning_rate": 0.0002604604418857409,
554
+ "loss": 0.8642,
555
+ "step": 182000
556
+ },
557
+ {
558
+ "epoch": 4.26,
559
+ "learning_rate": 0.00025999263991567087,
560
+ "loss": 0.8642,
561
+ "step": 184000
562
+ },
563
+ {
564
+ "epoch": 4.31,
565
+ "learning_rate": 0.00025952483794560084,
566
+ "loss": 0.8617,
567
+ "step": 186000
568
+ },
569
+ {
570
+ "epoch": 4.35,
571
+ "learning_rate": 0.0002590570359755308,
572
+ "loss": 0.8574,
573
+ "step": 188000
574
+ },
575
+ {
576
+ "epoch": 4.4,
577
+ "learning_rate": 0.0002585892340054608,
578
+ "loss": 0.8612,
579
+ "step": 190000
580
+ },
581
+ {
582
+ "epoch": 4.45,
583
+ "learning_rate": 0.00025812143203539074,
584
+ "loss": 0.8706,
585
+ "step": 192000
586
+ },
587
+ {
588
+ "epoch": 4.49,
589
+ "learning_rate": 0.0002576536300653207,
590
+ "loss": 0.8605,
591
+ "step": 194000
592
+ },
593
+ {
594
+ "epoch": 4.54,
595
+ "learning_rate": 0.0002571858280952507,
596
+ "loss": 0.8703,
597
+ "step": 196000
598
+ },
599
+ {
600
+ "epoch": 4.58,
601
+ "learning_rate": 0.00025671802612518065,
602
+ "loss": 0.8691,
603
+ "step": 198000
604
+ },
605
+ {
606
+ "epoch": 4.63,
607
+ "learning_rate": 0.0002562502241551106,
608
+ "loss": 0.873,
609
+ "step": 200000
610
+ },
611
+ {
612
+ "epoch": 4.68,
613
+ "learning_rate": 0.0002557824221850406,
614
+ "loss": 0.857,
615
+ "step": 202000
616
+ },
617
+ {
618
+ "epoch": 4.72,
619
+ "learning_rate": 0.00025531462021497056,
620
+ "loss": 0.8686,
621
+ "step": 204000
622
+ },
623
+ {
624
+ "epoch": 4.77,
625
+ "learning_rate": 0.0002548468182449005,
626
+ "loss": 0.868,
627
+ "step": 206000
628
+ },
629
+ {
630
+ "epoch": 4.82,
631
+ "learning_rate": 0.0002543790162748305,
632
+ "loss": 0.8756,
633
+ "step": 208000
634
+ },
635
+ {
636
+ "epoch": 4.86,
637
+ "learning_rate": 0.0002539112143047605,
638
+ "loss": 0.8656,
639
+ "step": 210000
640
+ },
641
+ {
642
+ "epoch": 4.91,
643
+ "learning_rate": 0.00025344341233469043,
644
+ "loss": 0.8725,
645
+ "step": 212000
646
+ },
647
+ {
648
+ "epoch": 4.96,
649
+ "learning_rate": 0.00025297561036462045,
650
+ "loss": 0.8675,
651
+ "step": 214000
652
+ },
653
+ {
654
+ "epoch": 5.0,
655
+ "learning_rate": 0.00025250780839455037,
656
+ "loss": 0.8644,
657
+ "step": 216000
658
+ },
659
+ {
660
+ "epoch": 5.05,
661
+ "learning_rate": 0.0002520400064244804,
662
+ "loss": 0.8397,
663
+ "step": 218000
664
+ },
665
+ {
666
+ "epoch": 5.09,
667
+ "learning_rate": 0.0002515722044544103,
668
+ "loss": 0.8374,
669
+ "step": 220000
670
+ },
671
+ {
672
+ "epoch": 5.14,
673
+ "learning_rate": 0.00025110440248434033,
674
+ "loss": 0.8364,
675
+ "step": 222000
676
+ },
677
+ {
678
+ "epoch": 5.19,
679
+ "learning_rate": 0.00025063660051427025,
680
+ "loss": 0.8454,
681
+ "step": 224000
682
+ },
683
+ {
684
+ "epoch": 5.23,
685
+ "learning_rate": 0.00025016879854420027,
686
+ "loss": 0.8367,
687
+ "step": 226000
688
+ },
689
+ {
690
+ "epoch": 5.28,
691
+ "learning_rate": 0.0002497009965741302,
692
+ "loss": 0.8275,
693
+ "step": 228000
694
+ },
695
+ {
696
+ "epoch": 5.33,
697
+ "learning_rate": 0.0002492331946040602,
698
+ "loss": 0.8398,
699
+ "step": 230000
700
+ },
701
+ {
702
+ "epoch": 5.37,
703
+ "learning_rate": 0.0002487653926339901,
704
+ "loss": 0.8403,
705
+ "step": 232000
706
+ },
707
+ {
708
+ "epoch": 5.42,
709
+ "learning_rate": 0.00024829759066392014,
710
+ "loss": 0.8409,
711
+ "step": 234000
712
+ },
713
+ {
714
+ "epoch": 5.46,
715
+ "learning_rate": 0.00024782978869385006,
716
+ "loss": 0.8366,
717
+ "step": 236000
718
+ },
719
+ {
720
+ "epoch": 5.51,
721
+ "learning_rate": 0.0002473619867237801,
722
+ "loss": 0.8381,
723
+ "step": 238000
724
+ },
725
+ {
726
+ "epoch": 5.56,
727
+ "learning_rate": 0.00024689418475371,
728
+ "loss": 0.842,
729
+ "step": 240000
730
+ },
731
+ {
732
+ "epoch": 5.6,
733
+ "learning_rate": 0.00024642638278364,
734
+ "loss": 0.843,
735
+ "step": 242000
736
+ },
737
+ {
738
+ "epoch": 5.65,
739
+ "learning_rate": 0.00024595858081357,
740
+ "loss": 0.8489,
741
+ "step": 244000
742
+ },
743
+ {
744
+ "epoch": 5.7,
745
+ "learning_rate": 0.00024549077884349996,
746
+ "loss": 0.8313,
747
+ "step": 246000
748
+ },
749
+ {
750
+ "epoch": 5.74,
751
+ "learning_rate": 0.0002450229768734299,
752
+ "loss": 0.8468,
753
+ "step": 248000
754
+ },
755
+ {
756
+ "epoch": 5.79,
757
+ "learning_rate": 0.0002445551749033599,
758
+ "loss": 0.8446,
759
+ "step": 250000
760
+ },
761
+ {
762
+ "epoch": 5.84,
763
+ "learning_rate": 0.00024408737293328986,
764
+ "loss": 0.8329,
765
+ "step": 252000
766
+ },
767
+ {
768
+ "epoch": 5.88,
769
+ "learning_rate": 0.00024361957096321983,
770
+ "loss": 0.8402,
771
+ "step": 254000
772
+ },
773
+ {
774
+ "epoch": 5.93,
775
+ "learning_rate": 0.0002431517689931498,
776
+ "loss": 0.8502,
777
+ "step": 256000
778
+ },
779
+ {
780
+ "epoch": 5.97,
781
+ "learning_rate": 0.00024268396702307977,
782
+ "loss": 0.8437,
783
+ "step": 258000
784
+ },
785
+ {
786
+ "epoch": 6.02,
787
+ "learning_rate": 0.00024221616505300974,
788
+ "loss": 0.8341,
789
+ "step": 260000
790
+ },
791
+ {
792
+ "epoch": 6.07,
793
+ "learning_rate": 0.0002417483630829397,
794
+ "loss": 0.8152,
795
+ "step": 262000
796
+ },
797
+ {
798
+ "epoch": 6.11,
799
+ "learning_rate": 0.00024128056111286968,
800
+ "loss": 0.8189,
801
+ "step": 264000
802
+ },
803
+ {
804
+ "epoch": 6.16,
805
+ "learning_rate": 0.00024081275914279965,
806
+ "loss": 0.8193,
807
+ "step": 266000
808
+ },
809
+ {
810
+ "epoch": 6.21,
811
+ "learning_rate": 0.00024034495717272961,
812
+ "loss": 0.825,
813
+ "step": 268000
814
+ },
815
+ {
816
+ "epoch": 6.25,
817
+ "learning_rate": 0.00023987715520265958,
818
+ "loss": 0.818,
819
+ "step": 270000
820
+ },
821
+ {
822
+ "epoch": 6.3,
823
+ "learning_rate": 0.00023940935323258955,
824
+ "loss": 0.8204,
825
+ "step": 272000
826
+ },
827
+ {
828
+ "epoch": 6.34,
829
+ "learning_rate": 0.00023894155126251952,
830
+ "loss": 0.823,
831
+ "step": 274000
832
+ },
833
+ {
834
+ "epoch": 6.39,
835
+ "learning_rate": 0.0002384737492924495,
836
+ "loss": 0.8179,
837
+ "step": 276000
838
+ },
839
+ {
840
+ "epoch": 6.44,
841
+ "learning_rate": 0.00023800594732237946,
842
+ "loss": 0.8152,
843
+ "step": 278000
844
+ },
845
+ {
846
+ "epoch": 6.48,
847
+ "learning_rate": 0.00023753814535230943,
848
+ "loss": 0.8178,
849
+ "step": 280000
850
+ },
851
+ {
852
+ "epoch": 6.53,
853
+ "learning_rate": 0.00023707034338223942,
854
+ "loss": 0.8212,
855
+ "step": 282000
856
+ },
857
+ {
858
+ "epoch": 6.58,
859
+ "learning_rate": 0.00023660254141216937,
860
+ "loss": 0.8161,
861
+ "step": 284000
862
+ },
863
+ {
864
+ "epoch": 6.62,
865
+ "learning_rate": 0.00023613473944209936,
866
+ "loss": 0.8162,
867
+ "step": 286000
868
+ },
869
+ {
870
+ "epoch": 6.67,
871
+ "learning_rate": 0.0002356669374720293,
872
+ "loss": 0.828,
873
+ "step": 288000
874
+ },
875
+ {
876
+ "epoch": 6.72,
877
+ "learning_rate": 0.0002351991355019593,
878
+ "loss": 0.823,
879
+ "step": 290000
880
+ },
881
+ {
882
+ "epoch": 6.76,
883
+ "learning_rate": 0.00023473133353188927,
884
+ "loss": 0.8156,
885
+ "step": 292000
886
+ },
887
+ {
888
+ "epoch": 6.81,
889
+ "learning_rate": 0.00023426353156181924,
890
+ "loss": 0.8176,
891
+ "step": 294000
892
+ },
893
+ {
894
+ "epoch": 6.85,
895
+ "learning_rate": 0.0002337957295917492,
896
+ "loss": 0.8226,
897
+ "step": 296000
898
+ },
899
+ {
900
+ "epoch": 6.9,
901
+ "learning_rate": 0.00023332792762167917,
902
+ "loss": 0.8189,
903
+ "step": 298000
904
+ },
905
+ {
906
+ "epoch": 6.95,
907
+ "learning_rate": 0.00023286012565160914,
908
+ "loss": 0.8082,
909
+ "step": 300000
910
+ },
911
+ {
912
+ "epoch": 6.99,
913
+ "learning_rate": 0.0002323923236815391,
914
+ "loss": 0.8257,
915
+ "step": 302000
916
+ },
917
+ {
918
+ "epoch": 7.04,
919
+ "learning_rate": 0.00023192452171146908,
920
+ "loss": 0.7925,
921
+ "step": 304000
922
+ },
923
+ {
924
+ "epoch": 7.09,
925
+ "learning_rate": 0.00023145671974139905,
926
+ "loss": 0.7902,
927
+ "step": 306000
928
+ },
929
+ {
930
+ "epoch": 7.13,
931
+ "learning_rate": 0.00023098891777132902,
932
+ "loss": 0.7994,
933
+ "step": 308000
934
+ },
935
+ {
936
+ "epoch": 7.18,
937
+ "learning_rate": 0.000230521115801259,
938
+ "loss": 0.8029,
939
+ "step": 310000
940
+ },
941
+ {
942
+ "epoch": 7.22,
943
+ "learning_rate": 0.00023005331383118896,
944
+ "loss": 0.7983,
945
+ "step": 312000
946
+ },
947
+ {
948
+ "epoch": 7.27,
949
+ "learning_rate": 0.00022958551186111893,
950
+ "loss": 0.797,
951
+ "step": 314000
952
+ },
953
+ {
954
+ "epoch": 7.32,
955
+ "learning_rate": 0.00022911770989104892,
956
+ "loss": 0.8009,
957
+ "step": 316000
958
+ },
959
+ {
960
+ "epoch": 7.36,
961
+ "learning_rate": 0.00022864990792097886,
962
+ "loss": 0.7979,
963
+ "step": 318000
964
+ },
965
+ {
966
+ "epoch": 7.41,
967
+ "learning_rate": 0.00022818210595090886,
968
+ "loss": 0.8055,
969
+ "step": 320000
970
+ },
971
+ {
972
+ "epoch": 7.46,
973
+ "learning_rate": 0.0002277143039808388,
974
+ "loss": 0.7971,
975
+ "step": 322000
976
+ },
977
+ {
978
+ "epoch": 7.5,
979
+ "learning_rate": 0.0002272465020107688,
980
+ "loss": 0.8066,
981
+ "step": 324000
982
+ },
983
+ {
984
+ "epoch": 7.55,
985
+ "learning_rate": 0.00022677870004069874,
986
+ "loss": 0.7975,
987
+ "step": 326000
988
+ },
989
+ {
990
+ "epoch": 7.6,
991
+ "learning_rate": 0.00022631089807062873,
992
+ "loss": 0.7912,
993
+ "step": 328000
994
+ },
995
+ {
996
+ "epoch": 7.64,
997
+ "learning_rate": 0.00022584309610055868,
998
+ "loss": 0.7988,
999
+ "step": 330000
1000
+ },
1001
+ {
1002
+ "epoch": 7.69,
1003
+ "learning_rate": 0.00022537529413048867,
1004
+ "loss": 0.7999,
1005
+ "step": 332000
1006
+ },
1007
+ {
1008
+ "epoch": 7.73,
1009
+ "learning_rate": 0.00022490749216041861,
1010
+ "loss": 0.8019,
1011
+ "step": 334000
1012
+ },
1013
+ {
1014
+ "epoch": 7.78,
1015
+ "learning_rate": 0.0002244396901903486,
1016
+ "loss": 0.8108,
1017
+ "step": 336000
1018
+ },
1019
+ {
1020
+ "epoch": 7.83,
1021
+ "learning_rate": 0.00022397188822027855,
1022
+ "loss": 0.8075,
1023
+ "step": 338000
1024
+ },
1025
+ {
1026
+ "epoch": 7.87,
1027
+ "learning_rate": 0.00022350408625020855,
1028
+ "loss": 0.7995,
1029
+ "step": 340000
1030
+ },
1031
+ {
1032
+ "epoch": 7.92,
1033
+ "learning_rate": 0.0002230362842801385,
1034
+ "loss": 0.802,
1035
+ "step": 342000
1036
+ },
1037
+ {
1038
+ "epoch": 7.97,
1039
+ "learning_rate": 0.00022256848231006848,
1040
+ "loss": 0.7998,
1041
+ "step": 344000
1042
+ },
1043
+ {
1044
+ "epoch": 8.01,
1045
+ "learning_rate": 0.00022210068033999843,
1046
+ "loss": 0.8023,
1047
+ "step": 346000
1048
+ },
1049
+ {
1050
+ "epoch": 8.06,
1051
+ "learning_rate": 0.00022163287836992842,
1052
+ "loss": 0.7748,
1053
+ "step": 348000
1054
+ },
1055
+ {
1056
+ "epoch": 8.1,
1057
+ "learning_rate": 0.00022116507639985836,
1058
+ "loss": 0.7839,
1059
+ "step": 350000
1060
+ },
1061
+ {
1062
+ "epoch": 8.15,
1063
+ "learning_rate": 0.00022069727442978836,
1064
+ "loss": 0.7806,
1065
+ "step": 352000
1066
+ },
1067
+ {
1068
+ "epoch": 8.2,
1069
+ "learning_rate": 0.00022022947245971836,
1070
+ "loss": 0.7775,
1071
+ "step": 354000
1072
+ },
1073
+ {
1074
+ "epoch": 8.24,
1075
+ "learning_rate": 0.0002197616704896483,
1076
+ "loss": 0.7734,
1077
+ "step": 356000
1078
+ },
1079
+ {
1080
+ "epoch": 8.29,
1081
+ "learning_rate": 0.0002192938685195783,
1082
+ "loss": 0.7728,
1083
+ "step": 358000
1084
+ },
1085
+ {
1086
+ "epoch": 8.34,
1087
+ "learning_rate": 0.00021882606654950824,
1088
+ "loss": 0.7879,
1089
+ "step": 360000
1090
+ },
1091
+ {
1092
+ "epoch": 8.38,
1093
+ "learning_rate": 0.00021835826457943823,
1094
+ "loss": 0.7891,
1095
+ "step": 362000
1096
+ },
1097
+ {
1098
+ "epoch": 8.43,
1099
+ "learning_rate": 0.00021789046260936817,
1100
+ "loss": 0.7922,
1101
+ "step": 364000
1102
+ },
1103
+ {
1104
+ "epoch": 8.48,
1105
+ "learning_rate": 0.00021742266063929817,
1106
+ "loss": 0.7837,
1107
+ "step": 366000
1108
+ },
1109
+ {
1110
+ "epoch": 8.52,
1111
+ "learning_rate": 0.0002169548586692281,
1112
+ "loss": 0.7838,
1113
+ "step": 368000
1114
+ },
1115
+ {
1116
+ "epoch": 8.57,
1117
+ "learning_rate": 0.0002164870566991581,
1118
+ "loss": 0.7797,
1119
+ "step": 370000
1120
+ },
1121
+ {
1122
+ "epoch": 8.61,
1123
+ "learning_rate": 0.00021601925472908805,
1124
+ "loss": 0.7818,
1125
+ "step": 372000
1126
+ },
1127
+ {
1128
+ "epoch": 8.66,
1129
+ "learning_rate": 0.00021555145275901804,
1130
+ "loss": 0.7838,
1131
+ "step": 374000
1132
+ },
1133
+ {
1134
+ "epoch": 8.71,
1135
+ "learning_rate": 0.000215083650788948,
1136
+ "loss": 0.7828,
1137
+ "step": 376000
1138
+ },
1139
+ {
1140
+ "epoch": 8.75,
1141
+ "learning_rate": 0.00021461584881887798,
1142
+ "loss": 0.7822,
1143
+ "step": 378000
1144
+ },
1145
+ {
1146
+ "epoch": 8.8,
1147
+ "learning_rate": 0.00021414804684880792,
1148
+ "loss": 0.7952,
1149
+ "step": 380000
1150
+ },
1151
+ {
1152
+ "epoch": 8.85,
1153
+ "learning_rate": 0.00021368024487873792,
1154
+ "loss": 0.7888,
1155
+ "step": 382000
1156
+ },
1157
+ {
1158
+ "epoch": 8.89,
1159
+ "learning_rate": 0.00021321244290866786,
1160
+ "loss": 0.7813,
1161
+ "step": 384000
1162
+ },
1163
+ {
1164
+ "epoch": 8.94,
1165
+ "learning_rate": 0.00021274464093859786,
1166
+ "loss": 0.7784,
1167
+ "step": 386000
1168
+ },
1169
+ {
1170
+ "epoch": 8.98,
1171
+ "learning_rate": 0.00021227683896852783,
1172
+ "loss": 0.776,
1173
+ "step": 388000
1174
+ },
1175
+ {
1176
+ "epoch": 9.03,
1177
+ "learning_rate": 0.0002118090369984578,
1178
+ "loss": 0.7686,
1179
+ "step": 390000
1180
+ },
1181
+ {
1182
+ "epoch": 9.08,
1183
+ "learning_rate": 0.0002113412350283878,
1184
+ "loss": 0.7576,
1185
+ "step": 392000
1186
+ },
1187
+ {
1188
+ "epoch": 9.12,
1189
+ "learning_rate": 0.00021087343305831773,
1190
+ "loss": 0.7611,
1191
+ "step": 394000
1192
+ },
1193
+ {
1194
+ "epoch": 9.17,
1195
+ "learning_rate": 0.00021040563108824773,
1196
+ "loss": 0.7698,
1197
+ "step": 396000
1198
+ },
1199
+ {
1200
+ "epoch": 9.22,
1201
+ "learning_rate": 0.00020993782911817767,
1202
+ "loss": 0.7632,
1203
+ "step": 398000
1204
+ },
1205
+ {
1206
+ "epoch": 9.26,
1207
+ "learning_rate": 0.00020947002714810767,
1208
+ "loss": 0.7725,
1209
+ "step": 400000
1210
+ },
1211
+ {
1212
+ "epoch": 9.31,
1213
+ "learning_rate": 0.0002090022251780376,
1214
+ "loss": 0.7706,
1215
+ "step": 402000
1216
+ },
1217
+ {
1218
+ "epoch": 9.35,
1219
+ "learning_rate": 0.0002085344232079676,
1220
+ "loss": 0.7709,
1221
+ "step": 404000
1222
+ },
1223
+ {
1224
+ "epoch": 9.4,
1225
+ "learning_rate": 0.00020806662123789755,
1226
+ "loss": 0.7651,
1227
+ "step": 406000
1228
+ },
1229
+ {
1230
+ "epoch": 9.45,
1231
+ "learning_rate": 0.00020759881926782754,
1232
+ "loss": 0.7657,
1233
+ "step": 408000
1234
+ },
1235
+ {
1236
+ "epoch": 9.49,
1237
+ "learning_rate": 0.00020713101729775748,
1238
+ "loss": 0.7589,
1239
+ "step": 410000
1240
+ },
1241
+ {
1242
+ "epoch": 9.54,
1243
+ "learning_rate": 0.00020666321532768748,
1244
+ "loss": 0.7683,
1245
+ "step": 412000
1246
+ },
1247
+ {
1248
+ "epoch": 9.59,
1249
+ "learning_rate": 0.00020619541335761742,
1250
+ "loss": 0.7684,
1251
+ "step": 414000
1252
+ },
1253
+ {
1254
+ "epoch": 9.63,
1255
+ "learning_rate": 0.00020572761138754742,
1256
+ "loss": 0.7756,
1257
+ "step": 416000
1258
+ },
1259
+ {
1260
+ "epoch": 9.68,
1261
+ "learning_rate": 0.00020525980941747736,
1262
+ "loss": 0.7653,
1263
+ "step": 418000
1264
+ },
1265
+ {
1266
+ "epoch": 9.73,
1267
+ "learning_rate": 0.00020479200744740736,
1268
+ "loss": 0.7718,
1269
+ "step": 420000
1270
+ },
1271
+ {
1272
+ "epoch": 9.77,
1273
+ "learning_rate": 0.0002043242054773373,
1274
+ "loss": 0.7676,
1275
+ "step": 422000
1276
+ },
1277
+ {
1278
+ "epoch": 9.82,
1279
+ "learning_rate": 0.0002038564035072673,
1280
+ "loss": 0.772,
1281
+ "step": 424000
1282
+ },
1283
+ {
1284
+ "epoch": 9.86,
1285
+ "learning_rate": 0.00020338860153719726,
1286
+ "loss": 0.766,
1287
+ "step": 426000
1288
+ },
1289
+ {
1290
+ "epoch": 9.91,
1291
+ "learning_rate": 0.00020292079956712723,
1292
+ "loss": 0.7739,
1293
+ "step": 428000
1294
+ },
1295
+ {
1296
+ "epoch": 9.96,
1297
+ "learning_rate": 0.0002024529975970572,
1298
+ "loss": 0.7743,
1299
+ "step": 430000
1300
+ },
1301
+ {
1302
+ "epoch": 10.0,
1303
+ "learning_rate": 0.00020198519562698717,
1304
+ "loss": 0.7719,
1305
+ "step": 432000
1306
+ },
1307
+ {
1308
+ "epoch": 10.05,
1309
+ "learning_rate": 0.00020151739365691714,
1310
+ "loss": 0.7443,
1311
+ "step": 434000
1312
+ },
1313
+ {
1314
+ "epoch": 10.1,
1315
+ "learning_rate": 0.0002010495916868471,
1316
+ "loss": 0.7573,
1317
+ "step": 436000
1318
+ },
1319
+ {
1320
+ "epoch": 10.14,
1321
+ "learning_rate": 0.00020058178971677708,
1322
+ "loss": 0.7546,
1323
+ "step": 438000
1324
+ },
1325
+ {
1326
+ "epoch": 10.19,
1327
+ "learning_rate": 0.00020011398774670704,
1328
+ "loss": 0.7516,
1329
+ "step": 440000
1330
+ },
1331
+ {
1332
+ "epoch": 10.23,
1333
+ "learning_rate": 0.000199646185776637,
1334
+ "loss": 0.7444,
1335
+ "step": 442000
1336
+ },
1337
+ {
1338
+ "epoch": 10.28,
1339
+ "learning_rate": 0.00019917838380656698,
1340
+ "loss": 0.7656,
1341
+ "step": 444000
1342
+ },
1343
+ {
1344
+ "epoch": 10.33,
1345
+ "learning_rate": 0.00019871058183649695,
1346
+ "loss": 0.7452,
1347
+ "step": 446000
1348
+ },
1349
+ {
1350
+ "epoch": 10.37,
1351
+ "learning_rate": 0.00019824277986642692,
1352
+ "loss": 0.7555,
1353
+ "step": 448000
1354
+ },
1355
+ {
1356
+ "epoch": 10.42,
1357
+ "learning_rate": 0.0001977749778963569,
1358
+ "loss": 0.7486,
1359
+ "step": 450000
1360
+ },
1361
+ {
1362
+ "epoch": 10.47,
1363
+ "learning_rate": 0.00019730717592628686,
1364
+ "loss": 0.7509,
1365
+ "step": 452000
1366
+ },
1367
+ {
1368
+ "epoch": 10.51,
1369
+ "learning_rate": 0.00019683937395621683,
1370
+ "loss": 0.7484,
1371
+ "step": 454000
1372
+ },
1373
+ {
1374
+ "epoch": 10.56,
1375
+ "learning_rate": 0.0001963715719861468,
1376
+ "loss": 0.7554,
1377
+ "step": 456000
1378
+ },
1379
+ {
1380
+ "epoch": 10.61,
1381
+ "learning_rate": 0.00019590377001607676,
1382
+ "loss": 0.7557,
1383
+ "step": 458000
1384
+ },
1385
+ {
1386
+ "epoch": 10.65,
1387
+ "learning_rate": 0.00019543596804600676,
1388
+ "loss": 0.7603,
1389
+ "step": 460000
1390
+ },
1391
+ {
1392
+ "epoch": 10.7,
1393
+ "learning_rate": 0.00019496816607593673,
1394
+ "loss": 0.7577,
1395
+ "step": 462000
1396
+ },
1397
+ {
1398
+ "epoch": 10.74,
1399
+ "learning_rate": 0.0001945003641058667,
1400
+ "loss": 0.7641,
1401
+ "step": 464000
1402
+ },
1403
+ {
1404
+ "epoch": 10.79,
1405
+ "learning_rate": 0.00019403256213579667,
1406
+ "loss": 0.7648,
1407
+ "step": 466000
1408
+ },
1409
+ {
1410
+ "epoch": 10.84,
1411
+ "learning_rate": 0.00019356476016572664,
1412
+ "loss": 0.755,
1413
+ "step": 468000
1414
+ },
1415
+ {
1416
+ "epoch": 10.88,
1417
+ "learning_rate": 0.0001930969581956566,
1418
+ "loss": 0.7445,
1419
+ "step": 470000
1420
+ },
1421
+ {
1422
+ "epoch": 10.93,
1423
+ "learning_rate": 0.00019262915622558657,
1424
+ "loss": 0.7614,
1425
+ "step": 472000
1426
+ },
1427
+ {
1428
+ "epoch": 10.98,
1429
+ "learning_rate": 0.00019216135425551654,
1430
+ "loss": 0.7526,
1431
+ "step": 474000
1432
+ },
1433
+ {
1434
+ "epoch": 11.02,
1435
+ "learning_rate": 0.0001916935522854465,
1436
+ "loss": 0.7493,
1437
+ "step": 476000
1438
+ },
1439
+ {
1440
+ "epoch": 11.07,
1441
+ "learning_rate": 0.00019122575031537648,
1442
+ "loss": 0.7299,
1443
+ "step": 478000
1444
+ },
1445
+ {
1446
+ "epoch": 11.11,
1447
+ "learning_rate": 0.00019075794834530645,
1448
+ "loss": 0.7379,
1449
+ "step": 480000
1450
+ },
1451
+ {
1452
+ "epoch": 11.16,
1453
+ "learning_rate": 0.00019029014637523642,
1454
+ "loss": 0.7365,
1455
+ "step": 482000
1456
+ },
1457
+ {
1458
+ "epoch": 11.21,
1459
+ "learning_rate": 0.00018982234440516639,
1460
+ "loss": 0.7402,
1461
+ "step": 484000
1462
+ },
1463
+ {
1464
+ "epoch": 11.25,
1465
+ "learning_rate": 0.00018935454243509636,
1466
+ "loss": 0.7409,
1467
+ "step": 486000
1468
+ },
1469
+ {
1470
+ "epoch": 11.3,
1471
+ "learning_rate": 0.00018888674046502632,
1472
+ "loss": 0.7294,
1473
+ "step": 488000
1474
+ },
1475
+ {
1476
+ "epoch": 11.35,
1477
+ "learning_rate": 0.0001884189384949563,
1478
+ "loss": 0.7467,
1479
+ "step": 490000
1480
+ },
1481
+ {
1482
+ "epoch": 11.39,
1483
+ "learning_rate": 0.00018795113652488626,
1484
+ "loss": 0.7357,
1485
+ "step": 492000
1486
+ },
1487
+ {
1488
+ "epoch": 11.44,
1489
+ "learning_rate": 0.00018748333455481623,
1490
+ "loss": 0.744,
1491
+ "step": 494000
1492
+ },
1493
+ {
1494
+ "epoch": 11.49,
1495
+ "learning_rate": 0.0001870155325847462,
1496
+ "loss": 0.741,
1497
+ "step": 496000
1498
+ },
1499
+ {
1500
+ "epoch": 11.53,
1501
+ "learning_rate": 0.0001865477306146762,
1502
+ "loss": 0.7404,
1503
+ "step": 498000
1504
+ },
1505
+ {
1506
+ "epoch": 11.58,
1507
+ "learning_rate": 0.00018607992864460614,
1508
+ "loss": 0.749,
1509
+ "step": 500000
1510
+ },
1511
+ {
1512
+ "epoch": 11.62,
1513
+ "learning_rate": 0.00018561212667453613,
1514
+ "loss": 0.7388,
1515
+ "step": 502000
1516
+ },
1517
+ {
1518
+ "epoch": 11.67,
1519
+ "learning_rate": 0.00018514432470446607,
1520
+ "loss": 0.742,
1521
+ "step": 504000
1522
+ },
1523
+ {
1524
+ "epoch": 11.72,
1525
+ "learning_rate": 0.00018467652273439607,
1526
+ "loss": 0.7481,
1527
+ "step": 506000
1528
+ },
1529
+ {
1530
+ "epoch": 11.76,
1531
+ "learning_rate": 0.000184208720764326,
1532
+ "loss": 0.7553,
1533
+ "step": 508000
1534
+ },
1535
+ {
1536
+ "epoch": 11.81,
1537
+ "learning_rate": 0.000183740918794256,
1538
+ "loss": 0.7457,
1539
+ "step": 510000
1540
+ },
1541
+ {
1542
+ "epoch": 11.86,
1543
+ "learning_rate": 0.00018327311682418595,
1544
+ "loss": 0.7447,
1545
+ "step": 512000
1546
+ },
1547
+ {
1548
+ "epoch": 11.9,
1549
+ "learning_rate": 0.00018280531485411595,
1550
+ "loss": 0.752,
1551
+ "step": 514000
1552
+ },
1553
+ {
1554
+ "epoch": 11.95,
1555
+ "learning_rate": 0.0001823375128840459,
1556
+ "loss": 0.7419,
1557
+ "step": 516000
1558
+ },
1559
+ {
1560
+ "epoch": 11.99,
1561
+ "learning_rate": 0.00018186971091397588,
1562
+ "loss": 0.7412,
1563
+ "step": 518000
1564
+ },
1565
+ {
1566
+ "epoch": 12.04,
1567
+ "learning_rate": 0.00018140190894390583,
1568
+ "loss": 0.7262,
1569
+ "step": 520000
1570
+ },
1571
+ {
1572
+ "epoch": 12.09,
1573
+ "learning_rate": 0.00018093410697383582,
1574
+ "loss": 0.7299,
1575
+ "step": 522000
1576
+ },
1577
+ {
1578
+ "epoch": 12.13,
1579
+ "learning_rate": 0.00018046630500376576,
1580
+ "loss": 0.7279,
1581
+ "step": 524000
1582
+ },
1583
+ {
1584
+ "epoch": 12.18,
1585
+ "learning_rate": 0.00017999850303369576,
1586
+ "loss": 0.7308,
1587
+ "step": 526000
1588
+ },
1589
+ {
1590
+ "epoch": 12.23,
1591
+ "learning_rate": 0.0001795307010636257,
1592
+ "loss": 0.7305,
1593
+ "step": 528000
1594
+ },
1595
+ {
1596
+ "epoch": 12.27,
1597
+ "learning_rate": 0.0001790628990935557,
1598
+ "loss": 0.7348,
1599
+ "step": 530000
1600
+ },
1601
+ {
1602
+ "epoch": 12.32,
1603
+ "learning_rate": 0.0001785950971234857,
1604
+ "loss": 0.7312,
1605
+ "step": 532000
1606
+ },
1607
+ {
1608
+ "epoch": 12.37,
1609
+ "learning_rate": 0.00017812729515341563,
1610
+ "loss": 0.7275,
1611
+ "step": 534000
1612
+ },
1613
+ {
1614
+ "epoch": 12.41,
1615
+ "learning_rate": 0.00017765949318334563,
1616
+ "loss": 0.7291,
1617
+ "step": 536000
1618
+ },
1619
+ {
1620
+ "epoch": 12.46,
1621
+ "learning_rate": 0.00017719169121327557,
1622
+ "loss": 0.7265,
1623
+ "step": 538000
1624
+ },
1625
+ {
1626
+ "epoch": 12.5,
1627
+ "learning_rate": 0.00017672388924320557,
1628
+ "loss": 0.7224,
1629
+ "step": 540000
1630
+ },
1631
+ {
1632
+ "epoch": 12.55,
1633
+ "learning_rate": 0.0001762560872731355,
1634
+ "loss": 0.7232,
1635
+ "step": 542000
1636
+ },
1637
+ {
1638
+ "epoch": 12.6,
1639
+ "learning_rate": 0.0001757882853030655,
1640
+ "loss": 0.7272,
1641
+ "step": 544000
1642
+ },
1643
+ {
1644
+ "epoch": 12.64,
1645
+ "learning_rate": 0.00017532048333299545,
1646
+ "loss": 0.7305,
1647
+ "step": 546000
1648
+ },
1649
+ {
1650
+ "epoch": 12.69,
1651
+ "learning_rate": 0.00017485268136292544,
1652
+ "loss": 0.7375,
1653
+ "step": 548000
1654
+ },
1655
+ {
1656
+ "epoch": 12.74,
1657
+ "learning_rate": 0.00017438487939285539,
1658
+ "loss": 0.7377,
1659
+ "step": 550000
1660
+ },
1661
+ {
1662
+ "epoch": 12.78,
1663
+ "learning_rate": 0.00017391707742278538,
1664
+ "loss": 0.7278,
1665
+ "step": 552000
1666
+ },
1667
+ {
1668
+ "epoch": 12.83,
1669
+ "learning_rate": 0.00017344927545271532,
1670
+ "loss": 0.7369,
1671
+ "step": 554000
1672
+ },
1673
+ {
1674
+ "epoch": 12.87,
1675
+ "learning_rate": 0.00017298147348264532,
1676
+ "loss": 0.7366,
1677
+ "step": 556000
1678
+ },
1679
+ {
1680
+ "epoch": 12.92,
1681
+ "learning_rate": 0.00017251367151257526,
1682
+ "loss": 0.736,
1683
+ "step": 558000
1684
+ },
1685
+ {
1686
+ "epoch": 12.97,
1687
+ "learning_rate": 0.00017204586954250526,
1688
+ "loss": 0.737,
1689
+ "step": 560000
1690
+ },
1691
+ {
1692
+ "epoch": 13.01,
1693
+ "learning_rate": 0.0001715780675724352,
1694
+ "loss": 0.7301,
1695
+ "step": 562000
1696
+ },
1697
+ {
1698
+ "epoch": 13.06,
1699
+ "learning_rate": 0.0001711102656023652,
1700
+ "loss": 0.7003,
1701
+ "step": 564000
1702
+ },
1703
+ {
1704
+ "epoch": 13.11,
1705
+ "learning_rate": 0.00017064246363229514,
1706
+ "loss": 0.7132,
1707
+ "step": 566000
1708
+ },
1709
+ {
1710
+ "epoch": 13.15,
1711
+ "learning_rate": 0.00017017466166222513,
1712
+ "loss": 0.7178,
1713
+ "step": 568000
1714
+ },
1715
+ {
1716
+ "epoch": 13.2,
1717
+ "learning_rate": 0.00016970685969215513,
1718
+ "loss": 0.7187,
1719
+ "step": 570000
1720
+ },
1721
+ {
1722
+ "epoch": 13.25,
1723
+ "learning_rate": 0.00016923905772208507,
1724
+ "loss": 0.7239,
1725
+ "step": 572000
1726
+ },
1727
+ {
1728
+ "epoch": 13.29,
1729
+ "learning_rate": 0.00016877125575201507,
1730
+ "loss": 0.7271,
1731
+ "step": 574000
1732
+ },
1733
+ {
1734
+ "epoch": 13.34,
1735
+ "learning_rate": 0.000168303453781945,
1736
+ "loss": 0.7208,
1737
+ "step": 576000
1738
+ },
1739
+ {
1740
+ "epoch": 13.38,
1741
+ "learning_rate": 0.000167835651811875,
1742
+ "loss": 0.7199,
1743
+ "step": 578000
1744
+ },
1745
+ {
1746
+ "epoch": 13.43,
1747
+ "learning_rate": 0.00016736784984180495,
1748
+ "loss": 0.7094,
1749
+ "step": 580000
1750
+ },
1751
+ {
1752
+ "epoch": 13.48,
1753
+ "learning_rate": 0.00016690004787173494,
1754
+ "loss": 0.7114,
1755
+ "step": 582000
1756
+ },
1757
+ {
1758
+ "epoch": 13.52,
1759
+ "learning_rate": 0.00016643224590166488,
1760
+ "loss": 0.7196,
1761
+ "step": 584000
1762
+ },
1763
+ {
1764
+ "epoch": 13.57,
1765
+ "learning_rate": 0.00016596444393159488,
1766
+ "loss": 0.7222,
1767
+ "step": 586000
1768
+ },
1769
+ {
1770
+ "epoch": 13.62,
1771
+ "learning_rate": 0.00016549664196152482,
1772
+ "loss": 0.7345,
1773
+ "step": 588000
1774
+ },
1775
+ {
1776
+ "epoch": 13.66,
1777
+ "learning_rate": 0.00016502883999145482,
1778
+ "loss": 0.7208,
1779
+ "step": 590000
1780
+ },
1781
+ {
1782
+ "epoch": 13.71,
1783
+ "learning_rate": 0.00016456103802138476,
1784
+ "loss": 0.7298,
1785
+ "step": 592000
1786
+ },
1787
+ {
1788
+ "epoch": 13.75,
1789
+ "learning_rate": 0.00016409323605131475,
1790
+ "loss": 0.7324,
1791
+ "step": 594000
1792
+ },
1793
+ {
1794
+ "epoch": 13.8,
1795
+ "learning_rate": 0.0001636254340812447,
1796
+ "loss": 0.7243,
1797
+ "step": 596000
1798
+ },
1799
+ {
1800
+ "epoch": 13.85,
1801
+ "learning_rate": 0.0001631576321111747,
1802
+ "loss": 0.7215,
1803
+ "step": 598000
1804
+ },
1805
+ {
1806
+ "epoch": 13.89,
1807
+ "learning_rate": 0.00016268983014110463,
1808
+ "loss": 0.7246,
1809
+ "step": 600000
1810
+ },
1811
+ {
1812
+ "epoch": 13.94,
1813
+ "learning_rate": 0.00016222202817103463,
1814
+ "loss": 0.7219,
1815
+ "step": 602000
1816
+ },
1817
+ {
1818
+ "epoch": 13.99,
1819
+ "learning_rate": 0.0001617542262009646,
1820
+ "loss": 0.7248,
1821
+ "step": 604000
1822
+ },
1823
+ {
1824
+ "epoch": 14.03,
1825
+ "learning_rate": 0.00016128642423089457,
1826
+ "loss": 0.7139,
1827
+ "step": 606000
1828
+ },
1829
+ {
1830
+ "epoch": 14.08,
1831
+ "learning_rate": 0.00016081862226082454,
1832
+ "loss": 0.7026,
1833
+ "step": 608000
1834
+ },
1835
+ {
1836
+ "epoch": 14.13,
1837
+ "learning_rate": 0.0001603508202907545,
1838
+ "loss": 0.7107,
1839
+ "step": 610000
1840
+ },
1841
+ {
1842
+ "epoch": 14.17,
1843
+ "learning_rate": 0.00015988301832068447,
1844
+ "loss": 0.7037,
1845
+ "step": 612000
1846
+ },
1847
+ {
1848
+ "epoch": 14.22,
1849
+ "learning_rate": 0.00015941521635061444,
1850
+ "loss": 0.7145,
1851
+ "step": 614000
1852
+ },
1853
+ {
1854
+ "epoch": 14.26,
1855
+ "learning_rate": 0.0001589474143805444,
1856
+ "loss": 0.7181,
1857
+ "step": 616000
1858
+ },
1859
+ {
1860
+ "epoch": 14.31,
1861
+ "learning_rate": 0.00015847961241047438,
1862
+ "loss": 0.7026,
1863
+ "step": 618000
1864
+ },
1865
+ {
1866
+ "epoch": 14.36,
1867
+ "learning_rate": 0.00015801181044040435,
1868
+ "loss": 0.7142,
1869
+ "step": 620000
1870
+ },
1871
+ {
1872
+ "epoch": 14.4,
1873
+ "learning_rate": 0.00015754400847033432,
1874
+ "loss": 0.7087,
1875
+ "step": 622000
1876
+ },
1877
+ {
1878
+ "epoch": 14.45,
1879
+ "learning_rate": 0.0001570762065002643,
1880
+ "loss": 0.7109,
1881
+ "step": 624000
1882
+ },
1883
+ {
1884
+ "epoch": 14.5,
1885
+ "learning_rate": 0.00015660840453019426,
1886
+ "loss": 0.7031,
1887
+ "step": 626000
1888
+ },
1889
+ {
1890
+ "epoch": 14.54,
1891
+ "learning_rate": 0.00015614060256012425,
1892
+ "loss": 0.7101,
1893
+ "step": 628000
1894
+ },
1895
+ {
1896
+ "epoch": 14.59,
1897
+ "learning_rate": 0.0001556728005900542,
1898
+ "loss": 0.7152,
1899
+ "step": 630000
1900
+ },
1901
+ {
1902
+ "epoch": 14.63,
1903
+ "learning_rate": 0.0001552049986199842,
1904
+ "loss": 0.7147,
1905
+ "step": 632000
1906
+ },
1907
+ {
1908
+ "epoch": 14.68,
1909
+ "learning_rate": 0.00015473719664991413,
1910
+ "loss": 0.7144,
1911
+ "step": 634000
1912
+ },
1913
+ {
1914
+ "epoch": 14.73,
1915
+ "learning_rate": 0.00015426939467984413,
1916
+ "loss": 0.7113,
1917
+ "step": 636000
1918
+ },
1919
+ {
1920
+ "epoch": 14.77,
1921
+ "learning_rate": 0.00015380159270977407,
1922
+ "loss": 0.7071,
1923
+ "step": 638000
1924
+ },
1925
+ {
1926
+ "epoch": 14.82,
1927
+ "learning_rate": 0.00015333379073970407,
1928
+ "loss": 0.7118,
1929
+ "step": 640000
1930
+ },
1931
+ {
1932
+ "epoch": 14.87,
1933
+ "learning_rate": 0.00015286598876963403,
1934
+ "loss": 0.7098,
1935
+ "step": 642000
1936
+ },
1937
+ {
1938
+ "epoch": 14.91,
1939
+ "learning_rate": 0.000152398186799564,
1940
+ "loss": 0.706,
1941
+ "step": 644000
1942
+ },
1943
+ {
1944
+ "epoch": 14.96,
1945
+ "learning_rate": 0.00015193038482949397,
1946
+ "loss": 0.709,
1947
+ "step": 646000
1948
+ },
1949
+ {
1950
+ "epoch": 15.01,
1951
+ "learning_rate": 0.00015146258285942394,
1952
+ "loss": 0.7087,
1953
+ "step": 648000
1954
+ },
1955
+ {
1956
+ "epoch": 15.05,
1957
+ "learning_rate": 0.0001509947808893539,
1958
+ "loss": 0.6983,
1959
+ "step": 650000
1960
+ },
1961
+ {
1962
+ "epoch": 15.1,
1963
+ "learning_rate": 0.00015052697891928388,
1964
+ "loss": 0.693,
1965
+ "step": 652000
1966
+ },
1967
+ {
1968
+ "epoch": 15.14,
1969
+ "learning_rate": 0.00015005917694921385,
1970
+ "loss": 0.6953,
1971
+ "step": 654000
1972
+ },
1973
+ {
1974
+ "epoch": 15.19,
1975
+ "learning_rate": 0.00014959137497914382,
1976
+ "loss": 0.6994,
1977
+ "step": 656000
1978
+ },
1979
+ {
1980
+ "epoch": 15.24,
1981
+ "learning_rate": 0.00014912357300907379,
1982
+ "loss": 0.6975,
1983
+ "step": 658000
1984
+ },
1985
+ {
1986
+ "epoch": 15.28,
1987
+ "learning_rate": 0.00014865577103900375,
1988
+ "loss": 0.7047,
1989
+ "step": 660000
1990
+ },
1991
+ {
1992
+ "epoch": 15.33,
1993
+ "learning_rate": 0.00014818796906893372,
1994
+ "loss": 0.6975,
1995
+ "step": 662000
1996
+ },
1997
+ {
1998
+ "epoch": 15.38,
1999
+ "learning_rate": 0.0001477201670988637,
2000
+ "loss": 0.704,
2001
+ "step": 664000
2002
+ },
2003
+ {
2004
+ "epoch": 15.42,
2005
+ "learning_rate": 0.00014725236512879366,
2006
+ "loss": 0.7042,
2007
+ "step": 666000
2008
+ },
2009
+ {
2010
+ "epoch": 15.47,
2011
+ "learning_rate": 0.00014678456315872363,
2012
+ "loss": 0.6917,
2013
+ "step": 668000
2014
+ },
2015
+ {
2016
+ "epoch": 15.51,
2017
+ "learning_rate": 0.0001463167611886536,
2018
+ "loss": 0.6914,
2019
+ "step": 670000
2020
+ },
2021
+ {
2022
+ "epoch": 15.56,
2023
+ "learning_rate": 0.00014584895921858357,
2024
+ "loss": 0.7018,
2025
+ "step": 672000
2026
+ },
2027
+ {
2028
+ "epoch": 15.61,
2029
+ "learning_rate": 0.00014538115724851354,
2030
+ "loss": 0.7016,
2031
+ "step": 674000
2032
+ },
2033
+ {
2034
+ "epoch": 15.65,
2035
+ "learning_rate": 0.0001449133552784435,
2036
+ "loss": 0.7078,
2037
+ "step": 676000
2038
+ },
2039
+ {
2040
+ "epoch": 15.7,
2041
+ "learning_rate": 0.00014444555330837347,
2042
+ "loss": 0.6932,
2043
+ "step": 678000
2044
+ },
2045
+ {
2046
+ "epoch": 15.75,
2047
+ "learning_rate": 0.00014397775133830344,
2048
+ "loss": 0.6964,
2049
+ "step": 680000
2050
+ },
2051
+ {
2052
+ "epoch": 15.79,
2053
+ "learning_rate": 0.0001435099493682334,
2054
+ "loss": 0.6997,
2055
+ "step": 682000
2056
+ },
2057
+ {
2058
+ "epoch": 15.84,
2059
+ "learning_rate": 0.00014304214739816338,
2060
+ "loss": 0.7065,
2061
+ "step": 684000
2062
+ },
2063
+ {
2064
+ "epoch": 15.88,
2065
+ "learning_rate": 0.00014257434542809335,
2066
+ "loss": 0.7047,
2067
+ "step": 686000
2068
+ },
2069
+ {
2070
+ "epoch": 15.93,
2071
+ "learning_rate": 0.00014210654345802332,
2072
+ "loss": 0.7154,
2073
+ "step": 688000
2074
+ },
2075
+ {
2076
+ "epoch": 15.98,
2077
+ "learning_rate": 0.0001416387414879533,
2078
+ "loss": 0.6993,
2079
+ "step": 690000
2080
+ },
2081
+ {
2082
+ "epoch": 16.02,
2083
+ "learning_rate": 0.00014117093951788326,
2084
+ "loss": 0.6969,
2085
+ "step": 692000
2086
+ },
2087
+ {
2088
+ "epoch": 16.07,
2089
+ "learning_rate": 0.00014070313754781325,
2090
+ "loss": 0.689,
2091
+ "step": 694000
2092
+ },
2093
+ {
2094
+ "epoch": 16.12,
2095
+ "learning_rate": 0.00014023533557774322,
2096
+ "loss": 0.6888,
2097
+ "step": 696000
2098
+ },
2099
+ {
2100
+ "epoch": 16.16,
2101
+ "learning_rate": 0.0001397675336076732,
2102
+ "loss": 0.6818,
2103
+ "step": 698000
2104
+ },
2105
+ {
2106
+ "epoch": 16.21,
2107
+ "learning_rate": 0.00013929973163760316,
2108
+ "loss": 0.693,
2109
+ "step": 700000
2110
+ },
2111
+ {
2112
+ "epoch": 16.26,
2113
+ "learning_rate": 0.00013883192966753313,
2114
+ "loss": 0.6909,
2115
+ "step": 702000
2116
+ },
2117
+ {
2118
+ "epoch": 16.3,
2119
+ "learning_rate": 0.0001383641276974631,
2120
+ "loss": 0.6873,
2121
+ "step": 704000
2122
+ },
2123
+ {
2124
+ "epoch": 16.35,
2125
+ "learning_rate": 0.00013789632572739307,
2126
+ "loss": 0.6906,
2127
+ "step": 706000
2128
+ },
2129
+ {
2130
+ "epoch": 16.39,
2131
+ "learning_rate": 0.00013742852375732303,
2132
+ "loss": 0.6866,
2133
+ "step": 708000
2134
+ },
2135
+ {
2136
+ "epoch": 16.44,
2137
+ "learning_rate": 0.000136960721787253,
2138
+ "loss": 0.701,
2139
+ "step": 710000
2140
+ },
2141
+ {
2142
+ "epoch": 16.49,
2143
+ "learning_rate": 0.00013649291981718297,
2144
+ "loss": 0.6937,
2145
+ "step": 712000
2146
+ },
2147
+ {
2148
+ "epoch": 16.53,
2149
+ "learning_rate": 0.00013602511784711294,
2150
+ "loss": 0.6907,
2151
+ "step": 714000
2152
+ },
2153
+ {
2154
+ "epoch": 16.58,
2155
+ "learning_rate": 0.0001355573158770429,
2156
+ "loss": 0.6897,
2157
+ "step": 716000
2158
+ },
2159
+ {
2160
+ "epoch": 16.63,
2161
+ "learning_rate": 0.00013508951390697288,
2162
+ "loss": 0.6952,
2163
+ "step": 718000
2164
+ },
2165
+ {
2166
+ "epoch": 16.67,
2167
+ "learning_rate": 0.00013462171193690285,
2168
+ "loss": 0.6865,
2169
+ "step": 720000
2170
+ },
2171
+ {
2172
+ "epoch": 16.72,
2173
+ "learning_rate": 0.00013415390996683282,
2174
+ "loss": 0.6935,
2175
+ "step": 722000
2176
+ },
2177
+ {
2178
+ "epoch": 16.76,
2179
+ "learning_rate": 0.00013368610799676278,
2180
+ "loss": 0.6919,
2181
+ "step": 724000
2182
+ },
2183
+ {
2184
+ "epoch": 16.81,
2185
+ "learning_rate": 0.00013321830602669275,
2186
+ "loss": 0.6904,
2187
+ "step": 726000
2188
+ },
2189
+ {
2190
+ "epoch": 16.86,
2191
+ "learning_rate": 0.00013275050405662272,
2192
+ "loss": 0.6964,
2193
+ "step": 728000
2194
+ },
2195
+ {
2196
+ "epoch": 16.9,
2197
+ "learning_rate": 0.00013228270208655272,
2198
+ "loss": 0.6943,
2199
+ "step": 730000
2200
+ },
2201
+ {
2202
+ "epoch": 16.95,
2203
+ "learning_rate": 0.0001318149001164827,
2204
+ "loss": 0.6949,
2205
+ "step": 732000
2206
+ },
2207
+ {
2208
+ "epoch": 17.0,
2209
+ "learning_rate": 0.00013134709814641266,
2210
+ "loss": 0.6943,
2211
+ "step": 734000
2212
+ },
2213
+ {
2214
+ "epoch": 17.04,
2215
+ "learning_rate": 0.00013087929617634263,
2216
+ "loss": 0.6851,
2217
+ "step": 736000
2218
+ },
2219
+ {
2220
+ "epoch": 17.09,
2221
+ "learning_rate": 0.0001304114942062726,
2222
+ "loss": 0.6802,
2223
+ "step": 738000
2224
+ },
2225
+ {
2226
+ "epoch": 17.14,
2227
+ "learning_rate": 0.00012994369223620256,
2228
+ "loss": 0.6801,
2229
+ "step": 740000
2230
+ },
2231
+ {
2232
+ "epoch": 17.18,
2233
+ "learning_rate": 0.00012947589026613253,
2234
+ "loss": 0.6756,
2235
+ "step": 742000
2236
+ },
2237
+ {
2238
+ "epoch": 17.23,
2239
+ "learning_rate": 0.0001290080882960625,
2240
+ "loss": 0.6824,
2241
+ "step": 744000
2242
+ },
2243
+ {
2244
+ "epoch": 17.27,
2245
+ "learning_rate": 0.00012854028632599247,
2246
+ "loss": 0.6894,
2247
+ "step": 746000
2248
+ },
2249
+ {
2250
+ "epoch": 17.32,
2251
+ "learning_rate": 0.00012807248435592244,
2252
+ "loss": 0.682,
2253
+ "step": 748000
2254
+ },
2255
+ {
2256
+ "epoch": 17.37,
2257
+ "learning_rate": 0.0001276046823858524,
2258
+ "loss": 0.6814,
2259
+ "step": 750000
2260
+ },
2261
+ {
2262
+ "epoch": 17.41,
2263
+ "learning_rate": 0.00012713688041578238,
2264
+ "loss": 0.6737,
2265
+ "step": 752000
2266
+ },
2267
+ {
2268
+ "epoch": 17.46,
2269
+ "learning_rate": 0.00012666907844571234,
2270
+ "loss": 0.6874,
2271
+ "step": 754000
2272
+ },
2273
+ {
2274
+ "epoch": 17.51,
2275
+ "learning_rate": 0.00012620127647564231,
2276
+ "loss": 0.6842,
2277
+ "step": 756000
2278
+ },
2279
+ {
2280
+ "epoch": 17.55,
2281
+ "learning_rate": 0.00012573347450557228,
2282
+ "loss": 0.6871,
2283
+ "step": 758000
2284
+ },
2285
+ {
2286
+ "epoch": 17.6,
2287
+ "learning_rate": 0.00012526567253550225,
2288
+ "loss": 0.6833,
2289
+ "step": 760000
2290
+ },
2291
+ {
2292
+ "epoch": 17.64,
2293
+ "learning_rate": 0.00012479787056543222,
2294
+ "loss": 0.6818,
2295
+ "step": 762000
2296
+ },
2297
+ {
2298
+ "epoch": 17.69,
2299
+ "learning_rate": 0.0001243300685953622,
2300
+ "loss": 0.6824,
2301
+ "step": 764000
2302
+ },
2303
+ {
2304
+ "epoch": 17.74,
2305
+ "learning_rate": 0.00012386226662529219,
2306
+ "loss": 0.684,
2307
+ "step": 766000
2308
+ },
2309
+ {
2310
+ "epoch": 17.78,
2311
+ "learning_rate": 0.00012339446465522215,
2312
+ "loss": 0.6822,
2313
+ "step": 768000
2314
+ },
2315
+ {
2316
+ "epoch": 17.83,
2317
+ "learning_rate": 0.00012292666268515212,
2318
+ "loss": 0.68,
2319
+ "step": 770000
2320
+ },
2321
+ {
2322
+ "epoch": 17.88,
2323
+ "learning_rate": 0.0001224588607150821,
2324
+ "loss": 0.6842,
2325
+ "step": 772000
2326
+ },
2327
+ {
2328
+ "epoch": 17.92,
2329
+ "learning_rate": 0.00012199105874501206,
2330
+ "loss": 0.6827,
2331
+ "step": 774000
2332
+ },
2333
+ {
2334
+ "epoch": 17.97,
2335
+ "learning_rate": 0.00012152325677494203,
2336
+ "loss": 0.6901,
2337
+ "step": 776000
2338
+ },
2339
+ {
2340
+ "epoch": 18.02,
2341
+ "learning_rate": 0.000121055454804872,
2342
+ "loss": 0.6781,
2343
+ "step": 778000
2344
+ },
2345
+ {
2346
+ "epoch": 18.06,
2347
+ "learning_rate": 0.00012058765283480197,
2348
+ "loss": 0.6768,
2349
+ "step": 780000
2350
+ },
2351
+ {
2352
+ "epoch": 18.11,
2353
+ "learning_rate": 0.00012011985086473194,
2354
+ "loss": 0.6704,
2355
+ "step": 782000
2356
+ },
2357
+ {
2358
+ "epoch": 18.15,
2359
+ "learning_rate": 0.0001196520488946619,
2360
+ "loss": 0.6767,
2361
+ "step": 784000
2362
+ },
2363
+ {
2364
+ "epoch": 18.2,
2365
+ "learning_rate": 0.00011918424692459187,
2366
+ "loss": 0.6696,
2367
+ "step": 786000
2368
+ },
2369
+ {
2370
+ "epoch": 18.25,
2371
+ "learning_rate": 0.00011871644495452184,
2372
+ "loss": 0.6717,
2373
+ "step": 788000
2374
+ },
2375
+ {
2376
+ "epoch": 18.29,
2377
+ "learning_rate": 0.00011824864298445181,
2378
+ "loss": 0.6666,
2379
+ "step": 790000
2380
+ },
2381
+ {
2382
+ "epoch": 18.34,
2383
+ "learning_rate": 0.00011778084101438178,
2384
+ "loss": 0.6681,
2385
+ "step": 792000
2386
+ },
2387
+ {
2388
+ "epoch": 18.39,
2389
+ "learning_rate": 0.00011731303904431175,
2390
+ "loss": 0.6688,
2391
+ "step": 794000
2392
+ },
2393
+ {
2394
+ "epoch": 18.43,
2395
+ "learning_rate": 0.00011684523707424172,
2396
+ "loss": 0.6809,
2397
+ "step": 796000
2398
+ },
2399
+ {
2400
+ "epoch": 18.48,
2401
+ "learning_rate": 0.00011637743510417169,
2402
+ "loss": 0.6704,
2403
+ "step": 798000
2404
+ },
2405
+ {
2406
+ "epoch": 18.52,
2407
+ "learning_rate": 0.00011590963313410166,
2408
+ "loss": 0.6732,
2409
+ "step": 800000
2410
+ },
2411
+ {
2412
+ "epoch": 18.57,
2413
+ "learning_rate": 0.00011544183116403164,
2414
+ "loss": 0.6688,
2415
+ "step": 802000
2416
+ },
2417
+ {
2418
+ "epoch": 18.62,
2419
+ "learning_rate": 0.00011497402919396161,
2420
+ "loss": 0.6767,
2421
+ "step": 804000
2422
+ },
2423
+ {
2424
+ "epoch": 18.66,
2425
+ "learning_rate": 0.00011450622722389158,
2426
+ "loss": 0.6721,
2427
+ "step": 806000
2428
+ },
2429
+ {
2430
+ "epoch": 18.71,
2431
+ "learning_rate": 0.00011403842525382154,
2432
+ "loss": 0.6716,
2433
+ "step": 808000
2434
+ },
2435
+ {
2436
+ "epoch": 18.76,
2437
+ "learning_rate": 0.00011357062328375151,
2438
+ "loss": 0.673,
2439
+ "step": 810000
2440
+ },
2441
+ {
2442
+ "epoch": 18.8,
2443
+ "learning_rate": 0.00011310282131368148,
2444
+ "loss": 0.6717,
2445
+ "step": 812000
2446
+ },
2447
+ {
2448
+ "epoch": 18.85,
2449
+ "learning_rate": 0.00011263501934361145,
2450
+ "loss": 0.6607,
2451
+ "step": 814000
2452
+ },
2453
+ {
2454
+ "epoch": 18.9,
2455
+ "learning_rate": 0.00011216721737354142,
2456
+ "loss": 0.6732,
2457
+ "step": 816000
2458
+ },
2459
+ {
2460
+ "epoch": 18.94,
2461
+ "learning_rate": 0.00011169941540347139,
2462
+ "loss": 0.6715,
2463
+ "step": 818000
2464
+ },
2465
+ {
2466
+ "epoch": 18.99,
2467
+ "learning_rate": 0.00011123161343340136,
2468
+ "loss": 0.678,
2469
+ "step": 820000
2470
+ },
2471
+ {
2472
+ "epoch": 19.03,
2473
+ "learning_rate": 0.00011076381146333133,
2474
+ "loss": 0.6618,
2475
+ "step": 822000
2476
+ },
2477
+ {
2478
+ "epoch": 19.08,
2479
+ "learning_rate": 0.0001102960094932613,
2480
+ "loss": 0.6589,
2481
+ "step": 824000
2482
+ },
2483
+ {
2484
+ "epoch": 19.13,
2485
+ "learning_rate": 0.00010982820752319126,
2486
+ "loss": 0.6624,
2487
+ "step": 826000
2488
+ },
2489
+ {
2490
+ "epoch": 19.17,
2491
+ "learning_rate": 0.00010936040555312123,
2492
+ "loss": 0.6618,
2493
+ "step": 828000
2494
+ },
2495
+ {
2496
+ "epoch": 19.22,
2497
+ "learning_rate": 0.0001088926035830512,
2498
+ "loss": 0.6666,
2499
+ "step": 830000
2500
+ },
2501
+ {
2502
+ "epoch": 19.27,
2503
+ "learning_rate": 0.00010842480161298117,
2504
+ "loss": 0.6645,
2505
+ "step": 832000
2506
+ },
2507
+ {
2508
+ "epoch": 19.31,
2509
+ "learning_rate": 0.00010795699964291114,
2510
+ "loss": 0.6667,
2511
+ "step": 834000
2512
+ },
2513
+ {
2514
+ "epoch": 19.36,
2515
+ "learning_rate": 0.00010748919767284111,
2516
+ "loss": 0.6649,
2517
+ "step": 836000
2518
+ },
2519
+ {
2520
+ "epoch": 19.4,
2521
+ "learning_rate": 0.0001070213957027711,
2522
+ "loss": 0.659,
2523
+ "step": 838000
2524
+ },
2525
+ {
2526
+ "epoch": 19.45,
2527
+ "learning_rate": 0.00010655359373270107,
2528
+ "loss": 0.6611,
2529
+ "step": 840000
2530
+ },
2531
+ {
2532
+ "epoch": 19.5,
2533
+ "learning_rate": 0.00010608579176263104,
2534
+ "loss": 0.6565,
2535
+ "step": 842000
2536
+ },
2537
+ {
2538
+ "epoch": 19.54,
2539
+ "learning_rate": 0.00010561798979256101,
2540
+ "loss": 0.6631,
2541
+ "step": 844000
2542
+ },
2543
+ {
2544
+ "epoch": 19.59,
2545
+ "learning_rate": 0.00010515018782249098,
2546
+ "loss": 0.6593,
2547
+ "step": 846000
2548
+ },
2549
+ {
2550
+ "epoch": 19.64,
2551
+ "learning_rate": 0.00010468238585242095,
2552
+ "loss": 0.6654,
2553
+ "step": 848000
2554
+ },
2555
+ {
2556
+ "epoch": 19.68,
2557
+ "learning_rate": 0.00010421458388235092,
2558
+ "loss": 0.6621,
2559
+ "step": 850000
2560
+ },
2561
+ {
2562
+ "epoch": 19.73,
2563
+ "learning_rate": 0.00010374678191228089,
2564
+ "loss": 0.661,
2565
+ "step": 852000
2566
+ },
2567
+ {
2568
+ "epoch": 19.78,
2569
+ "learning_rate": 0.00010327897994221086,
2570
+ "loss": 0.6515,
2571
+ "step": 854000
2572
+ },
2573
+ {
2574
+ "epoch": 19.82,
2575
+ "learning_rate": 0.00010281117797214082,
2576
+ "loss": 0.6614,
2577
+ "step": 856000
2578
+ },
2579
+ {
2580
+ "epoch": 19.87,
2581
+ "learning_rate": 0.0001023433760020708,
2582
+ "loss": 0.6616,
2583
+ "step": 858000
2584
+ },
2585
+ {
2586
+ "epoch": 19.91,
2587
+ "learning_rate": 0.00010187557403200076,
2588
+ "loss": 0.6598,
2589
+ "step": 860000
2590
+ },
2591
+ {
2592
+ "epoch": 19.96,
2593
+ "learning_rate": 0.00010140777206193073,
2594
+ "loss": 0.6616,
2595
+ "step": 862000
2596
+ },
2597
+ {
2598
+ "epoch": 20.01,
2599
+ "learning_rate": 0.0001009399700918607,
2600
+ "loss": 0.6679,
2601
+ "step": 864000
2602
+ },
2603
+ {
2604
+ "epoch": 20.05,
2605
+ "learning_rate": 0.00010047216812179067,
2606
+ "loss": 0.6518,
2607
+ "step": 866000
2608
+ },
2609
+ {
2610
+ "epoch": 20.1,
2611
+ "learning_rate": 0.00010000436615172064,
2612
+ "loss": 0.6463,
2613
+ "step": 868000
2614
+ },
2615
+ {
2616
+ "epoch": 20.15,
2617
+ "learning_rate": 9.95365641816506e-05,
2618
+ "loss": 0.6529,
2619
+ "step": 870000
2620
+ },
2621
+ {
2622
+ "epoch": 20.19,
2623
+ "learning_rate": 9.906876221158058e-05,
2624
+ "loss": 0.6463,
2625
+ "step": 872000
2626
+ },
2627
+ {
2628
+ "epoch": 20.24,
2629
+ "learning_rate": 9.860096024151056e-05,
2630
+ "loss": 0.6545,
2631
+ "step": 874000
2632
+ },
2633
+ {
2634
+ "epoch": 20.28,
2635
+ "learning_rate": 9.813315827144053e-05,
2636
+ "loss": 0.6531,
2637
+ "step": 876000
2638
+ },
2639
+ {
2640
+ "epoch": 20.33,
2641
+ "learning_rate": 9.76653563013705e-05,
2642
+ "loss": 0.6442,
2643
+ "step": 878000
2644
+ },
2645
+ {
2646
+ "epoch": 20.38,
2647
+ "learning_rate": 9.719755433130046e-05,
2648
+ "loss": 0.65,
2649
+ "step": 880000
2650
+ },
2651
+ {
2652
+ "epoch": 20.42,
2653
+ "learning_rate": 9.672975236123043e-05,
2654
+ "loss": 0.6518,
2655
+ "step": 882000
2656
+ },
2657
+ {
2658
+ "epoch": 20.47,
2659
+ "learning_rate": 9.62619503911604e-05,
2660
+ "loss": 0.6546,
2661
+ "step": 884000
2662
+ },
2663
+ {
2664
+ "epoch": 20.52,
2665
+ "learning_rate": 9.579414842109037e-05,
2666
+ "loss": 0.6494,
2667
+ "step": 886000
2668
+ },
2669
+ {
2670
+ "epoch": 20.56,
2671
+ "learning_rate": 9.532634645102035e-05,
2672
+ "loss": 0.654,
2673
+ "step": 888000
2674
+ },
2675
+ {
2676
+ "epoch": 20.61,
2677
+ "learning_rate": 9.485854448095032e-05,
2678
+ "loss": 0.6536,
2679
+ "step": 890000
2680
+ },
2681
+ {
2682
+ "epoch": 20.66,
2683
+ "learning_rate": 9.439074251088029e-05,
2684
+ "loss": 0.6547,
2685
+ "step": 892000
2686
+ },
2687
+ {
2688
+ "epoch": 20.7,
2689
+ "learning_rate": 9.392294054081026e-05,
2690
+ "loss": 0.6421,
2691
+ "step": 894000
2692
+ },
2693
+ {
2694
+ "epoch": 20.75,
2695
+ "learning_rate": 9.345513857074023e-05,
2696
+ "loss": 0.6506,
2697
+ "step": 896000
2698
+ },
2699
+ {
2700
+ "epoch": 20.79,
2701
+ "learning_rate": 9.29873366006702e-05,
2702
+ "loss": 0.6551,
2703
+ "step": 898000
2704
+ },
2705
+ {
2706
+ "epoch": 20.84,
2707
+ "learning_rate": 9.251953463060017e-05,
2708
+ "loss": 0.6542,
2709
+ "step": 900000
2710
+ },
2711
+ {
2712
+ "epoch": 20.89,
2713
+ "learning_rate": 9.205173266053014e-05,
2714
+ "loss": 0.6398,
2715
+ "step": 902000
2716
+ },
2717
+ {
2718
+ "epoch": 20.93,
2719
+ "learning_rate": 9.15839306904601e-05,
2720
+ "loss": 0.653,
2721
+ "step": 904000
2722
+ },
2723
+ {
2724
+ "epoch": 20.98,
2725
+ "learning_rate": 9.111612872039007e-05,
2726
+ "loss": 0.6476,
2727
+ "step": 906000
2728
+ },
2729
+ {
2730
+ "epoch": 21.03,
2731
+ "learning_rate": 9.064832675032004e-05,
2732
+ "loss": 0.6378,
2733
+ "step": 908000
2734
+ },
2735
+ {
2736
+ "epoch": 21.07,
2737
+ "learning_rate": 9.018052478025002e-05,
2738
+ "loss": 0.6413,
2739
+ "step": 910000
2740
+ },
2741
+ {
2742
+ "epoch": 21.12,
2743
+ "learning_rate": 8.971272281017999e-05,
2744
+ "loss": 0.6368,
2745
+ "step": 912000
2746
+ },
2747
+ {
2748
+ "epoch": 21.16,
2749
+ "learning_rate": 8.924492084010996e-05,
2750
+ "loss": 0.6366,
2751
+ "step": 914000
2752
+ },
2753
+ {
2754
+ "epoch": 21.21,
2755
+ "learning_rate": 8.877711887003993e-05,
2756
+ "loss": 0.6455,
2757
+ "step": 916000
2758
+ },
2759
+ {
2760
+ "epoch": 21.26,
2761
+ "learning_rate": 8.83093168999699e-05,
2762
+ "loss": 0.6448,
2763
+ "step": 918000
2764
+ },
2765
+ {
2766
+ "epoch": 21.3,
2767
+ "learning_rate": 8.784151492989987e-05,
2768
+ "loss": 0.6371,
2769
+ "step": 920000
2770
+ },
2771
+ {
2772
+ "epoch": 21.35,
2773
+ "learning_rate": 8.737371295982984e-05,
2774
+ "loss": 0.6457,
2775
+ "step": 922000
2776
+ },
2777
+ {
2778
+ "epoch": 21.4,
2779
+ "learning_rate": 8.69059109897598e-05,
2780
+ "loss": 0.6399,
2781
+ "step": 924000
2782
+ },
2783
+ {
2784
+ "epoch": 21.44,
2785
+ "learning_rate": 8.643810901968978e-05,
2786
+ "loss": 0.6389,
2787
+ "step": 926000
2788
+ },
2789
+ {
2790
+ "epoch": 21.49,
2791
+ "learning_rate": 8.597030704961974e-05,
2792
+ "loss": 0.6444,
2793
+ "step": 928000
2794
+ },
2795
+ {
2796
+ "epoch": 21.54,
2797
+ "learning_rate": 8.550250507954971e-05,
2798
+ "loss": 0.6346,
2799
+ "step": 930000
2800
+ },
2801
+ {
2802
+ "epoch": 21.58,
2803
+ "learning_rate": 8.503470310947968e-05,
2804
+ "loss": 0.6394,
2805
+ "step": 932000
2806
+ },
2807
+ {
2808
+ "epoch": 21.63,
2809
+ "learning_rate": 8.456690113940965e-05,
2810
+ "loss": 0.6397,
2811
+ "step": 934000
2812
+ },
2813
+ {
2814
+ "epoch": 21.67,
2815
+ "learning_rate": 8.409909916933962e-05,
2816
+ "loss": 0.6411,
2817
+ "step": 936000
2818
+ },
2819
+ {
2820
+ "epoch": 21.72,
2821
+ "learning_rate": 8.363129719926959e-05,
2822
+ "loss": 0.6383,
2823
+ "step": 938000
2824
+ },
2825
+ {
2826
+ "epoch": 21.77,
2827
+ "learning_rate": 8.316349522919956e-05,
2828
+ "loss": 0.6416,
2829
+ "step": 940000
2830
+ },
2831
+ {
2832
+ "epoch": 21.81,
2833
+ "learning_rate": 8.269569325912953e-05,
2834
+ "loss": 0.635,
2835
+ "step": 942000
2836
+ },
2837
+ {
2838
+ "epoch": 21.86,
2839
+ "learning_rate": 8.22278912890595e-05,
2840
+ "loss": 0.6371,
2841
+ "step": 944000
2842
+ },
2843
+ {
2844
+ "epoch": 21.91,
2845
+ "learning_rate": 8.176008931898949e-05,
2846
+ "loss": 0.6412,
2847
+ "step": 946000
2848
+ },
2849
+ {
2850
+ "epoch": 21.95,
2851
+ "learning_rate": 8.129228734891946e-05,
2852
+ "loss": 0.6414,
2853
+ "step": 948000
2854
+ },
2855
+ {
2856
+ "epoch": 22.0,
2857
+ "learning_rate": 8.082448537884943e-05,
2858
+ "loss": 0.6285,
2859
+ "step": 950000
2860
+ },
2861
+ {
2862
+ "epoch": 22.04,
2863
+ "learning_rate": 8.03566834087794e-05,
2864
+ "loss": 0.6285,
2865
+ "step": 952000
2866
+ },
2867
+ {
2868
+ "epoch": 22.09,
2869
+ "learning_rate": 7.988888143870937e-05,
2870
+ "loss": 0.6268,
2871
+ "step": 954000
2872
+ },
2873
+ {
2874
+ "epoch": 22.14,
2875
+ "learning_rate": 7.942107946863934e-05,
2876
+ "loss": 0.6251,
2877
+ "step": 956000
2878
+ },
2879
+ {
2880
+ "epoch": 22.18,
2881
+ "learning_rate": 7.89532774985693e-05,
2882
+ "loss": 0.6306,
2883
+ "step": 958000
2884
+ },
2885
+ {
2886
+ "epoch": 22.23,
2887
+ "learning_rate": 7.848547552849927e-05,
2888
+ "loss": 0.6283,
2889
+ "step": 960000
2890
+ },
2891
+ {
2892
+ "epoch": 22.28,
2893
+ "learning_rate": 7.801767355842924e-05,
2894
+ "loss": 0.6264,
2895
+ "step": 962000
2896
+ },
2897
+ {
2898
+ "epoch": 22.32,
2899
+ "learning_rate": 7.754987158835921e-05,
2900
+ "loss": 0.6279,
2901
+ "step": 964000
2902
+ },
2903
+ {
2904
+ "epoch": 22.37,
2905
+ "learning_rate": 7.708206961828918e-05,
2906
+ "loss": 0.6272,
2907
+ "step": 966000
2908
+ },
2909
+ {
2910
+ "epoch": 22.41,
2911
+ "learning_rate": 7.661426764821915e-05,
2912
+ "loss": 0.6355,
2913
+ "step": 968000
2914
+ },
2915
+ {
2916
+ "epoch": 22.46,
2917
+ "learning_rate": 7.614646567814912e-05,
2918
+ "loss": 0.6349,
2919
+ "step": 970000
2920
+ },
2921
+ {
2922
+ "epoch": 22.51,
2923
+ "learning_rate": 7.567866370807909e-05,
2924
+ "loss": 0.6281,
2925
+ "step": 972000
2926
+ },
2927
+ {
2928
+ "epoch": 22.55,
2929
+ "learning_rate": 7.521086173800905e-05,
2930
+ "loss": 0.6269,
2931
+ "step": 974000
2932
+ },
2933
+ {
2934
+ "epoch": 22.6,
2935
+ "learning_rate": 7.474305976793904e-05,
2936
+ "loss": 0.6221,
2937
+ "step": 976000
2938
+ },
2939
+ {
2940
+ "epoch": 22.65,
2941
+ "learning_rate": 7.4275257797869e-05,
2942
+ "loss": 0.6295,
2943
+ "step": 978000
2944
+ },
2945
+ {
2946
+ "epoch": 22.69,
2947
+ "learning_rate": 7.380745582779897e-05,
2948
+ "loss": 0.6265,
2949
+ "step": 980000
2950
+ },
2951
+ {
2952
+ "epoch": 22.74,
2953
+ "learning_rate": 7.333965385772894e-05,
2954
+ "loss": 0.6203,
2955
+ "step": 982000
2956
+ },
2957
+ {
2958
+ "epoch": 22.79,
2959
+ "learning_rate": 7.287185188765891e-05,
2960
+ "loss": 0.6306,
2961
+ "step": 984000
2962
+ },
2963
+ {
2964
+ "epoch": 22.83,
2965
+ "learning_rate": 7.240404991758888e-05,
2966
+ "loss": 0.6319,
2967
+ "step": 986000
2968
+ },
2969
+ {
2970
+ "epoch": 22.88,
2971
+ "learning_rate": 7.193624794751885e-05,
2972
+ "loss": 0.6211,
2973
+ "step": 988000
2974
+ },
2975
+ {
2976
+ "epoch": 22.92,
2977
+ "learning_rate": 7.146844597744882e-05,
2978
+ "loss": 0.6244,
2979
+ "step": 990000
2980
+ },
2981
+ {
2982
+ "epoch": 22.97,
2983
+ "learning_rate": 7.100064400737879e-05,
2984
+ "loss": 0.6262,
2985
+ "step": 992000
2986
+ },
2987
+ {
2988
+ "epoch": 23.02,
2989
+ "learning_rate": 7.053284203730876e-05,
2990
+ "loss": 0.6166,
2991
+ "step": 994000
2992
+ },
2993
+ {
2994
+ "epoch": 23.06,
2995
+ "learning_rate": 7.006504006723873e-05,
2996
+ "loss": 0.6166,
2997
+ "step": 996000
2998
+ },
2999
+ {
3000
+ "epoch": 23.11,
3001
+ "learning_rate": 6.95972380971687e-05,
3002
+ "loss": 0.6175,
3003
+ "step": 998000
3004
+ },
3005
+ {
3006
+ "epoch": 23.16,
3007
+ "learning_rate": 6.912943612709866e-05,
3008
+ "loss": 0.6151,
3009
+ "step": 1000000
3010
+ },
3011
+ {
3012
+ "epoch": 23.2,
3013
+ "learning_rate": 6.866163415702863e-05,
3014
+ "loss": 0.6153,
3015
+ "step": 1002000
3016
+ },
3017
+ {
3018
+ "epoch": 23.25,
3019
+ "learning_rate": 6.81938321869586e-05,
3020
+ "loss": 0.6212,
3021
+ "step": 1004000
3022
+ },
3023
+ {
3024
+ "epoch": 23.29,
3025
+ "learning_rate": 6.772603021688858e-05,
3026
+ "loss": 0.6161,
3027
+ "step": 1006000
3028
+ },
3029
+ {
3030
+ "epoch": 23.34,
3031
+ "learning_rate": 6.725822824681855e-05,
3032
+ "loss": 0.6158,
3033
+ "step": 1008000
3034
+ },
3035
+ {
3036
+ "epoch": 23.39,
3037
+ "learning_rate": 6.679042627674852e-05,
3038
+ "loss": 0.6089,
3039
+ "step": 1010000
3040
+ },
3041
+ {
3042
+ "epoch": 23.43,
3043
+ "learning_rate": 6.632262430667849e-05,
3044
+ "loss": 0.6166,
3045
+ "step": 1012000
3046
+ },
3047
+ {
3048
+ "epoch": 23.48,
3049
+ "learning_rate": 6.585482233660846e-05,
3050
+ "loss": 0.6134,
3051
+ "step": 1014000
3052
+ },
3053
+ {
3054
+ "epoch": 23.53,
3055
+ "learning_rate": 6.538702036653843e-05,
3056
+ "loss": 0.6171,
3057
+ "step": 1016000
3058
+ },
3059
+ {
3060
+ "epoch": 23.57,
3061
+ "learning_rate": 6.49192183964684e-05,
3062
+ "loss": 0.6122,
3063
+ "step": 1018000
3064
+ },
3065
+ {
3066
+ "epoch": 23.62,
3067
+ "learning_rate": 6.445141642639837e-05,
3068
+ "loss": 0.6176,
3069
+ "step": 1020000
3070
+ },
3071
+ {
3072
+ "epoch": 23.67,
3073
+ "learning_rate": 6.398361445632833e-05,
3074
+ "loss": 0.6146,
3075
+ "step": 1022000
3076
+ },
3077
+ {
3078
+ "epoch": 23.71,
3079
+ "learning_rate": 6.35158124862583e-05,
3080
+ "loss": 0.6069,
3081
+ "step": 1024000
3082
+ },
3083
+ {
3084
+ "epoch": 23.76,
3085
+ "learning_rate": 6.304801051618829e-05,
3086
+ "loss": 0.6169,
3087
+ "step": 1026000
3088
+ },
3089
+ {
3090
+ "epoch": 23.8,
3091
+ "learning_rate": 6.258020854611825e-05,
3092
+ "loss": 0.6222,
3093
+ "step": 1028000
3094
+ },
3095
+ {
3096
+ "epoch": 23.85,
3097
+ "learning_rate": 6.211240657604822e-05,
3098
+ "loss": 0.6152,
3099
+ "step": 1030000
3100
+ },
3101
+ {
3102
+ "epoch": 23.9,
3103
+ "learning_rate": 6.164460460597819e-05,
3104
+ "loss": 0.6181,
3105
+ "step": 1032000
3106
+ },
3107
+ {
3108
+ "epoch": 23.94,
3109
+ "learning_rate": 6.117680263590816e-05,
3110
+ "loss": 0.6123,
3111
+ "step": 1034000
3112
+ },
3113
+ {
3114
+ "epoch": 23.99,
3115
+ "learning_rate": 6.070900066583813e-05,
3116
+ "loss": 0.619,
3117
+ "step": 1036000
3118
+ },
3119
+ {
3120
+ "epoch": 24.04,
3121
+ "learning_rate": 6.02411986957681e-05,
3122
+ "loss": 0.6099,
3123
+ "step": 1038000
3124
+ },
3125
+ {
3126
+ "epoch": 24.08,
3127
+ "learning_rate": 5.977339672569807e-05,
3128
+ "loss": 0.6098,
3129
+ "step": 1040000
3130
+ },
3131
+ {
3132
+ "epoch": 24.13,
3133
+ "learning_rate": 5.930559475562804e-05,
3134
+ "loss": 0.5965,
3135
+ "step": 1042000
3136
+ },
3137
+ {
3138
+ "epoch": 24.17,
3139
+ "learning_rate": 5.883779278555802e-05,
3140
+ "loss": 0.6059,
3141
+ "step": 1044000
3142
+ },
3143
+ {
3144
+ "epoch": 24.22,
3145
+ "learning_rate": 5.836999081548799e-05,
3146
+ "loss": 0.6021,
3147
+ "step": 1046000
3148
+ },
3149
+ {
3150
+ "epoch": 24.27,
3151
+ "learning_rate": 5.790218884541796e-05,
3152
+ "loss": 0.6093,
3153
+ "step": 1048000
3154
+ },
3155
+ {
3156
+ "epoch": 24.31,
3157
+ "learning_rate": 5.7434386875347926e-05,
3158
+ "loss": 0.6031,
3159
+ "step": 1050000
3160
+ },
3161
+ {
3162
+ "epoch": 24.36,
3163
+ "learning_rate": 5.6966584905277895e-05,
3164
+ "loss": 0.6053,
3165
+ "step": 1052000
3166
+ },
3167
+ {
3168
+ "epoch": 24.41,
3169
+ "learning_rate": 5.6498782935207863e-05,
3170
+ "loss": 0.6036,
3171
+ "step": 1054000
3172
+ },
3173
+ {
3174
+ "epoch": 24.45,
3175
+ "learning_rate": 5.603098096513783e-05,
3176
+ "loss": 0.6011,
3177
+ "step": 1056000
3178
+ },
3179
+ {
3180
+ "epoch": 24.5,
3181
+ "learning_rate": 5.55631789950678e-05,
3182
+ "loss": 0.6035,
3183
+ "step": 1058000
3184
+ },
3185
+ {
3186
+ "epoch": 24.55,
3187
+ "learning_rate": 5.509537702499777e-05,
3188
+ "loss": 0.6066,
3189
+ "step": 1060000
3190
+ },
3191
+ {
3192
+ "epoch": 24.59,
3193
+ "learning_rate": 5.4627575054927746e-05,
3194
+ "loss": 0.6061,
3195
+ "step": 1062000
3196
+ },
3197
+ {
3198
+ "epoch": 24.64,
3199
+ "learning_rate": 5.4159773084857714e-05,
3200
+ "loss": 0.6027,
3201
+ "step": 1064000
3202
+ },
3203
+ {
3204
+ "epoch": 24.68,
3205
+ "learning_rate": 5.369197111478768e-05,
3206
+ "loss": 0.6,
3207
+ "step": 1066000
3208
+ },
3209
+ {
3210
+ "epoch": 24.73,
3211
+ "learning_rate": 5.322416914471765e-05,
3212
+ "loss": 0.6062,
3213
+ "step": 1068000
3214
+ },
3215
+ {
3216
+ "epoch": 24.78,
3217
+ "learning_rate": 5.275636717464762e-05,
3218
+ "loss": 0.6003,
3219
+ "step": 1070000
3220
+ },
3221
+ {
3222
+ "epoch": 24.82,
3223
+ "learning_rate": 5.228856520457759e-05,
3224
+ "loss": 0.5988,
3225
+ "step": 1072000
3226
+ },
3227
+ {
3228
+ "epoch": 24.87,
3229
+ "learning_rate": 5.182076323450756e-05,
3230
+ "loss": 0.6096,
3231
+ "step": 1074000
3232
+ },
3233
+ {
3234
+ "epoch": 24.92,
3235
+ "learning_rate": 5.135296126443753e-05,
3236
+ "loss": 0.5988,
3237
+ "step": 1076000
3238
+ },
3239
+ {
3240
+ "epoch": 24.96,
3241
+ "learning_rate": 5.08851592943675e-05,
3242
+ "loss": 0.6086,
3243
+ "step": 1078000
3244
+ },
3245
+ {
3246
+ "epoch": 25.01,
3247
+ "learning_rate": 5.041735732429748e-05,
3248
+ "loss": 0.5942,
3249
+ "step": 1080000
3250
+ },
3251
+ {
3252
+ "epoch": 25.05,
3253
+ "learning_rate": 4.994955535422745e-05,
3254
+ "loss": 0.5954,
3255
+ "step": 1082000
3256
+ },
3257
+ {
3258
+ "epoch": 25.1,
3259
+ "learning_rate": 4.9481753384157417e-05,
3260
+ "loss": 0.5948,
3261
+ "step": 1084000
3262
+ },
3263
+ {
3264
+ "epoch": 25.15,
3265
+ "learning_rate": 4.9013951414087385e-05,
3266
+ "loss": 0.5946,
3267
+ "step": 1086000
3268
+ },
3269
+ {
3270
+ "epoch": 25.19,
3271
+ "learning_rate": 4.8546149444017354e-05,
3272
+ "loss": 0.5938,
3273
+ "step": 1088000
3274
+ },
3275
+ {
3276
+ "epoch": 25.24,
3277
+ "learning_rate": 4.807834747394732e-05,
3278
+ "loss": 0.5961,
3279
+ "step": 1090000
3280
+ },
3281
+ {
3282
+ "epoch": 25.29,
3283
+ "learning_rate": 4.761054550387729e-05,
3284
+ "loss": 0.5947,
3285
+ "step": 1092000
3286
+ },
3287
+ {
3288
+ "epoch": 25.33,
3289
+ "learning_rate": 4.714274353380726e-05,
3290
+ "loss": 0.6019,
3291
+ "step": 1094000
3292
+ },
3293
+ {
3294
+ "epoch": 25.38,
3295
+ "learning_rate": 4.667494156373723e-05,
3296
+ "loss": 0.5927,
3297
+ "step": 1096000
3298
+ },
3299
+ {
3300
+ "epoch": 25.43,
3301
+ "learning_rate": 4.6207139593667205e-05,
3302
+ "loss": 0.5921,
3303
+ "step": 1098000
3304
+ },
3305
+ {
3306
+ "epoch": 25.47,
3307
+ "learning_rate": 4.5739337623597174e-05,
3308
+ "loss": 0.5954,
3309
+ "step": 1100000
3310
+ },
3311
+ {
3312
+ "epoch": 25.52,
3313
+ "learning_rate": 4.527153565352715e-05,
3314
+ "loss": 0.5926,
3315
+ "step": 1102000
3316
+ },
3317
+ {
3318
+ "epoch": 25.56,
3319
+ "learning_rate": 4.480373368345712e-05,
3320
+ "loss": 0.5963,
3321
+ "step": 1104000
3322
+ },
3323
+ {
3324
+ "epoch": 25.61,
3325
+ "learning_rate": 4.433593171338709e-05,
3326
+ "loss": 0.5902,
3327
+ "step": 1106000
3328
+ },
3329
+ {
3330
+ "epoch": 25.66,
3331
+ "learning_rate": 4.3868129743317056e-05,
3332
+ "loss": 0.5952,
3333
+ "step": 1108000
3334
+ },
3335
+ {
3336
+ "epoch": 25.7,
3337
+ "learning_rate": 4.3400327773247025e-05,
3338
+ "loss": 0.5878,
3339
+ "step": 1110000
3340
+ },
3341
+ {
3342
+ "epoch": 25.75,
3343
+ "learning_rate": 4.2932525803176994e-05,
3344
+ "loss": 0.5926,
3345
+ "step": 1112000
3346
+ },
3347
+ {
3348
+ "epoch": 25.8,
3349
+ "learning_rate": 4.246472383310696e-05,
3350
+ "loss": 0.5854,
3351
+ "step": 1114000
3352
+ },
3353
+ {
3354
+ "epoch": 25.84,
3355
+ "learning_rate": 4.199692186303694e-05,
3356
+ "loss": 0.5916,
3357
+ "step": 1116000
3358
+ },
3359
+ {
3360
+ "epoch": 25.89,
3361
+ "learning_rate": 4.152911989296691e-05,
3362
+ "loss": 0.5869,
3363
+ "step": 1118000
3364
+ },
3365
+ {
3366
+ "epoch": 25.93,
3367
+ "learning_rate": 4.1061317922896876e-05,
3368
+ "loss": 0.5913,
3369
+ "step": 1120000
3370
+ },
3371
+ {
3372
+ "epoch": 25.98,
3373
+ "learning_rate": 4.0593515952826845e-05,
3374
+ "loss": 0.5822,
3375
+ "step": 1122000
3376
+ },
3377
+ {
3378
+ "epoch": 26.03,
3379
+ "learning_rate": 4.0125713982756814e-05,
3380
+ "loss": 0.5831,
3381
+ "step": 1124000
3382
+ },
3383
+ {
3384
+ "epoch": 26.07,
3385
+ "learning_rate": 3.965791201268678e-05,
3386
+ "loss": 0.5847,
3387
+ "step": 1126000
3388
+ },
3389
+ {
3390
+ "epoch": 26.12,
3391
+ "learning_rate": 3.919011004261675e-05,
3392
+ "loss": 0.5828,
3393
+ "step": 1128000
3394
+ },
3395
+ {
3396
+ "epoch": 26.17,
3397
+ "learning_rate": 3.872230807254672e-05,
3398
+ "loss": 0.5825,
3399
+ "step": 1130000
3400
+ },
3401
+ {
3402
+ "epoch": 26.21,
3403
+ "learning_rate": 3.825450610247669e-05,
3404
+ "loss": 0.5848,
3405
+ "step": 1132000
3406
+ },
3407
+ {
3408
+ "epoch": 26.26,
3409
+ "learning_rate": 3.778670413240667e-05,
3410
+ "loss": 0.5866,
3411
+ "step": 1134000
3412
+ },
3413
+ {
3414
+ "epoch": 26.31,
3415
+ "learning_rate": 3.7318902162336634e-05,
3416
+ "loss": 0.5832,
3417
+ "step": 1136000
3418
+ },
3419
+ {
3420
+ "epoch": 26.35,
3421
+ "learning_rate": 3.685110019226661e-05,
3422
+ "loss": 0.58,
3423
+ "step": 1138000
3424
+ },
3425
+ {
3426
+ "epoch": 26.4,
3427
+ "learning_rate": 3.638329822219658e-05,
3428
+ "loss": 0.5767,
3429
+ "step": 1140000
3430
+ },
3431
+ {
3432
+ "epoch": 26.44,
3433
+ "learning_rate": 3.591549625212655e-05,
3434
+ "loss": 0.5792,
3435
+ "step": 1142000
3436
+ },
3437
+ {
3438
+ "epoch": 26.49,
3439
+ "learning_rate": 3.5447694282056516e-05,
3440
+ "loss": 0.5764,
3441
+ "step": 1144000
3442
+ },
3443
+ {
3444
+ "epoch": 26.54,
3445
+ "learning_rate": 3.4979892311986485e-05,
3446
+ "loss": 0.5794,
3447
+ "step": 1146000
3448
+ },
3449
+ {
3450
+ "epoch": 26.58,
3451
+ "learning_rate": 3.451209034191646e-05,
3452
+ "loss": 0.5738,
3453
+ "step": 1148000
3454
+ },
3455
+ {
3456
+ "epoch": 26.63,
3457
+ "learning_rate": 3.404428837184643e-05,
3458
+ "loss": 0.5822,
3459
+ "step": 1150000
3460
+ },
3461
+ {
3462
+ "epoch": 26.68,
3463
+ "learning_rate": 3.35764864017764e-05,
3464
+ "loss": 0.5734,
3465
+ "step": 1152000
3466
+ },
3467
+ {
3468
+ "epoch": 26.72,
3469
+ "learning_rate": 3.310868443170637e-05,
3470
+ "loss": 0.5794,
3471
+ "step": 1154000
3472
+ },
3473
+ {
3474
+ "epoch": 26.77,
3475
+ "learning_rate": 3.2640882461636336e-05,
3476
+ "loss": 0.5853,
3477
+ "step": 1156000
3478
+ },
3479
+ {
3480
+ "epoch": 26.81,
3481
+ "learning_rate": 3.2173080491566305e-05,
3482
+ "loss": 0.5842,
3483
+ "step": 1158000
3484
+ },
3485
+ {
3486
+ "epoch": 26.86,
3487
+ "learning_rate": 3.170527852149628e-05,
3488
+ "loss": 0.5847,
3489
+ "step": 1160000
3490
+ },
3491
+ {
3492
+ "epoch": 26.91,
3493
+ "learning_rate": 3.123747655142625e-05,
3494
+ "loss": 0.5786,
3495
+ "step": 1162000
3496
+ },
3497
+ {
3498
+ "epoch": 26.95,
3499
+ "learning_rate": 3.076967458135622e-05,
3500
+ "loss": 0.5818,
3501
+ "step": 1164000
3502
+ },
3503
+ {
3504
+ "epoch": 27.0,
3505
+ "learning_rate": 3.030187261128619e-05,
3506
+ "loss": 0.5722,
3507
+ "step": 1166000
3508
+ },
3509
+ {
3510
+ "epoch": 27.05,
3511
+ "learning_rate": 2.983407064121616e-05,
3512
+ "loss": 0.5726,
3513
+ "step": 1168000
3514
+ },
3515
+ {
3516
+ "epoch": 27.09,
3517
+ "learning_rate": 2.936626867114613e-05,
3518
+ "loss": 0.5745,
3519
+ "step": 1170000
3520
+ },
3521
+ {
3522
+ "epoch": 27.14,
3523
+ "learning_rate": 2.8898466701076097e-05,
3524
+ "loss": 0.5655,
3525
+ "step": 1172000
3526
+ },
3527
+ {
3528
+ "epoch": 27.19,
3529
+ "learning_rate": 2.843066473100607e-05,
3530
+ "loss": 0.5747,
3531
+ "step": 1174000
3532
+ },
3533
+ {
3534
+ "epoch": 27.23,
3535
+ "learning_rate": 2.7962862760936038e-05,
3536
+ "loss": 0.5734,
3537
+ "step": 1176000
3538
+ },
3539
+ {
3540
+ "epoch": 27.28,
3541
+ "learning_rate": 2.7495060790866007e-05,
3542
+ "loss": 0.5752,
3543
+ "step": 1178000
3544
+ },
3545
+ {
3546
+ "epoch": 27.32,
3547
+ "learning_rate": 2.7027258820795976e-05,
3548
+ "loss": 0.5784,
3549
+ "step": 1180000
3550
+ },
3551
+ {
3552
+ "epoch": 27.37,
3553
+ "learning_rate": 2.6559456850725945e-05,
3554
+ "loss": 0.5667,
3555
+ "step": 1182000
3556
+ },
3557
+ {
3558
+ "epoch": 27.42,
3559
+ "learning_rate": 2.609165488065592e-05,
3560
+ "loss": 0.5748,
3561
+ "step": 1184000
3562
+ },
3563
+ {
3564
+ "epoch": 27.46,
3565
+ "learning_rate": 2.562385291058589e-05,
3566
+ "loss": 0.5762,
3567
+ "step": 1186000
3568
+ },
3569
+ {
3570
+ "epoch": 27.51,
3571
+ "learning_rate": 2.5156050940515858e-05,
3572
+ "loss": 0.5783,
3573
+ "step": 1188000
3574
+ },
3575
+ {
3576
+ "epoch": 27.56,
3577
+ "learning_rate": 2.4688248970445827e-05,
3578
+ "loss": 0.5668,
3579
+ "step": 1190000
3580
+ },
3581
+ {
3582
+ "epoch": 27.6,
3583
+ "learning_rate": 2.42204470003758e-05,
3584
+ "loss": 0.5671,
3585
+ "step": 1192000
3586
+ },
3587
+ {
3588
+ "epoch": 27.65,
3589
+ "learning_rate": 2.3752645030305768e-05,
3590
+ "loss": 0.5688,
3591
+ "step": 1194000
3592
+ },
3593
+ {
3594
+ "epoch": 27.69,
3595
+ "learning_rate": 2.328484306023574e-05,
3596
+ "loss": 0.5643,
3597
+ "step": 1196000
3598
+ },
3599
+ {
3600
+ "epoch": 27.74,
3601
+ "learning_rate": 2.281704109016571e-05,
3602
+ "loss": 0.5688,
3603
+ "step": 1198000
3604
+ },
3605
+ {
3606
+ "epoch": 27.79,
3607
+ "learning_rate": 2.2349239120095678e-05,
3608
+ "loss": 0.5651,
3609
+ "step": 1200000
3610
+ },
3611
+ {
3612
+ "epoch": 27.83,
3613
+ "learning_rate": 2.188143715002565e-05,
3614
+ "loss": 0.5705,
3615
+ "step": 1202000
3616
+ },
3617
+ {
3618
+ "epoch": 27.88,
3619
+ "learning_rate": 2.141363517995562e-05,
3620
+ "loss": 0.5684,
3621
+ "step": 1204000
3622
+ },
3623
+ {
3624
+ "epoch": 27.93,
3625
+ "learning_rate": 2.0945833209885588e-05,
3626
+ "loss": 0.567,
3627
+ "step": 1206000
3628
+ },
3629
+ {
3630
+ "epoch": 27.97,
3631
+ "learning_rate": 2.0478031239815557e-05,
3632
+ "loss": 0.5711,
3633
+ "step": 1208000
3634
+ },
3635
+ {
3636
+ "epoch": 28.02,
3637
+ "learning_rate": 2.0010229269745533e-05,
3638
+ "loss": 0.5684,
3639
+ "step": 1210000
3640
+ },
3641
+ {
3642
+ "epoch": 28.06,
3643
+ "learning_rate": 1.95424272996755e-05,
3644
+ "loss": 0.5604,
3645
+ "step": 1212000
3646
+ },
3647
+ {
3648
+ "epoch": 28.11,
3649
+ "learning_rate": 1.907462532960547e-05,
3650
+ "loss": 0.5615,
3651
+ "step": 1214000
3652
+ },
3653
+ {
3654
+ "epoch": 28.16,
3655
+ "learning_rate": 1.860682335953544e-05,
3656
+ "loss": 0.5679,
3657
+ "step": 1216000
3658
+ },
3659
+ {
3660
+ "epoch": 28.2,
3661
+ "learning_rate": 1.813902138946541e-05,
3662
+ "loss": 0.5644,
3663
+ "step": 1218000
3664
+ },
3665
+ {
3666
+ "epoch": 28.25,
3667
+ "learning_rate": 1.767121941939538e-05,
3668
+ "loss": 0.5663,
3669
+ "step": 1220000
3670
+ },
3671
+ {
3672
+ "epoch": 28.3,
3673
+ "learning_rate": 1.720341744932535e-05,
3674
+ "loss": 0.5584,
3675
+ "step": 1222000
3676
+ },
3677
+ {
3678
+ "epoch": 28.34,
3679
+ "learning_rate": 1.6735615479255318e-05,
3680
+ "loss": 0.558,
3681
+ "step": 1224000
3682
+ },
3683
+ {
3684
+ "epoch": 28.39,
3685
+ "learning_rate": 1.626781350918529e-05,
3686
+ "loss": 0.5575,
3687
+ "step": 1226000
3688
+ },
3689
+ {
3690
+ "epoch": 28.44,
3691
+ "learning_rate": 1.580001153911526e-05,
3692
+ "loss": 0.5728,
3693
+ "step": 1228000
3694
+ },
3695
+ {
3696
+ "epoch": 28.48,
3697
+ "learning_rate": 1.533220956904523e-05,
3698
+ "loss": 0.5653,
3699
+ "step": 1230000
3700
+ },
3701
+ {
3702
+ "epoch": 28.53,
3703
+ "learning_rate": 1.48644075989752e-05,
3704
+ "loss": 0.5603,
3705
+ "step": 1232000
3706
+ },
3707
+ {
3708
+ "epoch": 28.57,
3709
+ "learning_rate": 1.4396605628905172e-05,
3710
+ "loss": 0.5613,
3711
+ "step": 1234000
3712
+ },
3713
+ {
3714
+ "epoch": 28.62,
3715
+ "learning_rate": 1.3928803658835141e-05,
3716
+ "loss": 0.5563,
3717
+ "step": 1236000
3718
+ },
3719
+ {
3720
+ "epoch": 28.67,
3721
+ "learning_rate": 1.346100168876511e-05,
3722
+ "loss": 0.5705,
3723
+ "step": 1238000
3724
+ },
3725
+ {
3726
+ "epoch": 28.71,
3727
+ "learning_rate": 1.299319971869508e-05,
3728
+ "loss": 0.5568,
3729
+ "step": 1240000
3730
+ },
3731
+ {
3732
+ "epoch": 28.76,
3733
+ "learning_rate": 1.252539774862505e-05,
3734
+ "loss": 0.5517,
3735
+ "step": 1242000
3736
+ },
3737
+ {
3738
+ "epoch": 28.81,
3739
+ "learning_rate": 1.2057595778555022e-05,
3740
+ "loss": 0.5647,
3741
+ "step": 1244000
3742
+ },
3743
+ {
3744
+ "epoch": 28.85,
3745
+ "learning_rate": 1.158979380848499e-05,
3746
+ "loss": 0.5551,
3747
+ "step": 1246000
3748
+ },
3749
+ {
3750
+ "epoch": 28.9,
3751
+ "learning_rate": 1.1121991838414961e-05,
3752
+ "loss": 0.5598,
3753
+ "step": 1248000
3754
+ },
3755
+ {
3756
+ "epoch": 28.94,
3757
+ "learning_rate": 1.0654189868344932e-05,
3758
+ "loss": 0.562,
3759
+ "step": 1250000
3760
+ },
3761
+ {
3762
+ "epoch": 28.99,
3763
+ "learning_rate": 1.0186387898274902e-05,
3764
+ "loss": 0.5563,
3765
+ "step": 1252000
3766
+ }
3767
+ ],
3768
+ "logging_steps": 2000,
3769
+ "max_steps": 1295550,
3770
+ "num_train_epochs": 30,
3771
+ "save_steps": 500,
3772
+ "total_flos": 2.559806921525875e+21,
3773
+ "trial_name": null,
3774
+ "trial_params": null
3775
+ }
checkpoint-1252379/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:15b4a29d89f83e50e17158ee0f4ffe9724daa5ff51dfccdeac7b5e2c6a3b564e
3
+ size 4600
checkpoint-1252379/vocab.json ADDED
@@ -0,0 +1,53 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "\b": 0,
3
+ "\"": 2,
4
+ "[PAD]": 50,
5
+ "[UNK]": 49,
6
+ "\\": 3,
7
+ "a": 4,
8
+ "b": 5,
9
+ "c": 6,
10
+ "d": 7,
11
+ "e": 8,
12
+ "f": 9,
13
+ "g": 10,
14
+ "h": 11,
15
+ "i": 12,
16
+ "j": 13,
17
+ "k": 14,
18
+ "l": 15,
19
+ "m": 16,
20
+ "n": 17,
21
+ "o": 18,
22
+ "p": 19,
23
+ "q": 20,
24
+ "r": 21,
25
+ "s": 22,
26
+ "t": 23,
27
+ "u": 24,
28
+ "v": 25,
29
+ "w": 26,
30
+ "x": 27,
31
+ "y": 28,
32
+ "z": 29,
33
+ "|": 1,
34
+ "ç": 30,
35
+ "é": 31,
36
+ "ş": 32,
37
+ "υ": 33,
38
+ "а": 34,
39
+ "в": 35,
40
+ "г": 36,
41
+ "д": 37,
42
+ "е": 38,
43
+ "к": 39,
44
+ "л": 40,
45
+ "н": 41,
46
+ "с": 42,
47
+ "ф": 43,
48
+ "ч": 44,
49
+ "​": 45,
50
+ "’": 46,
51
+ "⁄": 47,
52
+ "�": 48
53
+ }
checkpoint-1295550/added_tokens.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "</s>": 52,
3
+ "<s>": 51
4
+ }
checkpoint-1295550/config.json ADDED
@@ -0,0 +1,117 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/wav2vec2-large-xlsr-53",
3
+ "activation_dropout": 0.0,
4
+ "adapter_attn_dim": null,
5
+ "adapter_kernel_size": 3,
6
+ "adapter_stride": 2,
7
+ "add_adapter": false,
8
+ "apply_spec_augment": true,
9
+ "architectures": [
10
+ "Wav2Vec2ForCTC"
11
+ ],
12
+ "attention_dropout": 0.05,
13
+ "bos_token_id": 1,
14
+ "classifier_proj_size": 256,
15
+ "codevector_dim": 768,
16
+ "contrastive_logits_temperature": 0.1,
17
+ "conv_bias": true,
18
+ "conv_dim": [
19
+ 512,
20
+ 512,
21
+ 512,
22
+ 512,
23
+ 512,
24
+ 512,
25
+ 512
26
+ ],
27
+ "conv_kernel": [
28
+ 10,
29
+ 3,
30
+ 3,
31
+ 3,
32
+ 3,
33
+ 2,
34
+ 2
35
+ ],
36
+ "conv_stride": [
37
+ 5,
38
+ 2,
39
+ 2,
40
+ 2,
41
+ 2,
42
+ 2,
43
+ 2
44
+ ],
45
+ "ctc_loss_reduction": "mean",
46
+ "ctc_zero_infinity": true,
47
+ "diversity_loss_weight": 0.1,
48
+ "do_stable_layer_norm": true,
49
+ "eos_token_id": 2,
50
+ "feat_extract_activation": "gelu",
51
+ "feat_extract_dropout": 0.0,
52
+ "feat_extract_norm": "layer",
53
+ "feat_proj_dropout": 0.05,
54
+ "feat_quantizer_dropout": 0.0,
55
+ "final_dropout": 0.0,
56
+ "gradient_checkpointing": false,
57
+ "hidden_act": "gelu",
58
+ "hidden_dropout": 0.05,
59
+ "hidden_size": 1024,
60
+ "initializer_range": 0.02,
61
+ "intermediate_size": 4096,
62
+ "layer_norm_eps": 1e-05,
63
+ "layerdrop": 0.05,
64
+ "mask_channel_length": 10,
65
+ "mask_channel_min_space": 1,
66
+ "mask_channel_other": 0.0,
67
+ "mask_channel_prob": 0.0,
68
+ "mask_channel_selection": "static",
69
+ "mask_feature_length": 10,
70
+ "mask_feature_min_masks": 0,
71
+ "mask_feature_prob": 0.0,
72
+ "mask_time_length": 10,
73
+ "mask_time_min_masks": 2,
74
+ "mask_time_min_space": 1,
75
+ "mask_time_other": 0.0,
76
+ "mask_time_prob": 0.05,
77
+ "mask_time_selection": "static",
78
+ "model_type": "wav2vec2",
79
+ "num_adapter_layers": 3,
80
+ "num_attention_heads": 16,
81
+ "num_codevector_groups": 2,
82
+ "num_codevectors_per_group": 320,
83
+ "num_conv_pos_embedding_groups": 16,
84
+ "num_conv_pos_embeddings": 128,
85
+ "num_feat_extract_layers": 7,
86
+ "num_hidden_layers": 24,
87
+ "num_negatives": 100,
88
+ "output_hidden_size": 1024,
89
+ "pad_token_id": 50,
90
+ "proj_codevector_dim": 768,
91
+ "tdnn_dilation": [
92
+ 1,
93
+ 2,
94
+ 3,
95
+ 1,
96
+ 1
97
+ ],
98
+ "tdnn_dim": [
99
+ 512,
100
+ 512,
101
+ 512,
102
+ 512,
103
+ 1500
104
+ ],
105
+ "tdnn_kernel": [
106
+ 5,
107
+ 3,
108
+ 3,
109
+ 1,
110
+ 1
111
+ ],
112
+ "torch_dtype": "float32",
113
+ "transformers_version": "4.35.0",
114
+ "use_weighted_layer_sum": false,
115
+ "vocab_size": 53,
116
+ "xvector_output_dim": 512
117
+ }
checkpoint-1295550/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e8ca98ff5bd5909097ec35838b22fce9f53fe21dc0ce7d473d8ffc93958b9d50
3
+ size 1262024780
checkpoint-1295550/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:13dee4e6dfb55b4c448cc21226cf9d4f901828369b38840df2d29d1526710aa8
3
+ size 2490594358
checkpoint-1295550/preprocessor_config.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "feature_extractor_type": "Wav2Vec2FeatureExtractor",
4
+ "feature_size": 1,
5
+ "padding_side": "right",
6
+ "padding_value": 0,
7
+ "processor_class": "Wav2Vec2Processor",
8
+ "return_attention_mask": true,
9
+ "sampling_rate": 16000
10
+ }
checkpoint-1295550/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4f4974c63dcb679a0bf851196951f3f8e34acb833136b0b557ec58b36108ba19
3
+ size 15024
checkpoint-1295550/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ab6b8c7b0b424ee1408eb83b7d8e9804358a953e0a785e8c3b878455a1d11cc7
3
+ size 15024
checkpoint-1295550/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:07b7e1632e7647d7c277c56bb20c6687ad59d7873cdb9fe89502e3944feb6349
3
+ size 15024
checkpoint-1295550/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:92035b29dc2b2a4687c0e052eb3d470687804ab866afb2dfad17753d394a9083
3
+ size 15024
checkpoint-1295550/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7a45b8b26ba589ee3d49a213fd447479f51216d476db846a0e428fc4a8635723
3
+ size 1064
checkpoint-1295550/special_tokens_map.json ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": true,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "</s>",
11
+ "lstrip": false,
12
+ "normalized": true,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": {
17
+ "content": "[PAD]",
18
+ "lstrip": true,
19
+ "normalized": false,
20
+ "rstrip": true,
21
+ "single_word": false
22
+ },
23
+ "unk_token": {
24
+ "content": "[UNK]",
25
+ "lstrip": true,
26
+ "normalized": false,
27
+ "rstrip": true,
28
+ "single_word": false
29
+ }
30
+ }
checkpoint-1295550/tokenizer_config.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "49": {
4
+ "content": "[UNK]",
5
+ "lstrip": true,
6
+ "normalized": false,
7
+ "rstrip": true,
8
+ "single_word": false,
9
+ "special": false
10
+ },
11
+ "50": {
12
+ "content": "[PAD]",
13
+ "lstrip": true,
14
+ "normalized": false,
15
+ "rstrip": true,
16
+ "single_word": false,
17
+ "special": false
18
+ },
19
+ "51": {
20
+ "content": "<s>",
21
+ "lstrip": false,
22
+ "normalized": true,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "52": {
28
+ "content": "</s>",
29
+ "lstrip": false,
30
+ "normalized": true,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ }
35
+ },
36
+ "bos_token": "<s>",
37
+ "clean_up_tokenization_spaces": true,
38
+ "config": null,
39
+ "do_lower_case": false,
40
+ "eos_token": "</s>",
41
+ "model_max_length": 1000000000000000019884624838656,
42
+ "pad_token": "[PAD]",
43
+ "processor_class": "Wav2Vec2Processor",
44
+ "replace_word_delimiter_char": " ",
45
+ "target_lang": null,
46
+ "tokenizer_class": "Wav2Vec2CTCTokenizer",
47
+ "tokenizer_type": "wav2vec2",
48
+ "trust_remote_code": false,
49
+ "unk_token": "[UNK]",
50
+ "word_delimiter_token": "|"
51
+ }
checkpoint-1295550/trainer_state.json ADDED
@@ -0,0 +1,3901 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 29.99965266119415,
5
+ "eval_steps": 500,
6
+ "global_step": 1295550,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.05,
13
+ "learning_rate": 4.6310589688175354e-05,
14
+ "loss": 4.5055,
15
+ "step": 2000
16
+ },
17
+ {
18
+ "epoch": 0.09,
19
+ "learning_rate": 9.262117937635071e-05,
20
+ "loss": 1.4851,
21
+ "step": 4000
22
+ },
23
+ {
24
+ "epoch": 0.14,
25
+ "learning_rate": 0.00013893176906452608,
26
+ "loss": 1.1808,
27
+ "step": 6000
28
+ },
29
+ {
30
+ "epoch": 0.19,
31
+ "learning_rate": 0.00018524235875270141,
32
+ "loss": 1.132,
33
+ "step": 8000
34
+ },
35
+ {
36
+ "epoch": 0.23,
37
+ "learning_rate": 0.0002315529484408768,
38
+ "loss": 1.1349,
39
+ "step": 10000
40
+ },
41
+ {
42
+ "epoch": 0.28,
43
+ "learning_rate": 0.00027786353812905216,
44
+ "loss": 1.1422,
45
+ "step": 12000
46
+ },
47
+ {
48
+ "epoch": 0.32,
49
+ "learning_rate": 0.0002997558073716234,
50
+ "loss": 1.1513,
51
+ "step": 14000
52
+ },
53
+ {
54
+ "epoch": 0.37,
55
+ "learning_rate": 0.0002992880054015534,
56
+ "loss": 1.142,
57
+ "step": 16000
58
+ },
59
+ {
60
+ "epoch": 0.42,
61
+ "learning_rate": 0.00029882020343148334,
62
+ "loss": 1.1303,
63
+ "step": 18000
64
+ },
65
+ {
66
+ "epoch": 0.46,
67
+ "learning_rate": 0.0002983524014614133,
68
+ "loss": 1.1215,
69
+ "step": 20000
70
+ },
71
+ {
72
+ "epoch": 0.51,
73
+ "learning_rate": 0.0002978845994913433,
74
+ "loss": 1.109,
75
+ "step": 22000
76
+ },
77
+ {
78
+ "epoch": 0.56,
79
+ "learning_rate": 0.00029741679752127325,
80
+ "loss": 1.1034,
81
+ "step": 24000
82
+ },
83
+ {
84
+ "epoch": 0.6,
85
+ "learning_rate": 0.0002969489955512032,
86
+ "loss": 1.0998,
87
+ "step": 26000
88
+ },
89
+ {
90
+ "epoch": 0.65,
91
+ "learning_rate": 0.00029648119358113324,
92
+ "loss": 1.0979,
93
+ "step": 28000
94
+ },
95
+ {
96
+ "epoch": 0.69,
97
+ "learning_rate": 0.00029601339161106316,
98
+ "loss": 1.083,
99
+ "step": 30000
100
+ },
101
+ {
102
+ "epoch": 0.74,
103
+ "learning_rate": 0.0002955455896409932,
104
+ "loss": 1.0741,
105
+ "step": 32000
106
+ },
107
+ {
108
+ "epoch": 0.79,
109
+ "learning_rate": 0.0002950777876709231,
110
+ "loss": 1.0741,
111
+ "step": 34000
112
+ },
113
+ {
114
+ "epoch": 0.83,
115
+ "learning_rate": 0.0002946099857008531,
116
+ "loss": 1.0707,
117
+ "step": 36000
118
+ },
119
+ {
120
+ "epoch": 0.88,
121
+ "learning_rate": 0.00029414218373078303,
122
+ "loss": 1.0689,
123
+ "step": 38000
124
+ },
125
+ {
126
+ "epoch": 0.93,
127
+ "learning_rate": 0.00029367438176071306,
128
+ "loss": 1.0536,
129
+ "step": 40000
130
+ },
131
+ {
132
+ "epoch": 0.97,
133
+ "learning_rate": 0.00029320657979064297,
134
+ "loss": 1.0556,
135
+ "step": 42000
136
+ },
137
+ {
138
+ "epoch": 1.02,
139
+ "learning_rate": 0.000292738777820573,
140
+ "loss": 1.0358,
141
+ "step": 44000
142
+ },
143
+ {
144
+ "epoch": 1.07,
145
+ "learning_rate": 0.0002922709758505029,
146
+ "loss": 1.018,
147
+ "step": 46000
148
+ },
149
+ {
150
+ "epoch": 1.11,
151
+ "learning_rate": 0.00029180317388043293,
152
+ "loss": 1.0162,
153
+ "step": 48000
154
+ },
155
+ {
156
+ "epoch": 1.16,
157
+ "learning_rate": 0.00029133537191036285,
158
+ "loss": 1.017,
159
+ "step": 50000
160
+ },
161
+ {
162
+ "epoch": 1.2,
163
+ "learning_rate": 0.00029086756994029287,
164
+ "loss": 1.0145,
165
+ "step": 52000
166
+ },
167
+ {
168
+ "epoch": 1.25,
169
+ "learning_rate": 0.0002903997679702228,
170
+ "loss": 1.0103,
171
+ "step": 54000
172
+ },
173
+ {
174
+ "epoch": 1.3,
175
+ "learning_rate": 0.0002899319660001528,
176
+ "loss": 1.0046,
177
+ "step": 56000
178
+ },
179
+ {
180
+ "epoch": 1.34,
181
+ "learning_rate": 0.0002894641640300827,
182
+ "loss": 1.01,
183
+ "step": 58000
184
+ },
185
+ {
186
+ "epoch": 1.39,
187
+ "learning_rate": 0.00028899636206001274,
188
+ "loss": 1.0146,
189
+ "step": 60000
190
+ },
191
+ {
192
+ "epoch": 1.44,
193
+ "learning_rate": 0.00028852856008994266,
194
+ "loss": 1.0122,
195
+ "step": 62000
196
+ },
197
+ {
198
+ "epoch": 1.48,
199
+ "learning_rate": 0.0002880607581198727,
200
+ "loss": 0.9961,
201
+ "step": 64000
202
+ },
203
+ {
204
+ "epoch": 1.53,
205
+ "learning_rate": 0.00028759295614980265,
206
+ "loss": 0.9914,
207
+ "step": 66000
208
+ },
209
+ {
210
+ "epoch": 1.57,
211
+ "learning_rate": 0.0002871251541797326,
212
+ "loss": 0.9936,
213
+ "step": 68000
214
+ },
215
+ {
216
+ "epoch": 1.62,
217
+ "learning_rate": 0.0002866573522096626,
218
+ "loss": 0.9998,
219
+ "step": 70000
220
+ },
221
+ {
222
+ "epoch": 1.67,
223
+ "learning_rate": 0.00028618955023959256,
224
+ "loss": 0.9833,
225
+ "step": 72000
226
+ },
227
+ {
228
+ "epoch": 1.71,
229
+ "learning_rate": 0.0002857217482695225,
230
+ "loss": 0.9901,
231
+ "step": 74000
232
+ },
233
+ {
234
+ "epoch": 1.76,
235
+ "learning_rate": 0.0002852539462994525,
236
+ "loss": 0.9905,
237
+ "step": 76000
238
+ },
239
+ {
240
+ "epoch": 1.81,
241
+ "learning_rate": 0.00028478614432938246,
242
+ "loss": 0.9845,
243
+ "step": 78000
244
+ },
245
+ {
246
+ "epoch": 1.85,
247
+ "learning_rate": 0.00028431834235931243,
248
+ "loss": 0.9886,
249
+ "step": 80000
250
+ },
251
+ {
252
+ "epoch": 1.9,
253
+ "learning_rate": 0.0002838505403892424,
254
+ "loss": 0.9848,
255
+ "step": 82000
256
+ },
257
+ {
258
+ "epoch": 1.95,
259
+ "learning_rate": 0.00028338273841917237,
260
+ "loss": 0.9827,
261
+ "step": 84000
262
+ },
263
+ {
264
+ "epoch": 1.99,
265
+ "learning_rate": 0.00028291493644910234,
266
+ "loss": 0.9764,
267
+ "step": 86000
268
+ },
269
+ {
270
+ "epoch": 2.04,
271
+ "learning_rate": 0.0002824471344790323,
272
+ "loss": 0.9546,
273
+ "step": 88000
274
+ },
275
+ {
276
+ "epoch": 2.08,
277
+ "learning_rate": 0.0002819793325089623,
278
+ "loss": 0.9388,
279
+ "step": 90000
280
+ },
281
+ {
282
+ "epoch": 2.13,
283
+ "learning_rate": 0.00028151153053889225,
284
+ "loss": 0.9432,
285
+ "step": 92000
286
+ },
287
+ {
288
+ "epoch": 2.18,
289
+ "learning_rate": 0.0002810437285688222,
290
+ "loss": 0.9591,
291
+ "step": 94000
292
+ },
293
+ {
294
+ "epoch": 2.22,
295
+ "learning_rate": 0.0002805759265987522,
296
+ "loss": 0.9431,
297
+ "step": 96000
298
+ },
299
+ {
300
+ "epoch": 2.27,
301
+ "learning_rate": 0.00028010812462868215,
302
+ "loss": 0.9388,
303
+ "step": 98000
304
+ },
305
+ {
306
+ "epoch": 2.32,
307
+ "learning_rate": 0.0002796403226586121,
308
+ "loss": 0.9421,
309
+ "step": 100000
310
+ },
311
+ {
312
+ "epoch": 2.36,
313
+ "learning_rate": 0.0002791725206885421,
314
+ "loss": 0.9452,
315
+ "step": 102000
316
+ },
317
+ {
318
+ "epoch": 2.41,
319
+ "learning_rate": 0.0002787047187184721,
320
+ "loss": 0.9535,
321
+ "step": 104000
322
+ },
323
+ {
324
+ "epoch": 2.45,
325
+ "learning_rate": 0.00027823691674840203,
326
+ "loss": 0.9373,
327
+ "step": 106000
328
+ },
329
+ {
330
+ "epoch": 2.5,
331
+ "learning_rate": 0.00027776911477833205,
332
+ "loss": 0.9345,
333
+ "step": 108000
334
+ },
335
+ {
336
+ "epoch": 2.55,
337
+ "learning_rate": 0.00027730131280826197,
338
+ "loss": 0.9478,
339
+ "step": 110000
340
+ },
341
+ {
342
+ "epoch": 2.59,
343
+ "learning_rate": 0.000276833510838192,
344
+ "loss": 0.9445,
345
+ "step": 112000
346
+ },
347
+ {
348
+ "epoch": 2.64,
349
+ "learning_rate": 0.0002763657088681219,
350
+ "loss": 0.9297,
351
+ "step": 114000
352
+ },
353
+ {
354
+ "epoch": 2.69,
355
+ "learning_rate": 0.0002758979068980519,
356
+ "loss": 0.9294,
357
+ "step": 116000
358
+ },
359
+ {
360
+ "epoch": 2.73,
361
+ "learning_rate": 0.00027543010492798184,
362
+ "loss": 0.9273,
363
+ "step": 118000
364
+ },
365
+ {
366
+ "epoch": 2.78,
367
+ "learning_rate": 0.00027496230295791186,
368
+ "loss": 0.9242,
369
+ "step": 120000
370
+ },
371
+ {
372
+ "epoch": 2.83,
373
+ "learning_rate": 0.0002744945009878418,
374
+ "loss": 0.9228,
375
+ "step": 122000
376
+ },
377
+ {
378
+ "epoch": 2.87,
379
+ "learning_rate": 0.0002740266990177718,
380
+ "loss": 0.9316,
381
+ "step": 124000
382
+ },
383
+ {
384
+ "epoch": 2.92,
385
+ "learning_rate": 0.0002735588970477017,
386
+ "loss": 0.9448,
387
+ "step": 126000
388
+ },
389
+ {
390
+ "epoch": 2.96,
391
+ "learning_rate": 0.00027309109507763174,
392
+ "loss": 0.9253,
393
+ "step": 128000
394
+ },
395
+ {
396
+ "epoch": 3.01,
397
+ "learning_rate": 0.00027262329310756165,
398
+ "loss": 0.9197,
399
+ "step": 130000
400
+ },
401
+ {
402
+ "epoch": 3.06,
403
+ "learning_rate": 0.0002721554911374917,
404
+ "loss": 0.9025,
405
+ "step": 132000
406
+ },
407
+ {
408
+ "epoch": 3.1,
409
+ "learning_rate": 0.0002716876891674216,
410
+ "loss": 0.8983,
411
+ "step": 134000
412
+ },
413
+ {
414
+ "epoch": 3.15,
415
+ "learning_rate": 0.0002712198871973516,
416
+ "loss": 0.9013,
417
+ "step": 136000
418
+ },
419
+ {
420
+ "epoch": 3.2,
421
+ "learning_rate": 0.0002707520852272816,
422
+ "loss": 0.9019,
423
+ "step": 138000
424
+ },
425
+ {
426
+ "epoch": 3.24,
427
+ "learning_rate": 0.00027028428325721155,
428
+ "loss": 0.9016,
429
+ "step": 140000
430
+ },
431
+ {
432
+ "epoch": 3.29,
433
+ "learning_rate": 0.0002698164812871415,
434
+ "loss": 0.8979,
435
+ "step": 142000
436
+ },
437
+ {
438
+ "epoch": 3.33,
439
+ "learning_rate": 0.0002693486793170715,
440
+ "loss": 0.8961,
441
+ "step": 144000
442
+ },
443
+ {
444
+ "epoch": 3.38,
445
+ "learning_rate": 0.00026888087734700146,
446
+ "loss": 0.9007,
447
+ "step": 146000
448
+ },
449
+ {
450
+ "epoch": 3.43,
451
+ "learning_rate": 0.00026841307537693143,
452
+ "loss": 0.8951,
453
+ "step": 148000
454
+ },
455
+ {
456
+ "epoch": 3.47,
457
+ "learning_rate": 0.0002679452734068614,
458
+ "loss": 0.8926,
459
+ "step": 150000
460
+ },
461
+ {
462
+ "epoch": 3.52,
463
+ "learning_rate": 0.00026747747143679137,
464
+ "loss": 0.8924,
465
+ "step": 152000
466
+ },
467
+ {
468
+ "epoch": 3.57,
469
+ "learning_rate": 0.00026700966946672134,
470
+ "loss": 0.9044,
471
+ "step": 154000
472
+ },
473
+ {
474
+ "epoch": 3.61,
475
+ "learning_rate": 0.0002665418674966513,
476
+ "loss": 0.8952,
477
+ "step": 156000
478
+ },
479
+ {
480
+ "epoch": 3.66,
481
+ "learning_rate": 0.00026607406552658127,
482
+ "loss": 0.9001,
483
+ "step": 158000
484
+ },
485
+ {
486
+ "epoch": 3.7,
487
+ "learning_rate": 0.00026560626355651124,
488
+ "loss": 0.8898,
489
+ "step": 160000
490
+ },
491
+ {
492
+ "epoch": 3.75,
493
+ "learning_rate": 0.0002651384615864412,
494
+ "loss": 0.895,
495
+ "step": 162000
496
+ },
497
+ {
498
+ "epoch": 3.8,
499
+ "learning_rate": 0.0002646706596163712,
500
+ "loss": 0.9015,
501
+ "step": 164000
502
+ },
503
+ {
504
+ "epoch": 3.84,
505
+ "learning_rate": 0.00026420285764630115,
506
+ "loss": 0.892,
507
+ "step": 166000
508
+ },
509
+ {
510
+ "epoch": 3.89,
511
+ "learning_rate": 0.0002637350556762311,
512
+ "loss": 0.8903,
513
+ "step": 168000
514
+ },
515
+ {
516
+ "epoch": 3.94,
517
+ "learning_rate": 0.0002632672537061611,
518
+ "loss": 0.8916,
519
+ "step": 170000
520
+ },
521
+ {
522
+ "epoch": 3.98,
523
+ "learning_rate": 0.00026279945173609105,
524
+ "loss": 0.8941,
525
+ "step": 172000
526
+ },
527
+ {
528
+ "epoch": 4.03,
529
+ "learning_rate": 0.000262331649766021,
530
+ "loss": 0.8771,
531
+ "step": 174000
532
+ },
533
+ {
534
+ "epoch": 4.08,
535
+ "learning_rate": 0.000261863847795951,
536
+ "loss": 0.8716,
537
+ "step": 176000
538
+ },
539
+ {
540
+ "epoch": 4.12,
541
+ "learning_rate": 0.00026139604582588096,
542
+ "loss": 0.8632,
543
+ "step": 178000
544
+ },
545
+ {
546
+ "epoch": 4.17,
547
+ "learning_rate": 0.00026092824385581093,
548
+ "loss": 0.8573,
549
+ "step": 180000
550
+ },
551
+ {
552
+ "epoch": 4.21,
553
+ "learning_rate": 0.0002604604418857409,
554
+ "loss": 0.8642,
555
+ "step": 182000
556
+ },
557
+ {
558
+ "epoch": 4.26,
559
+ "learning_rate": 0.00025999263991567087,
560
+ "loss": 0.8642,
561
+ "step": 184000
562
+ },
563
+ {
564
+ "epoch": 4.31,
565
+ "learning_rate": 0.00025952483794560084,
566
+ "loss": 0.8617,
567
+ "step": 186000
568
+ },
569
+ {
570
+ "epoch": 4.35,
571
+ "learning_rate": 0.0002590570359755308,
572
+ "loss": 0.8574,
573
+ "step": 188000
574
+ },
575
+ {
576
+ "epoch": 4.4,
577
+ "learning_rate": 0.0002585892340054608,
578
+ "loss": 0.8612,
579
+ "step": 190000
580
+ },
581
+ {
582
+ "epoch": 4.45,
583
+ "learning_rate": 0.00025812143203539074,
584
+ "loss": 0.8706,
585
+ "step": 192000
586
+ },
587
+ {
588
+ "epoch": 4.49,
589
+ "learning_rate": 0.0002576536300653207,
590
+ "loss": 0.8605,
591
+ "step": 194000
592
+ },
593
+ {
594
+ "epoch": 4.54,
595
+ "learning_rate": 0.0002571858280952507,
596
+ "loss": 0.8703,
597
+ "step": 196000
598
+ },
599
+ {
600
+ "epoch": 4.58,
601
+ "learning_rate": 0.00025671802612518065,
602
+ "loss": 0.8691,
603
+ "step": 198000
604
+ },
605
+ {
606
+ "epoch": 4.63,
607
+ "learning_rate": 0.0002562502241551106,
608
+ "loss": 0.873,
609
+ "step": 200000
610
+ },
611
+ {
612
+ "epoch": 4.68,
613
+ "learning_rate": 0.0002557824221850406,
614
+ "loss": 0.857,
615
+ "step": 202000
616
+ },
617
+ {
618
+ "epoch": 4.72,
619
+ "learning_rate": 0.00025531462021497056,
620
+ "loss": 0.8686,
621
+ "step": 204000
622
+ },
623
+ {
624
+ "epoch": 4.77,
625
+ "learning_rate": 0.0002548468182449005,
626
+ "loss": 0.868,
627
+ "step": 206000
628
+ },
629
+ {
630
+ "epoch": 4.82,
631
+ "learning_rate": 0.0002543790162748305,
632
+ "loss": 0.8756,
633
+ "step": 208000
634
+ },
635
+ {
636
+ "epoch": 4.86,
637
+ "learning_rate": 0.0002539112143047605,
638
+ "loss": 0.8656,
639
+ "step": 210000
640
+ },
641
+ {
642
+ "epoch": 4.91,
643
+ "learning_rate": 0.00025344341233469043,
644
+ "loss": 0.8725,
645
+ "step": 212000
646
+ },
647
+ {
648
+ "epoch": 4.96,
649
+ "learning_rate": 0.00025297561036462045,
650
+ "loss": 0.8675,
651
+ "step": 214000
652
+ },
653
+ {
654
+ "epoch": 5.0,
655
+ "learning_rate": 0.00025250780839455037,
656
+ "loss": 0.8644,
657
+ "step": 216000
658
+ },
659
+ {
660
+ "epoch": 5.05,
661
+ "learning_rate": 0.0002520400064244804,
662
+ "loss": 0.8397,
663
+ "step": 218000
664
+ },
665
+ {
666
+ "epoch": 5.09,
667
+ "learning_rate": 0.0002515722044544103,
668
+ "loss": 0.8374,
669
+ "step": 220000
670
+ },
671
+ {
672
+ "epoch": 5.14,
673
+ "learning_rate": 0.00025110440248434033,
674
+ "loss": 0.8364,
675
+ "step": 222000
676
+ },
677
+ {
678
+ "epoch": 5.19,
679
+ "learning_rate": 0.00025063660051427025,
680
+ "loss": 0.8454,
681
+ "step": 224000
682
+ },
683
+ {
684
+ "epoch": 5.23,
685
+ "learning_rate": 0.00025016879854420027,
686
+ "loss": 0.8367,
687
+ "step": 226000
688
+ },
689
+ {
690
+ "epoch": 5.28,
691
+ "learning_rate": 0.0002497009965741302,
692
+ "loss": 0.8275,
693
+ "step": 228000
694
+ },
695
+ {
696
+ "epoch": 5.33,
697
+ "learning_rate": 0.0002492331946040602,
698
+ "loss": 0.8398,
699
+ "step": 230000
700
+ },
701
+ {
702
+ "epoch": 5.37,
703
+ "learning_rate": 0.0002487653926339901,
704
+ "loss": 0.8403,
705
+ "step": 232000
706
+ },
707
+ {
708
+ "epoch": 5.42,
709
+ "learning_rate": 0.00024829759066392014,
710
+ "loss": 0.8409,
711
+ "step": 234000
712
+ },
713
+ {
714
+ "epoch": 5.46,
715
+ "learning_rate": 0.00024782978869385006,
716
+ "loss": 0.8366,
717
+ "step": 236000
718
+ },
719
+ {
720
+ "epoch": 5.51,
721
+ "learning_rate": 0.0002473619867237801,
722
+ "loss": 0.8381,
723
+ "step": 238000
724
+ },
725
+ {
726
+ "epoch": 5.56,
727
+ "learning_rate": 0.00024689418475371,
728
+ "loss": 0.842,
729
+ "step": 240000
730
+ },
731
+ {
732
+ "epoch": 5.6,
733
+ "learning_rate": 0.00024642638278364,
734
+ "loss": 0.843,
735
+ "step": 242000
736
+ },
737
+ {
738
+ "epoch": 5.65,
739
+ "learning_rate": 0.00024595858081357,
740
+ "loss": 0.8489,
741
+ "step": 244000
742
+ },
743
+ {
744
+ "epoch": 5.7,
745
+ "learning_rate": 0.00024549077884349996,
746
+ "loss": 0.8313,
747
+ "step": 246000
748
+ },
749
+ {
750
+ "epoch": 5.74,
751
+ "learning_rate": 0.0002450229768734299,
752
+ "loss": 0.8468,
753
+ "step": 248000
754
+ },
755
+ {
756
+ "epoch": 5.79,
757
+ "learning_rate": 0.0002445551749033599,
758
+ "loss": 0.8446,
759
+ "step": 250000
760
+ },
761
+ {
762
+ "epoch": 5.84,
763
+ "learning_rate": 0.00024408737293328986,
764
+ "loss": 0.8329,
765
+ "step": 252000
766
+ },
767
+ {
768
+ "epoch": 5.88,
769
+ "learning_rate": 0.00024361957096321983,
770
+ "loss": 0.8402,
771
+ "step": 254000
772
+ },
773
+ {
774
+ "epoch": 5.93,
775
+ "learning_rate": 0.0002431517689931498,
776
+ "loss": 0.8502,
777
+ "step": 256000
778
+ },
779
+ {
780
+ "epoch": 5.97,
781
+ "learning_rate": 0.00024268396702307977,
782
+ "loss": 0.8437,
783
+ "step": 258000
784
+ },
785
+ {
786
+ "epoch": 6.02,
787
+ "learning_rate": 0.00024221616505300974,
788
+ "loss": 0.8341,
789
+ "step": 260000
790
+ },
791
+ {
792
+ "epoch": 6.07,
793
+ "learning_rate": 0.0002417483630829397,
794
+ "loss": 0.8152,
795
+ "step": 262000
796
+ },
797
+ {
798
+ "epoch": 6.11,
799
+ "learning_rate": 0.00024128056111286968,
800
+ "loss": 0.8189,
801
+ "step": 264000
802
+ },
803
+ {
804
+ "epoch": 6.16,
805
+ "learning_rate": 0.00024081275914279965,
806
+ "loss": 0.8193,
807
+ "step": 266000
808
+ },
809
+ {
810
+ "epoch": 6.21,
811
+ "learning_rate": 0.00024034495717272961,
812
+ "loss": 0.825,
813
+ "step": 268000
814
+ },
815
+ {
816
+ "epoch": 6.25,
817
+ "learning_rate": 0.00023987715520265958,
818
+ "loss": 0.818,
819
+ "step": 270000
820
+ },
821
+ {
822
+ "epoch": 6.3,
823
+ "learning_rate": 0.00023940935323258955,
824
+ "loss": 0.8204,
825
+ "step": 272000
826
+ },
827
+ {
828
+ "epoch": 6.34,
829
+ "learning_rate": 0.00023894155126251952,
830
+ "loss": 0.823,
831
+ "step": 274000
832
+ },
833
+ {
834
+ "epoch": 6.39,
835
+ "learning_rate": 0.0002384737492924495,
836
+ "loss": 0.8179,
837
+ "step": 276000
838
+ },
839
+ {
840
+ "epoch": 6.44,
841
+ "learning_rate": 0.00023800594732237946,
842
+ "loss": 0.8152,
843
+ "step": 278000
844
+ },
845
+ {
846
+ "epoch": 6.48,
847
+ "learning_rate": 0.00023753814535230943,
848
+ "loss": 0.8178,
849
+ "step": 280000
850
+ },
851
+ {
852
+ "epoch": 6.53,
853
+ "learning_rate": 0.00023707034338223942,
854
+ "loss": 0.8212,
855
+ "step": 282000
856
+ },
857
+ {
858
+ "epoch": 6.58,
859
+ "learning_rate": 0.00023660254141216937,
860
+ "loss": 0.8161,
861
+ "step": 284000
862
+ },
863
+ {
864
+ "epoch": 6.62,
865
+ "learning_rate": 0.00023613473944209936,
866
+ "loss": 0.8162,
867
+ "step": 286000
868
+ },
869
+ {
870
+ "epoch": 6.67,
871
+ "learning_rate": 0.0002356669374720293,
872
+ "loss": 0.828,
873
+ "step": 288000
874
+ },
875
+ {
876
+ "epoch": 6.72,
877
+ "learning_rate": 0.0002351991355019593,
878
+ "loss": 0.823,
879
+ "step": 290000
880
+ },
881
+ {
882
+ "epoch": 6.76,
883
+ "learning_rate": 0.00023473133353188927,
884
+ "loss": 0.8156,
885
+ "step": 292000
886
+ },
887
+ {
888
+ "epoch": 6.81,
889
+ "learning_rate": 0.00023426353156181924,
890
+ "loss": 0.8176,
891
+ "step": 294000
892
+ },
893
+ {
894
+ "epoch": 6.85,
895
+ "learning_rate": 0.0002337957295917492,
896
+ "loss": 0.8226,
897
+ "step": 296000
898
+ },
899
+ {
900
+ "epoch": 6.9,
901
+ "learning_rate": 0.00023332792762167917,
902
+ "loss": 0.8189,
903
+ "step": 298000
904
+ },
905
+ {
906
+ "epoch": 6.95,
907
+ "learning_rate": 0.00023286012565160914,
908
+ "loss": 0.8082,
909
+ "step": 300000
910
+ },
911
+ {
912
+ "epoch": 6.99,
913
+ "learning_rate": 0.0002323923236815391,
914
+ "loss": 0.8257,
915
+ "step": 302000
916
+ },
917
+ {
918
+ "epoch": 7.04,
919
+ "learning_rate": 0.00023192452171146908,
920
+ "loss": 0.7925,
921
+ "step": 304000
922
+ },
923
+ {
924
+ "epoch": 7.09,
925
+ "learning_rate": 0.00023145671974139905,
926
+ "loss": 0.7902,
927
+ "step": 306000
928
+ },
929
+ {
930
+ "epoch": 7.13,
931
+ "learning_rate": 0.00023098891777132902,
932
+ "loss": 0.7994,
933
+ "step": 308000
934
+ },
935
+ {
936
+ "epoch": 7.18,
937
+ "learning_rate": 0.000230521115801259,
938
+ "loss": 0.8029,
939
+ "step": 310000
940
+ },
941
+ {
942
+ "epoch": 7.22,
943
+ "learning_rate": 0.00023005331383118896,
944
+ "loss": 0.7983,
945
+ "step": 312000
946
+ },
947
+ {
948
+ "epoch": 7.27,
949
+ "learning_rate": 0.00022958551186111893,
950
+ "loss": 0.797,
951
+ "step": 314000
952
+ },
953
+ {
954
+ "epoch": 7.32,
955
+ "learning_rate": 0.00022911770989104892,
956
+ "loss": 0.8009,
957
+ "step": 316000
958
+ },
959
+ {
960
+ "epoch": 7.36,
961
+ "learning_rate": 0.00022864990792097886,
962
+ "loss": 0.7979,
963
+ "step": 318000
964
+ },
965
+ {
966
+ "epoch": 7.41,
967
+ "learning_rate": 0.00022818210595090886,
968
+ "loss": 0.8055,
969
+ "step": 320000
970
+ },
971
+ {
972
+ "epoch": 7.46,
973
+ "learning_rate": 0.0002277143039808388,
974
+ "loss": 0.7971,
975
+ "step": 322000
976
+ },
977
+ {
978
+ "epoch": 7.5,
979
+ "learning_rate": 0.0002272465020107688,
980
+ "loss": 0.8066,
981
+ "step": 324000
982
+ },
983
+ {
984
+ "epoch": 7.55,
985
+ "learning_rate": 0.00022677870004069874,
986
+ "loss": 0.7975,
987
+ "step": 326000
988
+ },
989
+ {
990
+ "epoch": 7.6,
991
+ "learning_rate": 0.00022631089807062873,
992
+ "loss": 0.7912,
993
+ "step": 328000
994
+ },
995
+ {
996
+ "epoch": 7.64,
997
+ "learning_rate": 0.00022584309610055868,
998
+ "loss": 0.7988,
999
+ "step": 330000
1000
+ },
1001
+ {
1002
+ "epoch": 7.69,
1003
+ "learning_rate": 0.00022537529413048867,
1004
+ "loss": 0.7999,
1005
+ "step": 332000
1006
+ },
1007
+ {
1008
+ "epoch": 7.73,
1009
+ "learning_rate": 0.00022490749216041861,
1010
+ "loss": 0.8019,
1011
+ "step": 334000
1012
+ },
1013
+ {
1014
+ "epoch": 7.78,
1015
+ "learning_rate": 0.0002244396901903486,
1016
+ "loss": 0.8108,
1017
+ "step": 336000
1018
+ },
1019
+ {
1020
+ "epoch": 7.83,
1021
+ "learning_rate": 0.00022397188822027855,
1022
+ "loss": 0.8075,
1023
+ "step": 338000
1024
+ },
1025
+ {
1026
+ "epoch": 7.87,
1027
+ "learning_rate": 0.00022350408625020855,
1028
+ "loss": 0.7995,
1029
+ "step": 340000
1030
+ },
1031
+ {
1032
+ "epoch": 7.92,
1033
+ "learning_rate": 0.0002230362842801385,
1034
+ "loss": 0.802,
1035
+ "step": 342000
1036
+ },
1037
+ {
1038
+ "epoch": 7.97,
1039
+ "learning_rate": 0.00022256848231006848,
1040
+ "loss": 0.7998,
1041
+ "step": 344000
1042
+ },
1043
+ {
1044
+ "epoch": 8.01,
1045
+ "learning_rate": 0.00022210068033999843,
1046
+ "loss": 0.8023,
1047
+ "step": 346000
1048
+ },
1049
+ {
1050
+ "epoch": 8.06,
1051
+ "learning_rate": 0.00022163287836992842,
1052
+ "loss": 0.7748,
1053
+ "step": 348000
1054
+ },
1055
+ {
1056
+ "epoch": 8.1,
1057
+ "learning_rate": 0.00022116507639985836,
1058
+ "loss": 0.7839,
1059
+ "step": 350000
1060
+ },
1061
+ {
1062
+ "epoch": 8.15,
1063
+ "learning_rate": 0.00022069727442978836,
1064
+ "loss": 0.7806,
1065
+ "step": 352000
1066
+ },
1067
+ {
1068
+ "epoch": 8.2,
1069
+ "learning_rate": 0.00022022947245971836,
1070
+ "loss": 0.7775,
1071
+ "step": 354000
1072
+ },
1073
+ {
1074
+ "epoch": 8.24,
1075
+ "learning_rate": 0.0002197616704896483,
1076
+ "loss": 0.7734,
1077
+ "step": 356000
1078
+ },
1079
+ {
1080
+ "epoch": 8.29,
1081
+ "learning_rate": 0.0002192938685195783,
1082
+ "loss": 0.7728,
1083
+ "step": 358000
1084
+ },
1085
+ {
1086
+ "epoch": 8.34,
1087
+ "learning_rate": 0.00021882606654950824,
1088
+ "loss": 0.7879,
1089
+ "step": 360000
1090
+ },
1091
+ {
1092
+ "epoch": 8.38,
1093
+ "learning_rate": 0.00021835826457943823,
1094
+ "loss": 0.7891,
1095
+ "step": 362000
1096
+ },
1097
+ {
1098
+ "epoch": 8.43,
1099
+ "learning_rate": 0.00021789046260936817,
1100
+ "loss": 0.7922,
1101
+ "step": 364000
1102
+ },
1103
+ {
1104
+ "epoch": 8.48,
1105
+ "learning_rate": 0.00021742266063929817,
1106
+ "loss": 0.7837,
1107
+ "step": 366000
1108
+ },
1109
+ {
1110
+ "epoch": 8.52,
1111
+ "learning_rate": 0.0002169548586692281,
1112
+ "loss": 0.7838,
1113
+ "step": 368000
1114
+ },
1115
+ {
1116
+ "epoch": 8.57,
1117
+ "learning_rate": 0.0002164870566991581,
1118
+ "loss": 0.7797,
1119
+ "step": 370000
1120
+ },
1121
+ {
1122
+ "epoch": 8.61,
1123
+ "learning_rate": 0.00021601925472908805,
1124
+ "loss": 0.7818,
1125
+ "step": 372000
1126
+ },
1127
+ {
1128
+ "epoch": 8.66,
1129
+ "learning_rate": 0.00021555145275901804,
1130
+ "loss": 0.7838,
1131
+ "step": 374000
1132
+ },
1133
+ {
1134
+ "epoch": 8.71,
1135
+ "learning_rate": 0.000215083650788948,
1136
+ "loss": 0.7828,
1137
+ "step": 376000
1138
+ },
1139
+ {
1140
+ "epoch": 8.75,
1141
+ "learning_rate": 0.00021461584881887798,
1142
+ "loss": 0.7822,
1143
+ "step": 378000
1144
+ },
1145
+ {
1146
+ "epoch": 8.8,
1147
+ "learning_rate": 0.00021414804684880792,
1148
+ "loss": 0.7952,
1149
+ "step": 380000
1150
+ },
1151
+ {
1152
+ "epoch": 8.85,
1153
+ "learning_rate": 0.00021368024487873792,
1154
+ "loss": 0.7888,
1155
+ "step": 382000
1156
+ },
1157
+ {
1158
+ "epoch": 8.89,
1159
+ "learning_rate": 0.00021321244290866786,
1160
+ "loss": 0.7813,
1161
+ "step": 384000
1162
+ },
1163
+ {
1164
+ "epoch": 8.94,
1165
+ "learning_rate": 0.00021274464093859786,
1166
+ "loss": 0.7784,
1167
+ "step": 386000
1168
+ },
1169
+ {
1170
+ "epoch": 8.98,
1171
+ "learning_rate": 0.00021227683896852783,
1172
+ "loss": 0.776,
1173
+ "step": 388000
1174
+ },
1175
+ {
1176
+ "epoch": 9.03,
1177
+ "learning_rate": 0.0002118090369984578,
1178
+ "loss": 0.7686,
1179
+ "step": 390000
1180
+ },
1181
+ {
1182
+ "epoch": 9.08,
1183
+ "learning_rate": 0.0002113412350283878,
1184
+ "loss": 0.7576,
1185
+ "step": 392000
1186
+ },
1187
+ {
1188
+ "epoch": 9.12,
1189
+ "learning_rate": 0.00021087343305831773,
1190
+ "loss": 0.7611,
1191
+ "step": 394000
1192
+ },
1193
+ {
1194
+ "epoch": 9.17,
1195
+ "learning_rate": 0.00021040563108824773,
1196
+ "loss": 0.7698,
1197
+ "step": 396000
1198
+ },
1199
+ {
1200
+ "epoch": 9.22,
1201
+ "learning_rate": 0.00020993782911817767,
1202
+ "loss": 0.7632,
1203
+ "step": 398000
1204
+ },
1205
+ {
1206
+ "epoch": 9.26,
1207
+ "learning_rate": 0.00020947002714810767,
1208
+ "loss": 0.7725,
1209
+ "step": 400000
1210
+ },
1211
+ {
1212
+ "epoch": 9.31,
1213
+ "learning_rate": 0.0002090022251780376,
1214
+ "loss": 0.7706,
1215
+ "step": 402000
1216
+ },
1217
+ {
1218
+ "epoch": 9.35,
1219
+ "learning_rate": 0.0002085344232079676,
1220
+ "loss": 0.7709,
1221
+ "step": 404000
1222
+ },
1223
+ {
1224
+ "epoch": 9.4,
1225
+ "learning_rate": 0.00020806662123789755,
1226
+ "loss": 0.7651,
1227
+ "step": 406000
1228
+ },
1229
+ {
1230
+ "epoch": 9.45,
1231
+ "learning_rate": 0.00020759881926782754,
1232
+ "loss": 0.7657,
1233
+ "step": 408000
1234
+ },
1235
+ {
1236
+ "epoch": 9.49,
1237
+ "learning_rate": 0.00020713101729775748,
1238
+ "loss": 0.7589,
1239
+ "step": 410000
1240
+ },
1241
+ {
1242
+ "epoch": 9.54,
1243
+ "learning_rate": 0.00020666321532768748,
1244
+ "loss": 0.7683,
1245
+ "step": 412000
1246
+ },
1247
+ {
1248
+ "epoch": 9.59,
1249
+ "learning_rate": 0.00020619541335761742,
1250
+ "loss": 0.7684,
1251
+ "step": 414000
1252
+ },
1253
+ {
1254
+ "epoch": 9.63,
1255
+ "learning_rate": 0.00020572761138754742,
1256
+ "loss": 0.7756,
1257
+ "step": 416000
1258
+ },
1259
+ {
1260
+ "epoch": 9.68,
1261
+ "learning_rate": 0.00020525980941747736,
1262
+ "loss": 0.7653,
1263
+ "step": 418000
1264
+ },
1265
+ {
1266
+ "epoch": 9.73,
1267
+ "learning_rate": 0.00020479200744740736,
1268
+ "loss": 0.7718,
1269
+ "step": 420000
1270
+ },
1271
+ {
1272
+ "epoch": 9.77,
1273
+ "learning_rate": 0.0002043242054773373,
1274
+ "loss": 0.7676,
1275
+ "step": 422000
1276
+ },
1277
+ {
1278
+ "epoch": 9.82,
1279
+ "learning_rate": 0.0002038564035072673,
1280
+ "loss": 0.772,
1281
+ "step": 424000
1282
+ },
1283
+ {
1284
+ "epoch": 9.86,
1285
+ "learning_rate": 0.00020338860153719726,
1286
+ "loss": 0.766,
1287
+ "step": 426000
1288
+ },
1289
+ {
1290
+ "epoch": 9.91,
1291
+ "learning_rate": 0.00020292079956712723,
1292
+ "loss": 0.7739,
1293
+ "step": 428000
1294
+ },
1295
+ {
1296
+ "epoch": 9.96,
1297
+ "learning_rate": 0.0002024529975970572,
1298
+ "loss": 0.7743,
1299
+ "step": 430000
1300
+ },
1301
+ {
1302
+ "epoch": 10.0,
1303
+ "learning_rate": 0.00020198519562698717,
1304
+ "loss": 0.7719,
1305
+ "step": 432000
1306
+ },
1307
+ {
1308
+ "epoch": 10.05,
1309
+ "learning_rate": 0.00020151739365691714,
1310
+ "loss": 0.7443,
1311
+ "step": 434000
1312
+ },
1313
+ {
1314
+ "epoch": 10.1,
1315
+ "learning_rate": 0.0002010495916868471,
1316
+ "loss": 0.7573,
1317
+ "step": 436000
1318
+ },
1319
+ {
1320
+ "epoch": 10.14,
1321
+ "learning_rate": 0.00020058178971677708,
1322
+ "loss": 0.7546,
1323
+ "step": 438000
1324
+ },
1325
+ {
1326
+ "epoch": 10.19,
1327
+ "learning_rate": 0.00020011398774670704,
1328
+ "loss": 0.7516,
1329
+ "step": 440000
1330
+ },
1331
+ {
1332
+ "epoch": 10.23,
1333
+ "learning_rate": 0.000199646185776637,
1334
+ "loss": 0.7444,
1335
+ "step": 442000
1336
+ },
1337
+ {
1338
+ "epoch": 10.28,
1339
+ "learning_rate": 0.00019917838380656698,
1340
+ "loss": 0.7656,
1341
+ "step": 444000
1342
+ },
1343
+ {
1344
+ "epoch": 10.33,
1345
+ "learning_rate": 0.00019871058183649695,
1346
+ "loss": 0.7452,
1347
+ "step": 446000
1348
+ },
1349
+ {
1350
+ "epoch": 10.37,
1351
+ "learning_rate": 0.00019824277986642692,
1352
+ "loss": 0.7555,
1353
+ "step": 448000
1354
+ },
1355
+ {
1356
+ "epoch": 10.42,
1357
+ "learning_rate": 0.0001977749778963569,
1358
+ "loss": 0.7486,
1359
+ "step": 450000
1360
+ },
1361
+ {
1362
+ "epoch": 10.47,
1363
+ "learning_rate": 0.00019730717592628686,
1364
+ "loss": 0.7509,
1365
+ "step": 452000
1366
+ },
1367
+ {
1368
+ "epoch": 10.51,
1369
+ "learning_rate": 0.00019683937395621683,
1370
+ "loss": 0.7484,
1371
+ "step": 454000
1372
+ },
1373
+ {
1374
+ "epoch": 10.56,
1375
+ "learning_rate": 0.0001963715719861468,
1376
+ "loss": 0.7554,
1377
+ "step": 456000
1378
+ },
1379
+ {
1380
+ "epoch": 10.61,
1381
+ "learning_rate": 0.00019590377001607676,
1382
+ "loss": 0.7557,
1383
+ "step": 458000
1384
+ },
1385
+ {
1386
+ "epoch": 10.65,
1387
+ "learning_rate": 0.00019543596804600676,
1388
+ "loss": 0.7603,
1389
+ "step": 460000
1390
+ },
1391
+ {
1392
+ "epoch": 10.7,
1393
+ "learning_rate": 0.00019496816607593673,
1394
+ "loss": 0.7577,
1395
+ "step": 462000
1396
+ },
1397
+ {
1398
+ "epoch": 10.74,
1399
+ "learning_rate": 0.0001945003641058667,
1400
+ "loss": 0.7641,
1401
+ "step": 464000
1402
+ },
1403
+ {
1404
+ "epoch": 10.79,
1405
+ "learning_rate": 0.00019403256213579667,
1406
+ "loss": 0.7648,
1407
+ "step": 466000
1408
+ },
1409
+ {
1410
+ "epoch": 10.84,
1411
+ "learning_rate": 0.00019356476016572664,
1412
+ "loss": 0.755,
1413
+ "step": 468000
1414
+ },
1415
+ {
1416
+ "epoch": 10.88,
1417
+ "learning_rate": 0.0001930969581956566,
1418
+ "loss": 0.7445,
1419
+ "step": 470000
1420
+ },
1421
+ {
1422
+ "epoch": 10.93,
1423
+ "learning_rate": 0.00019262915622558657,
1424
+ "loss": 0.7614,
1425
+ "step": 472000
1426
+ },
1427
+ {
1428
+ "epoch": 10.98,
1429
+ "learning_rate": 0.00019216135425551654,
1430
+ "loss": 0.7526,
1431
+ "step": 474000
1432
+ },
1433
+ {
1434
+ "epoch": 11.02,
1435
+ "learning_rate": 0.0001916935522854465,
1436
+ "loss": 0.7493,
1437
+ "step": 476000
1438
+ },
1439
+ {
1440
+ "epoch": 11.07,
1441
+ "learning_rate": 0.00019122575031537648,
1442
+ "loss": 0.7299,
1443
+ "step": 478000
1444
+ },
1445
+ {
1446
+ "epoch": 11.11,
1447
+ "learning_rate": 0.00019075794834530645,
1448
+ "loss": 0.7379,
1449
+ "step": 480000
1450
+ },
1451
+ {
1452
+ "epoch": 11.16,
1453
+ "learning_rate": 0.00019029014637523642,
1454
+ "loss": 0.7365,
1455
+ "step": 482000
1456
+ },
1457
+ {
1458
+ "epoch": 11.21,
1459
+ "learning_rate": 0.00018982234440516639,
1460
+ "loss": 0.7402,
1461
+ "step": 484000
1462
+ },
1463
+ {
1464
+ "epoch": 11.25,
1465
+ "learning_rate": 0.00018935454243509636,
1466
+ "loss": 0.7409,
1467
+ "step": 486000
1468
+ },
1469
+ {
1470
+ "epoch": 11.3,
1471
+ "learning_rate": 0.00018888674046502632,
1472
+ "loss": 0.7294,
1473
+ "step": 488000
1474
+ },
1475
+ {
1476
+ "epoch": 11.35,
1477
+ "learning_rate": 0.0001884189384949563,
1478
+ "loss": 0.7467,
1479
+ "step": 490000
1480
+ },
1481
+ {
1482
+ "epoch": 11.39,
1483
+ "learning_rate": 0.00018795113652488626,
1484
+ "loss": 0.7357,
1485
+ "step": 492000
1486
+ },
1487
+ {
1488
+ "epoch": 11.44,
1489
+ "learning_rate": 0.00018748333455481623,
1490
+ "loss": 0.744,
1491
+ "step": 494000
1492
+ },
1493
+ {
1494
+ "epoch": 11.49,
1495
+ "learning_rate": 0.0001870155325847462,
1496
+ "loss": 0.741,
1497
+ "step": 496000
1498
+ },
1499
+ {
1500
+ "epoch": 11.53,
1501
+ "learning_rate": 0.0001865477306146762,
1502
+ "loss": 0.7404,
1503
+ "step": 498000
1504
+ },
1505
+ {
1506
+ "epoch": 11.58,
1507
+ "learning_rate": 0.00018607992864460614,
1508
+ "loss": 0.749,
1509
+ "step": 500000
1510
+ },
1511
+ {
1512
+ "epoch": 11.62,
1513
+ "learning_rate": 0.00018561212667453613,
1514
+ "loss": 0.7388,
1515
+ "step": 502000
1516
+ },
1517
+ {
1518
+ "epoch": 11.67,
1519
+ "learning_rate": 0.00018514432470446607,
1520
+ "loss": 0.742,
1521
+ "step": 504000
1522
+ },
1523
+ {
1524
+ "epoch": 11.72,
1525
+ "learning_rate": 0.00018467652273439607,
1526
+ "loss": 0.7481,
1527
+ "step": 506000
1528
+ },
1529
+ {
1530
+ "epoch": 11.76,
1531
+ "learning_rate": 0.000184208720764326,
1532
+ "loss": 0.7553,
1533
+ "step": 508000
1534
+ },
1535
+ {
1536
+ "epoch": 11.81,
1537
+ "learning_rate": 0.000183740918794256,
1538
+ "loss": 0.7457,
1539
+ "step": 510000
1540
+ },
1541
+ {
1542
+ "epoch": 11.86,
1543
+ "learning_rate": 0.00018327311682418595,
1544
+ "loss": 0.7447,
1545
+ "step": 512000
1546
+ },
1547
+ {
1548
+ "epoch": 11.9,
1549
+ "learning_rate": 0.00018280531485411595,
1550
+ "loss": 0.752,
1551
+ "step": 514000
1552
+ },
1553
+ {
1554
+ "epoch": 11.95,
1555
+ "learning_rate": 0.0001823375128840459,
1556
+ "loss": 0.7419,
1557
+ "step": 516000
1558
+ },
1559
+ {
1560
+ "epoch": 11.99,
1561
+ "learning_rate": 0.00018186971091397588,
1562
+ "loss": 0.7412,
1563
+ "step": 518000
1564
+ },
1565
+ {
1566
+ "epoch": 12.04,
1567
+ "learning_rate": 0.00018140190894390583,
1568
+ "loss": 0.7262,
1569
+ "step": 520000
1570
+ },
1571
+ {
1572
+ "epoch": 12.09,
1573
+ "learning_rate": 0.00018093410697383582,
1574
+ "loss": 0.7299,
1575
+ "step": 522000
1576
+ },
1577
+ {
1578
+ "epoch": 12.13,
1579
+ "learning_rate": 0.00018046630500376576,
1580
+ "loss": 0.7279,
1581
+ "step": 524000
1582
+ },
1583
+ {
1584
+ "epoch": 12.18,
1585
+ "learning_rate": 0.00017999850303369576,
1586
+ "loss": 0.7308,
1587
+ "step": 526000
1588
+ },
1589
+ {
1590
+ "epoch": 12.23,
1591
+ "learning_rate": 0.0001795307010636257,
1592
+ "loss": 0.7305,
1593
+ "step": 528000
1594
+ },
1595
+ {
1596
+ "epoch": 12.27,
1597
+ "learning_rate": 0.0001790628990935557,
1598
+ "loss": 0.7348,
1599
+ "step": 530000
1600
+ },
1601
+ {
1602
+ "epoch": 12.32,
1603
+ "learning_rate": 0.0001785950971234857,
1604
+ "loss": 0.7312,
1605
+ "step": 532000
1606
+ },
1607
+ {
1608
+ "epoch": 12.37,
1609
+ "learning_rate": 0.00017812729515341563,
1610
+ "loss": 0.7275,
1611
+ "step": 534000
1612
+ },
1613
+ {
1614
+ "epoch": 12.41,
1615
+ "learning_rate": 0.00017765949318334563,
1616
+ "loss": 0.7291,
1617
+ "step": 536000
1618
+ },
1619
+ {
1620
+ "epoch": 12.46,
1621
+ "learning_rate": 0.00017719169121327557,
1622
+ "loss": 0.7265,
1623
+ "step": 538000
1624
+ },
1625
+ {
1626
+ "epoch": 12.5,
1627
+ "learning_rate": 0.00017672388924320557,
1628
+ "loss": 0.7224,
1629
+ "step": 540000
1630
+ },
1631
+ {
1632
+ "epoch": 12.55,
1633
+ "learning_rate": 0.0001762560872731355,
1634
+ "loss": 0.7232,
1635
+ "step": 542000
1636
+ },
1637
+ {
1638
+ "epoch": 12.6,
1639
+ "learning_rate": 0.0001757882853030655,
1640
+ "loss": 0.7272,
1641
+ "step": 544000
1642
+ },
1643
+ {
1644
+ "epoch": 12.64,
1645
+ "learning_rate": 0.00017532048333299545,
1646
+ "loss": 0.7305,
1647
+ "step": 546000
1648
+ },
1649
+ {
1650
+ "epoch": 12.69,
1651
+ "learning_rate": 0.00017485268136292544,
1652
+ "loss": 0.7375,
1653
+ "step": 548000
1654
+ },
1655
+ {
1656
+ "epoch": 12.74,
1657
+ "learning_rate": 0.00017438487939285539,
1658
+ "loss": 0.7377,
1659
+ "step": 550000
1660
+ },
1661
+ {
1662
+ "epoch": 12.78,
1663
+ "learning_rate": 0.00017391707742278538,
1664
+ "loss": 0.7278,
1665
+ "step": 552000
1666
+ },
1667
+ {
1668
+ "epoch": 12.83,
1669
+ "learning_rate": 0.00017344927545271532,
1670
+ "loss": 0.7369,
1671
+ "step": 554000
1672
+ },
1673
+ {
1674
+ "epoch": 12.87,
1675
+ "learning_rate": 0.00017298147348264532,
1676
+ "loss": 0.7366,
1677
+ "step": 556000
1678
+ },
1679
+ {
1680
+ "epoch": 12.92,
1681
+ "learning_rate": 0.00017251367151257526,
1682
+ "loss": 0.736,
1683
+ "step": 558000
1684
+ },
1685
+ {
1686
+ "epoch": 12.97,
1687
+ "learning_rate": 0.00017204586954250526,
1688
+ "loss": 0.737,
1689
+ "step": 560000
1690
+ },
1691
+ {
1692
+ "epoch": 13.01,
1693
+ "learning_rate": 0.0001715780675724352,
1694
+ "loss": 0.7301,
1695
+ "step": 562000
1696
+ },
1697
+ {
1698
+ "epoch": 13.06,
1699
+ "learning_rate": 0.0001711102656023652,
1700
+ "loss": 0.7003,
1701
+ "step": 564000
1702
+ },
1703
+ {
1704
+ "epoch": 13.11,
1705
+ "learning_rate": 0.00017064246363229514,
1706
+ "loss": 0.7132,
1707
+ "step": 566000
1708
+ },
1709
+ {
1710
+ "epoch": 13.15,
1711
+ "learning_rate": 0.00017017466166222513,
1712
+ "loss": 0.7178,
1713
+ "step": 568000
1714
+ },
1715
+ {
1716
+ "epoch": 13.2,
1717
+ "learning_rate": 0.00016970685969215513,
1718
+ "loss": 0.7187,
1719
+ "step": 570000
1720
+ },
1721
+ {
1722
+ "epoch": 13.25,
1723
+ "learning_rate": 0.00016923905772208507,
1724
+ "loss": 0.7239,
1725
+ "step": 572000
1726
+ },
1727
+ {
1728
+ "epoch": 13.29,
1729
+ "learning_rate": 0.00016877125575201507,
1730
+ "loss": 0.7271,
1731
+ "step": 574000
1732
+ },
1733
+ {
1734
+ "epoch": 13.34,
1735
+ "learning_rate": 0.000168303453781945,
1736
+ "loss": 0.7208,
1737
+ "step": 576000
1738
+ },
1739
+ {
1740
+ "epoch": 13.38,
1741
+ "learning_rate": 0.000167835651811875,
1742
+ "loss": 0.7199,
1743
+ "step": 578000
1744
+ },
1745
+ {
1746
+ "epoch": 13.43,
1747
+ "learning_rate": 0.00016736784984180495,
1748
+ "loss": 0.7094,
1749
+ "step": 580000
1750
+ },
1751
+ {
1752
+ "epoch": 13.48,
1753
+ "learning_rate": 0.00016690004787173494,
1754
+ "loss": 0.7114,
1755
+ "step": 582000
1756
+ },
1757
+ {
1758
+ "epoch": 13.52,
1759
+ "learning_rate": 0.00016643224590166488,
1760
+ "loss": 0.7196,
1761
+ "step": 584000
1762
+ },
1763
+ {
1764
+ "epoch": 13.57,
1765
+ "learning_rate": 0.00016596444393159488,
1766
+ "loss": 0.7222,
1767
+ "step": 586000
1768
+ },
1769
+ {
1770
+ "epoch": 13.62,
1771
+ "learning_rate": 0.00016549664196152482,
1772
+ "loss": 0.7345,
1773
+ "step": 588000
1774
+ },
1775
+ {
1776
+ "epoch": 13.66,
1777
+ "learning_rate": 0.00016502883999145482,
1778
+ "loss": 0.7208,
1779
+ "step": 590000
1780
+ },
1781
+ {
1782
+ "epoch": 13.71,
1783
+ "learning_rate": 0.00016456103802138476,
1784
+ "loss": 0.7298,
1785
+ "step": 592000
1786
+ },
1787
+ {
1788
+ "epoch": 13.75,
1789
+ "learning_rate": 0.00016409323605131475,
1790
+ "loss": 0.7324,
1791
+ "step": 594000
1792
+ },
1793
+ {
1794
+ "epoch": 13.8,
1795
+ "learning_rate": 0.0001636254340812447,
1796
+ "loss": 0.7243,
1797
+ "step": 596000
1798
+ },
1799
+ {
1800
+ "epoch": 13.85,
1801
+ "learning_rate": 0.0001631576321111747,
1802
+ "loss": 0.7215,
1803
+ "step": 598000
1804
+ },
1805
+ {
1806
+ "epoch": 13.89,
1807
+ "learning_rate": 0.00016268983014110463,
1808
+ "loss": 0.7246,
1809
+ "step": 600000
1810
+ },
1811
+ {
1812
+ "epoch": 13.94,
1813
+ "learning_rate": 0.00016222202817103463,
1814
+ "loss": 0.7219,
1815
+ "step": 602000
1816
+ },
1817
+ {
1818
+ "epoch": 13.99,
1819
+ "learning_rate": 0.0001617542262009646,
1820
+ "loss": 0.7248,
1821
+ "step": 604000
1822
+ },
1823
+ {
1824
+ "epoch": 14.03,
1825
+ "learning_rate": 0.00016128642423089457,
1826
+ "loss": 0.7139,
1827
+ "step": 606000
1828
+ },
1829
+ {
1830
+ "epoch": 14.08,
1831
+ "learning_rate": 0.00016081862226082454,
1832
+ "loss": 0.7026,
1833
+ "step": 608000
1834
+ },
1835
+ {
1836
+ "epoch": 14.13,
1837
+ "learning_rate": 0.0001603508202907545,
1838
+ "loss": 0.7107,
1839
+ "step": 610000
1840
+ },
1841
+ {
1842
+ "epoch": 14.17,
1843
+ "learning_rate": 0.00015988301832068447,
1844
+ "loss": 0.7037,
1845
+ "step": 612000
1846
+ },
1847
+ {
1848
+ "epoch": 14.22,
1849
+ "learning_rate": 0.00015941521635061444,
1850
+ "loss": 0.7145,
1851
+ "step": 614000
1852
+ },
1853
+ {
1854
+ "epoch": 14.26,
1855
+ "learning_rate": 0.0001589474143805444,
1856
+ "loss": 0.7181,
1857
+ "step": 616000
1858
+ },
1859
+ {
1860
+ "epoch": 14.31,
1861
+ "learning_rate": 0.00015847961241047438,
1862
+ "loss": 0.7026,
1863
+ "step": 618000
1864
+ },
1865
+ {
1866
+ "epoch": 14.36,
1867
+ "learning_rate": 0.00015801181044040435,
1868
+ "loss": 0.7142,
1869
+ "step": 620000
1870
+ },
1871
+ {
1872
+ "epoch": 14.4,
1873
+ "learning_rate": 0.00015754400847033432,
1874
+ "loss": 0.7087,
1875
+ "step": 622000
1876
+ },
1877
+ {
1878
+ "epoch": 14.45,
1879
+ "learning_rate": 0.0001570762065002643,
1880
+ "loss": 0.7109,
1881
+ "step": 624000
1882
+ },
1883
+ {
1884
+ "epoch": 14.5,
1885
+ "learning_rate": 0.00015660840453019426,
1886
+ "loss": 0.7031,
1887
+ "step": 626000
1888
+ },
1889
+ {
1890
+ "epoch": 14.54,
1891
+ "learning_rate": 0.00015614060256012425,
1892
+ "loss": 0.7101,
1893
+ "step": 628000
1894
+ },
1895
+ {
1896
+ "epoch": 14.59,
1897
+ "learning_rate": 0.0001556728005900542,
1898
+ "loss": 0.7152,
1899
+ "step": 630000
1900
+ },
1901
+ {
1902
+ "epoch": 14.63,
1903
+ "learning_rate": 0.0001552049986199842,
1904
+ "loss": 0.7147,
1905
+ "step": 632000
1906
+ },
1907
+ {
1908
+ "epoch": 14.68,
1909
+ "learning_rate": 0.00015473719664991413,
1910
+ "loss": 0.7144,
1911
+ "step": 634000
1912
+ },
1913
+ {
1914
+ "epoch": 14.73,
1915
+ "learning_rate": 0.00015426939467984413,
1916
+ "loss": 0.7113,
1917
+ "step": 636000
1918
+ },
1919
+ {
1920
+ "epoch": 14.77,
1921
+ "learning_rate": 0.00015380159270977407,
1922
+ "loss": 0.7071,
1923
+ "step": 638000
1924
+ },
1925
+ {
1926
+ "epoch": 14.82,
1927
+ "learning_rate": 0.00015333379073970407,
1928
+ "loss": 0.7118,
1929
+ "step": 640000
1930
+ },
1931
+ {
1932
+ "epoch": 14.87,
1933
+ "learning_rate": 0.00015286598876963403,
1934
+ "loss": 0.7098,
1935
+ "step": 642000
1936
+ },
1937
+ {
1938
+ "epoch": 14.91,
1939
+ "learning_rate": 0.000152398186799564,
1940
+ "loss": 0.706,
1941
+ "step": 644000
1942
+ },
1943
+ {
1944
+ "epoch": 14.96,
1945
+ "learning_rate": 0.00015193038482949397,
1946
+ "loss": 0.709,
1947
+ "step": 646000
1948
+ },
1949
+ {
1950
+ "epoch": 15.01,
1951
+ "learning_rate": 0.00015146258285942394,
1952
+ "loss": 0.7087,
1953
+ "step": 648000
1954
+ },
1955
+ {
1956
+ "epoch": 15.05,
1957
+ "learning_rate": 0.0001509947808893539,
1958
+ "loss": 0.6983,
1959
+ "step": 650000
1960
+ },
1961
+ {
1962
+ "epoch": 15.1,
1963
+ "learning_rate": 0.00015052697891928388,
1964
+ "loss": 0.693,
1965
+ "step": 652000
1966
+ },
1967
+ {
1968
+ "epoch": 15.14,
1969
+ "learning_rate": 0.00015005917694921385,
1970
+ "loss": 0.6953,
1971
+ "step": 654000
1972
+ },
1973
+ {
1974
+ "epoch": 15.19,
1975
+ "learning_rate": 0.00014959137497914382,
1976
+ "loss": 0.6994,
1977
+ "step": 656000
1978
+ },
1979
+ {
1980
+ "epoch": 15.24,
1981
+ "learning_rate": 0.00014912357300907379,
1982
+ "loss": 0.6975,
1983
+ "step": 658000
1984
+ },
1985
+ {
1986
+ "epoch": 15.28,
1987
+ "learning_rate": 0.00014865577103900375,
1988
+ "loss": 0.7047,
1989
+ "step": 660000
1990
+ },
1991
+ {
1992
+ "epoch": 15.33,
1993
+ "learning_rate": 0.00014818796906893372,
1994
+ "loss": 0.6975,
1995
+ "step": 662000
1996
+ },
1997
+ {
1998
+ "epoch": 15.38,
1999
+ "learning_rate": 0.0001477201670988637,
2000
+ "loss": 0.704,
2001
+ "step": 664000
2002
+ },
2003
+ {
2004
+ "epoch": 15.42,
2005
+ "learning_rate": 0.00014725236512879366,
2006
+ "loss": 0.7042,
2007
+ "step": 666000
2008
+ },
2009
+ {
2010
+ "epoch": 15.47,
2011
+ "learning_rate": 0.00014678456315872363,
2012
+ "loss": 0.6917,
2013
+ "step": 668000
2014
+ },
2015
+ {
2016
+ "epoch": 15.51,
2017
+ "learning_rate": 0.0001463167611886536,
2018
+ "loss": 0.6914,
2019
+ "step": 670000
2020
+ },
2021
+ {
2022
+ "epoch": 15.56,
2023
+ "learning_rate": 0.00014584895921858357,
2024
+ "loss": 0.7018,
2025
+ "step": 672000
2026
+ },
2027
+ {
2028
+ "epoch": 15.61,
2029
+ "learning_rate": 0.00014538115724851354,
2030
+ "loss": 0.7016,
2031
+ "step": 674000
2032
+ },
2033
+ {
2034
+ "epoch": 15.65,
2035
+ "learning_rate": 0.0001449133552784435,
2036
+ "loss": 0.7078,
2037
+ "step": 676000
2038
+ },
2039
+ {
2040
+ "epoch": 15.7,
2041
+ "learning_rate": 0.00014444555330837347,
2042
+ "loss": 0.6932,
2043
+ "step": 678000
2044
+ },
2045
+ {
2046
+ "epoch": 15.75,
2047
+ "learning_rate": 0.00014397775133830344,
2048
+ "loss": 0.6964,
2049
+ "step": 680000
2050
+ },
2051
+ {
2052
+ "epoch": 15.79,
2053
+ "learning_rate": 0.0001435099493682334,
2054
+ "loss": 0.6997,
2055
+ "step": 682000
2056
+ },
2057
+ {
2058
+ "epoch": 15.84,
2059
+ "learning_rate": 0.00014304214739816338,
2060
+ "loss": 0.7065,
2061
+ "step": 684000
2062
+ },
2063
+ {
2064
+ "epoch": 15.88,
2065
+ "learning_rate": 0.00014257434542809335,
2066
+ "loss": 0.7047,
2067
+ "step": 686000
2068
+ },
2069
+ {
2070
+ "epoch": 15.93,
2071
+ "learning_rate": 0.00014210654345802332,
2072
+ "loss": 0.7154,
2073
+ "step": 688000
2074
+ },
2075
+ {
2076
+ "epoch": 15.98,
2077
+ "learning_rate": 0.0001416387414879533,
2078
+ "loss": 0.6993,
2079
+ "step": 690000
2080
+ },
2081
+ {
2082
+ "epoch": 16.02,
2083
+ "learning_rate": 0.00014117093951788326,
2084
+ "loss": 0.6969,
2085
+ "step": 692000
2086
+ },
2087
+ {
2088
+ "epoch": 16.07,
2089
+ "learning_rate": 0.00014070313754781325,
2090
+ "loss": 0.689,
2091
+ "step": 694000
2092
+ },
2093
+ {
2094
+ "epoch": 16.12,
2095
+ "learning_rate": 0.00014023533557774322,
2096
+ "loss": 0.6888,
2097
+ "step": 696000
2098
+ },
2099
+ {
2100
+ "epoch": 16.16,
2101
+ "learning_rate": 0.0001397675336076732,
2102
+ "loss": 0.6818,
2103
+ "step": 698000
2104
+ },
2105
+ {
2106
+ "epoch": 16.21,
2107
+ "learning_rate": 0.00013929973163760316,
2108
+ "loss": 0.693,
2109
+ "step": 700000
2110
+ },
2111
+ {
2112
+ "epoch": 16.26,
2113
+ "learning_rate": 0.00013883192966753313,
2114
+ "loss": 0.6909,
2115
+ "step": 702000
2116
+ },
2117
+ {
2118
+ "epoch": 16.3,
2119
+ "learning_rate": 0.0001383641276974631,
2120
+ "loss": 0.6873,
2121
+ "step": 704000
2122
+ },
2123
+ {
2124
+ "epoch": 16.35,
2125
+ "learning_rate": 0.00013789632572739307,
2126
+ "loss": 0.6906,
2127
+ "step": 706000
2128
+ },
2129
+ {
2130
+ "epoch": 16.39,
2131
+ "learning_rate": 0.00013742852375732303,
2132
+ "loss": 0.6866,
2133
+ "step": 708000
2134
+ },
2135
+ {
2136
+ "epoch": 16.44,
2137
+ "learning_rate": 0.000136960721787253,
2138
+ "loss": 0.701,
2139
+ "step": 710000
2140
+ },
2141
+ {
2142
+ "epoch": 16.49,
2143
+ "learning_rate": 0.00013649291981718297,
2144
+ "loss": 0.6937,
2145
+ "step": 712000
2146
+ },
2147
+ {
2148
+ "epoch": 16.53,
2149
+ "learning_rate": 0.00013602511784711294,
2150
+ "loss": 0.6907,
2151
+ "step": 714000
2152
+ },
2153
+ {
2154
+ "epoch": 16.58,
2155
+ "learning_rate": 0.0001355573158770429,
2156
+ "loss": 0.6897,
2157
+ "step": 716000
2158
+ },
2159
+ {
2160
+ "epoch": 16.63,
2161
+ "learning_rate": 0.00013508951390697288,
2162
+ "loss": 0.6952,
2163
+ "step": 718000
2164
+ },
2165
+ {
2166
+ "epoch": 16.67,
2167
+ "learning_rate": 0.00013462171193690285,
2168
+ "loss": 0.6865,
2169
+ "step": 720000
2170
+ },
2171
+ {
2172
+ "epoch": 16.72,
2173
+ "learning_rate": 0.00013415390996683282,
2174
+ "loss": 0.6935,
2175
+ "step": 722000
2176
+ },
2177
+ {
2178
+ "epoch": 16.76,
2179
+ "learning_rate": 0.00013368610799676278,
2180
+ "loss": 0.6919,
2181
+ "step": 724000
2182
+ },
2183
+ {
2184
+ "epoch": 16.81,
2185
+ "learning_rate": 0.00013321830602669275,
2186
+ "loss": 0.6904,
2187
+ "step": 726000
2188
+ },
2189
+ {
2190
+ "epoch": 16.86,
2191
+ "learning_rate": 0.00013275050405662272,
2192
+ "loss": 0.6964,
2193
+ "step": 728000
2194
+ },
2195
+ {
2196
+ "epoch": 16.9,
2197
+ "learning_rate": 0.00013228270208655272,
2198
+ "loss": 0.6943,
2199
+ "step": 730000
2200
+ },
2201
+ {
2202
+ "epoch": 16.95,
2203
+ "learning_rate": 0.0001318149001164827,
2204
+ "loss": 0.6949,
2205
+ "step": 732000
2206
+ },
2207
+ {
2208
+ "epoch": 17.0,
2209
+ "learning_rate": 0.00013134709814641266,
2210
+ "loss": 0.6943,
2211
+ "step": 734000
2212
+ },
2213
+ {
2214
+ "epoch": 17.04,
2215
+ "learning_rate": 0.00013087929617634263,
2216
+ "loss": 0.6851,
2217
+ "step": 736000
2218
+ },
2219
+ {
2220
+ "epoch": 17.09,
2221
+ "learning_rate": 0.0001304114942062726,
2222
+ "loss": 0.6802,
2223
+ "step": 738000
2224
+ },
2225
+ {
2226
+ "epoch": 17.14,
2227
+ "learning_rate": 0.00012994369223620256,
2228
+ "loss": 0.6801,
2229
+ "step": 740000
2230
+ },
2231
+ {
2232
+ "epoch": 17.18,
2233
+ "learning_rate": 0.00012947589026613253,
2234
+ "loss": 0.6756,
2235
+ "step": 742000
2236
+ },
2237
+ {
2238
+ "epoch": 17.23,
2239
+ "learning_rate": 0.0001290080882960625,
2240
+ "loss": 0.6824,
2241
+ "step": 744000
2242
+ },
2243
+ {
2244
+ "epoch": 17.27,
2245
+ "learning_rate": 0.00012854028632599247,
2246
+ "loss": 0.6894,
2247
+ "step": 746000
2248
+ },
2249
+ {
2250
+ "epoch": 17.32,
2251
+ "learning_rate": 0.00012807248435592244,
2252
+ "loss": 0.682,
2253
+ "step": 748000
2254
+ },
2255
+ {
2256
+ "epoch": 17.37,
2257
+ "learning_rate": 0.0001276046823858524,
2258
+ "loss": 0.6814,
2259
+ "step": 750000
2260
+ },
2261
+ {
2262
+ "epoch": 17.41,
2263
+ "learning_rate": 0.00012713688041578238,
2264
+ "loss": 0.6737,
2265
+ "step": 752000
2266
+ },
2267
+ {
2268
+ "epoch": 17.46,
2269
+ "learning_rate": 0.00012666907844571234,
2270
+ "loss": 0.6874,
2271
+ "step": 754000
2272
+ },
2273
+ {
2274
+ "epoch": 17.51,
2275
+ "learning_rate": 0.00012620127647564231,
2276
+ "loss": 0.6842,
2277
+ "step": 756000
2278
+ },
2279
+ {
2280
+ "epoch": 17.55,
2281
+ "learning_rate": 0.00012573347450557228,
2282
+ "loss": 0.6871,
2283
+ "step": 758000
2284
+ },
2285
+ {
2286
+ "epoch": 17.6,
2287
+ "learning_rate": 0.00012526567253550225,
2288
+ "loss": 0.6833,
2289
+ "step": 760000
2290
+ },
2291
+ {
2292
+ "epoch": 17.64,
2293
+ "learning_rate": 0.00012479787056543222,
2294
+ "loss": 0.6818,
2295
+ "step": 762000
2296
+ },
2297
+ {
2298
+ "epoch": 17.69,
2299
+ "learning_rate": 0.0001243300685953622,
2300
+ "loss": 0.6824,
2301
+ "step": 764000
2302
+ },
2303
+ {
2304
+ "epoch": 17.74,
2305
+ "learning_rate": 0.00012386226662529219,
2306
+ "loss": 0.684,
2307
+ "step": 766000
2308
+ },
2309
+ {
2310
+ "epoch": 17.78,
2311
+ "learning_rate": 0.00012339446465522215,
2312
+ "loss": 0.6822,
2313
+ "step": 768000
2314
+ },
2315
+ {
2316
+ "epoch": 17.83,
2317
+ "learning_rate": 0.00012292666268515212,
2318
+ "loss": 0.68,
2319
+ "step": 770000
2320
+ },
2321
+ {
2322
+ "epoch": 17.88,
2323
+ "learning_rate": 0.0001224588607150821,
2324
+ "loss": 0.6842,
2325
+ "step": 772000
2326
+ },
2327
+ {
2328
+ "epoch": 17.92,
2329
+ "learning_rate": 0.00012199105874501206,
2330
+ "loss": 0.6827,
2331
+ "step": 774000
2332
+ },
2333
+ {
2334
+ "epoch": 17.97,
2335
+ "learning_rate": 0.00012152325677494203,
2336
+ "loss": 0.6901,
2337
+ "step": 776000
2338
+ },
2339
+ {
2340
+ "epoch": 18.02,
2341
+ "learning_rate": 0.000121055454804872,
2342
+ "loss": 0.6781,
2343
+ "step": 778000
2344
+ },
2345
+ {
2346
+ "epoch": 18.06,
2347
+ "learning_rate": 0.00012058765283480197,
2348
+ "loss": 0.6768,
2349
+ "step": 780000
2350
+ },
2351
+ {
2352
+ "epoch": 18.11,
2353
+ "learning_rate": 0.00012011985086473194,
2354
+ "loss": 0.6704,
2355
+ "step": 782000
2356
+ },
2357
+ {
2358
+ "epoch": 18.15,
2359
+ "learning_rate": 0.0001196520488946619,
2360
+ "loss": 0.6767,
2361
+ "step": 784000
2362
+ },
2363
+ {
2364
+ "epoch": 18.2,
2365
+ "learning_rate": 0.00011918424692459187,
2366
+ "loss": 0.6696,
2367
+ "step": 786000
2368
+ },
2369
+ {
2370
+ "epoch": 18.25,
2371
+ "learning_rate": 0.00011871644495452184,
2372
+ "loss": 0.6717,
2373
+ "step": 788000
2374
+ },
2375
+ {
2376
+ "epoch": 18.29,
2377
+ "learning_rate": 0.00011824864298445181,
2378
+ "loss": 0.6666,
2379
+ "step": 790000
2380
+ },
2381
+ {
2382
+ "epoch": 18.34,
2383
+ "learning_rate": 0.00011778084101438178,
2384
+ "loss": 0.6681,
2385
+ "step": 792000
2386
+ },
2387
+ {
2388
+ "epoch": 18.39,
2389
+ "learning_rate": 0.00011731303904431175,
2390
+ "loss": 0.6688,
2391
+ "step": 794000
2392
+ },
2393
+ {
2394
+ "epoch": 18.43,
2395
+ "learning_rate": 0.00011684523707424172,
2396
+ "loss": 0.6809,
2397
+ "step": 796000
2398
+ },
2399
+ {
2400
+ "epoch": 18.48,
2401
+ "learning_rate": 0.00011637743510417169,
2402
+ "loss": 0.6704,
2403
+ "step": 798000
2404
+ },
2405
+ {
2406
+ "epoch": 18.52,
2407
+ "learning_rate": 0.00011590963313410166,
2408
+ "loss": 0.6732,
2409
+ "step": 800000
2410
+ },
2411
+ {
2412
+ "epoch": 18.57,
2413
+ "learning_rate": 0.00011544183116403164,
2414
+ "loss": 0.6688,
2415
+ "step": 802000
2416
+ },
2417
+ {
2418
+ "epoch": 18.62,
2419
+ "learning_rate": 0.00011497402919396161,
2420
+ "loss": 0.6767,
2421
+ "step": 804000
2422
+ },
2423
+ {
2424
+ "epoch": 18.66,
2425
+ "learning_rate": 0.00011450622722389158,
2426
+ "loss": 0.6721,
2427
+ "step": 806000
2428
+ },
2429
+ {
2430
+ "epoch": 18.71,
2431
+ "learning_rate": 0.00011403842525382154,
2432
+ "loss": 0.6716,
2433
+ "step": 808000
2434
+ },
2435
+ {
2436
+ "epoch": 18.76,
2437
+ "learning_rate": 0.00011357062328375151,
2438
+ "loss": 0.673,
2439
+ "step": 810000
2440
+ },
2441
+ {
2442
+ "epoch": 18.8,
2443
+ "learning_rate": 0.00011310282131368148,
2444
+ "loss": 0.6717,
2445
+ "step": 812000
2446
+ },
2447
+ {
2448
+ "epoch": 18.85,
2449
+ "learning_rate": 0.00011263501934361145,
2450
+ "loss": 0.6607,
2451
+ "step": 814000
2452
+ },
2453
+ {
2454
+ "epoch": 18.9,
2455
+ "learning_rate": 0.00011216721737354142,
2456
+ "loss": 0.6732,
2457
+ "step": 816000
2458
+ },
2459
+ {
2460
+ "epoch": 18.94,
2461
+ "learning_rate": 0.00011169941540347139,
2462
+ "loss": 0.6715,
2463
+ "step": 818000
2464
+ },
2465
+ {
2466
+ "epoch": 18.99,
2467
+ "learning_rate": 0.00011123161343340136,
2468
+ "loss": 0.678,
2469
+ "step": 820000
2470
+ },
2471
+ {
2472
+ "epoch": 19.03,
2473
+ "learning_rate": 0.00011076381146333133,
2474
+ "loss": 0.6618,
2475
+ "step": 822000
2476
+ },
2477
+ {
2478
+ "epoch": 19.08,
2479
+ "learning_rate": 0.0001102960094932613,
2480
+ "loss": 0.6589,
2481
+ "step": 824000
2482
+ },
2483
+ {
2484
+ "epoch": 19.13,
2485
+ "learning_rate": 0.00010982820752319126,
2486
+ "loss": 0.6624,
2487
+ "step": 826000
2488
+ },
2489
+ {
2490
+ "epoch": 19.17,
2491
+ "learning_rate": 0.00010936040555312123,
2492
+ "loss": 0.6618,
2493
+ "step": 828000
2494
+ },
2495
+ {
2496
+ "epoch": 19.22,
2497
+ "learning_rate": 0.0001088926035830512,
2498
+ "loss": 0.6666,
2499
+ "step": 830000
2500
+ },
2501
+ {
2502
+ "epoch": 19.27,
2503
+ "learning_rate": 0.00010842480161298117,
2504
+ "loss": 0.6645,
2505
+ "step": 832000
2506
+ },
2507
+ {
2508
+ "epoch": 19.31,
2509
+ "learning_rate": 0.00010795699964291114,
2510
+ "loss": 0.6667,
2511
+ "step": 834000
2512
+ },
2513
+ {
2514
+ "epoch": 19.36,
2515
+ "learning_rate": 0.00010748919767284111,
2516
+ "loss": 0.6649,
2517
+ "step": 836000
2518
+ },
2519
+ {
2520
+ "epoch": 19.4,
2521
+ "learning_rate": 0.0001070213957027711,
2522
+ "loss": 0.659,
2523
+ "step": 838000
2524
+ },
2525
+ {
2526
+ "epoch": 19.45,
2527
+ "learning_rate": 0.00010655359373270107,
2528
+ "loss": 0.6611,
2529
+ "step": 840000
2530
+ },
2531
+ {
2532
+ "epoch": 19.5,
2533
+ "learning_rate": 0.00010608579176263104,
2534
+ "loss": 0.6565,
2535
+ "step": 842000
2536
+ },
2537
+ {
2538
+ "epoch": 19.54,
2539
+ "learning_rate": 0.00010561798979256101,
2540
+ "loss": 0.6631,
2541
+ "step": 844000
2542
+ },
2543
+ {
2544
+ "epoch": 19.59,
2545
+ "learning_rate": 0.00010515018782249098,
2546
+ "loss": 0.6593,
2547
+ "step": 846000
2548
+ },
2549
+ {
2550
+ "epoch": 19.64,
2551
+ "learning_rate": 0.00010468238585242095,
2552
+ "loss": 0.6654,
2553
+ "step": 848000
2554
+ },
2555
+ {
2556
+ "epoch": 19.68,
2557
+ "learning_rate": 0.00010421458388235092,
2558
+ "loss": 0.6621,
2559
+ "step": 850000
2560
+ },
2561
+ {
2562
+ "epoch": 19.73,
2563
+ "learning_rate": 0.00010374678191228089,
2564
+ "loss": 0.661,
2565
+ "step": 852000
2566
+ },
2567
+ {
2568
+ "epoch": 19.78,
2569
+ "learning_rate": 0.00010327897994221086,
2570
+ "loss": 0.6515,
2571
+ "step": 854000
2572
+ },
2573
+ {
2574
+ "epoch": 19.82,
2575
+ "learning_rate": 0.00010281117797214082,
2576
+ "loss": 0.6614,
2577
+ "step": 856000
2578
+ },
2579
+ {
2580
+ "epoch": 19.87,
2581
+ "learning_rate": 0.0001023433760020708,
2582
+ "loss": 0.6616,
2583
+ "step": 858000
2584
+ },
2585
+ {
2586
+ "epoch": 19.91,
2587
+ "learning_rate": 0.00010187557403200076,
2588
+ "loss": 0.6598,
2589
+ "step": 860000
2590
+ },
2591
+ {
2592
+ "epoch": 19.96,
2593
+ "learning_rate": 0.00010140777206193073,
2594
+ "loss": 0.6616,
2595
+ "step": 862000
2596
+ },
2597
+ {
2598
+ "epoch": 20.01,
2599
+ "learning_rate": 0.0001009399700918607,
2600
+ "loss": 0.6679,
2601
+ "step": 864000
2602
+ },
2603
+ {
2604
+ "epoch": 20.05,
2605
+ "learning_rate": 0.00010047216812179067,
2606
+ "loss": 0.6518,
2607
+ "step": 866000
2608
+ },
2609
+ {
2610
+ "epoch": 20.1,
2611
+ "learning_rate": 0.00010000436615172064,
2612
+ "loss": 0.6463,
2613
+ "step": 868000
2614
+ },
2615
+ {
2616
+ "epoch": 20.15,
2617
+ "learning_rate": 9.95365641816506e-05,
2618
+ "loss": 0.6529,
2619
+ "step": 870000
2620
+ },
2621
+ {
2622
+ "epoch": 20.19,
2623
+ "learning_rate": 9.906876221158058e-05,
2624
+ "loss": 0.6463,
2625
+ "step": 872000
2626
+ },
2627
+ {
2628
+ "epoch": 20.24,
2629
+ "learning_rate": 9.860096024151056e-05,
2630
+ "loss": 0.6545,
2631
+ "step": 874000
2632
+ },
2633
+ {
2634
+ "epoch": 20.28,
2635
+ "learning_rate": 9.813315827144053e-05,
2636
+ "loss": 0.6531,
2637
+ "step": 876000
2638
+ },
2639
+ {
2640
+ "epoch": 20.33,
2641
+ "learning_rate": 9.76653563013705e-05,
2642
+ "loss": 0.6442,
2643
+ "step": 878000
2644
+ },
2645
+ {
2646
+ "epoch": 20.38,
2647
+ "learning_rate": 9.719755433130046e-05,
2648
+ "loss": 0.65,
2649
+ "step": 880000
2650
+ },
2651
+ {
2652
+ "epoch": 20.42,
2653
+ "learning_rate": 9.672975236123043e-05,
2654
+ "loss": 0.6518,
2655
+ "step": 882000
2656
+ },
2657
+ {
2658
+ "epoch": 20.47,
2659
+ "learning_rate": 9.62619503911604e-05,
2660
+ "loss": 0.6546,
2661
+ "step": 884000
2662
+ },
2663
+ {
2664
+ "epoch": 20.52,
2665
+ "learning_rate": 9.579414842109037e-05,
2666
+ "loss": 0.6494,
2667
+ "step": 886000
2668
+ },
2669
+ {
2670
+ "epoch": 20.56,
2671
+ "learning_rate": 9.532634645102035e-05,
2672
+ "loss": 0.654,
2673
+ "step": 888000
2674
+ },
2675
+ {
2676
+ "epoch": 20.61,
2677
+ "learning_rate": 9.485854448095032e-05,
2678
+ "loss": 0.6536,
2679
+ "step": 890000
2680
+ },
2681
+ {
2682
+ "epoch": 20.66,
2683
+ "learning_rate": 9.439074251088029e-05,
2684
+ "loss": 0.6547,
2685
+ "step": 892000
2686
+ },
2687
+ {
2688
+ "epoch": 20.7,
2689
+ "learning_rate": 9.392294054081026e-05,
2690
+ "loss": 0.6421,
2691
+ "step": 894000
2692
+ },
2693
+ {
2694
+ "epoch": 20.75,
2695
+ "learning_rate": 9.345513857074023e-05,
2696
+ "loss": 0.6506,
2697
+ "step": 896000
2698
+ },
2699
+ {
2700
+ "epoch": 20.79,
2701
+ "learning_rate": 9.29873366006702e-05,
2702
+ "loss": 0.6551,
2703
+ "step": 898000
2704
+ },
2705
+ {
2706
+ "epoch": 20.84,
2707
+ "learning_rate": 9.251953463060017e-05,
2708
+ "loss": 0.6542,
2709
+ "step": 900000
2710
+ },
2711
+ {
2712
+ "epoch": 20.89,
2713
+ "learning_rate": 9.205173266053014e-05,
2714
+ "loss": 0.6398,
2715
+ "step": 902000
2716
+ },
2717
+ {
2718
+ "epoch": 20.93,
2719
+ "learning_rate": 9.15839306904601e-05,
2720
+ "loss": 0.653,
2721
+ "step": 904000
2722
+ },
2723
+ {
2724
+ "epoch": 20.98,
2725
+ "learning_rate": 9.111612872039007e-05,
2726
+ "loss": 0.6476,
2727
+ "step": 906000
2728
+ },
2729
+ {
2730
+ "epoch": 21.03,
2731
+ "learning_rate": 9.064832675032004e-05,
2732
+ "loss": 0.6378,
2733
+ "step": 908000
2734
+ },
2735
+ {
2736
+ "epoch": 21.07,
2737
+ "learning_rate": 9.018052478025002e-05,
2738
+ "loss": 0.6413,
2739
+ "step": 910000
2740
+ },
2741
+ {
2742
+ "epoch": 21.12,
2743
+ "learning_rate": 8.971272281017999e-05,
2744
+ "loss": 0.6368,
2745
+ "step": 912000
2746
+ },
2747
+ {
2748
+ "epoch": 21.16,
2749
+ "learning_rate": 8.924492084010996e-05,
2750
+ "loss": 0.6366,
2751
+ "step": 914000
2752
+ },
2753
+ {
2754
+ "epoch": 21.21,
2755
+ "learning_rate": 8.877711887003993e-05,
2756
+ "loss": 0.6455,
2757
+ "step": 916000
2758
+ },
2759
+ {
2760
+ "epoch": 21.26,
2761
+ "learning_rate": 8.83093168999699e-05,
2762
+ "loss": 0.6448,
2763
+ "step": 918000
2764
+ },
2765
+ {
2766
+ "epoch": 21.3,
2767
+ "learning_rate": 8.784151492989987e-05,
2768
+ "loss": 0.6371,
2769
+ "step": 920000
2770
+ },
2771
+ {
2772
+ "epoch": 21.35,
2773
+ "learning_rate": 8.737371295982984e-05,
2774
+ "loss": 0.6457,
2775
+ "step": 922000
2776
+ },
2777
+ {
2778
+ "epoch": 21.4,
2779
+ "learning_rate": 8.69059109897598e-05,
2780
+ "loss": 0.6399,
2781
+ "step": 924000
2782
+ },
2783
+ {
2784
+ "epoch": 21.44,
2785
+ "learning_rate": 8.643810901968978e-05,
2786
+ "loss": 0.6389,
2787
+ "step": 926000
2788
+ },
2789
+ {
2790
+ "epoch": 21.49,
2791
+ "learning_rate": 8.597030704961974e-05,
2792
+ "loss": 0.6444,
2793
+ "step": 928000
2794
+ },
2795
+ {
2796
+ "epoch": 21.54,
2797
+ "learning_rate": 8.550250507954971e-05,
2798
+ "loss": 0.6346,
2799
+ "step": 930000
2800
+ },
2801
+ {
2802
+ "epoch": 21.58,
2803
+ "learning_rate": 8.503470310947968e-05,
2804
+ "loss": 0.6394,
2805
+ "step": 932000
2806
+ },
2807
+ {
2808
+ "epoch": 21.63,
2809
+ "learning_rate": 8.456690113940965e-05,
2810
+ "loss": 0.6397,
2811
+ "step": 934000
2812
+ },
2813
+ {
2814
+ "epoch": 21.67,
2815
+ "learning_rate": 8.409909916933962e-05,
2816
+ "loss": 0.6411,
2817
+ "step": 936000
2818
+ },
2819
+ {
2820
+ "epoch": 21.72,
2821
+ "learning_rate": 8.363129719926959e-05,
2822
+ "loss": 0.6383,
2823
+ "step": 938000
2824
+ },
2825
+ {
2826
+ "epoch": 21.77,
2827
+ "learning_rate": 8.316349522919956e-05,
2828
+ "loss": 0.6416,
2829
+ "step": 940000
2830
+ },
2831
+ {
2832
+ "epoch": 21.81,
2833
+ "learning_rate": 8.269569325912953e-05,
2834
+ "loss": 0.635,
2835
+ "step": 942000
2836
+ },
2837
+ {
2838
+ "epoch": 21.86,
2839
+ "learning_rate": 8.22278912890595e-05,
2840
+ "loss": 0.6371,
2841
+ "step": 944000
2842
+ },
2843
+ {
2844
+ "epoch": 21.91,
2845
+ "learning_rate": 8.176008931898949e-05,
2846
+ "loss": 0.6412,
2847
+ "step": 946000
2848
+ },
2849
+ {
2850
+ "epoch": 21.95,
2851
+ "learning_rate": 8.129228734891946e-05,
2852
+ "loss": 0.6414,
2853
+ "step": 948000
2854
+ },
2855
+ {
2856
+ "epoch": 22.0,
2857
+ "learning_rate": 8.082448537884943e-05,
2858
+ "loss": 0.6285,
2859
+ "step": 950000
2860
+ },
2861
+ {
2862
+ "epoch": 22.04,
2863
+ "learning_rate": 8.03566834087794e-05,
2864
+ "loss": 0.6285,
2865
+ "step": 952000
2866
+ },
2867
+ {
2868
+ "epoch": 22.09,
2869
+ "learning_rate": 7.988888143870937e-05,
2870
+ "loss": 0.6268,
2871
+ "step": 954000
2872
+ },
2873
+ {
2874
+ "epoch": 22.14,
2875
+ "learning_rate": 7.942107946863934e-05,
2876
+ "loss": 0.6251,
2877
+ "step": 956000
2878
+ },
2879
+ {
2880
+ "epoch": 22.18,
2881
+ "learning_rate": 7.89532774985693e-05,
2882
+ "loss": 0.6306,
2883
+ "step": 958000
2884
+ },
2885
+ {
2886
+ "epoch": 22.23,
2887
+ "learning_rate": 7.848547552849927e-05,
2888
+ "loss": 0.6283,
2889
+ "step": 960000
2890
+ },
2891
+ {
2892
+ "epoch": 22.28,
2893
+ "learning_rate": 7.801767355842924e-05,
2894
+ "loss": 0.6264,
2895
+ "step": 962000
2896
+ },
2897
+ {
2898
+ "epoch": 22.32,
2899
+ "learning_rate": 7.754987158835921e-05,
2900
+ "loss": 0.6279,
2901
+ "step": 964000
2902
+ },
2903
+ {
2904
+ "epoch": 22.37,
2905
+ "learning_rate": 7.708206961828918e-05,
2906
+ "loss": 0.6272,
2907
+ "step": 966000
2908
+ },
2909
+ {
2910
+ "epoch": 22.41,
2911
+ "learning_rate": 7.661426764821915e-05,
2912
+ "loss": 0.6355,
2913
+ "step": 968000
2914
+ },
2915
+ {
2916
+ "epoch": 22.46,
2917
+ "learning_rate": 7.614646567814912e-05,
2918
+ "loss": 0.6349,
2919
+ "step": 970000
2920
+ },
2921
+ {
2922
+ "epoch": 22.51,
2923
+ "learning_rate": 7.567866370807909e-05,
2924
+ "loss": 0.6281,
2925
+ "step": 972000
2926
+ },
2927
+ {
2928
+ "epoch": 22.55,
2929
+ "learning_rate": 7.521086173800905e-05,
2930
+ "loss": 0.6269,
2931
+ "step": 974000
2932
+ },
2933
+ {
2934
+ "epoch": 22.6,
2935
+ "learning_rate": 7.474305976793904e-05,
2936
+ "loss": 0.6221,
2937
+ "step": 976000
2938
+ },
2939
+ {
2940
+ "epoch": 22.65,
2941
+ "learning_rate": 7.4275257797869e-05,
2942
+ "loss": 0.6295,
2943
+ "step": 978000
2944
+ },
2945
+ {
2946
+ "epoch": 22.69,
2947
+ "learning_rate": 7.380745582779897e-05,
2948
+ "loss": 0.6265,
2949
+ "step": 980000
2950
+ },
2951
+ {
2952
+ "epoch": 22.74,
2953
+ "learning_rate": 7.333965385772894e-05,
2954
+ "loss": 0.6203,
2955
+ "step": 982000
2956
+ },
2957
+ {
2958
+ "epoch": 22.79,
2959
+ "learning_rate": 7.287185188765891e-05,
2960
+ "loss": 0.6306,
2961
+ "step": 984000
2962
+ },
2963
+ {
2964
+ "epoch": 22.83,
2965
+ "learning_rate": 7.240404991758888e-05,
2966
+ "loss": 0.6319,
2967
+ "step": 986000
2968
+ },
2969
+ {
2970
+ "epoch": 22.88,
2971
+ "learning_rate": 7.193624794751885e-05,
2972
+ "loss": 0.6211,
2973
+ "step": 988000
2974
+ },
2975
+ {
2976
+ "epoch": 22.92,
2977
+ "learning_rate": 7.146844597744882e-05,
2978
+ "loss": 0.6244,
2979
+ "step": 990000
2980
+ },
2981
+ {
2982
+ "epoch": 22.97,
2983
+ "learning_rate": 7.100064400737879e-05,
2984
+ "loss": 0.6262,
2985
+ "step": 992000
2986
+ },
2987
+ {
2988
+ "epoch": 23.02,
2989
+ "learning_rate": 7.053284203730876e-05,
2990
+ "loss": 0.6166,
2991
+ "step": 994000
2992
+ },
2993
+ {
2994
+ "epoch": 23.06,
2995
+ "learning_rate": 7.006504006723873e-05,
2996
+ "loss": 0.6166,
2997
+ "step": 996000
2998
+ },
2999
+ {
3000
+ "epoch": 23.11,
3001
+ "learning_rate": 6.95972380971687e-05,
3002
+ "loss": 0.6175,
3003
+ "step": 998000
3004
+ },
3005
+ {
3006
+ "epoch": 23.16,
3007
+ "learning_rate": 6.912943612709866e-05,
3008
+ "loss": 0.6151,
3009
+ "step": 1000000
3010
+ },
3011
+ {
3012
+ "epoch": 23.2,
3013
+ "learning_rate": 6.866163415702863e-05,
3014
+ "loss": 0.6153,
3015
+ "step": 1002000
3016
+ },
3017
+ {
3018
+ "epoch": 23.25,
3019
+ "learning_rate": 6.81938321869586e-05,
3020
+ "loss": 0.6212,
3021
+ "step": 1004000
3022
+ },
3023
+ {
3024
+ "epoch": 23.29,
3025
+ "learning_rate": 6.772603021688858e-05,
3026
+ "loss": 0.6161,
3027
+ "step": 1006000
3028
+ },
3029
+ {
3030
+ "epoch": 23.34,
3031
+ "learning_rate": 6.725822824681855e-05,
3032
+ "loss": 0.6158,
3033
+ "step": 1008000
3034
+ },
3035
+ {
3036
+ "epoch": 23.39,
3037
+ "learning_rate": 6.679042627674852e-05,
3038
+ "loss": 0.6089,
3039
+ "step": 1010000
3040
+ },
3041
+ {
3042
+ "epoch": 23.43,
3043
+ "learning_rate": 6.632262430667849e-05,
3044
+ "loss": 0.6166,
3045
+ "step": 1012000
3046
+ },
3047
+ {
3048
+ "epoch": 23.48,
3049
+ "learning_rate": 6.585482233660846e-05,
3050
+ "loss": 0.6134,
3051
+ "step": 1014000
3052
+ },
3053
+ {
3054
+ "epoch": 23.53,
3055
+ "learning_rate": 6.538702036653843e-05,
3056
+ "loss": 0.6171,
3057
+ "step": 1016000
3058
+ },
3059
+ {
3060
+ "epoch": 23.57,
3061
+ "learning_rate": 6.49192183964684e-05,
3062
+ "loss": 0.6122,
3063
+ "step": 1018000
3064
+ },
3065
+ {
3066
+ "epoch": 23.62,
3067
+ "learning_rate": 6.445141642639837e-05,
3068
+ "loss": 0.6176,
3069
+ "step": 1020000
3070
+ },
3071
+ {
3072
+ "epoch": 23.67,
3073
+ "learning_rate": 6.398361445632833e-05,
3074
+ "loss": 0.6146,
3075
+ "step": 1022000
3076
+ },
3077
+ {
3078
+ "epoch": 23.71,
3079
+ "learning_rate": 6.35158124862583e-05,
3080
+ "loss": 0.6069,
3081
+ "step": 1024000
3082
+ },
3083
+ {
3084
+ "epoch": 23.76,
3085
+ "learning_rate": 6.304801051618829e-05,
3086
+ "loss": 0.6169,
3087
+ "step": 1026000
3088
+ },
3089
+ {
3090
+ "epoch": 23.8,
3091
+ "learning_rate": 6.258020854611825e-05,
3092
+ "loss": 0.6222,
3093
+ "step": 1028000
3094
+ },
3095
+ {
3096
+ "epoch": 23.85,
3097
+ "learning_rate": 6.211240657604822e-05,
3098
+ "loss": 0.6152,
3099
+ "step": 1030000
3100
+ },
3101
+ {
3102
+ "epoch": 23.9,
3103
+ "learning_rate": 6.164460460597819e-05,
3104
+ "loss": 0.6181,
3105
+ "step": 1032000
3106
+ },
3107
+ {
3108
+ "epoch": 23.94,
3109
+ "learning_rate": 6.117680263590816e-05,
3110
+ "loss": 0.6123,
3111
+ "step": 1034000
3112
+ },
3113
+ {
3114
+ "epoch": 23.99,
3115
+ "learning_rate": 6.070900066583813e-05,
3116
+ "loss": 0.619,
3117
+ "step": 1036000
3118
+ },
3119
+ {
3120
+ "epoch": 24.04,
3121
+ "learning_rate": 6.02411986957681e-05,
3122
+ "loss": 0.6099,
3123
+ "step": 1038000
3124
+ },
3125
+ {
3126
+ "epoch": 24.08,
3127
+ "learning_rate": 5.977339672569807e-05,
3128
+ "loss": 0.6098,
3129
+ "step": 1040000
3130
+ },
3131
+ {
3132
+ "epoch": 24.13,
3133
+ "learning_rate": 5.930559475562804e-05,
3134
+ "loss": 0.5965,
3135
+ "step": 1042000
3136
+ },
3137
+ {
3138
+ "epoch": 24.17,
3139
+ "learning_rate": 5.883779278555802e-05,
3140
+ "loss": 0.6059,
3141
+ "step": 1044000
3142
+ },
3143
+ {
3144
+ "epoch": 24.22,
3145
+ "learning_rate": 5.836999081548799e-05,
3146
+ "loss": 0.6021,
3147
+ "step": 1046000
3148
+ },
3149
+ {
3150
+ "epoch": 24.27,
3151
+ "learning_rate": 5.790218884541796e-05,
3152
+ "loss": 0.6093,
3153
+ "step": 1048000
3154
+ },
3155
+ {
3156
+ "epoch": 24.31,
3157
+ "learning_rate": 5.7434386875347926e-05,
3158
+ "loss": 0.6031,
3159
+ "step": 1050000
3160
+ },
3161
+ {
3162
+ "epoch": 24.36,
3163
+ "learning_rate": 5.6966584905277895e-05,
3164
+ "loss": 0.6053,
3165
+ "step": 1052000
3166
+ },
3167
+ {
3168
+ "epoch": 24.41,
3169
+ "learning_rate": 5.6498782935207863e-05,
3170
+ "loss": 0.6036,
3171
+ "step": 1054000
3172
+ },
3173
+ {
3174
+ "epoch": 24.45,
3175
+ "learning_rate": 5.603098096513783e-05,
3176
+ "loss": 0.6011,
3177
+ "step": 1056000
3178
+ },
3179
+ {
3180
+ "epoch": 24.5,
3181
+ "learning_rate": 5.55631789950678e-05,
3182
+ "loss": 0.6035,
3183
+ "step": 1058000
3184
+ },
3185
+ {
3186
+ "epoch": 24.55,
3187
+ "learning_rate": 5.509537702499777e-05,
3188
+ "loss": 0.6066,
3189
+ "step": 1060000
3190
+ },
3191
+ {
3192
+ "epoch": 24.59,
3193
+ "learning_rate": 5.4627575054927746e-05,
3194
+ "loss": 0.6061,
3195
+ "step": 1062000
3196
+ },
3197
+ {
3198
+ "epoch": 24.64,
3199
+ "learning_rate": 5.4159773084857714e-05,
3200
+ "loss": 0.6027,
3201
+ "step": 1064000
3202
+ },
3203
+ {
3204
+ "epoch": 24.68,
3205
+ "learning_rate": 5.369197111478768e-05,
3206
+ "loss": 0.6,
3207
+ "step": 1066000
3208
+ },
3209
+ {
3210
+ "epoch": 24.73,
3211
+ "learning_rate": 5.322416914471765e-05,
3212
+ "loss": 0.6062,
3213
+ "step": 1068000
3214
+ },
3215
+ {
3216
+ "epoch": 24.78,
3217
+ "learning_rate": 5.275636717464762e-05,
3218
+ "loss": 0.6003,
3219
+ "step": 1070000
3220
+ },
3221
+ {
3222
+ "epoch": 24.82,
3223
+ "learning_rate": 5.228856520457759e-05,
3224
+ "loss": 0.5988,
3225
+ "step": 1072000
3226
+ },
3227
+ {
3228
+ "epoch": 24.87,
3229
+ "learning_rate": 5.182076323450756e-05,
3230
+ "loss": 0.6096,
3231
+ "step": 1074000
3232
+ },
3233
+ {
3234
+ "epoch": 24.92,
3235
+ "learning_rate": 5.135296126443753e-05,
3236
+ "loss": 0.5988,
3237
+ "step": 1076000
3238
+ },
3239
+ {
3240
+ "epoch": 24.96,
3241
+ "learning_rate": 5.08851592943675e-05,
3242
+ "loss": 0.6086,
3243
+ "step": 1078000
3244
+ },
3245
+ {
3246
+ "epoch": 25.01,
3247
+ "learning_rate": 5.041735732429748e-05,
3248
+ "loss": 0.5942,
3249
+ "step": 1080000
3250
+ },
3251
+ {
3252
+ "epoch": 25.05,
3253
+ "learning_rate": 4.994955535422745e-05,
3254
+ "loss": 0.5954,
3255
+ "step": 1082000
3256
+ },
3257
+ {
3258
+ "epoch": 25.1,
3259
+ "learning_rate": 4.9481753384157417e-05,
3260
+ "loss": 0.5948,
3261
+ "step": 1084000
3262
+ },
3263
+ {
3264
+ "epoch": 25.15,
3265
+ "learning_rate": 4.9013951414087385e-05,
3266
+ "loss": 0.5946,
3267
+ "step": 1086000
3268
+ },
3269
+ {
3270
+ "epoch": 25.19,
3271
+ "learning_rate": 4.8546149444017354e-05,
3272
+ "loss": 0.5938,
3273
+ "step": 1088000
3274
+ },
3275
+ {
3276
+ "epoch": 25.24,
3277
+ "learning_rate": 4.807834747394732e-05,
3278
+ "loss": 0.5961,
3279
+ "step": 1090000
3280
+ },
3281
+ {
3282
+ "epoch": 25.29,
3283
+ "learning_rate": 4.761054550387729e-05,
3284
+ "loss": 0.5947,
3285
+ "step": 1092000
3286
+ },
3287
+ {
3288
+ "epoch": 25.33,
3289
+ "learning_rate": 4.714274353380726e-05,
3290
+ "loss": 0.6019,
3291
+ "step": 1094000
3292
+ },
3293
+ {
3294
+ "epoch": 25.38,
3295
+ "learning_rate": 4.667494156373723e-05,
3296
+ "loss": 0.5927,
3297
+ "step": 1096000
3298
+ },
3299
+ {
3300
+ "epoch": 25.43,
3301
+ "learning_rate": 4.6207139593667205e-05,
3302
+ "loss": 0.5921,
3303
+ "step": 1098000
3304
+ },
3305
+ {
3306
+ "epoch": 25.47,
3307
+ "learning_rate": 4.5739337623597174e-05,
3308
+ "loss": 0.5954,
3309
+ "step": 1100000
3310
+ },
3311
+ {
3312
+ "epoch": 25.52,
3313
+ "learning_rate": 4.527153565352715e-05,
3314
+ "loss": 0.5926,
3315
+ "step": 1102000
3316
+ },
3317
+ {
3318
+ "epoch": 25.56,
3319
+ "learning_rate": 4.480373368345712e-05,
3320
+ "loss": 0.5963,
3321
+ "step": 1104000
3322
+ },
3323
+ {
3324
+ "epoch": 25.61,
3325
+ "learning_rate": 4.433593171338709e-05,
3326
+ "loss": 0.5902,
3327
+ "step": 1106000
3328
+ },
3329
+ {
3330
+ "epoch": 25.66,
3331
+ "learning_rate": 4.3868129743317056e-05,
3332
+ "loss": 0.5952,
3333
+ "step": 1108000
3334
+ },
3335
+ {
3336
+ "epoch": 25.7,
3337
+ "learning_rate": 4.3400327773247025e-05,
3338
+ "loss": 0.5878,
3339
+ "step": 1110000
3340
+ },
3341
+ {
3342
+ "epoch": 25.75,
3343
+ "learning_rate": 4.2932525803176994e-05,
3344
+ "loss": 0.5926,
3345
+ "step": 1112000
3346
+ },
3347
+ {
3348
+ "epoch": 25.8,
3349
+ "learning_rate": 4.246472383310696e-05,
3350
+ "loss": 0.5854,
3351
+ "step": 1114000
3352
+ },
3353
+ {
3354
+ "epoch": 25.84,
3355
+ "learning_rate": 4.199692186303694e-05,
3356
+ "loss": 0.5916,
3357
+ "step": 1116000
3358
+ },
3359
+ {
3360
+ "epoch": 25.89,
3361
+ "learning_rate": 4.152911989296691e-05,
3362
+ "loss": 0.5869,
3363
+ "step": 1118000
3364
+ },
3365
+ {
3366
+ "epoch": 25.93,
3367
+ "learning_rate": 4.1061317922896876e-05,
3368
+ "loss": 0.5913,
3369
+ "step": 1120000
3370
+ },
3371
+ {
3372
+ "epoch": 25.98,
3373
+ "learning_rate": 4.0593515952826845e-05,
3374
+ "loss": 0.5822,
3375
+ "step": 1122000
3376
+ },
3377
+ {
3378
+ "epoch": 26.03,
3379
+ "learning_rate": 4.0125713982756814e-05,
3380
+ "loss": 0.5831,
3381
+ "step": 1124000
3382
+ },
3383
+ {
3384
+ "epoch": 26.07,
3385
+ "learning_rate": 3.965791201268678e-05,
3386
+ "loss": 0.5847,
3387
+ "step": 1126000
3388
+ },
3389
+ {
3390
+ "epoch": 26.12,
3391
+ "learning_rate": 3.919011004261675e-05,
3392
+ "loss": 0.5828,
3393
+ "step": 1128000
3394
+ },
3395
+ {
3396
+ "epoch": 26.17,
3397
+ "learning_rate": 3.872230807254672e-05,
3398
+ "loss": 0.5825,
3399
+ "step": 1130000
3400
+ },
3401
+ {
3402
+ "epoch": 26.21,
3403
+ "learning_rate": 3.825450610247669e-05,
3404
+ "loss": 0.5848,
3405
+ "step": 1132000
3406
+ },
3407
+ {
3408
+ "epoch": 26.26,
3409
+ "learning_rate": 3.778670413240667e-05,
3410
+ "loss": 0.5866,
3411
+ "step": 1134000
3412
+ },
3413
+ {
3414
+ "epoch": 26.31,
3415
+ "learning_rate": 3.7318902162336634e-05,
3416
+ "loss": 0.5832,
3417
+ "step": 1136000
3418
+ },
3419
+ {
3420
+ "epoch": 26.35,
3421
+ "learning_rate": 3.685110019226661e-05,
3422
+ "loss": 0.58,
3423
+ "step": 1138000
3424
+ },
3425
+ {
3426
+ "epoch": 26.4,
3427
+ "learning_rate": 3.638329822219658e-05,
3428
+ "loss": 0.5767,
3429
+ "step": 1140000
3430
+ },
3431
+ {
3432
+ "epoch": 26.44,
3433
+ "learning_rate": 3.591549625212655e-05,
3434
+ "loss": 0.5792,
3435
+ "step": 1142000
3436
+ },
3437
+ {
3438
+ "epoch": 26.49,
3439
+ "learning_rate": 3.5447694282056516e-05,
3440
+ "loss": 0.5764,
3441
+ "step": 1144000
3442
+ },
3443
+ {
3444
+ "epoch": 26.54,
3445
+ "learning_rate": 3.4979892311986485e-05,
3446
+ "loss": 0.5794,
3447
+ "step": 1146000
3448
+ },
3449
+ {
3450
+ "epoch": 26.58,
3451
+ "learning_rate": 3.451209034191646e-05,
3452
+ "loss": 0.5738,
3453
+ "step": 1148000
3454
+ },
3455
+ {
3456
+ "epoch": 26.63,
3457
+ "learning_rate": 3.404428837184643e-05,
3458
+ "loss": 0.5822,
3459
+ "step": 1150000
3460
+ },
3461
+ {
3462
+ "epoch": 26.68,
3463
+ "learning_rate": 3.35764864017764e-05,
3464
+ "loss": 0.5734,
3465
+ "step": 1152000
3466
+ },
3467
+ {
3468
+ "epoch": 26.72,
3469
+ "learning_rate": 3.310868443170637e-05,
3470
+ "loss": 0.5794,
3471
+ "step": 1154000
3472
+ },
3473
+ {
3474
+ "epoch": 26.77,
3475
+ "learning_rate": 3.2640882461636336e-05,
3476
+ "loss": 0.5853,
3477
+ "step": 1156000
3478
+ },
3479
+ {
3480
+ "epoch": 26.81,
3481
+ "learning_rate": 3.2173080491566305e-05,
3482
+ "loss": 0.5842,
3483
+ "step": 1158000
3484
+ },
3485
+ {
3486
+ "epoch": 26.86,
3487
+ "learning_rate": 3.170527852149628e-05,
3488
+ "loss": 0.5847,
3489
+ "step": 1160000
3490
+ },
3491
+ {
3492
+ "epoch": 26.91,
3493
+ "learning_rate": 3.123747655142625e-05,
3494
+ "loss": 0.5786,
3495
+ "step": 1162000
3496
+ },
3497
+ {
3498
+ "epoch": 26.95,
3499
+ "learning_rate": 3.076967458135622e-05,
3500
+ "loss": 0.5818,
3501
+ "step": 1164000
3502
+ },
3503
+ {
3504
+ "epoch": 27.0,
3505
+ "learning_rate": 3.030187261128619e-05,
3506
+ "loss": 0.5722,
3507
+ "step": 1166000
3508
+ },
3509
+ {
3510
+ "epoch": 27.05,
3511
+ "learning_rate": 2.983407064121616e-05,
3512
+ "loss": 0.5726,
3513
+ "step": 1168000
3514
+ },
3515
+ {
3516
+ "epoch": 27.09,
3517
+ "learning_rate": 2.936626867114613e-05,
3518
+ "loss": 0.5745,
3519
+ "step": 1170000
3520
+ },
3521
+ {
3522
+ "epoch": 27.14,
3523
+ "learning_rate": 2.8898466701076097e-05,
3524
+ "loss": 0.5655,
3525
+ "step": 1172000
3526
+ },
3527
+ {
3528
+ "epoch": 27.19,
3529
+ "learning_rate": 2.843066473100607e-05,
3530
+ "loss": 0.5747,
3531
+ "step": 1174000
3532
+ },
3533
+ {
3534
+ "epoch": 27.23,
3535
+ "learning_rate": 2.7962862760936038e-05,
3536
+ "loss": 0.5734,
3537
+ "step": 1176000
3538
+ },
3539
+ {
3540
+ "epoch": 27.28,
3541
+ "learning_rate": 2.7495060790866007e-05,
3542
+ "loss": 0.5752,
3543
+ "step": 1178000
3544
+ },
3545
+ {
3546
+ "epoch": 27.32,
3547
+ "learning_rate": 2.7027258820795976e-05,
3548
+ "loss": 0.5784,
3549
+ "step": 1180000
3550
+ },
3551
+ {
3552
+ "epoch": 27.37,
3553
+ "learning_rate": 2.6559456850725945e-05,
3554
+ "loss": 0.5667,
3555
+ "step": 1182000
3556
+ },
3557
+ {
3558
+ "epoch": 27.42,
3559
+ "learning_rate": 2.609165488065592e-05,
3560
+ "loss": 0.5748,
3561
+ "step": 1184000
3562
+ },
3563
+ {
3564
+ "epoch": 27.46,
3565
+ "learning_rate": 2.562385291058589e-05,
3566
+ "loss": 0.5762,
3567
+ "step": 1186000
3568
+ },
3569
+ {
3570
+ "epoch": 27.51,
3571
+ "learning_rate": 2.5156050940515858e-05,
3572
+ "loss": 0.5783,
3573
+ "step": 1188000
3574
+ },
3575
+ {
3576
+ "epoch": 27.56,
3577
+ "learning_rate": 2.4688248970445827e-05,
3578
+ "loss": 0.5668,
3579
+ "step": 1190000
3580
+ },
3581
+ {
3582
+ "epoch": 27.6,
3583
+ "learning_rate": 2.42204470003758e-05,
3584
+ "loss": 0.5671,
3585
+ "step": 1192000
3586
+ },
3587
+ {
3588
+ "epoch": 27.65,
3589
+ "learning_rate": 2.3752645030305768e-05,
3590
+ "loss": 0.5688,
3591
+ "step": 1194000
3592
+ },
3593
+ {
3594
+ "epoch": 27.69,
3595
+ "learning_rate": 2.328484306023574e-05,
3596
+ "loss": 0.5643,
3597
+ "step": 1196000
3598
+ },
3599
+ {
3600
+ "epoch": 27.74,
3601
+ "learning_rate": 2.281704109016571e-05,
3602
+ "loss": 0.5688,
3603
+ "step": 1198000
3604
+ },
3605
+ {
3606
+ "epoch": 27.79,
3607
+ "learning_rate": 2.2349239120095678e-05,
3608
+ "loss": 0.5651,
3609
+ "step": 1200000
3610
+ },
3611
+ {
3612
+ "epoch": 27.83,
3613
+ "learning_rate": 2.188143715002565e-05,
3614
+ "loss": 0.5705,
3615
+ "step": 1202000
3616
+ },
3617
+ {
3618
+ "epoch": 27.88,
3619
+ "learning_rate": 2.141363517995562e-05,
3620
+ "loss": 0.5684,
3621
+ "step": 1204000
3622
+ },
3623
+ {
3624
+ "epoch": 27.93,
3625
+ "learning_rate": 2.0945833209885588e-05,
3626
+ "loss": 0.567,
3627
+ "step": 1206000
3628
+ },
3629
+ {
3630
+ "epoch": 27.97,
3631
+ "learning_rate": 2.0478031239815557e-05,
3632
+ "loss": 0.5711,
3633
+ "step": 1208000
3634
+ },
3635
+ {
3636
+ "epoch": 28.02,
3637
+ "learning_rate": 2.0010229269745533e-05,
3638
+ "loss": 0.5684,
3639
+ "step": 1210000
3640
+ },
3641
+ {
3642
+ "epoch": 28.06,
3643
+ "learning_rate": 1.95424272996755e-05,
3644
+ "loss": 0.5604,
3645
+ "step": 1212000
3646
+ },
3647
+ {
3648
+ "epoch": 28.11,
3649
+ "learning_rate": 1.907462532960547e-05,
3650
+ "loss": 0.5615,
3651
+ "step": 1214000
3652
+ },
3653
+ {
3654
+ "epoch": 28.16,
3655
+ "learning_rate": 1.860682335953544e-05,
3656
+ "loss": 0.5679,
3657
+ "step": 1216000
3658
+ },
3659
+ {
3660
+ "epoch": 28.2,
3661
+ "learning_rate": 1.813902138946541e-05,
3662
+ "loss": 0.5644,
3663
+ "step": 1218000
3664
+ },
3665
+ {
3666
+ "epoch": 28.25,
3667
+ "learning_rate": 1.767121941939538e-05,
3668
+ "loss": 0.5663,
3669
+ "step": 1220000
3670
+ },
3671
+ {
3672
+ "epoch": 28.3,
3673
+ "learning_rate": 1.720341744932535e-05,
3674
+ "loss": 0.5584,
3675
+ "step": 1222000
3676
+ },
3677
+ {
3678
+ "epoch": 28.34,
3679
+ "learning_rate": 1.6735615479255318e-05,
3680
+ "loss": 0.558,
3681
+ "step": 1224000
3682
+ },
3683
+ {
3684
+ "epoch": 28.39,
3685
+ "learning_rate": 1.626781350918529e-05,
3686
+ "loss": 0.5575,
3687
+ "step": 1226000
3688
+ },
3689
+ {
3690
+ "epoch": 28.44,
3691
+ "learning_rate": 1.580001153911526e-05,
3692
+ "loss": 0.5728,
3693
+ "step": 1228000
3694
+ },
3695
+ {
3696
+ "epoch": 28.48,
3697
+ "learning_rate": 1.533220956904523e-05,
3698
+ "loss": 0.5653,
3699
+ "step": 1230000
3700
+ },
3701
+ {
3702
+ "epoch": 28.53,
3703
+ "learning_rate": 1.48644075989752e-05,
3704
+ "loss": 0.5603,
3705
+ "step": 1232000
3706
+ },
3707
+ {
3708
+ "epoch": 28.57,
3709
+ "learning_rate": 1.4396605628905172e-05,
3710
+ "loss": 0.5613,
3711
+ "step": 1234000
3712
+ },
3713
+ {
3714
+ "epoch": 28.62,
3715
+ "learning_rate": 1.3928803658835141e-05,
3716
+ "loss": 0.5563,
3717
+ "step": 1236000
3718
+ },
3719
+ {
3720
+ "epoch": 28.67,
3721
+ "learning_rate": 1.346100168876511e-05,
3722
+ "loss": 0.5705,
3723
+ "step": 1238000
3724
+ },
3725
+ {
3726
+ "epoch": 28.71,
3727
+ "learning_rate": 1.299319971869508e-05,
3728
+ "loss": 0.5568,
3729
+ "step": 1240000
3730
+ },
3731
+ {
3732
+ "epoch": 28.76,
3733
+ "learning_rate": 1.252539774862505e-05,
3734
+ "loss": 0.5517,
3735
+ "step": 1242000
3736
+ },
3737
+ {
3738
+ "epoch": 28.81,
3739
+ "learning_rate": 1.2057595778555022e-05,
3740
+ "loss": 0.5647,
3741
+ "step": 1244000
3742
+ },
3743
+ {
3744
+ "epoch": 28.85,
3745
+ "learning_rate": 1.158979380848499e-05,
3746
+ "loss": 0.5551,
3747
+ "step": 1246000
3748
+ },
3749
+ {
3750
+ "epoch": 28.9,
3751
+ "learning_rate": 1.1121991838414961e-05,
3752
+ "loss": 0.5598,
3753
+ "step": 1248000
3754
+ },
3755
+ {
3756
+ "epoch": 28.94,
3757
+ "learning_rate": 1.0654189868344932e-05,
3758
+ "loss": 0.562,
3759
+ "step": 1250000
3760
+ },
3761
+ {
3762
+ "epoch": 28.99,
3763
+ "learning_rate": 1.0186387898274902e-05,
3764
+ "loss": 0.5563,
3765
+ "step": 1252000
3766
+ },
3767
+ {
3768
+ "epoch": 29.04,
3769
+ "learning_rate": 9.718585928204871e-06,
3770
+ "loss": 0.5606,
3771
+ "step": 1254000
3772
+ },
3773
+ {
3774
+ "epoch": 29.08,
3775
+ "learning_rate": 9.250783958134842e-06,
3776
+ "loss": 0.5509,
3777
+ "step": 1256000
3778
+ },
3779
+ {
3780
+ "epoch": 29.13,
3781
+ "learning_rate": 8.782981988064812e-06,
3782
+ "loss": 0.551,
3783
+ "step": 1258000
3784
+ },
3785
+ {
3786
+ "epoch": 29.18,
3787
+ "learning_rate": 8.315180017994783e-06,
3788
+ "loss": 0.5548,
3789
+ "step": 1260000
3790
+ },
3791
+ {
3792
+ "epoch": 29.22,
3793
+ "learning_rate": 7.847378047924752e-06,
3794
+ "loss": 0.5562,
3795
+ "step": 1262000
3796
+ },
3797
+ {
3798
+ "epoch": 29.27,
3799
+ "learning_rate": 7.3795760778547214e-06,
3800
+ "loss": 0.5563,
3801
+ "step": 1264000
3802
+ },
3803
+ {
3804
+ "epoch": 29.32,
3805
+ "learning_rate": 6.911774107784692e-06,
3806
+ "loss": 0.5551,
3807
+ "step": 1266000
3808
+ },
3809
+ {
3810
+ "epoch": 29.36,
3811
+ "learning_rate": 6.443972137714662e-06,
3812
+ "loss": 0.555,
3813
+ "step": 1268000
3814
+ },
3815
+ {
3816
+ "epoch": 29.41,
3817
+ "learning_rate": 5.976170167644632e-06,
3818
+ "loss": 0.554,
3819
+ "step": 1270000
3820
+ },
3821
+ {
3822
+ "epoch": 29.45,
3823
+ "learning_rate": 5.508368197574602e-06,
3824
+ "loss": 0.5522,
3825
+ "step": 1272000
3826
+ },
3827
+ {
3828
+ "epoch": 29.5,
3829
+ "learning_rate": 5.0405662275045725e-06,
3830
+ "loss": 0.5522,
3831
+ "step": 1274000
3832
+ },
3833
+ {
3834
+ "epoch": 29.55,
3835
+ "learning_rate": 4.572764257434542e-06,
3836
+ "loss": 0.5601,
3837
+ "step": 1276000
3838
+ },
3839
+ {
3840
+ "epoch": 29.59,
3841
+ "learning_rate": 4.104962287364513e-06,
3842
+ "loss": 0.5578,
3843
+ "step": 1278000
3844
+ },
3845
+ {
3846
+ "epoch": 29.64,
3847
+ "learning_rate": 3.6371603172944825e-06,
3848
+ "loss": 0.5602,
3849
+ "step": 1280000
3850
+ },
3851
+ {
3852
+ "epoch": 29.69,
3853
+ "learning_rate": 3.1693583472244526e-06,
3854
+ "loss": 0.5517,
3855
+ "step": 1282000
3856
+ },
3857
+ {
3858
+ "epoch": 29.73,
3859
+ "learning_rate": 2.701556377154423e-06,
3860
+ "loss": 0.5538,
3861
+ "step": 1284000
3862
+ },
3863
+ {
3864
+ "epoch": 29.78,
3865
+ "learning_rate": 2.233754407084393e-06,
3866
+ "loss": 0.549,
3867
+ "step": 1286000
3868
+ },
3869
+ {
3870
+ "epoch": 29.82,
3871
+ "learning_rate": 1.765952437014363e-06,
3872
+ "loss": 0.5595,
3873
+ "step": 1288000
3874
+ },
3875
+ {
3876
+ "epoch": 29.87,
3877
+ "learning_rate": 1.2981504669443331e-06,
3878
+ "loss": 0.5546,
3879
+ "step": 1290000
3880
+ },
3881
+ {
3882
+ "epoch": 29.92,
3883
+ "learning_rate": 8.303484968743031e-07,
3884
+ "loss": 0.5547,
3885
+ "step": 1292000
3886
+ },
3887
+ {
3888
+ "epoch": 29.96,
3889
+ "learning_rate": 3.6254652680427316e-07,
3890
+ "loss": 0.5494,
3891
+ "step": 1294000
3892
+ }
3893
+ ],
3894
+ "logging_steps": 2000,
3895
+ "max_steps": 1295550,
3896
+ "num_train_epochs": 30,
3897
+ "save_steps": 500,
3898
+ "total_flos": 2.6480449905256835e+21,
3899
+ "trial_name": null,
3900
+ "trial_params": null
3901
+ }
checkpoint-1295550/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:15b4a29d89f83e50e17158ee0f4ffe9724daa5ff51dfccdeac7b5e2c6a3b564e
3
+ size 4600
checkpoint-1295550/vocab.json ADDED
@@ -0,0 +1,53 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "\b": 0,
3
+ "\"": 2,
4
+ "[PAD]": 50,
5
+ "[UNK]": 49,
6
+ "\\": 3,
7
+ "a": 4,
8
+ "b": 5,
9
+ "c": 6,
10
+ "d": 7,
11
+ "e": 8,
12
+ "f": 9,
13
+ "g": 10,
14
+ "h": 11,
15
+ "i": 12,
16
+ "j": 13,
17
+ "k": 14,
18
+ "l": 15,
19
+ "m": 16,
20
+ "n": 17,
21
+ "o": 18,
22
+ "p": 19,
23
+ "q": 20,
24
+ "r": 21,
25
+ "s": 22,
26
+ "t": 23,
27
+ "u": 24,
28
+ "v": 25,
29
+ "w": 26,
30
+ "x": 27,
31
+ "y": 28,
32
+ "z": 29,
33
+ "|": 1,
34
+ "ç": 30,
35
+ "é": 31,
36
+ "ş": 32,
37
+ "υ": 33,
38
+ "а": 34,
39
+ "в": 35,
40
+ "г": 36,
41
+ "д": 37,
42
+ "е": 38,
43
+ "к": 39,
44
+ "л": 40,
45
+ "н": 41,
46
+ "с": 42,
47
+ "ф": 43,
48
+ "ч": 44,
49
+ "​": 45,
50
+ "’": 46,
51
+ "⁄": 47,
52
+ "�": 48
53
+ }
config.json ADDED
@@ -0,0 +1,117 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/wav2vec2-large-xlsr-53",
3
+ "activation_dropout": 0.0,
4
+ "adapter_attn_dim": null,
5
+ "adapter_kernel_size": 3,
6
+ "adapter_stride": 2,
7
+ "add_adapter": false,
8
+ "apply_spec_augment": true,
9
+ "architectures": [
10
+ "Wav2Vec2ForCTC"
11
+ ],
12
+ "attention_dropout": 0.05,
13
+ "bos_token_id": 1,
14
+ "classifier_proj_size": 256,
15
+ "codevector_dim": 768,
16
+ "contrastive_logits_temperature": 0.1,
17
+ "conv_bias": true,
18
+ "conv_dim": [
19
+ 512,
20
+ 512,
21
+ 512,
22
+ 512,
23
+ 512,
24
+ 512,
25
+ 512
26
+ ],
27
+ "conv_kernel": [
28
+ 10,
29
+ 3,
30
+ 3,
31
+ 3,
32
+ 3,
33
+ 2,
34
+ 2
35
+ ],
36
+ "conv_stride": [
37
+ 5,
38
+ 2,
39
+ 2,
40
+ 2,
41
+ 2,
42
+ 2,
43
+ 2
44
+ ],
45
+ "ctc_loss_reduction": "mean",
46
+ "ctc_zero_infinity": true,
47
+ "diversity_loss_weight": 0.1,
48
+ "do_stable_layer_norm": true,
49
+ "eos_token_id": 2,
50
+ "feat_extract_activation": "gelu",
51
+ "feat_extract_dropout": 0.0,
52
+ "feat_extract_norm": "layer",
53
+ "feat_proj_dropout": 0.05,
54
+ "feat_quantizer_dropout": 0.0,
55
+ "final_dropout": 0.0,
56
+ "gradient_checkpointing": false,
57
+ "hidden_act": "gelu",
58
+ "hidden_dropout": 0.05,
59
+ "hidden_size": 1024,
60
+ "initializer_range": 0.02,
61
+ "intermediate_size": 4096,
62
+ "layer_norm_eps": 1e-05,
63
+ "layerdrop": 0.05,
64
+ "mask_channel_length": 10,
65
+ "mask_channel_min_space": 1,
66
+ "mask_channel_other": 0.0,
67
+ "mask_channel_prob": 0.0,
68
+ "mask_channel_selection": "static",
69
+ "mask_feature_length": 10,
70
+ "mask_feature_min_masks": 0,
71
+ "mask_feature_prob": 0.0,
72
+ "mask_time_length": 10,
73
+ "mask_time_min_masks": 2,
74
+ "mask_time_min_space": 1,
75
+ "mask_time_other": 0.0,
76
+ "mask_time_prob": 0.05,
77
+ "mask_time_selection": "static",
78
+ "model_type": "wav2vec2",
79
+ "num_adapter_layers": 3,
80
+ "num_attention_heads": 16,
81
+ "num_codevector_groups": 2,
82
+ "num_codevectors_per_group": 320,
83
+ "num_conv_pos_embedding_groups": 16,
84
+ "num_conv_pos_embeddings": 128,
85
+ "num_feat_extract_layers": 7,
86
+ "num_hidden_layers": 24,
87
+ "num_negatives": 100,
88
+ "output_hidden_size": 1024,
89
+ "pad_token_id": 50,
90
+ "proj_codevector_dim": 768,
91
+ "tdnn_dilation": [
92
+ 1,
93
+ 2,
94
+ 3,
95
+ 1,
96
+ 1
97
+ ],
98
+ "tdnn_dim": [
99
+ 512,
100
+ 512,
101
+ 512,
102
+ 512,
103
+ 1500
104
+ ],
105
+ "tdnn_kernel": [
106
+ 5,
107
+ 3,
108
+ 3,
109
+ 1,
110
+ 1
111
+ ],
112
+ "torch_dtype": "float32",
113
+ "transformers_version": "4.35.0",
114
+ "use_weighted_layer_sum": false,
115
+ "vocab_size": 53,
116
+ "xvector_output_dim": 512
117
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e8ca98ff5bd5909097ec35838b22fce9f53fe21dc0ce7d473d8ffc93958b9d50
3
+ size 1262024780
preprocessor_config.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "feature_extractor_type": "Wav2Vec2FeatureExtractor",
4
+ "feature_size": 1,
5
+ "padding_side": "right",
6
+ "padding_value": 0,
7
+ "processor_class": "Wav2Vec2Processor",
8
+ "return_attention_mask": true,
9
+ "sampling_rate": 16000
10
+ }
special_tokens_map.json ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": true,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "</s>",
11
+ "lstrip": false,
12
+ "normalized": true,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": {
17
+ "content": "[PAD]",
18
+ "lstrip": true,
19
+ "normalized": false,
20
+ "rstrip": true,
21
+ "single_word": false
22
+ },
23
+ "unk_token": {
24
+ "content": "[UNK]",
25
+ "lstrip": true,
26
+ "normalized": false,
27
+ "rstrip": true,
28
+ "single_word": false
29
+ }
30
+ }
tokenizer_config.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "49": {
4
+ "content": "[UNK]",
5
+ "lstrip": true,
6
+ "normalized": false,
7
+ "rstrip": true,
8
+ "single_word": false,
9
+ "special": false
10
+ },
11
+ "50": {
12
+ "content": "[PAD]",
13
+ "lstrip": true,
14
+ "normalized": false,
15
+ "rstrip": true,
16
+ "single_word": false,
17
+ "special": false
18
+ },
19
+ "51": {
20
+ "content": "<s>",
21
+ "lstrip": false,
22
+ "normalized": true,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "52": {
28
+ "content": "</s>",
29
+ "lstrip": false,
30
+ "normalized": true,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ }
35
+ },
36
+ "bos_token": "<s>",
37
+ "clean_up_tokenization_spaces": true,
38
+ "config": null,
39
+ "do_lower_case": false,
40
+ "eos_token": "</s>",
41
+ "model_max_length": 1000000000000000019884624838656,
42
+ "pad_token": "[PAD]",
43
+ "processor_class": "Wav2Vec2Processor",
44
+ "replace_word_delimiter_char": " ",
45
+ "target_lang": null,
46
+ "tokenizer_class": "Wav2Vec2CTCTokenizer",
47
+ "tokenizer_type": "wav2vec2",
48
+ "trust_remote_code": false,
49
+ "unk_token": "[UNK]",
50
+ "word_delimiter_token": "|"
51
+ }
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 30.0,
3
+ "train_loss": 0.7350748663054241,
4
+ "train_runtime": 658563.1153,
5
+ "train_samples": 690961,
6
+ "train_samples_per_second": 31.476,
7
+ "train_steps_per_second": 1.967
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,3910 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 29.99965266119415,
5
+ "eval_steps": 500,
6
+ "global_step": 1295550,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.05,
13
+ "learning_rate": 4.6310589688175354e-05,
14
+ "loss": 4.5055,
15
+ "step": 2000
16
+ },
17
+ {
18
+ "epoch": 0.09,
19
+ "learning_rate": 9.262117937635071e-05,
20
+ "loss": 1.4851,
21
+ "step": 4000
22
+ },
23
+ {
24
+ "epoch": 0.14,
25
+ "learning_rate": 0.00013893176906452608,
26
+ "loss": 1.1808,
27
+ "step": 6000
28
+ },
29
+ {
30
+ "epoch": 0.19,
31
+ "learning_rate": 0.00018524235875270141,
32
+ "loss": 1.132,
33
+ "step": 8000
34
+ },
35
+ {
36
+ "epoch": 0.23,
37
+ "learning_rate": 0.0002315529484408768,
38
+ "loss": 1.1349,
39
+ "step": 10000
40
+ },
41
+ {
42
+ "epoch": 0.28,
43
+ "learning_rate": 0.00027786353812905216,
44
+ "loss": 1.1422,
45
+ "step": 12000
46
+ },
47
+ {
48
+ "epoch": 0.32,
49
+ "learning_rate": 0.0002997558073716234,
50
+ "loss": 1.1513,
51
+ "step": 14000
52
+ },
53
+ {
54
+ "epoch": 0.37,
55
+ "learning_rate": 0.0002992880054015534,
56
+ "loss": 1.142,
57
+ "step": 16000
58
+ },
59
+ {
60
+ "epoch": 0.42,
61
+ "learning_rate": 0.00029882020343148334,
62
+ "loss": 1.1303,
63
+ "step": 18000
64
+ },
65
+ {
66
+ "epoch": 0.46,
67
+ "learning_rate": 0.0002983524014614133,
68
+ "loss": 1.1215,
69
+ "step": 20000
70
+ },
71
+ {
72
+ "epoch": 0.51,
73
+ "learning_rate": 0.0002978845994913433,
74
+ "loss": 1.109,
75
+ "step": 22000
76
+ },
77
+ {
78
+ "epoch": 0.56,
79
+ "learning_rate": 0.00029741679752127325,
80
+ "loss": 1.1034,
81
+ "step": 24000
82
+ },
83
+ {
84
+ "epoch": 0.6,
85
+ "learning_rate": 0.0002969489955512032,
86
+ "loss": 1.0998,
87
+ "step": 26000
88
+ },
89
+ {
90
+ "epoch": 0.65,
91
+ "learning_rate": 0.00029648119358113324,
92
+ "loss": 1.0979,
93
+ "step": 28000
94
+ },
95
+ {
96
+ "epoch": 0.69,
97
+ "learning_rate": 0.00029601339161106316,
98
+ "loss": 1.083,
99
+ "step": 30000
100
+ },
101
+ {
102
+ "epoch": 0.74,
103
+ "learning_rate": 0.0002955455896409932,
104
+ "loss": 1.0741,
105
+ "step": 32000
106
+ },
107
+ {
108
+ "epoch": 0.79,
109
+ "learning_rate": 0.0002950777876709231,
110
+ "loss": 1.0741,
111
+ "step": 34000
112
+ },
113
+ {
114
+ "epoch": 0.83,
115
+ "learning_rate": 0.0002946099857008531,
116
+ "loss": 1.0707,
117
+ "step": 36000
118
+ },
119
+ {
120
+ "epoch": 0.88,
121
+ "learning_rate": 0.00029414218373078303,
122
+ "loss": 1.0689,
123
+ "step": 38000
124
+ },
125
+ {
126
+ "epoch": 0.93,
127
+ "learning_rate": 0.00029367438176071306,
128
+ "loss": 1.0536,
129
+ "step": 40000
130
+ },
131
+ {
132
+ "epoch": 0.97,
133
+ "learning_rate": 0.00029320657979064297,
134
+ "loss": 1.0556,
135
+ "step": 42000
136
+ },
137
+ {
138
+ "epoch": 1.02,
139
+ "learning_rate": 0.000292738777820573,
140
+ "loss": 1.0358,
141
+ "step": 44000
142
+ },
143
+ {
144
+ "epoch": 1.07,
145
+ "learning_rate": 0.0002922709758505029,
146
+ "loss": 1.018,
147
+ "step": 46000
148
+ },
149
+ {
150
+ "epoch": 1.11,
151
+ "learning_rate": 0.00029180317388043293,
152
+ "loss": 1.0162,
153
+ "step": 48000
154
+ },
155
+ {
156
+ "epoch": 1.16,
157
+ "learning_rate": 0.00029133537191036285,
158
+ "loss": 1.017,
159
+ "step": 50000
160
+ },
161
+ {
162
+ "epoch": 1.2,
163
+ "learning_rate": 0.00029086756994029287,
164
+ "loss": 1.0145,
165
+ "step": 52000
166
+ },
167
+ {
168
+ "epoch": 1.25,
169
+ "learning_rate": 0.0002903997679702228,
170
+ "loss": 1.0103,
171
+ "step": 54000
172
+ },
173
+ {
174
+ "epoch": 1.3,
175
+ "learning_rate": 0.0002899319660001528,
176
+ "loss": 1.0046,
177
+ "step": 56000
178
+ },
179
+ {
180
+ "epoch": 1.34,
181
+ "learning_rate": 0.0002894641640300827,
182
+ "loss": 1.01,
183
+ "step": 58000
184
+ },
185
+ {
186
+ "epoch": 1.39,
187
+ "learning_rate": 0.00028899636206001274,
188
+ "loss": 1.0146,
189
+ "step": 60000
190
+ },
191
+ {
192
+ "epoch": 1.44,
193
+ "learning_rate": 0.00028852856008994266,
194
+ "loss": 1.0122,
195
+ "step": 62000
196
+ },
197
+ {
198
+ "epoch": 1.48,
199
+ "learning_rate": 0.0002880607581198727,
200
+ "loss": 0.9961,
201
+ "step": 64000
202
+ },
203
+ {
204
+ "epoch": 1.53,
205
+ "learning_rate": 0.00028759295614980265,
206
+ "loss": 0.9914,
207
+ "step": 66000
208
+ },
209
+ {
210
+ "epoch": 1.57,
211
+ "learning_rate": 0.0002871251541797326,
212
+ "loss": 0.9936,
213
+ "step": 68000
214
+ },
215
+ {
216
+ "epoch": 1.62,
217
+ "learning_rate": 0.0002866573522096626,
218
+ "loss": 0.9998,
219
+ "step": 70000
220
+ },
221
+ {
222
+ "epoch": 1.67,
223
+ "learning_rate": 0.00028618955023959256,
224
+ "loss": 0.9833,
225
+ "step": 72000
226
+ },
227
+ {
228
+ "epoch": 1.71,
229
+ "learning_rate": 0.0002857217482695225,
230
+ "loss": 0.9901,
231
+ "step": 74000
232
+ },
233
+ {
234
+ "epoch": 1.76,
235
+ "learning_rate": 0.0002852539462994525,
236
+ "loss": 0.9905,
237
+ "step": 76000
238
+ },
239
+ {
240
+ "epoch": 1.81,
241
+ "learning_rate": 0.00028478614432938246,
242
+ "loss": 0.9845,
243
+ "step": 78000
244
+ },
245
+ {
246
+ "epoch": 1.85,
247
+ "learning_rate": 0.00028431834235931243,
248
+ "loss": 0.9886,
249
+ "step": 80000
250
+ },
251
+ {
252
+ "epoch": 1.9,
253
+ "learning_rate": 0.0002838505403892424,
254
+ "loss": 0.9848,
255
+ "step": 82000
256
+ },
257
+ {
258
+ "epoch": 1.95,
259
+ "learning_rate": 0.00028338273841917237,
260
+ "loss": 0.9827,
261
+ "step": 84000
262
+ },
263
+ {
264
+ "epoch": 1.99,
265
+ "learning_rate": 0.00028291493644910234,
266
+ "loss": 0.9764,
267
+ "step": 86000
268
+ },
269
+ {
270
+ "epoch": 2.04,
271
+ "learning_rate": 0.0002824471344790323,
272
+ "loss": 0.9546,
273
+ "step": 88000
274
+ },
275
+ {
276
+ "epoch": 2.08,
277
+ "learning_rate": 0.0002819793325089623,
278
+ "loss": 0.9388,
279
+ "step": 90000
280
+ },
281
+ {
282
+ "epoch": 2.13,
283
+ "learning_rate": 0.00028151153053889225,
284
+ "loss": 0.9432,
285
+ "step": 92000
286
+ },
287
+ {
288
+ "epoch": 2.18,
289
+ "learning_rate": 0.0002810437285688222,
290
+ "loss": 0.9591,
291
+ "step": 94000
292
+ },
293
+ {
294
+ "epoch": 2.22,
295
+ "learning_rate": 0.0002805759265987522,
296
+ "loss": 0.9431,
297
+ "step": 96000
298
+ },
299
+ {
300
+ "epoch": 2.27,
301
+ "learning_rate": 0.00028010812462868215,
302
+ "loss": 0.9388,
303
+ "step": 98000
304
+ },
305
+ {
306
+ "epoch": 2.32,
307
+ "learning_rate": 0.0002796403226586121,
308
+ "loss": 0.9421,
309
+ "step": 100000
310
+ },
311
+ {
312
+ "epoch": 2.36,
313
+ "learning_rate": 0.0002791725206885421,
314
+ "loss": 0.9452,
315
+ "step": 102000
316
+ },
317
+ {
318
+ "epoch": 2.41,
319
+ "learning_rate": 0.0002787047187184721,
320
+ "loss": 0.9535,
321
+ "step": 104000
322
+ },
323
+ {
324
+ "epoch": 2.45,
325
+ "learning_rate": 0.00027823691674840203,
326
+ "loss": 0.9373,
327
+ "step": 106000
328
+ },
329
+ {
330
+ "epoch": 2.5,
331
+ "learning_rate": 0.00027776911477833205,
332
+ "loss": 0.9345,
333
+ "step": 108000
334
+ },
335
+ {
336
+ "epoch": 2.55,
337
+ "learning_rate": 0.00027730131280826197,
338
+ "loss": 0.9478,
339
+ "step": 110000
340
+ },
341
+ {
342
+ "epoch": 2.59,
343
+ "learning_rate": 0.000276833510838192,
344
+ "loss": 0.9445,
345
+ "step": 112000
346
+ },
347
+ {
348
+ "epoch": 2.64,
349
+ "learning_rate": 0.0002763657088681219,
350
+ "loss": 0.9297,
351
+ "step": 114000
352
+ },
353
+ {
354
+ "epoch": 2.69,
355
+ "learning_rate": 0.0002758979068980519,
356
+ "loss": 0.9294,
357
+ "step": 116000
358
+ },
359
+ {
360
+ "epoch": 2.73,
361
+ "learning_rate": 0.00027543010492798184,
362
+ "loss": 0.9273,
363
+ "step": 118000
364
+ },
365
+ {
366
+ "epoch": 2.78,
367
+ "learning_rate": 0.00027496230295791186,
368
+ "loss": 0.9242,
369
+ "step": 120000
370
+ },
371
+ {
372
+ "epoch": 2.83,
373
+ "learning_rate": 0.0002744945009878418,
374
+ "loss": 0.9228,
375
+ "step": 122000
376
+ },
377
+ {
378
+ "epoch": 2.87,
379
+ "learning_rate": 0.0002740266990177718,
380
+ "loss": 0.9316,
381
+ "step": 124000
382
+ },
383
+ {
384
+ "epoch": 2.92,
385
+ "learning_rate": 0.0002735588970477017,
386
+ "loss": 0.9448,
387
+ "step": 126000
388
+ },
389
+ {
390
+ "epoch": 2.96,
391
+ "learning_rate": 0.00027309109507763174,
392
+ "loss": 0.9253,
393
+ "step": 128000
394
+ },
395
+ {
396
+ "epoch": 3.01,
397
+ "learning_rate": 0.00027262329310756165,
398
+ "loss": 0.9197,
399
+ "step": 130000
400
+ },
401
+ {
402
+ "epoch": 3.06,
403
+ "learning_rate": 0.0002721554911374917,
404
+ "loss": 0.9025,
405
+ "step": 132000
406
+ },
407
+ {
408
+ "epoch": 3.1,
409
+ "learning_rate": 0.0002716876891674216,
410
+ "loss": 0.8983,
411
+ "step": 134000
412
+ },
413
+ {
414
+ "epoch": 3.15,
415
+ "learning_rate": 0.0002712198871973516,
416
+ "loss": 0.9013,
417
+ "step": 136000
418
+ },
419
+ {
420
+ "epoch": 3.2,
421
+ "learning_rate": 0.0002707520852272816,
422
+ "loss": 0.9019,
423
+ "step": 138000
424
+ },
425
+ {
426
+ "epoch": 3.24,
427
+ "learning_rate": 0.00027028428325721155,
428
+ "loss": 0.9016,
429
+ "step": 140000
430
+ },
431
+ {
432
+ "epoch": 3.29,
433
+ "learning_rate": 0.0002698164812871415,
434
+ "loss": 0.8979,
435
+ "step": 142000
436
+ },
437
+ {
438
+ "epoch": 3.33,
439
+ "learning_rate": 0.0002693486793170715,
440
+ "loss": 0.8961,
441
+ "step": 144000
442
+ },
443
+ {
444
+ "epoch": 3.38,
445
+ "learning_rate": 0.00026888087734700146,
446
+ "loss": 0.9007,
447
+ "step": 146000
448
+ },
449
+ {
450
+ "epoch": 3.43,
451
+ "learning_rate": 0.00026841307537693143,
452
+ "loss": 0.8951,
453
+ "step": 148000
454
+ },
455
+ {
456
+ "epoch": 3.47,
457
+ "learning_rate": 0.0002679452734068614,
458
+ "loss": 0.8926,
459
+ "step": 150000
460
+ },
461
+ {
462
+ "epoch": 3.52,
463
+ "learning_rate": 0.00026747747143679137,
464
+ "loss": 0.8924,
465
+ "step": 152000
466
+ },
467
+ {
468
+ "epoch": 3.57,
469
+ "learning_rate": 0.00026700966946672134,
470
+ "loss": 0.9044,
471
+ "step": 154000
472
+ },
473
+ {
474
+ "epoch": 3.61,
475
+ "learning_rate": 0.0002665418674966513,
476
+ "loss": 0.8952,
477
+ "step": 156000
478
+ },
479
+ {
480
+ "epoch": 3.66,
481
+ "learning_rate": 0.00026607406552658127,
482
+ "loss": 0.9001,
483
+ "step": 158000
484
+ },
485
+ {
486
+ "epoch": 3.7,
487
+ "learning_rate": 0.00026560626355651124,
488
+ "loss": 0.8898,
489
+ "step": 160000
490
+ },
491
+ {
492
+ "epoch": 3.75,
493
+ "learning_rate": 0.0002651384615864412,
494
+ "loss": 0.895,
495
+ "step": 162000
496
+ },
497
+ {
498
+ "epoch": 3.8,
499
+ "learning_rate": 0.0002646706596163712,
500
+ "loss": 0.9015,
501
+ "step": 164000
502
+ },
503
+ {
504
+ "epoch": 3.84,
505
+ "learning_rate": 0.00026420285764630115,
506
+ "loss": 0.892,
507
+ "step": 166000
508
+ },
509
+ {
510
+ "epoch": 3.89,
511
+ "learning_rate": 0.0002637350556762311,
512
+ "loss": 0.8903,
513
+ "step": 168000
514
+ },
515
+ {
516
+ "epoch": 3.94,
517
+ "learning_rate": 0.0002632672537061611,
518
+ "loss": 0.8916,
519
+ "step": 170000
520
+ },
521
+ {
522
+ "epoch": 3.98,
523
+ "learning_rate": 0.00026279945173609105,
524
+ "loss": 0.8941,
525
+ "step": 172000
526
+ },
527
+ {
528
+ "epoch": 4.03,
529
+ "learning_rate": 0.000262331649766021,
530
+ "loss": 0.8771,
531
+ "step": 174000
532
+ },
533
+ {
534
+ "epoch": 4.08,
535
+ "learning_rate": 0.000261863847795951,
536
+ "loss": 0.8716,
537
+ "step": 176000
538
+ },
539
+ {
540
+ "epoch": 4.12,
541
+ "learning_rate": 0.00026139604582588096,
542
+ "loss": 0.8632,
543
+ "step": 178000
544
+ },
545
+ {
546
+ "epoch": 4.17,
547
+ "learning_rate": 0.00026092824385581093,
548
+ "loss": 0.8573,
549
+ "step": 180000
550
+ },
551
+ {
552
+ "epoch": 4.21,
553
+ "learning_rate": 0.0002604604418857409,
554
+ "loss": 0.8642,
555
+ "step": 182000
556
+ },
557
+ {
558
+ "epoch": 4.26,
559
+ "learning_rate": 0.00025999263991567087,
560
+ "loss": 0.8642,
561
+ "step": 184000
562
+ },
563
+ {
564
+ "epoch": 4.31,
565
+ "learning_rate": 0.00025952483794560084,
566
+ "loss": 0.8617,
567
+ "step": 186000
568
+ },
569
+ {
570
+ "epoch": 4.35,
571
+ "learning_rate": 0.0002590570359755308,
572
+ "loss": 0.8574,
573
+ "step": 188000
574
+ },
575
+ {
576
+ "epoch": 4.4,
577
+ "learning_rate": 0.0002585892340054608,
578
+ "loss": 0.8612,
579
+ "step": 190000
580
+ },
581
+ {
582
+ "epoch": 4.45,
583
+ "learning_rate": 0.00025812143203539074,
584
+ "loss": 0.8706,
585
+ "step": 192000
586
+ },
587
+ {
588
+ "epoch": 4.49,
589
+ "learning_rate": 0.0002576536300653207,
590
+ "loss": 0.8605,
591
+ "step": 194000
592
+ },
593
+ {
594
+ "epoch": 4.54,
595
+ "learning_rate": 0.0002571858280952507,
596
+ "loss": 0.8703,
597
+ "step": 196000
598
+ },
599
+ {
600
+ "epoch": 4.58,
601
+ "learning_rate": 0.00025671802612518065,
602
+ "loss": 0.8691,
603
+ "step": 198000
604
+ },
605
+ {
606
+ "epoch": 4.63,
607
+ "learning_rate": 0.0002562502241551106,
608
+ "loss": 0.873,
609
+ "step": 200000
610
+ },
611
+ {
612
+ "epoch": 4.68,
613
+ "learning_rate": 0.0002557824221850406,
614
+ "loss": 0.857,
615
+ "step": 202000
616
+ },
617
+ {
618
+ "epoch": 4.72,
619
+ "learning_rate": 0.00025531462021497056,
620
+ "loss": 0.8686,
621
+ "step": 204000
622
+ },
623
+ {
624
+ "epoch": 4.77,
625
+ "learning_rate": 0.0002548468182449005,
626
+ "loss": 0.868,
627
+ "step": 206000
628
+ },
629
+ {
630
+ "epoch": 4.82,
631
+ "learning_rate": 0.0002543790162748305,
632
+ "loss": 0.8756,
633
+ "step": 208000
634
+ },
635
+ {
636
+ "epoch": 4.86,
637
+ "learning_rate": 0.0002539112143047605,
638
+ "loss": 0.8656,
639
+ "step": 210000
640
+ },
641
+ {
642
+ "epoch": 4.91,
643
+ "learning_rate": 0.00025344341233469043,
644
+ "loss": 0.8725,
645
+ "step": 212000
646
+ },
647
+ {
648
+ "epoch": 4.96,
649
+ "learning_rate": 0.00025297561036462045,
650
+ "loss": 0.8675,
651
+ "step": 214000
652
+ },
653
+ {
654
+ "epoch": 5.0,
655
+ "learning_rate": 0.00025250780839455037,
656
+ "loss": 0.8644,
657
+ "step": 216000
658
+ },
659
+ {
660
+ "epoch": 5.05,
661
+ "learning_rate": 0.0002520400064244804,
662
+ "loss": 0.8397,
663
+ "step": 218000
664
+ },
665
+ {
666
+ "epoch": 5.09,
667
+ "learning_rate": 0.0002515722044544103,
668
+ "loss": 0.8374,
669
+ "step": 220000
670
+ },
671
+ {
672
+ "epoch": 5.14,
673
+ "learning_rate": 0.00025110440248434033,
674
+ "loss": 0.8364,
675
+ "step": 222000
676
+ },
677
+ {
678
+ "epoch": 5.19,
679
+ "learning_rate": 0.00025063660051427025,
680
+ "loss": 0.8454,
681
+ "step": 224000
682
+ },
683
+ {
684
+ "epoch": 5.23,
685
+ "learning_rate": 0.00025016879854420027,
686
+ "loss": 0.8367,
687
+ "step": 226000
688
+ },
689
+ {
690
+ "epoch": 5.28,
691
+ "learning_rate": 0.0002497009965741302,
692
+ "loss": 0.8275,
693
+ "step": 228000
694
+ },
695
+ {
696
+ "epoch": 5.33,
697
+ "learning_rate": 0.0002492331946040602,
698
+ "loss": 0.8398,
699
+ "step": 230000
700
+ },
701
+ {
702
+ "epoch": 5.37,
703
+ "learning_rate": 0.0002487653926339901,
704
+ "loss": 0.8403,
705
+ "step": 232000
706
+ },
707
+ {
708
+ "epoch": 5.42,
709
+ "learning_rate": 0.00024829759066392014,
710
+ "loss": 0.8409,
711
+ "step": 234000
712
+ },
713
+ {
714
+ "epoch": 5.46,
715
+ "learning_rate": 0.00024782978869385006,
716
+ "loss": 0.8366,
717
+ "step": 236000
718
+ },
719
+ {
720
+ "epoch": 5.51,
721
+ "learning_rate": 0.0002473619867237801,
722
+ "loss": 0.8381,
723
+ "step": 238000
724
+ },
725
+ {
726
+ "epoch": 5.56,
727
+ "learning_rate": 0.00024689418475371,
728
+ "loss": 0.842,
729
+ "step": 240000
730
+ },
731
+ {
732
+ "epoch": 5.6,
733
+ "learning_rate": 0.00024642638278364,
734
+ "loss": 0.843,
735
+ "step": 242000
736
+ },
737
+ {
738
+ "epoch": 5.65,
739
+ "learning_rate": 0.00024595858081357,
740
+ "loss": 0.8489,
741
+ "step": 244000
742
+ },
743
+ {
744
+ "epoch": 5.7,
745
+ "learning_rate": 0.00024549077884349996,
746
+ "loss": 0.8313,
747
+ "step": 246000
748
+ },
749
+ {
750
+ "epoch": 5.74,
751
+ "learning_rate": 0.0002450229768734299,
752
+ "loss": 0.8468,
753
+ "step": 248000
754
+ },
755
+ {
756
+ "epoch": 5.79,
757
+ "learning_rate": 0.0002445551749033599,
758
+ "loss": 0.8446,
759
+ "step": 250000
760
+ },
761
+ {
762
+ "epoch": 5.84,
763
+ "learning_rate": 0.00024408737293328986,
764
+ "loss": 0.8329,
765
+ "step": 252000
766
+ },
767
+ {
768
+ "epoch": 5.88,
769
+ "learning_rate": 0.00024361957096321983,
770
+ "loss": 0.8402,
771
+ "step": 254000
772
+ },
773
+ {
774
+ "epoch": 5.93,
775
+ "learning_rate": 0.0002431517689931498,
776
+ "loss": 0.8502,
777
+ "step": 256000
778
+ },
779
+ {
780
+ "epoch": 5.97,
781
+ "learning_rate": 0.00024268396702307977,
782
+ "loss": 0.8437,
783
+ "step": 258000
784
+ },
785
+ {
786
+ "epoch": 6.02,
787
+ "learning_rate": 0.00024221616505300974,
788
+ "loss": 0.8341,
789
+ "step": 260000
790
+ },
791
+ {
792
+ "epoch": 6.07,
793
+ "learning_rate": 0.0002417483630829397,
794
+ "loss": 0.8152,
795
+ "step": 262000
796
+ },
797
+ {
798
+ "epoch": 6.11,
799
+ "learning_rate": 0.00024128056111286968,
800
+ "loss": 0.8189,
801
+ "step": 264000
802
+ },
803
+ {
804
+ "epoch": 6.16,
805
+ "learning_rate": 0.00024081275914279965,
806
+ "loss": 0.8193,
807
+ "step": 266000
808
+ },
809
+ {
810
+ "epoch": 6.21,
811
+ "learning_rate": 0.00024034495717272961,
812
+ "loss": 0.825,
813
+ "step": 268000
814
+ },
815
+ {
816
+ "epoch": 6.25,
817
+ "learning_rate": 0.00023987715520265958,
818
+ "loss": 0.818,
819
+ "step": 270000
820
+ },
821
+ {
822
+ "epoch": 6.3,
823
+ "learning_rate": 0.00023940935323258955,
824
+ "loss": 0.8204,
825
+ "step": 272000
826
+ },
827
+ {
828
+ "epoch": 6.34,
829
+ "learning_rate": 0.00023894155126251952,
830
+ "loss": 0.823,
831
+ "step": 274000
832
+ },
833
+ {
834
+ "epoch": 6.39,
835
+ "learning_rate": 0.0002384737492924495,
836
+ "loss": 0.8179,
837
+ "step": 276000
838
+ },
839
+ {
840
+ "epoch": 6.44,
841
+ "learning_rate": 0.00023800594732237946,
842
+ "loss": 0.8152,
843
+ "step": 278000
844
+ },
845
+ {
846
+ "epoch": 6.48,
847
+ "learning_rate": 0.00023753814535230943,
848
+ "loss": 0.8178,
849
+ "step": 280000
850
+ },
851
+ {
852
+ "epoch": 6.53,
853
+ "learning_rate": 0.00023707034338223942,
854
+ "loss": 0.8212,
855
+ "step": 282000
856
+ },
857
+ {
858
+ "epoch": 6.58,
859
+ "learning_rate": 0.00023660254141216937,
860
+ "loss": 0.8161,
861
+ "step": 284000
862
+ },
863
+ {
864
+ "epoch": 6.62,
865
+ "learning_rate": 0.00023613473944209936,
866
+ "loss": 0.8162,
867
+ "step": 286000
868
+ },
869
+ {
870
+ "epoch": 6.67,
871
+ "learning_rate": 0.0002356669374720293,
872
+ "loss": 0.828,
873
+ "step": 288000
874
+ },
875
+ {
876
+ "epoch": 6.72,
877
+ "learning_rate": 0.0002351991355019593,
878
+ "loss": 0.823,
879
+ "step": 290000
880
+ },
881
+ {
882
+ "epoch": 6.76,
883
+ "learning_rate": 0.00023473133353188927,
884
+ "loss": 0.8156,
885
+ "step": 292000
886
+ },
887
+ {
888
+ "epoch": 6.81,
889
+ "learning_rate": 0.00023426353156181924,
890
+ "loss": 0.8176,
891
+ "step": 294000
892
+ },
893
+ {
894
+ "epoch": 6.85,
895
+ "learning_rate": 0.0002337957295917492,
896
+ "loss": 0.8226,
897
+ "step": 296000
898
+ },
899
+ {
900
+ "epoch": 6.9,
901
+ "learning_rate": 0.00023332792762167917,
902
+ "loss": 0.8189,
903
+ "step": 298000
904
+ },
905
+ {
906
+ "epoch": 6.95,
907
+ "learning_rate": 0.00023286012565160914,
908
+ "loss": 0.8082,
909
+ "step": 300000
910
+ },
911
+ {
912
+ "epoch": 6.99,
913
+ "learning_rate": 0.0002323923236815391,
914
+ "loss": 0.8257,
915
+ "step": 302000
916
+ },
917
+ {
918
+ "epoch": 7.04,
919
+ "learning_rate": 0.00023192452171146908,
920
+ "loss": 0.7925,
921
+ "step": 304000
922
+ },
923
+ {
924
+ "epoch": 7.09,
925
+ "learning_rate": 0.00023145671974139905,
926
+ "loss": 0.7902,
927
+ "step": 306000
928
+ },
929
+ {
930
+ "epoch": 7.13,
931
+ "learning_rate": 0.00023098891777132902,
932
+ "loss": 0.7994,
933
+ "step": 308000
934
+ },
935
+ {
936
+ "epoch": 7.18,
937
+ "learning_rate": 0.000230521115801259,
938
+ "loss": 0.8029,
939
+ "step": 310000
940
+ },
941
+ {
942
+ "epoch": 7.22,
943
+ "learning_rate": 0.00023005331383118896,
944
+ "loss": 0.7983,
945
+ "step": 312000
946
+ },
947
+ {
948
+ "epoch": 7.27,
949
+ "learning_rate": 0.00022958551186111893,
950
+ "loss": 0.797,
951
+ "step": 314000
952
+ },
953
+ {
954
+ "epoch": 7.32,
955
+ "learning_rate": 0.00022911770989104892,
956
+ "loss": 0.8009,
957
+ "step": 316000
958
+ },
959
+ {
960
+ "epoch": 7.36,
961
+ "learning_rate": 0.00022864990792097886,
962
+ "loss": 0.7979,
963
+ "step": 318000
964
+ },
965
+ {
966
+ "epoch": 7.41,
967
+ "learning_rate": 0.00022818210595090886,
968
+ "loss": 0.8055,
969
+ "step": 320000
970
+ },
971
+ {
972
+ "epoch": 7.46,
973
+ "learning_rate": 0.0002277143039808388,
974
+ "loss": 0.7971,
975
+ "step": 322000
976
+ },
977
+ {
978
+ "epoch": 7.5,
979
+ "learning_rate": 0.0002272465020107688,
980
+ "loss": 0.8066,
981
+ "step": 324000
982
+ },
983
+ {
984
+ "epoch": 7.55,
985
+ "learning_rate": 0.00022677870004069874,
986
+ "loss": 0.7975,
987
+ "step": 326000
988
+ },
989
+ {
990
+ "epoch": 7.6,
991
+ "learning_rate": 0.00022631089807062873,
992
+ "loss": 0.7912,
993
+ "step": 328000
994
+ },
995
+ {
996
+ "epoch": 7.64,
997
+ "learning_rate": 0.00022584309610055868,
998
+ "loss": 0.7988,
999
+ "step": 330000
1000
+ },
1001
+ {
1002
+ "epoch": 7.69,
1003
+ "learning_rate": 0.00022537529413048867,
1004
+ "loss": 0.7999,
1005
+ "step": 332000
1006
+ },
1007
+ {
1008
+ "epoch": 7.73,
1009
+ "learning_rate": 0.00022490749216041861,
1010
+ "loss": 0.8019,
1011
+ "step": 334000
1012
+ },
1013
+ {
1014
+ "epoch": 7.78,
1015
+ "learning_rate": 0.0002244396901903486,
1016
+ "loss": 0.8108,
1017
+ "step": 336000
1018
+ },
1019
+ {
1020
+ "epoch": 7.83,
1021
+ "learning_rate": 0.00022397188822027855,
1022
+ "loss": 0.8075,
1023
+ "step": 338000
1024
+ },
1025
+ {
1026
+ "epoch": 7.87,
1027
+ "learning_rate": 0.00022350408625020855,
1028
+ "loss": 0.7995,
1029
+ "step": 340000
1030
+ },
1031
+ {
1032
+ "epoch": 7.92,
1033
+ "learning_rate": 0.0002230362842801385,
1034
+ "loss": 0.802,
1035
+ "step": 342000
1036
+ },
1037
+ {
1038
+ "epoch": 7.97,
1039
+ "learning_rate": 0.00022256848231006848,
1040
+ "loss": 0.7998,
1041
+ "step": 344000
1042
+ },
1043
+ {
1044
+ "epoch": 8.01,
1045
+ "learning_rate": 0.00022210068033999843,
1046
+ "loss": 0.8023,
1047
+ "step": 346000
1048
+ },
1049
+ {
1050
+ "epoch": 8.06,
1051
+ "learning_rate": 0.00022163287836992842,
1052
+ "loss": 0.7748,
1053
+ "step": 348000
1054
+ },
1055
+ {
1056
+ "epoch": 8.1,
1057
+ "learning_rate": 0.00022116507639985836,
1058
+ "loss": 0.7839,
1059
+ "step": 350000
1060
+ },
1061
+ {
1062
+ "epoch": 8.15,
1063
+ "learning_rate": 0.00022069727442978836,
1064
+ "loss": 0.7806,
1065
+ "step": 352000
1066
+ },
1067
+ {
1068
+ "epoch": 8.2,
1069
+ "learning_rate": 0.00022022947245971836,
1070
+ "loss": 0.7775,
1071
+ "step": 354000
1072
+ },
1073
+ {
1074
+ "epoch": 8.24,
1075
+ "learning_rate": 0.0002197616704896483,
1076
+ "loss": 0.7734,
1077
+ "step": 356000
1078
+ },
1079
+ {
1080
+ "epoch": 8.29,
1081
+ "learning_rate": 0.0002192938685195783,
1082
+ "loss": 0.7728,
1083
+ "step": 358000
1084
+ },
1085
+ {
1086
+ "epoch": 8.34,
1087
+ "learning_rate": 0.00021882606654950824,
1088
+ "loss": 0.7879,
1089
+ "step": 360000
1090
+ },
1091
+ {
1092
+ "epoch": 8.38,
1093
+ "learning_rate": 0.00021835826457943823,
1094
+ "loss": 0.7891,
1095
+ "step": 362000
1096
+ },
1097
+ {
1098
+ "epoch": 8.43,
1099
+ "learning_rate": 0.00021789046260936817,
1100
+ "loss": 0.7922,
1101
+ "step": 364000
1102
+ },
1103
+ {
1104
+ "epoch": 8.48,
1105
+ "learning_rate": 0.00021742266063929817,
1106
+ "loss": 0.7837,
1107
+ "step": 366000
1108
+ },
1109
+ {
1110
+ "epoch": 8.52,
1111
+ "learning_rate": 0.0002169548586692281,
1112
+ "loss": 0.7838,
1113
+ "step": 368000
1114
+ },
1115
+ {
1116
+ "epoch": 8.57,
1117
+ "learning_rate": 0.0002164870566991581,
1118
+ "loss": 0.7797,
1119
+ "step": 370000
1120
+ },
1121
+ {
1122
+ "epoch": 8.61,
1123
+ "learning_rate": 0.00021601925472908805,
1124
+ "loss": 0.7818,
1125
+ "step": 372000
1126
+ },
1127
+ {
1128
+ "epoch": 8.66,
1129
+ "learning_rate": 0.00021555145275901804,
1130
+ "loss": 0.7838,
1131
+ "step": 374000
1132
+ },
1133
+ {
1134
+ "epoch": 8.71,
1135
+ "learning_rate": 0.000215083650788948,
1136
+ "loss": 0.7828,
1137
+ "step": 376000
1138
+ },
1139
+ {
1140
+ "epoch": 8.75,
1141
+ "learning_rate": 0.00021461584881887798,
1142
+ "loss": 0.7822,
1143
+ "step": 378000
1144
+ },
1145
+ {
1146
+ "epoch": 8.8,
1147
+ "learning_rate": 0.00021414804684880792,
1148
+ "loss": 0.7952,
1149
+ "step": 380000
1150
+ },
1151
+ {
1152
+ "epoch": 8.85,
1153
+ "learning_rate": 0.00021368024487873792,
1154
+ "loss": 0.7888,
1155
+ "step": 382000
1156
+ },
1157
+ {
1158
+ "epoch": 8.89,
1159
+ "learning_rate": 0.00021321244290866786,
1160
+ "loss": 0.7813,
1161
+ "step": 384000
1162
+ },
1163
+ {
1164
+ "epoch": 8.94,
1165
+ "learning_rate": 0.00021274464093859786,
1166
+ "loss": 0.7784,
1167
+ "step": 386000
1168
+ },
1169
+ {
1170
+ "epoch": 8.98,
1171
+ "learning_rate": 0.00021227683896852783,
1172
+ "loss": 0.776,
1173
+ "step": 388000
1174
+ },
1175
+ {
1176
+ "epoch": 9.03,
1177
+ "learning_rate": 0.0002118090369984578,
1178
+ "loss": 0.7686,
1179
+ "step": 390000
1180
+ },
1181
+ {
1182
+ "epoch": 9.08,
1183
+ "learning_rate": 0.0002113412350283878,
1184
+ "loss": 0.7576,
1185
+ "step": 392000
1186
+ },
1187
+ {
1188
+ "epoch": 9.12,
1189
+ "learning_rate": 0.00021087343305831773,
1190
+ "loss": 0.7611,
1191
+ "step": 394000
1192
+ },
1193
+ {
1194
+ "epoch": 9.17,
1195
+ "learning_rate": 0.00021040563108824773,
1196
+ "loss": 0.7698,
1197
+ "step": 396000
1198
+ },
1199
+ {
1200
+ "epoch": 9.22,
1201
+ "learning_rate": 0.00020993782911817767,
1202
+ "loss": 0.7632,
1203
+ "step": 398000
1204
+ },
1205
+ {
1206
+ "epoch": 9.26,
1207
+ "learning_rate": 0.00020947002714810767,
1208
+ "loss": 0.7725,
1209
+ "step": 400000
1210
+ },
1211
+ {
1212
+ "epoch": 9.31,
1213
+ "learning_rate": 0.0002090022251780376,
1214
+ "loss": 0.7706,
1215
+ "step": 402000
1216
+ },
1217
+ {
1218
+ "epoch": 9.35,
1219
+ "learning_rate": 0.0002085344232079676,
1220
+ "loss": 0.7709,
1221
+ "step": 404000
1222
+ },
1223
+ {
1224
+ "epoch": 9.4,
1225
+ "learning_rate": 0.00020806662123789755,
1226
+ "loss": 0.7651,
1227
+ "step": 406000
1228
+ },
1229
+ {
1230
+ "epoch": 9.45,
1231
+ "learning_rate": 0.00020759881926782754,
1232
+ "loss": 0.7657,
1233
+ "step": 408000
1234
+ },
1235
+ {
1236
+ "epoch": 9.49,
1237
+ "learning_rate": 0.00020713101729775748,
1238
+ "loss": 0.7589,
1239
+ "step": 410000
1240
+ },
1241
+ {
1242
+ "epoch": 9.54,
1243
+ "learning_rate": 0.00020666321532768748,
1244
+ "loss": 0.7683,
1245
+ "step": 412000
1246
+ },
1247
+ {
1248
+ "epoch": 9.59,
1249
+ "learning_rate": 0.00020619541335761742,
1250
+ "loss": 0.7684,
1251
+ "step": 414000
1252
+ },
1253
+ {
1254
+ "epoch": 9.63,
1255
+ "learning_rate": 0.00020572761138754742,
1256
+ "loss": 0.7756,
1257
+ "step": 416000
1258
+ },
1259
+ {
1260
+ "epoch": 9.68,
1261
+ "learning_rate": 0.00020525980941747736,
1262
+ "loss": 0.7653,
1263
+ "step": 418000
1264
+ },
1265
+ {
1266
+ "epoch": 9.73,
1267
+ "learning_rate": 0.00020479200744740736,
1268
+ "loss": 0.7718,
1269
+ "step": 420000
1270
+ },
1271
+ {
1272
+ "epoch": 9.77,
1273
+ "learning_rate": 0.0002043242054773373,
1274
+ "loss": 0.7676,
1275
+ "step": 422000
1276
+ },
1277
+ {
1278
+ "epoch": 9.82,
1279
+ "learning_rate": 0.0002038564035072673,
1280
+ "loss": 0.772,
1281
+ "step": 424000
1282
+ },
1283
+ {
1284
+ "epoch": 9.86,
1285
+ "learning_rate": 0.00020338860153719726,
1286
+ "loss": 0.766,
1287
+ "step": 426000
1288
+ },
1289
+ {
1290
+ "epoch": 9.91,
1291
+ "learning_rate": 0.00020292079956712723,
1292
+ "loss": 0.7739,
1293
+ "step": 428000
1294
+ },
1295
+ {
1296
+ "epoch": 9.96,
1297
+ "learning_rate": 0.0002024529975970572,
1298
+ "loss": 0.7743,
1299
+ "step": 430000
1300
+ },
1301
+ {
1302
+ "epoch": 10.0,
1303
+ "learning_rate": 0.00020198519562698717,
1304
+ "loss": 0.7719,
1305
+ "step": 432000
1306
+ },
1307
+ {
1308
+ "epoch": 10.05,
1309
+ "learning_rate": 0.00020151739365691714,
1310
+ "loss": 0.7443,
1311
+ "step": 434000
1312
+ },
1313
+ {
1314
+ "epoch": 10.1,
1315
+ "learning_rate": 0.0002010495916868471,
1316
+ "loss": 0.7573,
1317
+ "step": 436000
1318
+ },
1319
+ {
1320
+ "epoch": 10.14,
1321
+ "learning_rate": 0.00020058178971677708,
1322
+ "loss": 0.7546,
1323
+ "step": 438000
1324
+ },
1325
+ {
1326
+ "epoch": 10.19,
1327
+ "learning_rate": 0.00020011398774670704,
1328
+ "loss": 0.7516,
1329
+ "step": 440000
1330
+ },
1331
+ {
1332
+ "epoch": 10.23,
1333
+ "learning_rate": 0.000199646185776637,
1334
+ "loss": 0.7444,
1335
+ "step": 442000
1336
+ },
1337
+ {
1338
+ "epoch": 10.28,
1339
+ "learning_rate": 0.00019917838380656698,
1340
+ "loss": 0.7656,
1341
+ "step": 444000
1342
+ },
1343
+ {
1344
+ "epoch": 10.33,
1345
+ "learning_rate": 0.00019871058183649695,
1346
+ "loss": 0.7452,
1347
+ "step": 446000
1348
+ },
1349
+ {
1350
+ "epoch": 10.37,
1351
+ "learning_rate": 0.00019824277986642692,
1352
+ "loss": 0.7555,
1353
+ "step": 448000
1354
+ },
1355
+ {
1356
+ "epoch": 10.42,
1357
+ "learning_rate": 0.0001977749778963569,
1358
+ "loss": 0.7486,
1359
+ "step": 450000
1360
+ },
1361
+ {
1362
+ "epoch": 10.47,
1363
+ "learning_rate": 0.00019730717592628686,
1364
+ "loss": 0.7509,
1365
+ "step": 452000
1366
+ },
1367
+ {
1368
+ "epoch": 10.51,
1369
+ "learning_rate": 0.00019683937395621683,
1370
+ "loss": 0.7484,
1371
+ "step": 454000
1372
+ },
1373
+ {
1374
+ "epoch": 10.56,
1375
+ "learning_rate": 0.0001963715719861468,
1376
+ "loss": 0.7554,
1377
+ "step": 456000
1378
+ },
1379
+ {
1380
+ "epoch": 10.61,
1381
+ "learning_rate": 0.00019590377001607676,
1382
+ "loss": 0.7557,
1383
+ "step": 458000
1384
+ },
1385
+ {
1386
+ "epoch": 10.65,
1387
+ "learning_rate": 0.00019543596804600676,
1388
+ "loss": 0.7603,
1389
+ "step": 460000
1390
+ },
1391
+ {
1392
+ "epoch": 10.7,
1393
+ "learning_rate": 0.00019496816607593673,
1394
+ "loss": 0.7577,
1395
+ "step": 462000
1396
+ },
1397
+ {
1398
+ "epoch": 10.74,
1399
+ "learning_rate": 0.0001945003641058667,
1400
+ "loss": 0.7641,
1401
+ "step": 464000
1402
+ },
1403
+ {
1404
+ "epoch": 10.79,
1405
+ "learning_rate": 0.00019403256213579667,
1406
+ "loss": 0.7648,
1407
+ "step": 466000
1408
+ },
1409
+ {
1410
+ "epoch": 10.84,
1411
+ "learning_rate": 0.00019356476016572664,
1412
+ "loss": 0.755,
1413
+ "step": 468000
1414
+ },
1415
+ {
1416
+ "epoch": 10.88,
1417
+ "learning_rate": 0.0001930969581956566,
1418
+ "loss": 0.7445,
1419
+ "step": 470000
1420
+ },
1421
+ {
1422
+ "epoch": 10.93,
1423
+ "learning_rate": 0.00019262915622558657,
1424
+ "loss": 0.7614,
1425
+ "step": 472000
1426
+ },
1427
+ {
1428
+ "epoch": 10.98,
1429
+ "learning_rate": 0.00019216135425551654,
1430
+ "loss": 0.7526,
1431
+ "step": 474000
1432
+ },
1433
+ {
1434
+ "epoch": 11.02,
1435
+ "learning_rate": 0.0001916935522854465,
1436
+ "loss": 0.7493,
1437
+ "step": 476000
1438
+ },
1439
+ {
1440
+ "epoch": 11.07,
1441
+ "learning_rate": 0.00019122575031537648,
1442
+ "loss": 0.7299,
1443
+ "step": 478000
1444
+ },
1445
+ {
1446
+ "epoch": 11.11,
1447
+ "learning_rate": 0.00019075794834530645,
1448
+ "loss": 0.7379,
1449
+ "step": 480000
1450
+ },
1451
+ {
1452
+ "epoch": 11.16,
1453
+ "learning_rate": 0.00019029014637523642,
1454
+ "loss": 0.7365,
1455
+ "step": 482000
1456
+ },
1457
+ {
1458
+ "epoch": 11.21,
1459
+ "learning_rate": 0.00018982234440516639,
1460
+ "loss": 0.7402,
1461
+ "step": 484000
1462
+ },
1463
+ {
1464
+ "epoch": 11.25,
1465
+ "learning_rate": 0.00018935454243509636,
1466
+ "loss": 0.7409,
1467
+ "step": 486000
1468
+ },
1469
+ {
1470
+ "epoch": 11.3,
1471
+ "learning_rate": 0.00018888674046502632,
1472
+ "loss": 0.7294,
1473
+ "step": 488000
1474
+ },
1475
+ {
1476
+ "epoch": 11.35,
1477
+ "learning_rate": 0.0001884189384949563,
1478
+ "loss": 0.7467,
1479
+ "step": 490000
1480
+ },
1481
+ {
1482
+ "epoch": 11.39,
1483
+ "learning_rate": 0.00018795113652488626,
1484
+ "loss": 0.7357,
1485
+ "step": 492000
1486
+ },
1487
+ {
1488
+ "epoch": 11.44,
1489
+ "learning_rate": 0.00018748333455481623,
1490
+ "loss": 0.744,
1491
+ "step": 494000
1492
+ },
1493
+ {
1494
+ "epoch": 11.49,
1495
+ "learning_rate": 0.0001870155325847462,
1496
+ "loss": 0.741,
1497
+ "step": 496000
1498
+ },
1499
+ {
1500
+ "epoch": 11.53,
1501
+ "learning_rate": 0.0001865477306146762,
1502
+ "loss": 0.7404,
1503
+ "step": 498000
1504
+ },
1505
+ {
1506
+ "epoch": 11.58,
1507
+ "learning_rate": 0.00018607992864460614,
1508
+ "loss": 0.749,
1509
+ "step": 500000
1510
+ },
1511
+ {
1512
+ "epoch": 11.62,
1513
+ "learning_rate": 0.00018561212667453613,
1514
+ "loss": 0.7388,
1515
+ "step": 502000
1516
+ },
1517
+ {
1518
+ "epoch": 11.67,
1519
+ "learning_rate": 0.00018514432470446607,
1520
+ "loss": 0.742,
1521
+ "step": 504000
1522
+ },
1523
+ {
1524
+ "epoch": 11.72,
1525
+ "learning_rate": 0.00018467652273439607,
1526
+ "loss": 0.7481,
1527
+ "step": 506000
1528
+ },
1529
+ {
1530
+ "epoch": 11.76,
1531
+ "learning_rate": 0.000184208720764326,
1532
+ "loss": 0.7553,
1533
+ "step": 508000
1534
+ },
1535
+ {
1536
+ "epoch": 11.81,
1537
+ "learning_rate": 0.000183740918794256,
1538
+ "loss": 0.7457,
1539
+ "step": 510000
1540
+ },
1541
+ {
1542
+ "epoch": 11.86,
1543
+ "learning_rate": 0.00018327311682418595,
1544
+ "loss": 0.7447,
1545
+ "step": 512000
1546
+ },
1547
+ {
1548
+ "epoch": 11.9,
1549
+ "learning_rate": 0.00018280531485411595,
1550
+ "loss": 0.752,
1551
+ "step": 514000
1552
+ },
1553
+ {
1554
+ "epoch": 11.95,
1555
+ "learning_rate": 0.0001823375128840459,
1556
+ "loss": 0.7419,
1557
+ "step": 516000
1558
+ },
1559
+ {
1560
+ "epoch": 11.99,
1561
+ "learning_rate": 0.00018186971091397588,
1562
+ "loss": 0.7412,
1563
+ "step": 518000
1564
+ },
1565
+ {
1566
+ "epoch": 12.04,
1567
+ "learning_rate": 0.00018140190894390583,
1568
+ "loss": 0.7262,
1569
+ "step": 520000
1570
+ },
1571
+ {
1572
+ "epoch": 12.09,
1573
+ "learning_rate": 0.00018093410697383582,
1574
+ "loss": 0.7299,
1575
+ "step": 522000
1576
+ },
1577
+ {
1578
+ "epoch": 12.13,
1579
+ "learning_rate": 0.00018046630500376576,
1580
+ "loss": 0.7279,
1581
+ "step": 524000
1582
+ },
1583
+ {
1584
+ "epoch": 12.18,
1585
+ "learning_rate": 0.00017999850303369576,
1586
+ "loss": 0.7308,
1587
+ "step": 526000
1588
+ },
1589
+ {
1590
+ "epoch": 12.23,
1591
+ "learning_rate": 0.0001795307010636257,
1592
+ "loss": 0.7305,
1593
+ "step": 528000
1594
+ },
1595
+ {
1596
+ "epoch": 12.27,
1597
+ "learning_rate": 0.0001790628990935557,
1598
+ "loss": 0.7348,
1599
+ "step": 530000
1600
+ },
1601
+ {
1602
+ "epoch": 12.32,
1603
+ "learning_rate": 0.0001785950971234857,
1604
+ "loss": 0.7312,
1605
+ "step": 532000
1606
+ },
1607
+ {
1608
+ "epoch": 12.37,
1609
+ "learning_rate": 0.00017812729515341563,
1610
+ "loss": 0.7275,
1611
+ "step": 534000
1612
+ },
1613
+ {
1614
+ "epoch": 12.41,
1615
+ "learning_rate": 0.00017765949318334563,
1616
+ "loss": 0.7291,
1617
+ "step": 536000
1618
+ },
1619
+ {
1620
+ "epoch": 12.46,
1621
+ "learning_rate": 0.00017719169121327557,
1622
+ "loss": 0.7265,
1623
+ "step": 538000
1624
+ },
1625
+ {
1626
+ "epoch": 12.5,
1627
+ "learning_rate": 0.00017672388924320557,
1628
+ "loss": 0.7224,
1629
+ "step": 540000
1630
+ },
1631
+ {
1632
+ "epoch": 12.55,
1633
+ "learning_rate": 0.0001762560872731355,
1634
+ "loss": 0.7232,
1635
+ "step": 542000
1636
+ },
1637
+ {
1638
+ "epoch": 12.6,
1639
+ "learning_rate": 0.0001757882853030655,
1640
+ "loss": 0.7272,
1641
+ "step": 544000
1642
+ },
1643
+ {
1644
+ "epoch": 12.64,
1645
+ "learning_rate": 0.00017532048333299545,
1646
+ "loss": 0.7305,
1647
+ "step": 546000
1648
+ },
1649
+ {
1650
+ "epoch": 12.69,
1651
+ "learning_rate": 0.00017485268136292544,
1652
+ "loss": 0.7375,
1653
+ "step": 548000
1654
+ },
1655
+ {
1656
+ "epoch": 12.74,
1657
+ "learning_rate": 0.00017438487939285539,
1658
+ "loss": 0.7377,
1659
+ "step": 550000
1660
+ },
1661
+ {
1662
+ "epoch": 12.78,
1663
+ "learning_rate": 0.00017391707742278538,
1664
+ "loss": 0.7278,
1665
+ "step": 552000
1666
+ },
1667
+ {
1668
+ "epoch": 12.83,
1669
+ "learning_rate": 0.00017344927545271532,
1670
+ "loss": 0.7369,
1671
+ "step": 554000
1672
+ },
1673
+ {
1674
+ "epoch": 12.87,
1675
+ "learning_rate": 0.00017298147348264532,
1676
+ "loss": 0.7366,
1677
+ "step": 556000
1678
+ },
1679
+ {
1680
+ "epoch": 12.92,
1681
+ "learning_rate": 0.00017251367151257526,
1682
+ "loss": 0.736,
1683
+ "step": 558000
1684
+ },
1685
+ {
1686
+ "epoch": 12.97,
1687
+ "learning_rate": 0.00017204586954250526,
1688
+ "loss": 0.737,
1689
+ "step": 560000
1690
+ },
1691
+ {
1692
+ "epoch": 13.01,
1693
+ "learning_rate": 0.0001715780675724352,
1694
+ "loss": 0.7301,
1695
+ "step": 562000
1696
+ },
1697
+ {
1698
+ "epoch": 13.06,
1699
+ "learning_rate": 0.0001711102656023652,
1700
+ "loss": 0.7003,
1701
+ "step": 564000
1702
+ },
1703
+ {
1704
+ "epoch": 13.11,
1705
+ "learning_rate": 0.00017064246363229514,
1706
+ "loss": 0.7132,
1707
+ "step": 566000
1708
+ },
1709
+ {
1710
+ "epoch": 13.15,
1711
+ "learning_rate": 0.00017017466166222513,
1712
+ "loss": 0.7178,
1713
+ "step": 568000
1714
+ },
1715
+ {
1716
+ "epoch": 13.2,
1717
+ "learning_rate": 0.00016970685969215513,
1718
+ "loss": 0.7187,
1719
+ "step": 570000
1720
+ },
1721
+ {
1722
+ "epoch": 13.25,
1723
+ "learning_rate": 0.00016923905772208507,
1724
+ "loss": 0.7239,
1725
+ "step": 572000
1726
+ },
1727
+ {
1728
+ "epoch": 13.29,
1729
+ "learning_rate": 0.00016877125575201507,
1730
+ "loss": 0.7271,
1731
+ "step": 574000
1732
+ },
1733
+ {
1734
+ "epoch": 13.34,
1735
+ "learning_rate": 0.000168303453781945,
1736
+ "loss": 0.7208,
1737
+ "step": 576000
1738
+ },
1739
+ {
1740
+ "epoch": 13.38,
1741
+ "learning_rate": 0.000167835651811875,
1742
+ "loss": 0.7199,
1743
+ "step": 578000
1744
+ },
1745
+ {
1746
+ "epoch": 13.43,
1747
+ "learning_rate": 0.00016736784984180495,
1748
+ "loss": 0.7094,
1749
+ "step": 580000
1750
+ },
1751
+ {
1752
+ "epoch": 13.48,
1753
+ "learning_rate": 0.00016690004787173494,
1754
+ "loss": 0.7114,
1755
+ "step": 582000
1756
+ },
1757
+ {
1758
+ "epoch": 13.52,
1759
+ "learning_rate": 0.00016643224590166488,
1760
+ "loss": 0.7196,
1761
+ "step": 584000
1762
+ },
1763
+ {
1764
+ "epoch": 13.57,
1765
+ "learning_rate": 0.00016596444393159488,
1766
+ "loss": 0.7222,
1767
+ "step": 586000
1768
+ },
1769
+ {
1770
+ "epoch": 13.62,
1771
+ "learning_rate": 0.00016549664196152482,
1772
+ "loss": 0.7345,
1773
+ "step": 588000
1774
+ },
1775
+ {
1776
+ "epoch": 13.66,
1777
+ "learning_rate": 0.00016502883999145482,
1778
+ "loss": 0.7208,
1779
+ "step": 590000
1780
+ },
1781
+ {
1782
+ "epoch": 13.71,
1783
+ "learning_rate": 0.00016456103802138476,
1784
+ "loss": 0.7298,
1785
+ "step": 592000
1786
+ },
1787
+ {
1788
+ "epoch": 13.75,
1789
+ "learning_rate": 0.00016409323605131475,
1790
+ "loss": 0.7324,
1791
+ "step": 594000
1792
+ },
1793
+ {
1794
+ "epoch": 13.8,
1795
+ "learning_rate": 0.0001636254340812447,
1796
+ "loss": 0.7243,
1797
+ "step": 596000
1798
+ },
1799
+ {
1800
+ "epoch": 13.85,
1801
+ "learning_rate": 0.0001631576321111747,
1802
+ "loss": 0.7215,
1803
+ "step": 598000
1804
+ },
1805
+ {
1806
+ "epoch": 13.89,
1807
+ "learning_rate": 0.00016268983014110463,
1808
+ "loss": 0.7246,
1809
+ "step": 600000
1810
+ },
1811
+ {
1812
+ "epoch": 13.94,
1813
+ "learning_rate": 0.00016222202817103463,
1814
+ "loss": 0.7219,
1815
+ "step": 602000
1816
+ },
1817
+ {
1818
+ "epoch": 13.99,
1819
+ "learning_rate": 0.0001617542262009646,
1820
+ "loss": 0.7248,
1821
+ "step": 604000
1822
+ },
1823
+ {
1824
+ "epoch": 14.03,
1825
+ "learning_rate": 0.00016128642423089457,
1826
+ "loss": 0.7139,
1827
+ "step": 606000
1828
+ },
1829
+ {
1830
+ "epoch": 14.08,
1831
+ "learning_rate": 0.00016081862226082454,
1832
+ "loss": 0.7026,
1833
+ "step": 608000
1834
+ },
1835
+ {
1836
+ "epoch": 14.13,
1837
+ "learning_rate": 0.0001603508202907545,
1838
+ "loss": 0.7107,
1839
+ "step": 610000
1840
+ },
1841
+ {
1842
+ "epoch": 14.17,
1843
+ "learning_rate": 0.00015988301832068447,
1844
+ "loss": 0.7037,
1845
+ "step": 612000
1846
+ },
1847
+ {
1848
+ "epoch": 14.22,
1849
+ "learning_rate": 0.00015941521635061444,
1850
+ "loss": 0.7145,
1851
+ "step": 614000
1852
+ },
1853
+ {
1854
+ "epoch": 14.26,
1855
+ "learning_rate": 0.0001589474143805444,
1856
+ "loss": 0.7181,
1857
+ "step": 616000
1858
+ },
1859
+ {
1860
+ "epoch": 14.31,
1861
+ "learning_rate": 0.00015847961241047438,
1862
+ "loss": 0.7026,
1863
+ "step": 618000
1864
+ },
1865
+ {
1866
+ "epoch": 14.36,
1867
+ "learning_rate": 0.00015801181044040435,
1868
+ "loss": 0.7142,
1869
+ "step": 620000
1870
+ },
1871
+ {
1872
+ "epoch": 14.4,
1873
+ "learning_rate": 0.00015754400847033432,
1874
+ "loss": 0.7087,
1875
+ "step": 622000
1876
+ },
1877
+ {
1878
+ "epoch": 14.45,
1879
+ "learning_rate": 0.0001570762065002643,
1880
+ "loss": 0.7109,
1881
+ "step": 624000
1882
+ },
1883
+ {
1884
+ "epoch": 14.5,
1885
+ "learning_rate": 0.00015660840453019426,
1886
+ "loss": 0.7031,
1887
+ "step": 626000
1888
+ },
1889
+ {
1890
+ "epoch": 14.54,
1891
+ "learning_rate": 0.00015614060256012425,
1892
+ "loss": 0.7101,
1893
+ "step": 628000
1894
+ },
1895
+ {
1896
+ "epoch": 14.59,
1897
+ "learning_rate": 0.0001556728005900542,
1898
+ "loss": 0.7152,
1899
+ "step": 630000
1900
+ },
1901
+ {
1902
+ "epoch": 14.63,
1903
+ "learning_rate": 0.0001552049986199842,
1904
+ "loss": 0.7147,
1905
+ "step": 632000
1906
+ },
1907
+ {
1908
+ "epoch": 14.68,
1909
+ "learning_rate": 0.00015473719664991413,
1910
+ "loss": 0.7144,
1911
+ "step": 634000
1912
+ },
1913
+ {
1914
+ "epoch": 14.73,
1915
+ "learning_rate": 0.00015426939467984413,
1916
+ "loss": 0.7113,
1917
+ "step": 636000
1918
+ },
1919
+ {
1920
+ "epoch": 14.77,
1921
+ "learning_rate": 0.00015380159270977407,
1922
+ "loss": 0.7071,
1923
+ "step": 638000
1924
+ },
1925
+ {
1926
+ "epoch": 14.82,
1927
+ "learning_rate": 0.00015333379073970407,
1928
+ "loss": 0.7118,
1929
+ "step": 640000
1930
+ },
1931
+ {
1932
+ "epoch": 14.87,
1933
+ "learning_rate": 0.00015286598876963403,
1934
+ "loss": 0.7098,
1935
+ "step": 642000
1936
+ },
1937
+ {
1938
+ "epoch": 14.91,
1939
+ "learning_rate": 0.000152398186799564,
1940
+ "loss": 0.706,
1941
+ "step": 644000
1942
+ },
1943
+ {
1944
+ "epoch": 14.96,
1945
+ "learning_rate": 0.00015193038482949397,
1946
+ "loss": 0.709,
1947
+ "step": 646000
1948
+ },
1949
+ {
1950
+ "epoch": 15.01,
1951
+ "learning_rate": 0.00015146258285942394,
1952
+ "loss": 0.7087,
1953
+ "step": 648000
1954
+ },
1955
+ {
1956
+ "epoch": 15.05,
1957
+ "learning_rate": 0.0001509947808893539,
1958
+ "loss": 0.6983,
1959
+ "step": 650000
1960
+ },
1961
+ {
1962
+ "epoch": 15.1,
1963
+ "learning_rate": 0.00015052697891928388,
1964
+ "loss": 0.693,
1965
+ "step": 652000
1966
+ },
1967
+ {
1968
+ "epoch": 15.14,
1969
+ "learning_rate": 0.00015005917694921385,
1970
+ "loss": 0.6953,
1971
+ "step": 654000
1972
+ },
1973
+ {
1974
+ "epoch": 15.19,
1975
+ "learning_rate": 0.00014959137497914382,
1976
+ "loss": 0.6994,
1977
+ "step": 656000
1978
+ },
1979
+ {
1980
+ "epoch": 15.24,
1981
+ "learning_rate": 0.00014912357300907379,
1982
+ "loss": 0.6975,
1983
+ "step": 658000
1984
+ },
1985
+ {
1986
+ "epoch": 15.28,
1987
+ "learning_rate": 0.00014865577103900375,
1988
+ "loss": 0.7047,
1989
+ "step": 660000
1990
+ },
1991
+ {
1992
+ "epoch": 15.33,
1993
+ "learning_rate": 0.00014818796906893372,
1994
+ "loss": 0.6975,
1995
+ "step": 662000
1996
+ },
1997
+ {
1998
+ "epoch": 15.38,
1999
+ "learning_rate": 0.0001477201670988637,
2000
+ "loss": 0.704,
2001
+ "step": 664000
2002
+ },
2003
+ {
2004
+ "epoch": 15.42,
2005
+ "learning_rate": 0.00014725236512879366,
2006
+ "loss": 0.7042,
2007
+ "step": 666000
2008
+ },
2009
+ {
2010
+ "epoch": 15.47,
2011
+ "learning_rate": 0.00014678456315872363,
2012
+ "loss": 0.6917,
2013
+ "step": 668000
2014
+ },
2015
+ {
2016
+ "epoch": 15.51,
2017
+ "learning_rate": 0.0001463167611886536,
2018
+ "loss": 0.6914,
2019
+ "step": 670000
2020
+ },
2021
+ {
2022
+ "epoch": 15.56,
2023
+ "learning_rate": 0.00014584895921858357,
2024
+ "loss": 0.7018,
2025
+ "step": 672000
2026
+ },
2027
+ {
2028
+ "epoch": 15.61,
2029
+ "learning_rate": 0.00014538115724851354,
2030
+ "loss": 0.7016,
2031
+ "step": 674000
2032
+ },
2033
+ {
2034
+ "epoch": 15.65,
2035
+ "learning_rate": 0.0001449133552784435,
2036
+ "loss": 0.7078,
2037
+ "step": 676000
2038
+ },
2039
+ {
2040
+ "epoch": 15.7,
2041
+ "learning_rate": 0.00014444555330837347,
2042
+ "loss": 0.6932,
2043
+ "step": 678000
2044
+ },
2045
+ {
2046
+ "epoch": 15.75,
2047
+ "learning_rate": 0.00014397775133830344,
2048
+ "loss": 0.6964,
2049
+ "step": 680000
2050
+ },
2051
+ {
2052
+ "epoch": 15.79,
2053
+ "learning_rate": 0.0001435099493682334,
2054
+ "loss": 0.6997,
2055
+ "step": 682000
2056
+ },
2057
+ {
2058
+ "epoch": 15.84,
2059
+ "learning_rate": 0.00014304214739816338,
2060
+ "loss": 0.7065,
2061
+ "step": 684000
2062
+ },
2063
+ {
2064
+ "epoch": 15.88,
2065
+ "learning_rate": 0.00014257434542809335,
2066
+ "loss": 0.7047,
2067
+ "step": 686000
2068
+ },
2069
+ {
2070
+ "epoch": 15.93,
2071
+ "learning_rate": 0.00014210654345802332,
2072
+ "loss": 0.7154,
2073
+ "step": 688000
2074
+ },
2075
+ {
2076
+ "epoch": 15.98,
2077
+ "learning_rate": 0.0001416387414879533,
2078
+ "loss": 0.6993,
2079
+ "step": 690000
2080
+ },
2081
+ {
2082
+ "epoch": 16.02,
2083
+ "learning_rate": 0.00014117093951788326,
2084
+ "loss": 0.6969,
2085
+ "step": 692000
2086
+ },
2087
+ {
2088
+ "epoch": 16.07,
2089
+ "learning_rate": 0.00014070313754781325,
2090
+ "loss": 0.689,
2091
+ "step": 694000
2092
+ },
2093
+ {
2094
+ "epoch": 16.12,
2095
+ "learning_rate": 0.00014023533557774322,
2096
+ "loss": 0.6888,
2097
+ "step": 696000
2098
+ },
2099
+ {
2100
+ "epoch": 16.16,
2101
+ "learning_rate": 0.0001397675336076732,
2102
+ "loss": 0.6818,
2103
+ "step": 698000
2104
+ },
2105
+ {
2106
+ "epoch": 16.21,
2107
+ "learning_rate": 0.00013929973163760316,
2108
+ "loss": 0.693,
2109
+ "step": 700000
2110
+ },
2111
+ {
2112
+ "epoch": 16.26,
2113
+ "learning_rate": 0.00013883192966753313,
2114
+ "loss": 0.6909,
2115
+ "step": 702000
2116
+ },
2117
+ {
2118
+ "epoch": 16.3,
2119
+ "learning_rate": 0.0001383641276974631,
2120
+ "loss": 0.6873,
2121
+ "step": 704000
2122
+ },
2123
+ {
2124
+ "epoch": 16.35,
2125
+ "learning_rate": 0.00013789632572739307,
2126
+ "loss": 0.6906,
2127
+ "step": 706000
2128
+ },
2129
+ {
2130
+ "epoch": 16.39,
2131
+ "learning_rate": 0.00013742852375732303,
2132
+ "loss": 0.6866,
2133
+ "step": 708000
2134
+ },
2135
+ {
2136
+ "epoch": 16.44,
2137
+ "learning_rate": 0.000136960721787253,
2138
+ "loss": 0.701,
2139
+ "step": 710000
2140
+ },
2141
+ {
2142
+ "epoch": 16.49,
2143
+ "learning_rate": 0.00013649291981718297,
2144
+ "loss": 0.6937,
2145
+ "step": 712000
2146
+ },
2147
+ {
2148
+ "epoch": 16.53,
2149
+ "learning_rate": 0.00013602511784711294,
2150
+ "loss": 0.6907,
2151
+ "step": 714000
2152
+ },
2153
+ {
2154
+ "epoch": 16.58,
2155
+ "learning_rate": 0.0001355573158770429,
2156
+ "loss": 0.6897,
2157
+ "step": 716000
2158
+ },
2159
+ {
2160
+ "epoch": 16.63,
2161
+ "learning_rate": 0.00013508951390697288,
2162
+ "loss": 0.6952,
2163
+ "step": 718000
2164
+ },
2165
+ {
2166
+ "epoch": 16.67,
2167
+ "learning_rate": 0.00013462171193690285,
2168
+ "loss": 0.6865,
2169
+ "step": 720000
2170
+ },
2171
+ {
2172
+ "epoch": 16.72,
2173
+ "learning_rate": 0.00013415390996683282,
2174
+ "loss": 0.6935,
2175
+ "step": 722000
2176
+ },
2177
+ {
2178
+ "epoch": 16.76,
2179
+ "learning_rate": 0.00013368610799676278,
2180
+ "loss": 0.6919,
2181
+ "step": 724000
2182
+ },
2183
+ {
2184
+ "epoch": 16.81,
2185
+ "learning_rate": 0.00013321830602669275,
2186
+ "loss": 0.6904,
2187
+ "step": 726000
2188
+ },
2189
+ {
2190
+ "epoch": 16.86,
2191
+ "learning_rate": 0.00013275050405662272,
2192
+ "loss": 0.6964,
2193
+ "step": 728000
2194
+ },
2195
+ {
2196
+ "epoch": 16.9,
2197
+ "learning_rate": 0.00013228270208655272,
2198
+ "loss": 0.6943,
2199
+ "step": 730000
2200
+ },
2201
+ {
2202
+ "epoch": 16.95,
2203
+ "learning_rate": 0.0001318149001164827,
2204
+ "loss": 0.6949,
2205
+ "step": 732000
2206
+ },
2207
+ {
2208
+ "epoch": 17.0,
2209
+ "learning_rate": 0.00013134709814641266,
2210
+ "loss": 0.6943,
2211
+ "step": 734000
2212
+ },
2213
+ {
2214
+ "epoch": 17.04,
2215
+ "learning_rate": 0.00013087929617634263,
2216
+ "loss": 0.6851,
2217
+ "step": 736000
2218
+ },
2219
+ {
2220
+ "epoch": 17.09,
2221
+ "learning_rate": 0.0001304114942062726,
2222
+ "loss": 0.6802,
2223
+ "step": 738000
2224
+ },
2225
+ {
2226
+ "epoch": 17.14,
2227
+ "learning_rate": 0.00012994369223620256,
2228
+ "loss": 0.6801,
2229
+ "step": 740000
2230
+ },
2231
+ {
2232
+ "epoch": 17.18,
2233
+ "learning_rate": 0.00012947589026613253,
2234
+ "loss": 0.6756,
2235
+ "step": 742000
2236
+ },
2237
+ {
2238
+ "epoch": 17.23,
2239
+ "learning_rate": 0.0001290080882960625,
2240
+ "loss": 0.6824,
2241
+ "step": 744000
2242
+ },
2243
+ {
2244
+ "epoch": 17.27,
2245
+ "learning_rate": 0.00012854028632599247,
2246
+ "loss": 0.6894,
2247
+ "step": 746000
2248
+ },
2249
+ {
2250
+ "epoch": 17.32,
2251
+ "learning_rate": 0.00012807248435592244,
2252
+ "loss": 0.682,
2253
+ "step": 748000
2254
+ },
2255
+ {
2256
+ "epoch": 17.37,
2257
+ "learning_rate": 0.0001276046823858524,
2258
+ "loss": 0.6814,
2259
+ "step": 750000
2260
+ },
2261
+ {
2262
+ "epoch": 17.41,
2263
+ "learning_rate": 0.00012713688041578238,
2264
+ "loss": 0.6737,
2265
+ "step": 752000
2266
+ },
2267
+ {
2268
+ "epoch": 17.46,
2269
+ "learning_rate": 0.00012666907844571234,
2270
+ "loss": 0.6874,
2271
+ "step": 754000
2272
+ },
2273
+ {
2274
+ "epoch": 17.51,
2275
+ "learning_rate": 0.00012620127647564231,
2276
+ "loss": 0.6842,
2277
+ "step": 756000
2278
+ },
2279
+ {
2280
+ "epoch": 17.55,
2281
+ "learning_rate": 0.00012573347450557228,
2282
+ "loss": 0.6871,
2283
+ "step": 758000
2284
+ },
2285
+ {
2286
+ "epoch": 17.6,
2287
+ "learning_rate": 0.00012526567253550225,
2288
+ "loss": 0.6833,
2289
+ "step": 760000
2290
+ },
2291
+ {
2292
+ "epoch": 17.64,
2293
+ "learning_rate": 0.00012479787056543222,
2294
+ "loss": 0.6818,
2295
+ "step": 762000
2296
+ },
2297
+ {
2298
+ "epoch": 17.69,
2299
+ "learning_rate": 0.0001243300685953622,
2300
+ "loss": 0.6824,
2301
+ "step": 764000
2302
+ },
2303
+ {
2304
+ "epoch": 17.74,
2305
+ "learning_rate": 0.00012386226662529219,
2306
+ "loss": 0.684,
2307
+ "step": 766000
2308
+ },
2309
+ {
2310
+ "epoch": 17.78,
2311
+ "learning_rate": 0.00012339446465522215,
2312
+ "loss": 0.6822,
2313
+ "step": 768000
2314
+ },
2315
+ {
2316
+ "epoch": 17.83,
2317
+ "learning_rate": 0.00012292666268515212,
2318
+ "loss": 0.68,
2319
+ "step": 770000
2320
+ },
2321
+ {
2322
+ "epoch": 17.88,
2323
+ "learning_rate": 0.0001224588607150821,
2324
+ "loss": 0.6842,
2325
+ "step": 772000
2326
+ },
2327
+ {
2328
+ "epoch": 17.92,
2329
+ "learning_rate": 0.00012199105874501206,
2330
+ "loss": 0.6827,
2331
+ "step": 774000
2332
+ },
2333
+ {
2334
+ "epoch": 17.97,
2335
+ "learning_rate": 0.00012152325677494203,
2336
+ "loss": 0.6901,
2337
+ "step": 776000
2338
+ },
2339
+ {
2340
+ "epoch": 18.02,
2341
+ "learning_rate": 0.000121055454804872,
2342
+ "loss": 0.6781,
2343
+ "step": 778000
2344
+ },
2345
+ {
2346
+ "epoch": 18.06,
2347
+ "learning_rate": 0.00012058765283480197,
2348
+ "loss": 0.6768,
2349
+ "step": 780000
2350
+ },
2351
+ {
2352
+ "epoch": 18.11,
2353
+ "learning_rate": 0.00012011985086473194,
2354
+ "loss": 0.6704,
2355
+ "step": 782000
2356
+ },
2357
+ {
2358
+ "epoch": 18.15,
2359
+ "learning_rate": 0.0001196520488946619,
2360
+ "loss": 0.6767,
2361
+ "step": 784000
2362
+ },
2363
+ {
2364
+ "epoch": 18.2,
2365
+ "learning_rate": 0.00011918424692459187,
2366
+ "loss": 0.6696,
2367
+ "step": 786000
2368
+ },
2369
+ {
2370
+ "epoch": 18.25,
2371
+ "learning_rate": 0.00011871644495452184,
2372
+ "loss": 0.6717,
2373
+ "step": 788000
2374
+ },
2375
+ {
2376
+ "epoch": 18.29,
2377
+ "learning_rate": 0.00011824864298445181,
2378
+ "loss": 0.6666,
2379
+ "step": 790000
2380
+ },
2381
+ {
2382
+ "epoch": 18.34,
2383
+ "learning_rate": 0.00011778084101438178,
2384
+ "loss": 0.6681,
2385
+ "step": 792000
2386
+ },
2387
+ {
2388
+ "epoch": 18.39,
2389
+ "learning_rate": 0.00011731303904431175,
2390
+ "loss": 0.6688,
2391
+ "step": 794000
2392
+ },
2393
+ {
2394
+ "epoch": 18.43,
2395
+ "learning_rate": 0.00011684523707424172,
2396
+ "loss": 0.6809,
2397
+ "step": 796000
2398
+ },
2399
+ {
2400
+ "epoch": 18.48,
2401
+ "learning_rate": 0.00011637743510417169,
2402
+ "loss": 0.6704,
2403
+ "step": 798000
2404
+ },
2405
+ {
2406
+ "epoch": 18.52,
2407
+ "learning_rate": 0.00011590963313410166,
2408
+ "loss": 0.6732,
2409
+ "step": 800000
2410
+ },
2411
+ {
2412
+ "epoch": 18.57,
2413
+ "learning_rate": 0.00011544183116403164,
2414
+ "loss": 0.6688,
2415
+ "step": 802000
2416
+ },
2417
+ {
2418
+ "epoch": 18.62,
2419
+ "learning_rate": 0.00011497402919396161,
2420
+ "loss": 0.6767,
2421
+ "step": 804000
2422
+ },
2423
+ {
2424
+ "epoch": 18.66,
2425
+ "learning_rate": 0.00011450622722389158,
2426
+ "loss": 0.6721,
2427
+ "step": 806000
2428
+ },
2429
+ {
2430
+ "epoch": 18.71,
2431
+ "learning_rate": 0.00011403842525382154,
2432
+ "loss": 0.6716,
2433
+ "step": 808000
2434
+ },
2435
+ {
2436
+ "epoch": 18.76,
2437
+ "learning_rate": 0.00011357062328375151,
2438
+ "loss": 0.673,
2439
+ "step": 810000
2440
+ },
2441
+ {
2442
+ "epoch": 18.8,
2443
+ "learning_rate": 0.00011310282131368148,
2444
+ "loss": 0.6717,
2445
+ "step": 812000
2446
+ },
2447
+ {
2448
+ "epoch": 18.85,
2449
+ "learning_rate": 0.00011263501934361145,
2450
+ "loss": 0.6607,
2451
+ "step": 814000
2452
+ },
2453
+ {
2454
+ "epoch": 18.9,
2455
+ "learning_rate": 0.00011216721737354142,
2456
+ "loss": 0.6732,
2457
+ "step": 816000
2458
+ },
2459
+ {
2460
+ "epoch": 18.94,
2461
+ "learning_rate": 0.00011169941540347139,
2462
+ "loss": 0.6715,
2463
+ "step": 818000
2464
+ },
2465
+ {
2466
+ "epoch": 18.99,
2467
+ "learning_rate": 0.00011123161343340136,
2468
+ "loss": 0.678,
2469
+ "step": 820000
2470
+ },
2471
+ {
2472
+ "epoch": 19.03,
2473
+ "learning_rate": 0.00011076381146333133,
2474
+ "loss": 0.6618,
2475
+ "step": 822000
2476
+ },
2477
+ {
2478
+ "epoch": 19.08,
2479
+ "learning_rate": 0.0001102960094932613,
2480
+ "loss": 0.6589,
2481
+ "step": 824000
2482
+ },
2483
+ {
2484
+ "epoch": 19.13,
2485
+ "learning_rate": 0.00010982820752319126,
2486
+ "loss": 0.6624,
2487
+ "step": 826000
2488
+ },
2489
+ {
2490
+ "epoch": 19.17,
2491
+ "learning_rate": 0.00010936040555312123,
2492
+ "loss": 0.6618,
2493
+ "step": 828000
2494
+ },
2495
+ {
2496
+ "epoch": 19.22,
2497
+ "learning_rate": 0.0001088926035830512,
2498
+ "loss": 0.6666,
2499
+ "step": 830000
2500
+ },
2501
+ {
2502
+ "epoch": 19.27,
2503
+ "learning_rate": 0.00010842480161298117,
2504
+ "loss": 0.6645,
2505
+ "step": 832000
2506
+ },
2507
+ {
2508
+ "epoch": 19.31,
2509
+ "learning_rate": 0.00010795699964291114,
2510
+ "loss": 0.6667,
2511
+ "step": 834000
2512
+ },
2513
+ {
2514
+ "epoch": 19.36,
2515
+ "learning_rate": 0.00010748919767284111,
2516
+ "loss": 0.6649,
2517
+ "step": 836000
2518
+ },
2519
+ {
2520
+ "epoch": 19.4,
2521
+ "learning_rate": 0.0001070213957027711,
2522
+ "loss": 0.659,
2523
+ "step": 838000
2524
+ },
2525
+ {
2526
+ "epoch": 19.45,
2527
+ "learning_rate": 0.00010655359373270107,
2528
+ "loss": 0.6611,
2529
+ "step": 840000
2530
+ },
2531
+ {
2532
+ "epoch": 19.5,
2533
+ "learning_rate": 0.00010608579176263104,
2534
+ "loss": 0.6565,
2535
+ "step": 842000
2536
+ },
2537
+ {
2538
+ "epoch": 19.54,
2539
+ "learning_rate": 0.00010561798979256101,
2540
+ "loss": 0.6631,
2541
+ "step": 844000
2542
+ },
2543
+ {
2544
+ "epoch": 19.59,
2545
+ "learning_rate": 0.00010515018782249098,
2546
+ "loss": 0.6593,
2547
+ "step": 846000
2548
+ },
2549
+ {
2550
+ "epoch": 19.64,
2551
+ "learning_rate": 0.00010468238585242095,
2552
+ "loss": 0.6654,
2553
+ "step": 848000
2554
+ },
2555
+ {
2556
+ "epoch": 19.68,
2557
+ "learning_rate": 0.00010421458388235092,
2558
+ "loss": 0.6621,
2559
+ "step": 850000
2560
+ },
2561
+ {
2562
+ "epoch": 19.73,
2563
+ "learning_rate": 0.00010374678191228089,
2564
+ "loss": 0.661,
2565
+ "step": 852000
2566
+ },
2567
+ {
2568
+ "epoch": 19.78,
2569
+ "learning_rate": 0.00010327897994221086,
2570
+ "loss": 0.6515,
2571
+ "step": 854000
2572
+ },
2573
+ {
2574
+ "epoch": 19.82,
2575
+ "learning_rate": 0.00010281117797214082,
2576
+ "loss": 0.6614,
2577
+ "step": 856000
2578
+ },
2579
+ {
2580
+ "epoch": 19.87,
2581
+ "learning_rate": 0.0001023433760020708,
2582
+ "loss": 0.6616,
2583
+ "step": 858000
2584
+ },
2585
+ {
2586
+ "epoch": 19.91,
2587
+ "learning_rate": 0.00010187557403200076,
2588
+ "loss": 0.6598,
2589
+ "step": 860000
2590
+ },
2591
+ {
2592
+ "epoch": 19.96,
2593
+ "learning_rate": 0.00010140777206193073,
2594
+ "loss": 0.6616,
2595
+ "step": 862000
2596
+ },
2597
+ {
2598
+ "epoch": 20.01,
2599
+ "learning_rate": 0.0001009399700918607,
2600
+ "loss": 0.6679,
2601
+ "step": 864000
2602
+ },
2603
+ {
2604
+ "epoch": 20.05,
2605
+ "learning_rate": 0.00010047216812179067,
2606
+ "loss": 0.6518,
2607
+ "step": 866000
2608
+ },
2609
+ {
2610
+ "epoch": 20.1,
2611
+ "learning_rate": 0.00010000436615172064,
2612
+ "loss": 0.6463,
2613
+ "step": 868000
2614
+ },
2615
+ {
2616
+ "epoch": 20.15,
2617
+ "learning_rate": 9.95365641816506e-05,
2618
+ "loss": 0.6529,
2619
+ "step": 870000
2620
+ },
2621
+ {
2622
+ "epoch": 20.19,
2623
+ "learning_rate": 9.906876221158058e-05,
2624
+ "loss": 0.6463,
2625
+ "step": 872000
2626
+ },
2627
+ {
2628
+ "epoch": 20.24,
2629
+ "learning_rate": 9.860096024151056e-05,
2630
+ "loss": 0.6545,
2631
+ "step": 874000
2632
+ },
2633
+ {
2634
+ "epoch": 20.28,
2635
+ "learning_rate": 9.813315827144053e-05,
2636
+ "loss": 0.6531,
2637
+ "step": 876000
2638
+ },
2639
+ {
2640
+ "epoch": 20.33,
2641
+ "learning_rate": 9.76653563013705e-05,
2642
+ "loss": 0.6442,
2643
+ "step": 878000
2644
+ },
2645
+ {
2646
+ "epoch": 20.38,
2647
+ "learning_rate": 9.719755433130046e-05,
2648
+ "loss": 0.65,
2649
+ "step": 880000
2650
+ },
2651
+ {
2652
+ "epoch": 20.42,
2653
+ "learning_rate": 9.672975236123043e-05,
2654
+ "loss": 0.6518,
2655
+ "step": 882000
2656
+ },
2657
+ {
2658
+ "epoch": 20.47,
2659
+ "learning_rate": 9.62619503911604e-05,
2660
+ "loss": 0.6546,
2661
+ "step": 884000
2662
+ },
2663
+ {
2664
+ "epoch": 20.52,
2665
+ "learning_rate": 9.579414842109037e-05,
2666
+ "loss": 0.6494,
2667
+ "step": 886000
2668
+ },
2669
+ {
2670
+ "epoch": 20.56,
2671
+ "learning_rate": 9.532634645102035e-05,
2672
+ "loss": 0.654,
2673
+ "step": 888000
2674
+ },
2675
+ {
2676
+ "epoch": 20.61,
2677
+ "learning_rate": 9.485854448095032e-05,
2678
+ "loss": 0.6536,
2679
+ "step": 890000
2680
+ },
2681
+ {
2682
+ "epoch": 20.66,
2683
+ "learning_rate": 9.439074251088029e-05,
2684
+ "loss": 0.6547,
2685
+ "step": 892000
2686
+ },
2687
+ {
2688
+ "epoch": 20.7,
2689
+ "learning_rate": 9.392294054081026e-05,
2690
+ "loss": 0.6421,
2691
+ "step": 894000
2692
+ },
2693
+ {
2694
+ "epoch": 20.75,
2695
+ "learning_rate": 9.345513857074023e-05,
2696
+ "loss": 0.6506,
2697
+ "step": 896000
2698
+ },
2699
+ {
2700
+ "epoch": 20.79,
2701
+ "learning_rate": 9.29873366006702e-05,
2702
+ "loss": 0.6551,
2703
+ "step": 898000
2704
+ },
2705
+ {
2706
+ "epoch": 20.84,
2707
+ "learning_rate": 9.251953463060017e-05,
2708
+ "loss": 0.6542,
2709
+ "step": 900000
2710
+ },
2711
+ {
2712
+ "epoch": 20.89,
2713
+ "learning_rate": 9.205173266053014e-05,
2714
+ "loss": 0.6398,
2715
+ "step": 902000
2716
+ },
2717
+ {
2718
+ "epoch": 20.93,
2719
+ "learning_rate": 9.15839306904601e-05,
2720
+ "loss": 0.653,
2721
+ "step": 904000
2722
+ },
2723
+ {
2724
+ "epoch": 20.98,
2725
+ "learning_rate": 9.111612872039007e-05,
2726
+ "loss": 0.6476,
2727
+ "step": 906000
2728
+ },
2729
+ {
2730
+ "epoch": 21.03,
2731
+ "learning_rate": 9.064832675032004e-05,
2732
+ "loss": 0.6378,
2733
+ "step": 908000
2734
+ },
2735
+ {
2736
+ "epoch": 21.07,
2737
+ "learning_rate": 9.018052478025002e-05,
2738
+ "loss": 0.6413,
2739
+ "step": 910000
2740
+ },
2741
+ {
2742
+ "epoch": 21.12,
2743
+ "learning_rate": 8.971272281017999e-05,
2744
+ "loss": 0.6368,
2745
+ "step": 912000
2746
+ },
2747
+ {
2748
+ "epoch": 21.16,
2749
+ "learning_rate": 8.924492084010996e-05,
2750
+ "loss": 0.6366,
2751
+ "step": 914000
2752
+ },
2753
+ {
2754
+ "epoch": 21.21,
2755
+ "learning_rate": 8.877711887003993e-05,
2756
+ "loss": 0.6455,
2757
+ "step": 916000
2758
+ },
2759
+ {
2760
+ "epoch": 21.26,
2761
+ "learning_rate": 8.83093168999699e-05,
2762
+ "loss": 0.6448,
2763
+ "step": 918000
2764
+ },
2765
+ {
2766
+ "epoch": 21.3,
2767
+ "learning_rate": 8.784151492989987e-05,
2768
+ "loss": 0.6371,
2769
+ "step": 920000
2770
+ },
2771
+ {
2772
+ "epoch": 21.35,
2773
+ "learning_rate": 8.737371295982984e-05,
2774
+ "loss": 0.6457,
2775
+ "step": 922000
2776
+ },
2777
+ {
2778
+ "epoch": 21.4,
2779
+ "learning_rate": 8.69059109897598e-05,
2780
+ "loss": 0.6399,
2781
+ "step": 924000
2782
+ },
2783
+ {
2784
+ "epoch": 21.44,
2785
+ "learning_rate": 8.643810901968978e-05,
2786
+ "loss": 0.6389,
2787
+ "step": 926000
2788
+ },
2789
+ {
2790
+ "epoch": 21.49,
2791
+ "learning_rate": 8.597030704961974e-05,
2792
+ "loss": 0.6444,
2793
+ "step": 928000
2794
+ },
2795
+ {
2796
+ "epoch": 21.54,
2797
+ "learning_rate": 8.550250507954971e-05,
2798
+ "loss": 0.6346,
2799
+ "step": 930000
2800
+ },
2801
+ {
2802
+ "epoch": 21.58,
2803
+ "learning_rate": 8.503470310947968e-05,
2804
+ "loss": 0.6394,
2805
+ "step": 932000
2806
+ },
2807
+ {
2808
+ "epoch": 21.63,
2809
+ "learning_rate": 8.456690113940965e-05,
2810
+ "loss": 0.6397,
2811
+ "step": 934000
2812
+ },
2813
+ {
2814
+ "epoch": 21.67,
2815
+ "learning_rate": 8.409909916933962e-05,
2816
+ "loss": 0.6411,
2817
+ "step": 936000
2818
+ },
2819
+ {
2820
+ "epoch": 21.72,
2821
+ "learning_rate": 8.363129719926959e-05,
2822
+ "loss": 0.6383,
2823
+ "step": 938000
2824
+ },
2825
+ {
2826
+ "epoch": 21.77,
2827
+ "learning_rate": 8.316349522919956e-05,
2828
+ "loss": 0.6416,
2829
+ "step": 940000
2830
+ },
2831
+ {
2832
+ "epoch": 21.81,
2833
+ "learning_rate": 8.269569325912953e-05,
2834
+ "loss": 0.635,
2835
+ "step": 942000
2836
+ },
2837
+ {
2838
+ "epoch": 21.86,
2839
+ "learning_rate": 8.22278912890595e-05,
2840
+ "loss": 0.6371,
2841
+ "step": 944000
2842
+ },
2843
+ {
2844
+ "epoch": 21.91,
2845
+ "learning_rate": 8.176008931898949e-05,
2846
+ "loss": 0.6412,
2847
+ "step": 946000
2848
+ },
2849
+ {
2850
+ "epoch": 21.95,
2851
+ "learning_rate": 8.129228734891946e-05,
2852
+ "loss": 0.6414,
2853
+ "step": 948000
2854
+ },
2855
+ {
2856
+ "epoch": 22.0,
2857
+ "learning_rate": 8.082448537884943e-05,
2858
+ "loss": 0.6285,
2859
+ "step": 950000
2860
+ },
2861
+ {
2862
+ "epoch": 22.04,
2863
+ "learning_rate": 8.03566834087794e-05,
2864
+ "loss": 0.6285,
2865
+ "step": 952000
2866
+ },
2867
+ {
2868
+ "epoch": 22.09,
2869
+ "learning_rate": 7.988888143870937e-05,
2870
+ "loss": 0.6268,
2871
+ "step": 954000
2872
+ },
2873
+ {
2874
+ "epoch": 22.14,
2875
+ "learning_rate": 7.942107946863934e-05,
2876
+ "loss": 0.6251,
2877
+ "step": 956000
2878
+ },
2879
+ {
2880
+ "epoch": 22.18,
2881
+ "learning_rate": 7.89532774985693e-05,
2882
+ "loss": 0.6306,
2883
+ "step": 958000
2884
+ },
2885
+ {
2886
+ "epoch": 22.23,
2887
+ "learning_rate": 7.848547552849927e-05,
2888
+ "loss": 0.6283,
2889
+ "step": 960000
2890
+ },
2891
+ {
2892
+ "epoch": 22.28,
2893
+ "learning_rate": 7.801767355842924e-05,
2894
+ "loss": 0.6264,
2895
+ "step": 962000
2896
+ },
2897
+ {
2898
+ "epoch": 22.32,
2899
+ "learning_rate": 7.754987158835921e-05,
2900
+ "loss": 0.6279,
2901
+ "step": 964000
2902
+ },
2903
+ {
2904
+ "epoch": 22.37,
2905
+ "learning_rate": 7.708206961828918e-05,
2906
+ "loss": 0.6272,
2907
+ "step": 966000
2908
+ },
2909
+ {
2910
+ "epoch": 22.41,
2911
+ "learning_rate": 7.661426764821915e-05,
2912
+ "loss": 0.6355,
2913
+ "step": 968000
2914
+ },
2915
+ {
2916
+ "epoch": 22.46,
2917
+ "learning_rate": 7.614646567814912e-05,
2918
+ "loss": 0.6349,
2919
+ "step": 970000
2920
+ },
2921
+ {
2922
+ "epoch": 22.51,
2923
+ "learning_rate": 7.567866370807909e-05,
2924
+ "loss": 0.6281,
2925
+ "step": 972000
2926
+ },
2927
+ {
2928
+ "epoch": 22.55,
2929
+ "learning_rate": 7.521086173800905e-05,
2930
+ "loss": 0.6269,
2931
+ "step": 974000
2932
+ },
2933
+ {
2934
+ "epoch": 22.6,
2935
+ "learning_rate": 7.474305976793904e-05,
2936
+ "loss": 0.6221,
2937
+ "step": 976000
2938
+ },
2939
+ {
2940
+ "epoch": 22.65,
2941
+ "learning_rate": 7.4275257797869e-05,
2942
+ "loss": 0.6295,
2943
+ "step": 978000
2944
+ },
2945
+ {
2946
+ "epoch": 22.69,
2947
+ "learning_rate": 7.380745582779897e-05,
2948
+ "loss": 0.6265,
2949
+ "step": 980000
2950
+ },
2951
+ {
2952
+ "epoch": 22.74,
2953
+ "learning_rate": 7.333965385772894e-05,
2954
+ "loss": 0.6203,
2955
+ "step": 982000
2956
+ },
2957
+ {
2958
+ "epoch": 22.79,
2959
+ "learning_rate": 7.287185188765891e-05,
2960
+ "loss": 0.6306,
2961
+ "step": 984000
2962
+ },
2963
+ {
2964
+ "epoch": 22.83,
2965
+ "learning_rate": 7.240404991758888e-05,
2966
+ "loss": 0.6319,
2967
+ "step": 986000
2968
+ },
2969
+ {
2970
+ "epoch": 22.88,
2971
+ "learning_rate": 7.193624794751885e-05,
2972
+ "loss": 0.6211,
2973
+ "step": 988000
2974
+ },
2975
+ {
2976
+ "epoch": 22.92,
2977
+ "learning_rate": 7.146844597744882e-05,
2978
+ "loss": 0.6244,
2979
+ "step": 990000
2980
+ },
2981
+ {
2982
+ "epoch": 22.97,
2983
+ "learning_rate": 7.100064400737879e-05,
2984
+ "loss": 0.6262,
2985
+ "step": 992000
2986
+ },
2987
+ {
2988
+ "epoch": 23.02,
2989
+ "learning_rate": 7.053284203730876e-05,
2990
+ "loss": 0.6166,
2991
+ "step": 994000
2992
+ },
2993
+ {
2994
+ "epoch": 23.06,
2995
+ "learning_rate": 7.006504006723873e-05,
2996
+ "loss": 0.6166,
2997
+ "step": 996000
2998
+ },
2999
+ {
3000
+ "epoch": 23.11,
3001
+ "learning_rate": 6.95972380971687e-05,
3002
+ "loss": 0.6175,
3003
+ "step": 998000
3004
+ },
3005
+ {
3006
+ "epoch": 23.16,
3007
+ "learning_rate": 6.912943612709866e-05,
3008
+ "loss": 0.6151,
3009
+ "step": 1000000
3010
+ },
3011
+ {
3012
+ "epoch": 23.2,
3013
+ "learning_rate": 6.866163415702863e-05,
3014
+ "loss": 0.6153,
3015
+ "step": 1002000
3016
+ },
3017
+ {
3018
+ "epoch": 23.25,
3019
+ "learning_rate": 6.81938321869586e-05,
3020
+ "loss": 0.6212,
3021
+ "step": 1004000
3022
+ },
3023
+ {
3024
+ "epoch": 23.29,
3025
+ "learning_rate": 6.772603021688858e-05,
3026
+ "loss": 0.6161,
3027
+ "step": 1006000
3028
+ },
3029
+ {
3030
+ "epoch": 23.34,
3031
+ "learning_rate": 6.725822824681855e-05,
3032
+ "loss": 0.6158,
3033
+ "step": 1008000
3034
+ },
3035
+ {
3036
+ "epoch": 23.39,
3037
+ "learning_rate": 6.679042627674852e-05,
3038
+ "loss": 0.6089,
3039
+ "step": 1010000
3040
+ },
3041
+ {
3042
+ "epoch": 23.43,
3043
+ "learning_rate": 6.632262430667849e-05,
3044
+ "loss": 0.6166,
3045
+ "step": 1012000
3046
+ },
3047
+ {
3048
+ "epoch": 23.48,
3049
+ "learning_rate": 6.585482233660846e-05,
3050
+ "loss": 0.6134,
3051
+ "step": 1014000
3052
+ },
3053
+ {
3054
+ "epoch": 23.53,
3055
+ "learning_rate": 6.538702036653843e-05,
3056
+ "loss": 0.6171,
3057
+ "step": 1016000
3058
+ },
3059
+ {
3060
+ "epoch": 23.57,
3061
+ "learning_rate": 6.49192183964684e-05,
3062
+ "loss": 0.6122,
3063
+ "step": 1018000
3064
+ },
3065
+ {
3066
+ "epoch": 23.62,
3067
+ "learning_rate": 6.445141642639837e-05,
3068
+ "loss": 0.6176,
3069
+ "step": 1020000
3070
+ },
3071
+ {
3072
+ "epoch": 23.67,
3073
+ "learning_rate": 6.398361445632833e-05,
3074
+ "loss": 0.6146,
3075
+ "step": 1022000
3076
+ },
3077
+ {
3078
+ "epoch": 23.71,
3079
+ "learning_rate": 6.35158124862583e-05,
3080
+ "loss": 0.6069,
3081
+ "step": 1024000
3082
+ },
3083
+ {
3084
+ "epoch": 23.76,
3085
+ "learning_rate": 6.304801051618829e-05,
3086
+ "loss": 0.6169,
3087
+ "step": 1026000
3088
+ },
3089
+ {
3090
+ "epoch": 23.8,
3091
+ "learning_rate": 6.258020854611825e-05,
3092
+ "loss": 0.6222,
3093
+ "step": 1028000
3094
+ },
3095
+ {
3096
+ "epoch": 23.85,
3097
+ "learning_rate": 6.211240657604822e-05,
3098
+ "loss": 0.6152,
3099
+ "step": 1030000
3100
+ },
3101
+ {
3102
+ "epoch": 23.9,
3103
+ "learning_rate": 6.164460460597819e-05,
3104
+ "loss": 0.6181,
3105
+ "step": 1032000
3106
+ },
3107
+ {
3108
+ "epoch": 23.94,
3109
+ "learning_rate": 6.117680263590816e-05,
3110
+ "loss": 0.6123,
3111
+ "step": 1034000
3112
+ },
3113
+ {
3114
+ "epoch": 23.99,
3115
+ "learning_rate": 6.070900066583813e-05,
3116
+ "loss": 0.619,
3117
+ "step": 1036000
3118
+ },
3119
+ {
3120
+ "epoch": 24.04,
3121
+ "learning_rate": 6.02411986957681e-05,
3122
+ "loss": 0.6099,
3123
+ "step": 1038000
3124
+ },
3125
+ {
3126
+ "epoch": 24.08,
3127
+ "learning_rate": 5.977339672569807e-05,
3128
+ "loss": 0.6098,
3129
+ "step": 1040000
3130
+ },
3131
+ {
3132
+ "epoch": 24.13,
3133
+ "learning_rate": 5.930559475562804e-05,
3134
+ "loss": 0.5965,
3135
+ "step": 1042000
3136
+ },
3137
+ {
3138
+ "epoch": 24.17,
3139
+ "learning_rate": 5.883779278555802e-05,
3140
+ "loss": 0.6059,
3141
+ "step": 1044000
3142
+ },
3143
+ {
3144
+ "epoch": 24.22,
3145
+ "learning_rate": 5.836999081548799e-05,
3146
+ "loss": 0.6021,
3147
+ "step": 1046000
3148
+ },
3149
+ {
3150
+ "epoch": 24.27,
3151
+ "learning_rate": 5.790218884541796e-05,
3152
+ "loss": 0.6093,
3153
+ "step": 1048000
3154
+ },
3155
+ {
3156
+ "epoch": 24.31,
3157
+ "learning_rate": 5.7434386875347926e-05,
3158
+ "loss": 0.6031,
3159
+ "step": 1050000
3160
+ },
3161
+ {
3162
+ "epoch": 24.36,
3163
+ "learning_rate": 5.6966584905277895e-05,
3164
+ "loss": 0.6053,
3165
+ "step": 1052000
3166
+ },
3167
+ {
3168
+ "epoch": 24.41,
3169
+ "learning_rate": 5.6498782935207863e-05,
3170
+ "loss": 0.6036,
3171
+ "step": 1054000
3172
+ },
3173
+ {
3174
+ "epoch": 24.45,
3175
+ "learning_rate": 5.603098096513783e-05,
3176
+ "loss": 0.6011,
3177
+ "step": 1056000
3178
+ },
3179
+ {
3180
+ "epoch": 24.5,
3181
+ "learning_rate": 5.55631789950678e-05,
3182
+ "loss": 0.6035,
3183
+ "step": 1058000
3184
+ },
3185
+ {
3186
+ "epoch": 24.55,
3187
+ "learning_rate": 5.509537702499777e-05,
3188
+ "loss": 0.6066,
3189
+ "step": 1060000
3190
+ },
3191
+ {
3192
+ "epoch": 24.59,
3193
+ "learning_rate": 5.4627575054927746e-05,
3194
+ "loss": 0.6061,
3195
+ "step": 1062000
3196
+ },
3197
+ {
3198
+ "epoch": 24.64,
3199
+ "learning_rate": 5.4159773084857714e-05,
3200
+ "loss": 0.6027,
3201
+ "step": 1064000
3202
+ },
3203
+ {
3204
+ "epoch": 24.68,
3205
+ "learning_rate": 5.369197111478768e-05,
3206
+ "loss": 0.6,
3207
+ "step": 1066000
3208
+ },
3209
+ {
3210
+ "epoch": 24.73,
3211
+ "learning_rate": 5.322416914471765e-05,
3212
+ "loss": 0.6062,
3213
+ "step": 1068000
3214
+ },
3215
+ {
3216
+ "epoch": 24.78,
3217
+ "learning_rate": 5.275636717464762e-05,
3218
+ "loss": 0.6003,
3219
+ "step": 1070000
3220
+ },
3221
+ {
3222
+ "epoch": 24.82,
3223
+ "learning_rate": 5.228856520457759e-05,
3224
+ "loss": 0.5988,
3225
+ "step": 1072000
3226
+ },
3227
+ {
3228
+ "epoch": 24.87,
3229
+ "learning_rate": 5.182076323450756e-05,
3230
+ "loss": 0.6096,
3231
+ "step": 1074000
3232
+ },
3233
+ {
3234
+ "epoch": 24.92,
3235
+ "learning_rate": 5.135296126443753e-05,
3236
+ "loss": 0.5988,
3237
+ "step": 1076000
3238
+ },
3239
+ {
3240
+ "epoch": 24.96,
3241
+ "learning_rate": 5.08851592943675e-05,
3242
+ "loss": 0.6086,
3243
+ "step": 1078000
3244
+ },
3245
+ {
3246
+ "epoch": 25.01,
3247
+ "learning_rate": 5.041735732429748e-05,
3248
+ "loss": 0.5942,
3249
+ "step": 1080000
3250
+ },
3251
+ {
3252
+ "epoch": 25.05,
3253
+ "learning_rate": 4.994955535422745e-05,
3254
+ "loss": 0.5954,
3255
+ "step": 1082000
3256
+ },
3257
+ {
3258
+ "epoch": 25.1,
3259
+ "learning_rate": 4.9481753384157417e-05,
3260
+ "loss": 0.5948,
3261
+ "step": 1084000
3262
+ },
3263
+ {
3264
+ "epoch": 25.15,
3265
+ "learning_rate": 4.9013951414087385e-05,
3266
+ "loss": 0.5946,
3267
+ "step": 1086000
3268
+ },
3269
+ {
3270
+ "epoch": 25.19,
3271
+ "learning_rate": 4.8546149444017354e-05,
3272
+ "loss": 0.5938,
3273
+ "step": 1088000
3274
+ },
3275
+ {
3276
+ "epoch": 25.24,
3277
+ "learning_rate": 4.807834747394732e-05,
3278
+ "loss": 0.5961,
3279
+ "step": 1090000
3280
+ },
3281
+ {
3282
+ "epoch": 25.29,
3283
+ "learning_rate": 4.761054550387729e-05,
3284
+ "loss": 0.5947,
3285
+ "step": 1092000
3286
+ },
3287
+ {
3288
+ "epoch": 25.33,
3289
+ "learning_rate": 4.714274353380726e-05,
3290
+ "loss": 0.6019,
3291
+ "step": 1094000
3292
+ },
3293
+ {
3294
+ "epoch": 25.38,
3295
+ "learning_rate": 4.667494156373723e-05,
3296
+ "loss": 0.5927,
3297
+ "step": 1096000
3298
+ },
3299
+ {
3300
+ "epoch": 25.43,
3301
+ "learning_rate": 4.6207139593667205e-05,
3302
+ "loss": 0.5921,
3303
+ "step": 1098000
3304
+ },
3305
+ {
3306
+ "epoch": 25.47,
3307
+ "learning_rate": 4.5739337623597174e-05,
3308
+ "loss": 0.5954,
3309
+ "step": 1100000
3310
+ },
3311
+ {
3312
+ "epoch": 25.52,
3313
+ "learning_rate": 4.527153565352715e-05,
3314
+ "loss": 0.5926,
3315
+ "step": 1102000
3316
+ },
3317
+ {
3318
+ "epoch": 25.56,
3319
+ "learning_rate": 4.480373368345712e-05,
3320
+ "loss": 0.5963,
3321
+ "step": 1104000
3322
+ },
3323
+ {
3324
+ "epoch": 25.61,
3325
+ "learning_rate": 4.433593171338709e-05,
3326
+ "loss": 0.5902,
3327
+ "step": 1106000
3328
+ },
3329
+ {
3330
+ "epoch": 25.66,
3331
+ "learning_rate": 4.3868129743317056e-05,
3332
+ "loss": 0.5952,
3333
+ "step": 1108000
3334
+ },
3335
+ {
3336
+ "epoch": 25.7,
3337
+ "learning_rate": 4.3400327773247025e-05,
3338
+ "loss": 0.5878,
3339
+ "step": 1110000
3340
+ },
3341
+ {
3342
+ "epoch": 25.75,
3343
+ "learning_rate": 4.2932525803176994e-05,
3344
+ "loss": 0.5926,
3345
+ "step": 1112000
3346
+ },
3347
+ {
3348
+ "epoch": 25.8,
3349
+ "learning_rate": 4.246472383310696e-05,
3350
+ "loss": 0.5854,
3351
+ "step": 1114000
3352
+ },
3353
+ {
3354
+ "epoch": 25.84,
3355
+ "learning_rate": 4.199692186303694e-05,
3356
+ "loss": 0.5916,
3357
+ "step": 1116000
3358
+ },
3359
+ {
3360
+ "epoch": 25.89,
3361
+ "learning_rate": 4.152911989296691e-05,
3362
+ "loss": 0.5869,
3363
+ "step": 1118000
3364
+ },
3365
+ {
3366
+ "epoch": 25.93,
3367
+ "learning_rate": 4.1061317922896876e-05,
3368
+ "loss": 0.5913,
3369
+ "step": 1120000
3370
+ },
3371
+ {
3372
+ "epoch": 25.98,
3373
+ "learning_rate": 4.0593515952826845e-05,
3374
+ "loss": 0.5822,
3375
+ "step": 1122000
3376
+ },
3377
+ {
3378
+ "epoch": 26.03,
3379
+ "learning_rate": 4.0125713982756814e-05,
3380
+ "loss": 0.5831,
3381
+ "step": 1124000
3382
+ },
3383
+ {
3384
+ "epoch": 26.07,
3385
+ "learning_rate": 3.965791201268678e-05,
3386
+ "loss": 0.5847,
3387
+ "step": 1126000
3388
+ },
3389
+ {
3390
+ "epoch": 26.12,
3391
+ "learning_rate": 3.919011004261675e-05,
3392
+ "loss": 0.5828,
3393
+ "step": 1128000
3394
+ },
3395
+ {
3396
+ "epoch": 26.17,
3397
+ "learning_rate": 3.872230807254672e-05,
3398
+ "loss": 0.5825,
3399
+ "step": 1130000
3400
+ },
3401
+ {
3402
+ "epoch": 26.21,
3403
+ "learning_rate": 3.825450610247669e-05,
3404
+ "loss": 0.5848,
3405
+ "step": 1132000
3406
+ },
3407
+ {
3408
+ "epoch": 26.26,
3409
+ "learning_rate": 3.778670413240667e-05,
3410
+ "loss": 0.5866,
3411
+ "step": 1134000
3412
+ },
3413
+ {
3414
+ "epoch": 26.31,
3415
+ "learning_rate": 3.7318902162336634e-05,
3416
+ "loss": 0.5832,
3417
+ "step": 1136000
3418
+ },
3419
+ {
3420
+ "epoch": 26.35,
3421
+ "learning_rate": 3.685110019226661e-05,
3422
+ "loss": 0.58,
3423
+ "step": 1138000
3424
+ },
3425
+ {
3426
+ "epoch": 26.4,
3427
+ "learning_rate": 3.638329822219658e-05,
3428
+ "loss": 0.5767,
3429
+ "step": 1140000
3430
+ },
3431
+ {
3432
+ "epoch": 26.44,
3433
+ "learning_rate": 3.591549625212655e-05,
3434
+ "loss": 0.5792,
3435
+ "step": 1142000
3436
+ },
3437
+ {
3438
+ "epoch": 26.49,
3439
+ "learning_rate": 3.5447694282056516e-05,
3440
+ "loss": 0.5764,
3441
+ "step": 1144000
3442
+ },
3443
+ {
3444
+ "epoch": 26.54,
3445
+ "learning_rate": 3.4979892311986485e-05,
3446
+ "loss": 0.5794,
3447
+ "step": 1146000
3448
+ },
3449
+ {
3450
+ "epoch": 26.58,
3451
+ "learning_rate": 3.451209034191646e-05,
3452
+ "loss": 0.5738,
3453
+ "step": 1148000
3454
+ },
3455
+ {
3456
+ "epoch": 26.63,
3457
+ "learning_rate": 3.404428837184643e-05,
3458
+ "loss": 0.5822,
3459
+ "step": 1150000
3460
+ },
3461
+ {
3462
+ "epoch": 26.68,
3463
+ "learning_rate": 3.35764864017764e-05,
3464
+ "loss": 0.5734,
3465
+ "step": 1152000
3466
+ },
3467
+ {
3468
+ "epoch": 26.72,
3469
+ "learning_rate": 3.310868443170637e-05,
3470
+ "loss": 0.5794,
3471
+ "step": 1154000
3472
+ },
3473
+ {
3474
+ "epoch": 26.77,
3475
+ "learning_rate": 3.2640882461636336e-05,
3476
+ "loss": 0.5853,
3477
+ "step": 1156000
3478
+ },
3479
+ {
3480
+ "epoch": 26.81,
3481
+ "learning_rate": 3.2173080491566305e-05,
3482
+ "loss": 0.5842,
3483
+ "step": 1158000
3484
+ },
3485
+ {
3486
+ "epoch": 26.86,
3487
+ "learning_rate": 3.170527852149628e-05,
3488
+ "loss": 0.5847,
3489
+ "step": 1160000
3490
+ },
3491
+ {
3492
+ "epoch": 26.91,
3493
+ "learning_rate": 3.123747655142625e-05,
3494
+ "loss": 0.5786,
3495
+ "step": 1162000
3496
+ },
3497
+ {
3498
+ "epoch": 26.95,
3499
+ "learning_rate": 3.076967458135622e-05,
3500
+ "loss": 0.5818,
3501
+ "step": 1164000
3502
+ },
3503
+ {
3504
+ "epoch": 27.0,
3505
+ "learning_rate": 3.030187261128619e-05,
3506
+ "loss": 0.5722,
3507
+ "step": 1166000
3508
+ },
3509
+ {
3510
+ "epoch": 27.05,
3511
+ "learning_rate": 2.983407064121616e-05,
3512
+ "loss": 0.5726,
3513
+ "step": 1168000
3514
+ },
3515
+ {
3516
+ "epoch": 27.09,
3517
+ "learning_rate": 2.936626867114613e-05,
3518
+ "loss": 0.5745,
3519
+ "step": 1170000
3520
+ },
3521
+ {
3522
+ "epoch": 27.14,
3523
+ "learning_rate": 2.8898466701076097e-05,
3524
+ "loss": 0.5655,
3525
+ "step": 1172000
3526
+ },
3527
+ {
3528
+ "epoch": 27.19,
3529
+ "learning_rate": 2.843066473100607e-05,
3530
+ "loss": 0.5747,
3531
+ "step": 1174000
3532
+ },
3533
+ {
3534
+ "epoch": 27.23,
3535
+ "learning_rate": 2.7962862760936038e-05,
3536
+ "loss": 0.5734,
3537
+ "step": 1176000
3538
+ },
3539
+ {
3540
+ "epoch": 27.28,
3541
+ "learning_rate": 2.7495060790866007e-05,
3542
+ "loss": 0.5752,
3543
+ "step": 1178000
3544
+ },
3545
+ {
3546
+ "epoch": 27.32,
3547
+ "learning_rate": 2.7027258820795976e-05,
3548
+ "loss": 0.5784,
3549
+ "step": 1180000
3550
+ },
3551
+ {
3552
+ "epoch": 27.37,
3553
+ "learning_rate": 2.6559456850725945e-05,
3554
+ "loss": 0.5667,
3555
+ "step": 1182000
3556
+ },
3557
+ {
3558
+ "epoch": 27.42,
3559
+ "learning_rate": 2.609165488065592e-05,
3560
+ "loss": 0.5748,
3561
+ "step": 1184000
3562
+ },
3563
+ {
3564
+ "epoch": 27.46,
3565
+ "learning_rate": 2.562385291058589e-05,
3566
+ "loss": 0.5762,
3567
+ "step": 1186000
3568
+ },
3569
+ {
3570
+ "epoch": 27.51,
3571
+ "learning_rate": 2.5156050940515858e-05,
3572
+ "loss": 0.5783,
3573
+ "step": 1188000
3574
+ },
3575
+ {
3576
+ "epoch": 27.56,
3577
+ "learning_rate": 2.4688248970445827e-05,
3578
+ "loss": 0.5668,
3579
+ "step": 1190000
3580
+ },
3581
+ {
3582
+ "epoch": 27.6,
3583
+ "learning_rate": 2.42204470003758e-05,
3584
+ "loss": 0.5671,
3585
+ "step": 1192000
3586
+ },
3587
+ {
3588
+ "epoch": 27.65,
3589
+ "learning_rate": 2.3752645030305768e-05,
3590
+ "loss": 0.5688,
3591
+ "step": 1194000
3592
+ },
3593
+ {
3594
+ "epoch": 27.69,
3595
+ "learning_rate": 2.328484306023574e-05,
3596
+ "loss": 0.5643,
3597
+ "step": 1196000
3598
+ },
3599
+ {
3600
+ "epoch": 27.74,
3601
+ "learning_rate": 2.281704109016571e-05,
3602
+ "loss": 0.5688,
3603
+ "step": 1198000
3604
+ },
3605
+ {
3606
+ "epoch": 27.79,
3607
+ "learning_rate": 2.2349239120095678e-05,
3608
+ "loss": 0.5651,
3609
+ "step": 1200000
3610
+ },
3611
+ {
3612
+ "epoch": 27.83,
3613
+ "learning_rate": 2.188143715002565e-05,
3614
+ "loss": 0.5705,
3615
+ "step": 1202000
3616
+ },
3617
+ {
3618
+ "epoch": 27.88,
3619
+ "learning_rate": 2.141363517995562e-05,
3620
+ "loss": 0.5684,
3621
+ "step": 1204000
3622
+ },
3623
+ {
3624
+ "epoch": 27.93,
3625
+ "learning_rate": 2.0945833209885588e-05,
3626
+ "loss": 0.567,
3627
+ "step": 1206000
3628
+ },
3629
+ {
3630
+ "epoch": 27.97,
3631
+ "learning_rate": 2.0478031239815557e-05,
3632
+ "loss": 0.5711,
3633
+ "step": 1208000
3634
+ },
3635
+ {
3636
+ "epoch": 28.02,
3637
+ "learning_rate": 2.0010229269745533e-05,
3638
+ "loss": 0.5684,
3639
+ "step": 1210000
3640
+ },
3641
+ {
3642
+ "epoch": 28.06,
3643
+ "learning_rate": 1.95424272996755e-05,
3644
+ "loss": 0.5604,
3645
+ "step": 1212000
3646
+ },
3647
+ {
3648
+ "epoch": 28.11,
3649
+ "learning_rate": 1.907462532960547e-05,
3650
+ "loss": 0.5615,
3651
+ "step": 1214000
3652
+ },
3653
+ {
3654
+ "epoch": 28.16,
3655
+ "learning_rate": 1.860682335953544e-05,
3656
+ "loss": 0.5679,
3657
+ "step": 1216000
3658
+ },
3659
+ {
3660
+ "epoch": 28.2,
3661
+ "learning_rate": 1.813902138946541e-05,
3662
+ "loss": 0.5644,
3663
+ "step": 1218000
3664
+ },
3665
+ {
3666
+ "epoch": 28.25,
3667
+ "learning_rate": 1.767121941939538e-05,
3668
+ "loss": 0.5663,
3669
+ "step": 1220000
3670
+ },
3671
+ {
3672
+ "epoch": 28.3,
3673
+ "learning_rate": 1.720341744932535e-05,
3674
+ "loss": 0.5584,
3675
+ "step": 1222000
3676
+ },
3677
+ {
3678
+ "epoch": 28.34,
3679
+ "learning_rate": 1.6735615479255318e-05,
3680
+ "loss": 0.558,
3681
+ "step": 1224000
3682
+ },
3683
+ {
3684
+ "epoch": 28.39,
3685
+ "learning_rate": 1.626781350918529e-05,
3686
+ "loss": 0.5575,
3687
+ "step": 1226000
3688
+ },
3689
+ {
3690
+ "epoch": 28.44,
3691
+ "learning_rate": 1.580001153911526e-05,
3692
+ "loss": 0.5728,
3693
+ "step": 1228000
3694
+ },
3695
+ {
3696
+ "epoch": 28.48,
3697
+ "learning_rate": 1.533220956904523e-05,
3698
+ "loss": 0.5653,
3699
+ "step": 1230000
3700
+ },
3701
+ {
3702
+ "epoch": 28.53,
3703
+ "learning_rate": 1.48644075989752e-05,
3704
+ "loss": 0.5603,
3705
+ "step": 1232000
3706
+ },
3707
+ {
3708
+ "epoch": 28.57,
3709
+ "learning_rate": 1.4396605628905172e-05,
3710
+ "loss": 0.5613,
3711
+ "step": 1234000
3712
+ },
3713
+ {
3714
+ "epoch": 28.62,
3715
+ "learning_rate": 1.3928803658835141e-05,
3716
+ "loss": 0.5563,
3717
+ "step": 1236000
3718
+ },
3719
+ {
3720
+ "epoch": 28.67,
3721
+ "learning_rate": 1.346100168876511e-05,
3722
+ "loss": 0.5705,
3723
+ "step": 1238000
3724
+ },
3725
+ {
3726
+ "epoch": 28.71,
3727
+ "learning_rate": 1.299319971869508e-05,
3728
+ "loss": 0.5568,
3729
+ "step": 1240000
3730
+ },
3731
+ {
3732
+ "epoch": 28.76,
3733
+ "learning_rate": 1.252539774862505e-05,
3734
+ "loss": 0.5517,
3735
+ "step": 1242000
3736
+ },
3737
+ {
3738
+ "epoch": 28.81,
3739
+ "learning_rate": 1.2057595778555022e-05,
3740
+ "loss": 0.5647,
3741
+ "step": 1244000
3742
+ },
3743
+ {
3744
+ "epoch": 28.85,
3745
+ "learning_rate": 1.158979380848499e-05,
3746
+ "loss": 0.5551,
3747
+ "step": 1246000
3748
+ },
3749
+ {
3750
+ "epoch": 28.9,
3751
+ "learning_rate": 1.1121991838414961e-05,
3752
+ "loss": 0.5598,
3753
+ "step": 1248000
3754
+ },
3755
+ {
3756
+ "epoch": 28.94,
3757
+ "learning_rate": 1.0654189868344932e-05,
3758
+ "loss": 0.562,
3759
+ "step": 1250000
3760
+ },
3761
+ {
3762
+ "epoch": 28.99,
3763
+ "learning_rate": 1.0186387898274902e-05,
3764
+ "loss": 0.5563,
3765
+ "step": 1252000
3766
+ },
3767
+ {
3768
+ "epoch": 29.04,
3769
+ "learning_rate": 9.718585928204871e-06,
3770
+ "loss": 0.5606,
3771
+ "step": 1254000
3772
+ },
3773
+ {
3774
+ "epoch": 29.08,
3775
+ "learning_rate": 9.250783958134842e-06,
3776
+ "loss": 0.5509,
3777
+ "step": 1256000
3778
+ },
3779
+ {
3780
+ "epoch": 29.13,
3781
+ "learning_rate": 8.782981988064812e-06,
3782
+ "loss": 0.551,
3783
+ "step": 1258000
3784
+ },
3785
+ {
3786
+ "epoch": 29.18,
3787
+ "learning_rate": 8.315180017994783e-06,
3788
+ "loss": 0.5548,
3789
+ "step": 1260000
3790
+ },
3791
+ {
3792
+ "epoch": 29.22,
3793
+ "learning_rate": 7.847378047924752e-06,
3794
+ "loss": 0.5562,
3795
+ "step": 1262000
3796
+ },
3797
+ {
3798
+ "epoch": 29.27,
3799
+ "learning_rate": 7.3795760778547214e-06,
3800
+ "loss": 0.5563,
3801
+ "step": 1264000
3802
+ },
3803
+ {
3804
+ "epoch": 29.32,
3805
+ "learning_rate": 6.911774107784692e-06,
3806
+ "loss": 0.5551,
3807
+ "step": 1266000
3808
+ },
3809
+ {
3810
+ "epoch": 29.36,
3811
+ "learning_rate": 6.443972137714662e-06,
3812
+ "loss": 0.555,
3813
+ "step": 1268000
3814
+ },
3815
+ {
3816
+ "epoch": 29.41,
3817
+ "learning_rate": 5.976170167644632e-06,
3818
+ "loss": 0.554,
3819
+ "step": 1270000
3820
+ },
3821
+ {
3822
+ "epoch": 29.45,
3823
+ "learning_rate": 5.508368197574602e-06,
3824
+ "loss": 0.5522,
3825
+ "step": 1272000
3826
+ },
3827
+ {
3828
+ "epoch": 29.5,
3829
+ "learning_rate": 5.0405662275045725e-06,
3830
+ "loss": 0.5522,
3831
+ "step": 1274000
3832
+ },
3833
+ {
3834
+ "epoch": 29.55,
3835
+ "learning_rate": 4.572764257434542e-06,
3836
+ "loss": 0.5601,
3837
+ "step": 1276000
3838
+ },
3839
+ {
3840
+ "epoch": 29.59,
3841
+ "learning_rate": 4.104962287364513e-06,
3842
+ "loss": 0.5578,
3843
+ "step": 1278000
3844
+ },
3845
+ {
3846
+ "epoch": 29.64,
3847
+ "learning_rate": 3.6371603172944825e-06,
3848
+ "loss": 0.5602,
3849
+ "step": 1280000
3850
+ },
3851
+ {
3852
+ "epoch": 29.69,
3853
+ "learning_rate": 3.1693583472244526e-06,
3854
+ "loss": 0.5517,
3855
+ "step": 1282000
3856
+ },
3857
+ {
3858
+ "epoch": 29.73,
3859
+ "learning_rate": 2.701556377154423e-06,
3860
+ "loss": 0.5538,
3861
+ "step": 1284000
3862
+ },
3863
+ {
3864
+ "epoch": 29.78,
3865
+ "learning_rate": 2.233754407084393e-06,
3866
+ "loss": 0.549,
3867
+ "step": 1286000
3868
+ },
3869
+ {
3870
+ "epoch": 29.82,
3871
+ "learning_rate": 1.765952437014363e-06,
3872
+ "loss": 0.5595,
3873
+ "step": 1288000
3874
+ },
3875
+ {
3876
+ "epoch": 29.87,
3877
+ "learning_rate": 1.2981504669443331e-06,
3878
+ "loss": 0.5546,
3879
+ "step": 1290000
3880
+ },
3881
+ {
3882
+ "epoch": 29.92,
3883
+ "learning_rate": 8.303484968743031e-07,
3884
+ "loss": 0.5547,
3885
+ "step": 1292000
3886
+ },
3887
+ {
3888
+ "epoch": 29.96,
3889
+ "learning_rate": 3.6254652680427316e-07,
3890
+ "loss": 0.5494,
3891
+ "step": 1294000
3892
+ },
3893
+ {
3894
+ "epoch": 30.0,
3895
+ "step": 1295550,
3896
+ "total_flos": 2.6480449905256835e+21,
3897
+ "train_loss": 0.7350748663054241,
3898
+ "train_runtime": 658563.1153,
3899
+ "train_samples_per_second": 31.476,
3900
+ "train_steps_per_second": 1.967
3901
+ }
3902
+ ],
3903
+ "logging_steps": 2000,
3904
+ "max_steps": 1295550,
3905
+ "num_train_epochs": 30,
3906
+ "save_steps": 500,
3907
+ "total_flos": 2.6480449905256835e+21,
3908
+ "trial_name": null,
3909
+ "trial_params": null
3910
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:15b4a29d89f83e50e17158ee0f4ffe9724daa5ff51dfccdeac7b5e2c6a3b564e
3
+ size 4600
vocab.json ADDED
@@ -0,0 +1,53 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "\b": 0,
3
+ "\"": 2,
4
+ "[PAD]": 50,
5
+ "[UNK]": 49,
6
+ "\\": 3,
7
+ "a": 4,
8
+ "b": 5,
9
+ "c": 6,
10
+ "d": 7,
11
+ "e": 8,
12
+ "f": 9,
13
+ "g": 10,
14
+ "h": 11,
15
+ "i": 12,
16
+ "j": 13,
17
+ "k": 14,
18
+ "l": 15,
19
+ "m": 16,
20
+ "n": 17,
21
+ "o": 18,
22
+ "p": 19,
23
+ "q": 20,
24
+ "r": 21,
25
+ "s": 22,
26
+ "t": 23,
27
+ "u": 24,
28
+ "v": 25,
29
+ "w": 26,
30
+ "x": 27,
31
+ "y": 28,
32
+ "z": 29,
33
+ "|": 1,
34
+ "ç": 30,
35
+ "é": 31,
36
+ "ş": 32,
37
+ "υ": 33,
38
+ "а": 34,
39
+ "в": 35,
40
+ "г": 36,
41
+ "д": 37,
42
+ "е": 38,
43
+ "к": 39,
44
+ "л": 40,
45
+ "н": 41,
46
+ "с": 42,
47
+ "ф": 43,
48
+ "ч": 44,
49
+ "​": 45,
50
+ "’": 46,
51
+ "⁄": 47,
52
+ "�": 48
53
+ }