Hugging Face's logo

duongttr
/
vietnamese-jd-generation-gpt2

Text Generation
Transformers
PyTorch
gpt2
Generated from Trainer
Eval Results
text-generation-inference
Model card Files Community
duongttr commited on
Commit
731e90e
·
1 Parent(s): 9eee936

Upload folder using huggingface_hub

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. README.md +72 -0
  2. added_tokens.json +7 -0
  3. all_results.json +15 -0
  4. checkpoint-14000/added_tokens.json +7 -0
  5. checkpoint-14000/config.json +41 -0
  6. checkpoint-14000/generation_config.json +6 -0
  7. checkpoint-14000/merges.txt +0 -0
  8. checkpoint-14000/optimizer.pt +3 -0
  9. checkpoint-14000/pytorch_model.bin +3 -0
  10. checkpoint-14000/rng_state.pth +3 -0
  11. checkpoint-14000/scheduler.pt +3 -0
  12. checkpoint-14000/special_tokens_map.json +6 -0
  13. checkpoint-14000/tokenizer.json +0 -0
  14. checkpoint-14000/tokenizer_config.json +9 -0
  15. checkpoint-14000/trainer_state.json +0 -0
  16. checkpoint-14000/training_args.bin +3 -0
  17. checkpoint-14000/vocab.json +0 -0
  18. checkpoint-21000/added_tokens.json +7 -0
  19. checkpoint-21000/config.json +41 -0
  20. checkpoint-21000/generation_config.json +6 -0
  21. checkpoint-21000/merges.txt +0 -0
  22. checkpoint-21000/optimizer.pt +3 -0
  23. checkpoint-21000/pytorch_model.bin +3 -0
  24. checkpoint-21000/rng_state.pth +3 -0
  25. checkpoint-21000/scheduler.pt +3 -0
  26. checkpoint-21000/special_tokens_map.json +6 -0
  27. checkpoint-21000/tokenizer.json +0 -0
  28. checkpoint-21000/tokenizer_config.json +9 -0
  29. checkpoint-21000/trainer_state.json +0 -0
  30. checkpoint-21000/training_args.bin +3 -0
  31. checkpoint-21000/vocab.json +0 -0
  32. checkpoint-7000/added_tokens.json +7 -0
  33. checkpoint-7000/config.json +41 -0
  34. checkpoint-7000/generation_config.json +6 -0
  35. checkpoint-7000/merges.txt +0 -0
  36. checkpoint-7000/optimizer.pt +3 -0
  37. checkpoint-7000/pytorch_model.bin +3 -0
  38. checkpoint-7000/rng_state.pth +3 -0
  39. checkpoint-7000/scheduler.pt +3 -0
  40. checkpoint-7000/special_tokens_map.json +6 -0
  41. checkpoint-7000/tokenizer.json +0 -0
  42. checkpoint-7000/tokenizer_config.json +9 -0
  43. checkpoint-7000/trainer_state.json +4219 -0
  44. checkpoint-7000/training_args.bin +3 -0
  45. checkpoint-7000/vocab.json +0 -0
  46. config.json +41 -0
  47. eval_results.json +10 -0
  48. generation_config.json +6 -0
  49. merges.txt +0 -0
  50. pytorch_model.bin +3 -0
README.md ADDED
@@ -0,0 +1,72 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ base_model: chronopt-research/vietnamese-gpt2-medium
4
+ tags:
5
+ - generated_from_trainer
6
+ datasets:
7
+ - duongttr/JD-Data-56k-clean
8
+ metrics:
9
+ - accuracy
10
+ model-index:
11
+ - name: results
12
+ results:
13
+ - task:
14
+ name: Causal Language Modeling
15
+ type: text-generation
16
+ dataset:
17
+ name: duongttr/JD-Data-56k-clean
18
+ type: duongttr/JD-Data-56k-clean
19
+ metrics:
20
+ - name: Accuracy
21
+ type: accuracy
22
+ value: 0.8983410461950713
23
+ ---
24
+
25
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
26
+ should probably proofread and complete it, then remove this comment. -->
27
+
28
+ # results
29
+
30
+ This model is a fine-tuned version of [chronopt-research/vietnamese-gpt2-medium](https://huggingface.co/chronopt-research/vietnamese-gpt2-medium) on the duongttr/JD-Data-56k-clean dataset.
31
+ It achieves the following results on the evaluation set:
32
+ - Loss: 0.5219
33
+ - Accuracy: 0.8983
34
+
35
+ ## Model description
36
+
37
+ More information needed
38
+
39
+ ## Intended uses & limitations
40
+
41
+ More information needed
42
+
43
+ ## Training and evaluation data
44
+
45
+ More information needed
46
+
47
+ ## Training procedure
48
+
49
+ ### Training hyperparameters
50
+
51
+ The following hyperparameters were used during training:
52
+ - learning_rate: 0.0001
53
+ - train_batch_size: 8
54
+ - eval_batch_size: 8
55
+ - seed: 42
56
+ - gradient_accumulation_steps: 4
57
+ - total_train_batch_size: 32
58
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
59
+ - lr_scheduler_type: linear
60
+ - lr_scheduler_warmup_ratio: 0.2
61
+ - num_epochs: 5.0
62
+
63
+ ### Training results
64
+
65
+
66
+
67
+ ### Framework versions
68
+
69
+ - Transformers 4.33.0.dev0
70
+ - Pytorch 2.0.1
71
+ - Datasets 2.14.4
72
+ - Tokenizers 0.13.3
added_tokens.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "<|beginofbnf|>": 50261,
3
+ "<|beginofdesc|>": 50259,
4
+ "<|beginofreq|>": 50260,
5
+ "<|endoftext|>": 50257,
6
+ "<|pad|>": 50258
7
+ }
all_results.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 5.0,
3
+ "eval_accuracy": 0.8983410461950713,
4
+ "eval_loss": 0.5218655467033386,
5
+ "eval_runtime": 103.6013,
6
+ "eval_samples": 8529,
7
+ "eval_samples_per_second": 82.325,
8
+ "eval_steps_per_second": 10.299,
9
+ "perplexity": 1.6851684796072373,
10
+ "train_loss": 0.44226374260195156,
11
+ "train_runtime": 20787.5605,
12
+ "train_samples": 162006,
13
+ "train_samples_per_second": 38.967,
14
+ "train_steps_per_second": 1.218
15
+ }
checkpoint-14000/added_tokens.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "<|beginofbnf|>": 50261,
3
+ "<|beginofdesc|>": 50259,
4
+ "<|beginofreq|>": 50260,
5
+ "<|endoftext|>": 50257,
6
+ "<|pad|>": 50258
7
+ }
checkpoint-14000/config.json ADDED
@@ -0,0 +1,41 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "chronopt-research/vietnamese-gpt2-medium",
3
+ "activation_function": "gelu_new",
4
+ "architectures": [
5
+ "GPT2LMHeadModel"
6
+ ],
7
+ "attn_pdrop": 0.0,
8
+ "bos_token_id": 50256,
9
+ "embd_pdrop": 0.0,
10
+ "eos_token_id": 50256,
11
+ "initializer_range": 0.02,
12
+ "layer_norm_epsilon": 1e-05,
13
+ "model_type": "gpt2",
14
+ "n_ctx": 1024,
15
+ "n_embd": 1024,
16
+ "n_head": 16,
17
+ "n_inner": null,
18
+ "n_layer": 24,
19
+ "n_positions": 1024,
20
+ "n_special": 0,
21
+ "predict_special_tokens": true,
22
+ "reorder_and_upcast_attn": false,
23
+ "resid_pdrop": 0.0,
24
+ "scale_attn_by_inverse_layer_idx": false,
25
+ "scale_attn_weights": true,
26
+ "summary_activation": null,
27
+ "summary_first_dropout": 0.1,
28
+ "summary_proj_to_labels": true,
29
+ "summary_type": "cls_index",
30
+ "summary_use_proj": true,
31
+ "task_specific_params": {
32
+ "text-generation": {
33
+ "do_sample": true,
34
+ "max_length": 50
35
+ }
36
+ },
37
+ "torch_dtype": "float32",
38
+ "transformers_version": "4.33.0.dev0",
39
+ "use_cache": true,
40
+ "vocab_size": 50262
41
+ }
checkpoint-14000/generation_config.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 50256,
4
+ "eos_token_id": 50256,
5
+ "transformers_version": "4.33.0.dev0"
6
+ }
checkpoint-14000/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-14000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cb6c335ccf2da550680f697a7ce8d0bd9993c3b4fa941eb8916968de022809b1
3
+ size 2838869765
checkpoint-14000/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7c6b0fd4b8e6848dc6b2f357c2e55b1a167a47b6f61cfac60c0e72768ba6fdb1
3
+ size 1419408349
checkpoint-14000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7508d4b8dd267de5cc58e972da25236687927651336a28f292c92f7f23951475
3
+ size 14575
checkpoint-14000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5f6d0e9540ada78794e9c38c490b66b029b699f5a692ebe943e5565b97ce6de7
3
+ size 627
checkpoint-14000/special_tokens_map.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<|endoftext|>",
3
+ "eos_token": "<|endoftext|>",
4
+ "pad_token": "<|pad|>",
5
+ "unk_token": "<|endoftext|>"
6
+ }
checkpoint-14000/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-14000/tokenizer_config.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "bos_token": "<|endoftext|>",
4
+ "clean_up_tokenization_spaces": true,
5
+ "eos_token": "<|endoftext|>",
6
+ "model_max_length": 1000000000000000019884624838656,
7
+ "tokenizer_class": "GPT2Tokenizer",
8
+ "unk_token": "<|endoftext|>"
9
+ }
checkpoint-14000/trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-14000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6b861ebe1aabc6404f5654d57dd93aa3f72512ce343953af0f357d2575b8b9d4
3
+ size 4027
checkpoint-14000/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-21000/added_tokens.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "<|beginofbnf|>": 50261,
3
+ "<|beginofdesc|>": 50259,
4
+ "<|beginofreq|>": 50260,
5
+ "<|endoftext|>": 50257,
6
+ "<|pad|>": 50258
7
+ }
checkpoint-21000/config.json ADDED
@@ -0,0 +1,41 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "chronopt-research/vietnamese-gpt2-medium",
3
+ "activation_function": "gelu_new",
4
+ "architectures": [
5
+ "GPT2LMHeadModel"
6
+ ],
7
+ "attn_pdrop": 0.0,
8
+ "bos_token_id": 50256,
9
+ "embd_pdrop": 0.0,
10
+ "eos_token_id": 50256,
11
+ "initializer_range": 0.02,
12
+ "layer_norm_epsilon": 1e-05,
13
+ "model_type": "gpt2",
14
+ "n_ctx": 1024,
15
+ "n_embd": 1024,
16
+ "n_head": 16,
17
+ "n_inner": null,
18
+ "n_layer": 24,
19
+ "n_positions": 1024,
20
+ "n_special": 0,
21
+ "predict_special_tokens": true,
22
+ "reorder_and_upcast_attn": false,
23
+ "resid_pdrop": 0.0,
24
+ "scale_attn_by_inverse_layer_idx": false,
25
+ "scale_attn_weights": true,
26
+ "summary_activation": null,
27
+ "summary_first_dropout": 0.1,
28
+ "summary_proj_to_labels": true,
29
+ "summary_type": "cls_index",
30
+ "summary_use_proj": true,
31
+ "task_specific_params": {
32
+ "text-generation": {
33
+ "do_sample": true,
34
+ "max_length": 50
35
+ }
36
+ },
37
+ "torch_dtype": "float32",
38
+ "transformers_version": "4.33.0.dev0",
39
+ "use_cache": true,
40
+ "vocab_size": 50262
41
+ }
checkpoint-21000/generation_config.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 50256,
4
+ "eos_token_id": 50256,
5
+ "transformers_version": "4.33.0.dev0"
6
+ }
checkpoint-21000/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-21000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6ae77ed2c98f97acd28f38a152051733f0452c4dd3cebcfdeaa0ff4a4fad0f95
3
+ size 2838869765
checkpoint-21000/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b7de88fbbe56b7e649823e28544b54ddbeab3113f5f4b455823fba55903999a3
3
+ size 1419408349
checkpoint-21000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:08d1f90efd2ec5917f1161087472df1d694f3bfb61fcf103c3c900b90c19e167
3
+ size 14575
checkpoint-21000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:72ae46dd54e44476d15d951e65c0cc740b0fb7529d9c859264e9edb83085b471
3
+ size 627
checkpoint-21000/special_tokens_map.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<|endoftext|>",
3
+ "eos_token": "<|endoftext|>",
4
+ "pad_token": "<|pad|>",
5
+ "unk_token": "<|endoftext|>"
6
+ }
checkpoint-21000/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-21000/tokenizer_config.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "bos_token": "<|endoftext|>",
4
+ "clean_up_tokenization_spaces": true,
5
+ "eos_token": "<|endoftext|>",
6
+ "model_max_length": 1000000000000000019884624838656,
7
+ "tokenizer_class": "GPT2Tokenizer",
8
+ "unk_token": "<|endoftext|>"
9
+ }
checkpoint-21000/trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-21000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6b861ebe1aabc6404f5654d57dd93aa3f72512ce343953af0f357d2575b8b9d4
3
+ size 4027
checkpoint-21000/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-7000/added_tokens.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "<|beginofbnf|>": 50261,
3
+ "<|beginofdesc|>": 50259,
4
+ "<|beginofreq|>": 50260,
5
+ "<|endoftext|>": 50257,
6
+ "<|pad|>": 50258
7
+ }
checkpoint-7000/config.json ADDED
@@ -0,0 +1,41 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "chronopt-research/vietnamese-gpt2-medium",
3
+ "activation_function": "gelu_new",
4
+ "architectures": [
5
+ "GPT2LMHeadModel"
6
+ ],
7
+ "attn_pdrop": 0.0,
8
+ "bos_token_id": 50256,
9
+ "embd_pdrop": 0.0,
10
+ "eos_token_id": 50256,
11
+ "initializer_range": 0.02,
12
+ "layer_norm_epsilon": 1e-05,
13
+ "model_type": "gpt2",
14
+ "n_ctx": 1024,
15
+ "n_embd": 1024,
16
+ "n_head": 16,
17
+ "n_inner": null,
18
+ "n_layer": 24,
19
+ "n_positions": 1024,
20
+ "n_special": 0,
21
+ "predict_special_tokens": true,
22
+ "reorder_and_upcast_attn": false,
23
+ "resid_pdrop": 0.0,
24
+ "scale_attn_by_inverse_layer_idx": false,
25
+ "scale_attn_weights": true,
26
+ "summary_activation": null,
27
+ "summary_first_dropout": 0.1,
28
+ "summary_proj_to_labels": true,
29
+ "summary_type": "cls_index",
30
+ "summary_use_proj": true,
31
+ "task_specific_params": {
32
+ "text-generation": {
33
+ "do_sample": true,
34
+ "max_length": 50
35
+ }
36
+ },
37
+ "torch_dtype": "float32",
38
+ "transformers_version": "4.33.0.dev0",
39
+ "use_cache": true,
40
+ "vocab_size": 50262
41
+ }
checkpoint-7000/generation_config.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 50256,
4
+ "eos_token_id": 50256,
5
+ "transformers_version": "4.33.0.dev0"
6
+ }
checkpoint-7000/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-7000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2650e46d604ec4abf8695cd90f7fa67bc051f6b3889795cd4282b9c8c52ff211
3
+ size 2838869765
checkpoint-7000/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4885080a35a85c9d621c8137f42dda9b3f7ad8d8da08d1ee12f59ba81270a5a9
3
+ size 1419408349
checkpoint-7000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:972139d83957a9cf2600cb6eeca17287d7a5377c33a53500ae7e13fe830ad36b
3
+ size 14575
checkpoint-7000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b1c27f7eaacd33410949bff85c7040487cfc6756d39a62f3e2757560db0fe6c8
3
+ size 627
checkpoint-7000/special_tokens_map.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<|endoftext|>",
3
+ "eos_token": "<|endoftext|>",
4
+ "pad_token": "<|pad|>",
5
+ "unk_token": "<|endoftext|>"
6
+ }
checkpoint-7000/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-7000/tokenizer_config.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "bos_token": "<|endoftext|>",
4
+ "clean_up_tokenization_spaces": true,
5
+ "eos_token": "<|endoftext|>",
6
+ "model_max_length": 1000000000000000019884624838656,
7
+ "tokenizer_class": "GPT2Tokenizer",
8
+ "unk_token": "<|endoftext|>"
9
+ }
checkpoint-7000/trainer_state.json ADDED
@@ -0,0 +1,4219 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 1.38264777048047,
5
+ "eval_steps": 7000.0,
6
+ "global_step": 7000,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.0,
13
+ "learning_rate": 1.975503753457132e-07,
14
+ "loss": 7.9807,
15
+ "step": 10
16
+ },
17
+ {
18
+ "epoch": 0.0,
19
+ "learning_rate": 3.951007506914264e-07,
20
+ "loss": 7.3025,
21
+ "step": 20
22
+ },
23
+ {
24
+ "epoch": 0.01,
25
+ "learning_rate": 5.926511260371394e-07,
26
+ "loss": 6.442,
27
+ "step": 30
28
+ },
29
+ {
30
+ "epoch": 0.01,
31
+ "learning_rate": 7.902015013828528e-07,
32
+ "loss": 4.9928,
33
+ "step": 40
34
+ },
35
+ {
36
+ "epoch": 0.01,
37
+ "learning_rate": 9.877518767285659e-07,
38
+ "loss": 3.0368,
39
+ "step": 50
40
+ },
41
+ {
42
+ "epoch": 0.01,
43
+ "learning_rate": 1.1853022520742789e-06,
44
+ "loss": 1.5885,
45
+ "step": 60
46
+ },
47
+ {
48
+ "epoch": 0.01,
49
+ "learning_rate": 1.382852627419992e-06,
50
+ "loss": 1.118,
51
+ "step": 70
52
+ },
53
+ {
54
+ "epoch": 0.02,
55
+ "learning_rate": 1.5804030027657055e-06,
56
+ "loss": 1.0564,
57
+ "step": 80
58
+ },
59
+ {
60
+ "epoch": 0.02,
61
+ "learning_rate": 1.7779533781114185e-06,
62
+ "loss": 1.1117,
63
+ "step": 90
64
+ },
65
+ {
66
+ "epoch": 0.02,
67
+ "learning_rate": 1.9755037534571318e-06,
68
+ "loss": 1.1398,
69
+ "step": 100
70
+ },
71
+ {
72
+ "epoch": 0.02,
73
+ "learning_rate": 2.173054128802845e-06,
74
+ "loss": 0.9834,
75
+ "step": 110
76
+ },
77
+ {
78
+ "epoch": 0.02,
79
+ "learning_rate": 2.3706045041485578e-06,
80
+ "loss": 1.0677,
81
+ "step": 120
82
+ },
83
+ {
84
+ "epoch": 0.03,
85
+ "learning_rate": 2.5681548794942714e-06,
86
+ "loss": 0.9412,
87
+ "step": 130
88
+ },
89
+ {
90
+ "epoch": 0.03,
91
+ "learning_rate": 2.765705254839984e-06,
92
+ "loss": 0.9347,
93
+ "step": 140
94
+ },
95
+ {
96
+ "epoch": 0.03,
97
+ "learning_rate": 2.9632556301856974e-06,
98
+ "loss": 0.9574,
99
+ "step": 150
100
+ },
101
+ {
102
+ "epoch": 0.03,
103
+ "learning_rate": 3.160806005531411e-06,
104
+ "loss": 0.9889,
105
+ "step": 160
106
+ },
107
+ {
108
+ "epoch": 0.03,
109
+ "learning_rate": 3.3583563808771234e-06,
110
+ "loss": 0.9102,
111
+ "step": 170
112
+ },
113
+ {
114
+ "epoch": 0.04,
115
+ "learning_rate": 3.555906756222837e-06,
116
+ "loss": 0.9191,
117
+ "step": 180
118
+ },
119
+ {
120
+ "epoch": 0.04,
121
+ "learning_rate": 3.7534571315685503e-06,
122
+ "loss": 0.9766,
123
+ "step": 190
124
+ },
125
+ {
126
+ "epoch": 0.04,
127
+ "learning_rate": 3.9510075069142635e-06,
128
+ "loss": 0.9135,
129
+ "step": 200
130
+ },
131
+ {
132
+ "epoch": 0.04,
133
+ "learning_rate": 4.148557882259976e-06,
134
+ "loss": 0.8757,
135
+ "step": 210
136
+ },
137
+ {
138
+ "epoch": 0.04,
139
+ "learning_rate": 4.34610825760569e-06,
140
+ "loss": 0.9067,
141
+ "step": 220
142
+ },
143
+ {
144
+ "epoch": 0.05,
145
+ "learning_rate": 4.543658632951403e-06,
146
+ "loss": 0.9219,
147
+ "step": 230
148
+ },
149
+ {
150
+ "epoch": 0.05,
151
+ "learning_rate": 4.7412090082971155e-06,
152
+ "loss": 0.8996,
153
+ "step": 240
154
+ },
155
+ {
156
+ "epoch": 0.05,
157
+ "learning_rate": 4.938759383642829e-06,
158
+ "loss": 0.9467,
159
+ "step": 250
160
+ },
161
+ {
162
+ "epoch": 0.05,
163
+ "learning_rate": 5.136309758988543e-06,
164
+ "loss": 0.8877,
165
+ "step": 260
166
+ },
167
+ {
168
+ "epoch": 0.05,
169
+ "learning_rate": 5.333860134334256e-06,
170
+ "loss": 0.8398,
171
+ "step": 270
172
+ },
173
+ {
174
+ "epoch": 0.06,
175
+ "learning_rate": 5.531410509679968e-06,
176
+ "loss": 0.8764,
177
+ "step": 280
178
+ },
179
+ {
180
+ "epoch": 0.06,
181
+ "learning_rate": 5.728960885025682e-06,
182
+ "loss": 0.8809,
183
+ "step": 290
184
+ },
185
+ {
186
+ "epoch": 0.06,
187
+ "learning_rate": 5.926511260371395e-06,
188
+ "loss": 0.8699,
189
+ "step": 300
190
+ },
191
+ {
192
+ "epoch": 0.06,
193
+ "learning_rate": 6.124061635717108e-06,
194
+ "loss": 0.8377,
195
+ "step": 310
196
+ },
197
+ {
198
+ "epoch": 0.06,
199
+ "learning_rate": 6.321612011062822e-06,
200
+ "loss": 0.8943,
201
+ "step": 320
202
+ },
203
+ {
204
+ "epoch": 0.07,
205
+ "learning_rate": 6.519162386408534e-06,
206
+ "loss": 0.8393,
207
+ "step": 330
208
+ },
209
+ {
210
+ "epoch": 0.07,
211
+ "learning_rate": 6.716712761754247e-06,
212
+ "loss": 0.8192,
213
+ "step": 340
214
+ },
215
+ {
216
+ "epoch": 0.07,
217
+ "learning_rate": 6.9142631370999605e-06,
218
+ "loss": 0.903,
219
+ "step": 350
220
+ },
221
+ {
222
+ "epoch": 0.07,
223
+ "learning_rate": 7.111813512445674e-06,
224
+ "loss": 0.8917,
225
+ "step": 360
226
+ },
227
+ {
228
+ "epoch": 0.07,
229
+ "learning_rate": 7.309363887791387e-06,
230
+ "loss": 0.916,
231
+ "step": 370
232
+ },
233
+ {
234
+ "epoch": 0.08,
235
+ "learning_rate": 7.506914263137101e-06,
236
+ "loss": 0.864,
237
+ "step": 380
238
+ },
239
+ {
240
+ "epoch": 0.08,
241
+ "learning_rate": 7.704464638482814e-06,
242
+ "loss": 0.7809,
243
+ "step": 390
244
+ },
245
+ {
246
+ "epoch": 0.08,
247
+ "learning_rate": 7.902015013828527e-06,
248
+ "loss": 0.885,
249
+ "step": 400
250
+ },
251
+ {
252
+ "epoch": 0.08,
253
+ "learning_rate": 8.09956538917424e-06,
254
+ "loss": 0.8183,
255
+ "step": 410
256
+ },
257
+ {
258
+ "epoch": 0.08,
259
+ "learning_rate": 8.297115764519953e-06,
260
+ "loss": 0.7897,
261
+ "step": 420
262
+ },
263
+ {
264
+ "epoch": 0.08,
265
+ "learning_rate": 8.494666139865665e-06,
266
+ "loss": 0.8315,
267
+ "step": 430
268
+ },
269
+ {
270
+ "epoch": 0.09,
271
+ "learning_rate": 8.69221651521138e-06,
272
+ "loss": 0.8266,
273
+ "step": 440
274
+ },
275
+ {
276
+ "epoch": 0.09,
277
+ "learning_rate": 8.889766890557093e-06,
278
+ "loss": 0.8755,
279
+ "step": 450
280
+ },
281
+ {
282
+ "epoch": 0.09,
283
+ "learning_rate": 9.087317265902805e-06,
284
+ "loss": 0.8103,
285
+ "step": 460
286
+ },
287
+ {
288
+ "epoch": 0.09,
289
+ "learning_rate": 9.28486764124852e-06,
290
+ "loss": 0.8226,
291
+ "step": 470
292
+ },
293
+ {
294
+ "epoch": 0.09,
295
+ "learning_rate": 9.482418016594231e-06,
296
+ "loss": 0.8174,
297
+ "step": 480
298
+ },
299
+ {
300
+ "epoch": 0.1,
301
+ "learning_rate": 9.679968391939946e-06,
302
+ "loss": 0.8336,
303
+ "step": 490
304
+ },
305
+ {
306
+ "epoch": 0.1,
307
+ "learning_rate": 9.877518767285658e-06,
308
+ "loss": 0.8214,
309
+ "step": 500
310
+ },
311
+ {
312
+ "epoch": 0.1,
313
+ "learning_rate": 1.0075069142631371e-05,
314
+ "loss": 0.8534,
315
+ "step": 510
316
+ },
317
+ {
318
+ "epoch": 0.1,
319
+ "learning_rate": 1.0272619517977086e-05,
320
+ "loss": 0.8148,
321
+ "step": 520
322
+ },
323
+ {
324
+ "epoch": 0.1,
325
+ "learning_rate": 1.0470169893322798e-05,
326
+ "loss": 0.8032,
327
+ "step": 530
328
+ },
329
+ {
330
+ "epoch": 0.11,
331
+ "learning_rate": 1.0667720268668511e-05,
332
+ "loss": 0.7561,
333
+ "step": 540
334
+ },
335
+ {
336
+ "epoch": 0.11,
337
+ "learning_rate": 1.0865270644014224e-05,
338
+ "loss": 0.8219,
339
+ "step": 550
340
+ },
341
+ {
342
+ "epoch": 0.11,
343
+ "learning_rate": 1.1062821019359937e-05,
344
+ "loss": 0.8108,
345
+ "step": 560
346
+ },
347
+ {
348
+ "epoch": 0.11,
349
+ "learning_rate": 1.126037139470565e-05,
350
+ "loss": 0.7673,
351
+ "step": 570
352
+ },
353
+ {
354
+ "epoch": 0.11,
355
+ "learning_rate": 1.1457921770051364e-05,
356
+ "loss": 0.7731,
357
+ "step": 580
358
+ },
359
+ {
360
+ "epoch": 0.12,
361
+ "learning_rate": 1.1655472145397077e-05,
362
+ "loss": 0.8135,
363
+ "step": 590
364
+ },
365
+ {
366
+ "epoch": 0.12,
367
+ "learning_rate": 1.185302252074279e-05,
368
+ "loss": 0.8525,
369
+ "step": 600
370
+ },
371
+ {
372
+ "epoch": 0.12,
373
+ "learning_rate": 1.2050572896088504e-05,
374
+ "loss": 0.8473,
375
+ "step": 610
376
+ },
377
+ {
378
+ "epoch": 0.12,
379
+ "learning_rate": 1.2248123271434215e-05,
380
+ "loss": 0.7937,
381
+ "step": 620
382
+ },
383
+ {
384
+ "epoch": 0.12,
385
+ "learning_rate": 1.244567364677993e-05,
386
+ "loss": 0.7924,
387
+ "step": 630
388
+ },
389
+ {
390
+ "epoch": 0.13,
391
+ "learning_rate": 1.2643224022125644e-05,
392
+ "loss": 0.7778,
393
+ "step": 640
394
+ },
395
+ {
396
+ "epoch": 0.13,
397
+ "learning_rate": 1.2840774397471355e-05,
398
+ "loss": 0.8282,
399
+ "step": 650
400
+ },
401
+ {
402
+ "epoch": 0.13,
403
+ "learning_rate": 1.3038324772817068e-05,
404
+ "loss": 0.811,
405
+ "step": 660
406
+ },
407
+ {
408
+ "epoch": 0.13,
409
+ "learning_rate": 1.3235875148162783e-05,
410
+ "loss": 0.7167,
411
+ "step": 670
412
+ },
413
+ {
414
+ "epoch": 0.13,
415
+ "learning_rate": 1.3433425523508494e-05,
416
+ "loss": 0.8321,
417
+ "step": 680
418
+ },
419
+ {
420
+ "epoch": 0.14,
421
+ "learning_rate": 1.363097589885421e-05,
422
+ "loss": 0.7477,
423
+ "step": 690
424
+ },
425
+ {
426
+ "epoch": 0.14,
427
+ "learning_rate": 1.3828526274199921e-05,
428
+ "loss": 0.8232,
429
+ "step": 700
430
+ },
431
+ {
432
+ "epoch": 0.14,
433
+ "learning_rate": 1.4026076649545635e-05,
434
+ "loss": 0.799,
435
+ "step": 710
436
+ },
437
+ {
438
+ "epoch": 0.14,
439
+ "learning_rate": 1.4223627024891348e-05,
440
+ "loss": 0.8554,
441
+ "step": 720
442
+ },
443
+ {
444
+ "epoch": 0.14,
445
+ "learning_rate": 1.442117740023706e-05,
446
+ "loss": 0.778,
447
+ "step": 730
448
+ },
449
+ {
450
+ "epoch": 0.15,
451
+ "learning_rate": 1.4618727775582774e-05,
452
+ "loss": 0.8337,
453
+ "step": 740
454
+ },
455
+ {
456
+ "epoch": 0.15,
457
+ "learning_rate": 1.4816278150928487e-05,
458
+ "loss": 0.7793,
459
+ "step": 750
460
+ },
461
+ {
462
+ "epoch": 0.15,
463
+ "learning_rate": 1.5013828526274201e-05,
464
+ "loss": 0.7812,
465
+ "step": 760
466
+ },
467
+ {
468
+ "epoch": 0.15,
469
+ "learning_rate": 1.5211378901619914e-05,
470
+ "loss": 0.7837,
471
+ "step": 770
472
+ },
473
+ {
474
+ "epoch": 0.15,
475
+ "learning_rate": 1.540892927696563e-05,
476
+ "loss": 0.7714,
477
+ "step": 780
478
+ },
479
+ {
480
+ "epoch": 0.16,
481
+ "learning_rate": 1.560647965231134e-05,
482
+ "loss": 0.74,
483
+ "step": 790
484
+ },
485
+ {
486
+ "epoch": 0.16,
487
+ "learning_rate": 1.5804030027657054e-05,
488
+ "loss": 0.7725,
489
+ "step": 800
490
+ },
491
+ {
492
+ "epoch": 0.16,
493
+ "learning_rate": 1.6001580403002767e-05,
494
+ "loss": 0.7618,
495
+ "step": 810
496
+ },
497
+ {
498
+ "epoch": 0.16,
499
+ "learning_rate": 1.619913077834848e-05,
500
+ "loss": 0.767,
501
+ "step": 820
502
+ },
503
+ {
504
+ "epoch": 0.16,
505
+ "learning_rate": 1.6396681153694192e-05,
506
+ "loss": 0.8257,
507
+ "step": 830
508
+ },
509
+ {
510
+ "epoch": 0.17,
511
+ "learning_rate": 1.6594231529039905e-05,
512
+ "loss": 0.8116,
513
+ "step": 840
514
+ },
515
+ {
516
+ "epoch": 0.17,
517
+ "learning_rate": 1.679178190438562e-05,
518
+ "loss": 0.7649,
519
+ "step": 850
520
+ },
521
+ {
522
+ "epoch": 0.17,
523
+ "learning_rate": 1.698933227973133e-05,
524
+ "loss": 0.7616,
525
+ "step": 860
526
+ },
527
+ {
528
+ "epoch": 0.17,
529
+ "learning_rate": 1.7186882655077047e-05,
530
+ "loss": 0.7219,
531
+ "step": 870
532
+ },
533
+ {
534
+ "epoch": 0.17,
535
+ "learning_rate": 1.738443303042276e-05,
536
+ "loss": 0.7532,
537
+ "step": 880
538
+ },
539
+ {
540
+ "epoch": 0.18,
541
+ "learning_rate": 1.758198340576847e-05,
542
+ "loss": 0.7122,
543
+ "step": 890
544
+ },
545
+ {
546
+ "epoch": 0.18,
547
+ "learning_rate": 1.7779533781114185e-05,
548
+ "loss": 0.8242,
549
+ "step": 900
550
+ },
551
+ {
552
+ "epoch": 0.18,
553
+ "learning_rate": 1.7977084156459898e-05,
554
+ "loss": 0.7554,
555
+ "step": 910
556
+ },
557
+ {
558
+ "epoch": 0.18,
559
+ "learning_rate": 1.817463453180561e-05,
560
+ "loss": 0.7588,
561
+ "step": 920
562
+ },
563
+ {
564
+ "epoch": 0.18,
565
+ "learning_rate": 1.8372184907151324e-05,
566
+ "loss": 0.7404,
567
+ "step": 930
568
+ },
569
+ {
570
+ "epoch": 0.19,
571
+ "learning_rate": 1.856973528249704e-05,
572
+ "loss": 0.7907,
573
+ "step": 940
574
+ },
575
+ {
576
+ "epoch": 0.19,
577
+ "learning_rate": 1.876728565784275e-05,
578
+ "loss": 0.7775,
579
+ "step": 950
580
+ },
581
+ {
582
+ "epoch": 0.19,
583
+ "learning_rate": 1.8964836033188462e-05,
584
+ "loss": 0.7544,
585
+ "step": 960
586
+ },
587
+ {
588
+ "epoch": 0.19,
589
+ "learning_rate": 1.9162386408534178e-05,
590
+ "loss": 0.8053,
591
+ "step": 970
592
+ },
593
+ {
594
+ "epoch": 0.19,
595
+ "learning_rate": 1.935993678387989e-05,
596
+ "loss": 0.7226,
597
+ "step": 980
598
+ },
599
+ {
600
+ "epoch": 0.2,
601
+ "learning_rate": 1.9557487159225604e-05,
602
+ "loss": 0.7482,
603
+ "step": 990
604
+ },
605
+ {
606
+ "epoch": 0.2,
607
+ "learning_rate": 1.9755037534571317e-05,
608
+ "loss": 0.7171,
609
+ "step": 1000
610
+ },
611
+ {
612
+ "epoch": 0.2,
613
+ "learning_rate": 1.995258790991703e-05,
614
+ "loss": 0.7421,
615
+ "step": 1010
616
+ },
617
+ {
618
+ "epoch": 0.2,
619
+ "learning_rate": 2.0150138285262742e-05,
620
+ "loss": 0.7356,
621
+ "step": 1020
622
+ },
623
+ {
624
+ "epoch": 0.2,
625
+ "learning_rate": 2.0347688660608455e-05,
626
+ "loss": 0.713,
627
+ "step": 1030
628
+ },
629
+ {
630
+ "epoch": 0.21,
631
+ "learning_rate": 2.054523903595417e-05,
632
+ "loss": 0.7705,
633
+ "step": 1040
634
+ },
635
+ {
636
+ "epoch": 0.21,
637
+ "learning_rate": 2.074278941129988e-05,
638
+ "loss": 0.7501,
639
+ "step": 1050
640
+ },
641
+ {
642
+ "epoch": 0.21,
643
+ "learning_rate": 2.0940339786645597e-05,
644
+ "loss": 0.7572,
645
+ "step": 1060
646
+ },
647
+ {
648
+ "epoch": 0.21,
649
+ "learning_rate": 2.113789016199131e-05,
650
+ "loss": 0.7382,
651
+ "step": 1070
652
+ },
653
+ {
654
+ "epoch": 0.21,
655
+ "learning_rate": 2.1335440537337022e-05,
656
+ "loss": 0.7596,
657
+ "step": 1080
658
+ },
659
+ {
660
+ "epoch": 0.22,
661
+ "learning_rate": 2.1532990912682735e-05,
662
+ "loss": 0.7063,
663
+ "step": 1090
664
+ },
665
+ {
666
+ "epoch": 0.22,
667
+ "learning_rate": 2.1730541288028448e-05,
668
+ "loss": 0.7239,
669
+ "step": 1100
670
+ },
671
+ {
672
+ "epoch": 0.22,
673
+ "learning_rate": 2.192809166337416e-05,
674
+ "loss": 0.7648,
675
+ "step": 1110
676
+ },
677
+ {
678
+ "epoch": 0.22,
679
+ "learning_rate": 2.2125642038719874e-05,
680
+ "loss": 0.7539,
681
+ "step": 1120
682
+ },
683
+ {
684
+ "epoch": 0.22,
685
+ "learning_rate": 2.232319241406559e-05,
686
+ "loss": 0.7498,
687
+ "step": 1130
688
+ },
689
+ {
690
+ "epoch": 0.23,
691
+ "learning_rate": 2.25207427894113e-05,
692
+ "loss": 0.7647,
693
+ "step": 1140
694
+ },
695
+ {
696
+ "epoch": 0.23,
697
+ "learning_rate": 2.2718293164757015e-05,
698
+ "loss": 0.7657,
699
+ "step": 1150
700
+ },
701
+ {
702
+ "epoch": 0.23,
703
+ "learning_rate": 2.2915843540102728e-05,
704
+ "loss": 0.7451,
705
+ "step": 1160
706
+ },
707
+ {
708
+ "epoch": 0.23,
709
+ "learning_rate": 2.3113393915448438e-05,
710
+ "loss": 0.7528,
711
+ "step": 1170
712
+ },
713
+ {
714
+ "epoch": 0.23,
715
+ "learning_rate": 2.3310944290794154e-05,
716
+ "loss": 0.7548,
717
+ "step": 1180
718
+ },
719
+ {
720
+ "epoch": 0.24,
721
+ "learning_rate": 2.3508494666139867e-05,
722
+ "loss": 0.7616,
723
+ "step": 1190
724
+ },
725
+ {
726
+ "epoch": 0.24,
727
+ "learning_rate": 2.370604504148558e-05,
728
+ "loss": 0.7654,
729
+ "step": 1200
730
+ },
731
+ {
732
+ "epoch": 0.24,
733
+ "learning_rate": 2.3903595416831292e-05,
734
+ "loss": 0.7681,
735
+ "step": 1210
736
+ },
737
+ {
738
+ "epoch": 0.24,
739
+ "learning_rate": 2.410114579217701e-05,
740
+ "loss": 0.7302,
741
+ "step": 1220
742
+ },
743
+ {
744
+ "epoch": 0.24,
745
+ "learning_rate": 2.4298696167522718e-05,
746
+ "loss": 0.7516,
747
+ "step": 1230
748
+ },
749
+ {
750
+ "epoch": 0.24,
751
+ "learning_rate": 2.449624654286843e-05,
752
+ "loss": 0.7331,
753
+ "step": 1240
754
+ },
755
+ {
756
+ "epoch": 0.25,
757
+ "learning_rate": 2.4693796918214147e-05,
758
+ "loss": 0.7709,
759
+ "step": 1250
760
+ },
761
+ {
762
+ "epoch": 0.25,
763
+ "learning_rate": 2.489134729355986e-05,
764
+ "loss": 0.696,
765
+ "step": 1260
766
+ },
767
+ {
768
+ "epoch": 0.25,
769
+ "learning_rate": 2.508889766890557e-05,
770
+ "loss": 0.7537,
771
+ "step": 1270
772
+ },
773
+ {
774
+ "epoch": 0.25,
775
+ "learning_rate": 2.528644804425129e-05,
776
+ "loss": 0.7789,
777
+ "step": 1280
778
+ },
779
+ {
780
+ "epoch": 0.25,
781
+ "learning_rate": 2.5483998419596998e-05,
782
+ "loss": 0.7277,
783
+ "step": 1290
784
+ },
785
+ {
786
+ "epoch": 0.26,
787
+ "learning_rate": 2.568154879494271e-05,
788
+ "loss": 0.6968,
789
+ "step": 1300
790
+ },
791
+ {
792
+ "epoch": 0.26,
793
+ "learning_rate": 2.5879099170288423e-05,
794
+ "loss": 0.7329,
795
+ "step": 1310
796
+ },
797
+ {
798
+ "epoch": 0.26,
799
+ "learning_rate": 2.6076649545634136e-05,
800
+ "loss": 0.7501,
801
+ "step": 1320
802
+ },
803
+ {
804
+ "epoch": 0.26,
805
+ "learning_rate": 2.6274199920979852e-05,
806
+ "loss": 0.7224,
807
+ "step": 1330
808
+ },
809
+ {
810
+ "epoch": 0.26,
811
+ "learning_rate": 2.6471750296325565e-05,
812
+ "loss": 0.7276,
813
+ "step": 1340
814
+ },
815
+ {
816
+ "epoch": 0.27,
817
+ "learning_rate": 2.6669300671671278e-05,
818
+ "loss": 0.6827,
819
+ "step": 1350
820
+ },
821
+ {
822
+ "epoch": 0.27,
823
+ "learning_rate": 2.6866851047016987e-05,
824
+ "loss": 0.7762,
825
+ "step": 1360
826
+ },
827
+ {
828
+ "epoch": 0.27,
829
+ "learning_rate": 2.7064401422362707e-05,
830
+ "loss": 0.7124,
831
+ "step": 1370
832
+ },
833
+ {
834
+ "epoch": 0.27,
835
+ "learning_rate": 2.726195179770842e-05,
836
+ "loss": 0.6812,
837
+ "step": 1380
838
+ },
839
+ {
840
+ "epoch": 0.27,
841
+ "learning_rate": 2.745950217305413e-05,
842
+ "loss": 0.6922,
843
+ "step": 1390
844
+ },
845
+ {
846
+ "epoch": 0.28,
847
+ "learning_rate": 2.7657052548399842e-05,
848
+ "loss": 0.7216,
849
+ "step": 1400
850
+ },
851
+ {
852
+ "epoch": 0.28,
853
+ "learning_rate": 2.7854602923745555e-05,
854
+ "loss": 0.6899,
855
+ "step": 1410
856
+ },
857
+ {
858
+ "epoch": 0.28,
859
+ "learning_rate": 2.805215329909127e-05,
860
+ "loss": 0.7433,
861
+ "step": 1420
862
+ },
863
+ {
864
+ "epoch": 0.28,
865
+ "learning_rate": 2.8249703674436984e-05,
866
+ "loss": 0.7103,
867
+ "step": 1430
868
+ },
869
+ {
870
+ "epoch": 0.28,
871
+ "learning_rate": 2.8447254049782697e-05,
872
+ "loss": 0.7348,
873
+ "step": 1440
874
+ },
875
+ {
876
+ "epoch": 0.29,
877
+ "learning_rate": 2.8644804425128406e-05,
878
+ "loss": 0.8088,
879
+ "step": 1450
880
+ },
881
+ {
882
+ "epoch": 0.29,
883
+ "learning_rate": 2.884235480047412e-05,
884
+ "loss": 0.7432,
885
+ "step": 1460
886
+ },
887
+ {
888
+ "epoch": 0.29,
889
+ "learning_rate": 2.903990517581984e-05,
890
+ "loss": 0.7381,
891
+ "step": 1470
892
+ },
893
+ {
894
+ "epoch": 0.29,
895
+ "learning_rate": 2.9237455551165548e-05,
896
+ "loss": 0.7195,
897
+ "step": 1480
898
+ },
899
+ {
900
+ "epoch": 0.29,
901
+ "learning_rate": 2.943500592651126e-05,
902
+ "loss": 0.7038,
903
+ "step": 1490
904
+ },
905
+ {
906
+ "epoch": 0.3,
907
+ "learning_rate": 2.9632556301856973e-05,
908
+ "loss": 0.6643,
909
+ "step": 1500
910
+ },
911
+ {
912
+ "epoch": 0.3,
913
+ "learning_rate": 2.983010667720269e-05,
914
+ "loss": 0.7159,
915
+ "step": 1510
916
+ },
917
+ {
918
+ "epoch": 0.3,
919
+ "learning_rate": 3.0027657052548402e-05,
920
+ "loss": 0.6977,
921
+ "step": 1520
922
+ },
923
+ {
924
+ "epoch": 0.3,
925
+ "learning_rate": 3.0225207427894115e-05,
926
+ "loss": 0.713,
927
+ "step": 1530
928
+ },
929
+ {
930
+ "epoch": 0.3,
931
+ "learning_rate": 3.0422757803239828e-05,
932
+ "loss": 0.6713,
933
+ "step": 1540
934
+ },
935
+ {
936
+ "epoch": 0.31,
937
+ "learning_rate": 3.062030817858554e-05,
938
+ "loss": 0.6948,
939
+ "step": 1550
940
+ },
941
+ {
942
+ "epoch": 0.31,
943
+ "learning_rate": 3.081785855393126e-05,
944
+ "loss": 0.6964,
945
+ "step": 1560
946
+ },
947
+ {
948
+ "epoch": 0.31,
949
+ "learning_rate": 3.101540892927697e-05,
950
+ "loss": 0.7344,
951
+ "step": 1570
952
+ },
953
+ {
954
+ "epoch": 0.31,
955
+ "learning_rate": 3.121295930462268e-05,
956
+ "loss": 0.6921,
957
+ "step": 1580
958
+ },
959
+ {
960
+ "epoch": 0.31,
961
+ "learning_rate": 3.141050967996839e-05,
962
+ "loss": 0.7013,
963
+ "step": 1590
964
+ },
965
+ {
966
+ "epoch": 0.32,
967
+ "learning_rate": 3.160806005531411e-05,
968
+ "loss": 0.6525,
969
+ "step": 1600
970
+ },
971
+ {
972
+ "epoch": 0.32,
973
+ "learning_rate": 3.180561043065982e-05,
974
+ "loss": 0.6412,
975
+ "step": 1610
976
+ },
977
+ {
978
+ "epoch": 0.32,
979
+ "learning_rate": 3.2003160806005534e-05,
980
+ "loss": 0.6697,
981
+ "step": 1620
982
+ },
983
+ {
984
+ "epoch": 0.32,
985
+ "learning_rate": 3.2200711181351246e-05,
986
+ "loss": 0.7215,
987
+ "step": 1630
988
+ },
989
+ {
990
+ "epoch": 0.32,
991
+ "learning_rate": 3.239826155669696e-05,
992
+ "loss": 0.6918,
993
+ "step": 1640
994
+ },
995
+ {
996
+ "epoch": 0.33,
997
+ "learning_rate": 3.259581193204267e-05,
998
+ "loss": 0.7667,
999
+ "step": 1650
1000
+ },
1001
+ {
1002
+ "epoch": 0.33,
1003
+ "learning_rate": 3.2793362307388385e-05,
1004
+ "loss": 0.69,
1005
+ "step": 1660
1006
+ },
1007
+ {
1008
+ "epoch": 0.33,
1009
+ "learning_rate": 3.29909126827341e-05,
1010
+ "loss": 0.6899,
1011
+ "step": 1670
1012
+ },
1013
+ {
1014
+ "epoch": 0.33,
1015
+ "learning_rate": 3.318846305807981e-05,
1016
+ "loss": 0.663,
1017
+ "step": 1680
1018
+ },
1019
+ {
1020
+ "epoch": 0.33,
1021
+ "learning_rate": 3.338601343342552e-05,
1022
+ "loss": 0.7416,
1023
+ "step": 1690
1024
+ },
1025
+ {
1026
+ "epoch": 0.34,
1027
+ "learning_rate": 3.358356380877124e-05,
1028
+ "loss": 0.689,
1029
+ "step": 1700
1030
+ },
1031
+ {
1032
+ "epoch": 0.34,
1033
+ "learning_rate": 3.378111418411695e-05,
1034
+ "loss": 0.6992,
1035
+ "step": 1710
1036
+ },
1037
+ {
1038
+ "epoch": 0.34,
1039
+ "learning_rate": 3.397866455946266e-05,
1040
+ "loss": 0.6602,
1041
+ "step": 1720
1042
+ },
1043
+ {
1044
+ "epoch": 0.34,
1045
+ "learning_rate": 3.4176214934808374e-05,
1046
+ "loss": 0.6961,
1047
+ "step": 1730
1048
+ },
1049
+ {
1050
+ "epoch": 0.34,
1051
+ "learning_rate": 3.4373765310154094e-05,
1052
+ "loss": 0.7106,
1053
+ "step": 1740
1054
+ },
1055
+ {
1056
+ "epoch": 0.35,
1057
+ "learning_rate": 3.457131568549981e-05,
1058
+ "loss": 0.6957,
1059
+ "step": 1750
1060
+ },
1061
+ {
1062
+ "epoch": 0.35,
1063
+ "learning_rate": 3.476886606084552e-05,
1064
+ "loss": 0.6981,
1065
+ "step": 1760
1066
+ },
1067
+ {
1068
+ "epoch": 0.35,
1069
+ "learning_rate": 3.496641643619123e-05,
1070
+ "loss": 0.7585,
1071
+ "step": 1770
1072
+ },
1073
+ {
1074
+ "epoch": 0.35,
1075
+ "learning_rate": 3.516396681153694e-05,
1076
+ "loss": 0.6903,
1077
+ "step": 1780
1078
+ },
1079
+ {
1080
+ "epoch": 0.35,
1081
+ "learning_rate": 3.536151718688266e-05,
1082
+ "loss": 0.6592,
1083
+ "step": 1790
1084
+ },
1085
+ {
1086
+ "epoch": 0.36,
1087
+ "learning_rate": 3.555906756222837e-05,
1088
+ "loss": 0.732,
1089
+ "step": 1800
1090
+ },
1091
+ {
1092
+ "epoch": 0.36,
1093
+ "learning_rate": 3.5756617937574084e-05,
1094
+ "loss": 0.703,
1095
+ "step": 1810
1096
+ },
1097
+ {
1098
+ "epoch": 0.36,
1099
+ "learning_rate": 3.5954168312919796e-05,
1100
+ "loss": 0.7397,
1101
+ "step": 1820
1102
+ },
1103
+ {
1104
+ "epoch": 0.36,
1105
+ "learning_rate": 3.615171868826551e-05,
1106
+ "loss": 0.6665,
1107
+ "step": 1830
1108
+ },
1109
+ {
1110
+ "epoch": 0.36,
1111
+ "learning_rate": 3.634926906361122e-05,
1112
+ "loss": 0.7053,
1113
+ "step": 1840
1114
+ },
1115
+ {
1116
+ "epoch": 0.37,
1117
+ "learning_rate": 3.6546819438956935e-05,
1118
+ "loss": 0.7008,
1119
+ "step": 1850
1120
+ },
1121
+ {
1122
+ "epoch": 0.37,
1123
+ "learning_rate": 3.674436981430265e-05,
1124
+ "loss": 0.7181,
1125
+ "step": 1860
1126
+ },
1127
+ {
1128
+ "epoch": 0.37,
1129
+ "learning_rate": 3.694192018964836e-05,
1130
+ "loss": 0.7223,
1131
+ "step": 1870
1132
+ },
1133
+ {
1134
+ "epoch": 0.37,
1135
+ "learning_rate": 3.713947056499408e-05,
1136
+ "loss": 0.7027,
1137
+ "step": 1880
1138
+ },
1139
+ {
1140
+ "epoch": 0.37,
1141
+ "learning_rate": 3.733702094033979e-05,
1142
+ "loss": 0.6727,
1143
+ "step": 1890
1144
+ },
1145
+ {
1146
+ "epoch": 0.38,
1147
+ "learning_rate": 3.75345713156855e-05,
1148
+ "loss": 0.6756,
1149
+ "step": 1900
1150
+ },
1151
+ {
1152
+ "epoch": 0.38,
1153
+ "learning_rate": 3.773212169103121e-05,
1154
+ "loss": 0.7386,
1155
+ "step": 1910
1156
+ },
1157
+ {
1158
+ "epoch": 0.38,
1159
+ "learning_rate": 3.7929672066376924e-05,
1160
+ "loss": 0.6824,
1161
+ "step": 1920
1162
+ },
1163
+ {
1164
+ "epoch": 0.38,
1165
+ "learning_rate": 3.8127222441722644e-05,
1166
+ "loss": 0.7292,
1167
+ "step": 1930
1168
+ },
1169
+ {
1170
+ "epoch": 0.38,
1171
+ "learning_rate": 3.8324772817068357e-05,
1172
+ "loss": 0.716,
1173
+ "step": 1940
1174
+ },
1175
+ {
1176
+ "epoch": 0.39,
1177
+ "learning_rate": 3.852232319241407e-05,
1178
+ "loss": 0.657,
1179
+ "step": 1950
1180
+ },
1181
+ {
1182
+ "epoch": 0.39,
1183
+ "learning_rate": 3.871987356775978e-05,
1184
+ "loss": 0.6604,
1185
+ "step": 1960
1186
+ },
1187
+ {
1188
+ "epoch": 0.39,
1189
+ "learning_rate": 3.891742394310549e-05,
1190
+ "loss": 0.679,
1191
+ "step": 1970
1192
+ },
1193
+ {
1194
+ "epoch": 0.39,
1195
+ "learning_rate": 3.911497431845121e-05,
1196
+ "loss": 0.6932,
1197
+ "step": 1980
1198
+ },
1199
+ {
1200
+ "epoch": 0.39,
1201
+ "learning_rate": 3.931252469379692e-05,
1202
+ "loss": 0.7374,
1203
+ "step": 1990
1204
+ },
1205
+ {
1206
+ "epoch": 0.4,
1207
+ "learning_rate": 3.951007506914263e-05,
1208
+ "loss": 0.7186,
1209
+ "step": 2000
1210
+ },
1211
+ {
1212
+ "epoch": 0.4,
1213
+ "learning_rate": 3.9707625444488346e-05,
1214
+ "loss": 0.7086,
1215
+ "step": 2010
1216
+ },
1217
+ {
1218
+ "epoch": 0.4,
1219
+ "learning_rate": 3.990517581983406e-05,
1220
+ "loss": 0.6731,
1221
+ "step": 2020
1222
+ },
1223
+ {
1224
+ "epoch": 0.4,
1225
+ "learning_rate": 4.010272619517977e-05,
1226
+ "loss": 0.69,
1227
+ "step": 2030
1228
+ },
1229
+ {
1230
+ "epoch": 0.4,
1231
+ "learning_rate": 4.0300276570525485e-05,
1232
+ "loss": 0.6567,
1233
+ "step": 2040
1234
+ },
1235
+ {
1236
+ "epoch": 0.4,
1237
+ "learning_rate": 4.04978269458712e-05,
1238
+ "loss": 0.6974,
1239
+ "step": 2050
1240
+ },
1241
+ {
1242
+ "epoch": 0.41,
1243
+ "learning_rate": 4.069537732121691e-05,
1244
+ "loss": 0.7603,
1245
+ "step": 2060
1246
+ },
1247
+ {
1248
+ "epoch": 0.41,
1249
+ "learning_rate": 4.089292769656263e-05,
1250
+ "loss": 0.7012,
1251
+ "step": 2070
1252
+ },
1253
+ {
1254
+ "epoch": 0.41,
1255
+ "learning_rate": 4.109047807190834e-05,
1256
+ "loss": 0.7585,
1257
+ "step": 2080
1258
+ },
1259
+ {
1260
+ "epoch": 0.41,
1261
+ "learning_rate": 4.128802844725405e-05,
1262
+ "loss": 0.7086,
1263
+ "step": 2090
1264
+ },
1265
+ {
1266
+ "epoch": 0.41,
1267
+ "learning_rate": 4.148557882259976e-05,
1268
+ "loss": 0.7345,
1269
+ "step": 2100
1270
+ },
1271
+ {
1272
+ "epoch": 0.42,
1273
+ "learning_rate": 4.168312919794548e-05,
1274
+ "loss": 0.6888,
1275
+ "step": 2110
1276
+ },
1277
+ {
1278
+ "epoch": 0.42,
1279
+ "learning_rate": 4.1880679573291194e-05,
1280
+ "loss": 0.6637,
1281
+ "step": 2120
1282
+ },
1283
+ {
1284
+ "epoch": 0.42,
1285
+ "learning_rate": 4.2078229948636906e-05,
1286
+ "loss": 0.6936,
1287
+ "step": 2130
1288
+ },
1289
+ {
1290
+ "epoch": 0.42,
1291
+ "learning_rate": 4.227578032398262e-05,
1292
+ "loss": 0.6844,
1293
+ "step": 2140
1294
+ },
1295
+ {
1296
+ "epoch": 0.42,
1297
+ "learning_rate": 4.2473330699328325e-05,
1298
+ "loss": 0.709,
1299
+ "step": 2150
1300
+ },
1301
+ {
1302
+ "epoch": 0.43,
1303
+ "learning_rate": 4.2670881074674045e-05,
1304
+ "loss": 0.6244,
1305
+ "step": 2160
1306
+ },
1307
+ {
1308
+ "epoch": 0.43,
1309
+ "learning_rate": 4.286843145001976e-05,
1310
+ "loss": 0.6936,
1311
+ "step": 2170
1312
+ },
1313
+ {
1314
+ "epoch": 0.43,
1315
+ "learning_rate": 4.306598182536547e-05,
1316
+ "loss": 0.7073,
1317
+ "step": 2180
1318
+ },
1319
+ {
1320
+ "epoch": 0.43,
1321
+ "learning_rate": 4.326353220071118e-05,
1322
+ "loss": 0.7202,
1323
+ "step": 2190
1324
+ },
1325
+ {
1326
+ "epoch": 0.43,
1327
+ "learning_rate": 4.3461082576056896e-05,
1328
+ "loss": 0.6506,
1329
+ "step": 2200
1330
+ },
1331
+ {
1332
+ "epoch": 0.44,
1333
+ "learning_rate": 4.365863295140261e-05,
1334
+ "loss": 0.6339,
1335
+ "step": 2210
1336
+ },
1337
+ {
1338
+ "epoch": 0.44,
1339
+ "learning_rate": 4.385618332674832e-05,
1340
+ "loss": 0.6635,
1341
+ "step": 2220
1342
+ },
1343
+ {
1344
+ "epoch": 0.44,
1345
+ "learning_rate": 4.4053733702094034e-05,
1346
+ "loss": 0.6674,
1347
+ "step": 2230
1348
+ },
1349
+ {
1350
+ "epoch": 0.44,
1351
+ "learning_rate": 4.425128407743975e-05,
1352
+ "loss": 0.6857,
1353
+ "step": 2240
1354
+ },
1355
+ {
1356
+ "epoch": 0.44,
1357
+ "learning_rate": 4.444883445278547e-05,
1358
+ "loss": 0.7223,
1359
+ "step": 2250
1360
+ },
1361
+ {
1362
+ "epoch": 0.45,
1363
+ "learning_rate": 4.464638482813118e-05,
1364
+ "loss": 0.6716,
1365
+ "step": 2260
1366
+ },
1367
+ {
1368
+ "epoch": 0.45,
1369
+ "learning_rate": 4.4843935203476886e-05,
1370
+ "loss": 0.7094,
1371
+ "step": 2270
1372
+ },
1373
+ {
1374
+ "epoch": 0.45,
1375
+ "learning_rate": 4.50414855788226e-05,
1376
+ "loss": 0.6477,
1377
+ "step": 2280
1378
+ },
1379
+ {
1380
+ "epoch": 0.45,
1381
+ "learning_rate": 4.523903595416831e-05,
1382
+ "loss": 0.6409,
1383
+ "step": 2290
1384
+ },
1385
+ {
1386
+ "epoch": 0.45,
1387
+ "learning_rate": 4.543658632951403e-05,
1388
+ "loss": 0.6682,
1389
+ "step": 2300
1390
+ },
1391
+ {
1392
+ "epoch": 0.46,
1393
+ "learning_rate": 4.5634136704859744e-05,
1394
+ "loss": 0.6727,
1395
+ "step": 2310
1396
+ },
1397
+ {
1398
+ "epoch": 0.46,
1399
+ "learning_rate": 4.5831687080205456e-05,
1400
+ "loss": 0.6348,
1401
+ "step": 2320
1402
+ },
1403
+ {
1404
+ "epoch": 0.46,
1405
+ "learning_rate": 4.602923745555117e-05,
1406
+ "loss": 0.683,
1407
+ "step": 2330
1408
+ },
1409
+ {
1410
+ "epoch": 0.46,
1411
+ "learning_rate": 4.6226787830896875e-05,
1412
+ "loss": 0.5872,
1413
+ "step": 2340
1414
+ },
1415
+ {
1416
+ "epoch": 0.46,
1417
+ "learning_rate": 4.6424338206242595e-05,
1418
+ "loss": 0.7156,
1419
+ "step": 2350
1420
+ },
1421
+ {
1422
+ "epoch": 0.47,
1423
+ "learning_rate": 4.662188858158831e-05,
1424
+ "loss": 0.6441,
1425
+ "step": 2360
1426
+ },
1427
+ {
1428
+ "epoch": 0.47,
1429
+ "learning_rate": 4.681943895693402e-05,
1430
+ "loss": 0.647,
1431
+ "step": 2370
1432
+ },
1433
+ {
1434
+ "epoch": 0.47,
1435
+ "learning_rate": 4.701698933227973e-05,
1436
+ "loss": 0.7303,
1437
+ "step": 2380
1438
+ },
1439
+ {
1440
+ "epoch": 0.47,
1441
+ "learning_rate": 4.7214539707625446e-05,
1442
+ "loss": 0.6304,
1443
+ "step": 2390
1444
+ },
1445
+ {
1446
+ "epoch": 0.47,
1447
+ "learning_rate": 4.741209008297116e-05,
1448
+ "loss": 0.6577,
1449
+ "step": 2400
1450
+ },
1451
+ {
1452
+ "epoch": 0.48,
1453
+ "learning_rate": 4.760964045831687e-05,
1454
+ "loss": 0.6941,
1455
+ "step": 2410
1456
+ },
1457
+ {
1458
+ "epoch": 0.48,
1459
+ "learning_rate": 4.7807190833662584e-05,
1460
+ "loss": 0.6309,
1461
+ "step": 2420
1462
+ },
1463
+ {
1464
+ "epoch": 0.48,
1465
+ "learning_rate": 4.80047412090083e-05,
1466
+ "loss": 0.6584,
1467
+ "step": 2430
1468
+ },
1469
+ {
1470
+ "epoch": 0.48,
1471
+ "learning_rate": 4.820229158435402e-05,
1472
+ "loss": 0.6387,
1473
+ "step": 2440
1474
+ },
1475
+ {
1476
+ "epoch": 0.48,
1477
+ "learning_rate": 4.839984195969973e-05,
1478
+ "loss": 0.6602,
1479
+ "step": 2450
1480
+ },
1481
+ {
1482
+ "epoch": 0.49,
1483
+ "learning_rate": 4.8597392335045435e-05,
1484
+ "loss": 0.7043,
1485
+ "step": 2460
1486
+ },
1487
+ {
1488
+ "epoch": 0.49,
1489
+ "learning_rate": 4.879494271039115e-05,
1490
+ "loss": 0.711,
1491
+ "step": 2470
1492
+ },
1493
+ {
1494
+ "epoch": 0.49,
1495
+ "learning_rate": 4.899249308573686e-05,
1496
+ "loss": 0.6393,
1497
+ "step": 2480
1498
+ },
1499
+ {
1500
+ "epoch": 0.49,
1501
+ "learning_rate": 4.919004346108258e-05,
1502
+ "loss": 0.7015,
1503
+ "step": 2490
1504
+ },
1505
+ {
1506
+ "epoch": 0.49,
1507
+ "learning_rate": 4.9387593836428293e-05,
1508
+ "loss": 0.6153,
1509
+ "step": 2500
1510
+ },
1511
+ {
1512
+ "epoch": 0.5,
1513
+ "learning_rate": 4.9585144211774006e-05,
1514
+ "loss": 0.6787,
1515
+ "step": 2510
1516
+ },
1517
+ {
1518
+ "epoch": 0.5,
1519
+ "learning_rate": 4.978269458711972e-05,
1520
+ "loss": 0.6316,
1521
+ "step": 2520
1522
+ },
1523
+ {
1524
+ "epoch": 0.5,
1525
+ "learning_rate": 4.998024496246543e-05,
1526
+ "loss": 0.6494,
1527
+ "step": 2530
1528
+ },
1529
+ {
1530
+ "epoch": 0.5,
1531
+ "learning_rate": 5.017779533781114e-05,
1532
+ "loss": 0.6925,
1533
+ "step": 2540
1534
+ },
1535
+ {
1536
+ "epoch": 0.5,
1537
+ "learning_rate": 5.0375345713156864e-05,
1538
+ "loss": 0.7249,
1539
+ "step": 2550
1540
+ },
1541
+ {
1542
+ "epoch": 0.51,
1543
+ "learning_rate": 5.057289608850258e-05,
1544
+ "loss": 0.6799,
1545
+ "step": 2560
1546
+ },
1547
+ {
1548
+ "epoch": 0.51,
1549
+ "learning_rate": 5.077044646384829e-05,
1550
+ "loss": 0.7094,
1551
+ "step": 2570
1552
+ },
1553
+ {
1554
+ "epoch": 0.51,
1555
+ "learning_rate": 5.0967996839193996e-05,
1556
+ "loss": 0.7007,
1557
+ "step": 2580
1558
+ },
1559
+ {
1560
+ "epoch": 0.51,
1561
+ "learning_rate": 5.116554721453971e-05,
1562
+ "loss": 0.7456,
1563
+ "step": 2590
1564
+ },
1565
+ {
1566
+ "epoch": 0.51,
1567
+ "learning_rate": 5.136309758988542e-05,
1568
+ "loss": 0.6767,
1569
+ "step": 2600
1570
+ },
1571
+ {
1572
+ "epoch": 0.52,
1573
+ "learning_rate": 5.1560647965231134e-05,
1574
+ "loss": 0.7366,
1575
+ "step": 2610
1576
+ },
1577
+ {
1578
+ "epoch": 0.52,
1579
+ "learning_rate": 5.175819834057685e-05,
1580
+ "loss": 0.6579,
1581
+ "step": 2620
1582
+ },
1583
+ {
1584
+ "epoch": 0.52,
1585
+ "learning_rate": 5.195574871592256e-05,
1586
+ "loss": 0.6985,
1587
+ "step": 2630
1588
+ },
1589
+ {
1590
+ "epoch": 0.52,
1591
+ "learning_rate": 5.215329909126827e-05,
1592
+ "loss": 0.7202,
1593
+ "step": 2640
1594
+ },
1595
+ {
1596
+ "epoch": 0.52,
1597
+ "learning_rate": 5.235084946661399e-05,
1598
+ "loss": 0.6233,
1599
+ "step": 2650
1600
+ },
1601
+ {
1602
+ "epoch": 0.53,
1603
+ "learning_rate": 5.2548399841959705e-05,
1604
+ "loss": 0.6271,
1605
+ "step": 2660
1606
+ },
1607
+ {
1608
+ "epoch": 0.53,
1609
+ "learning_rate": 5.274595021730542e-05,
1610
+ "loss": 0.697,
1611
+ "step": 2670
1612
+ },
1613
+ {
1614
+ "epoch": 0.53,
1615
+ "learning_rate": 5.294350059265113e-05,
1616
+ "loss": 0.6596,
1617
+ "step": 2680
1618
+ },
1619
+ {
1620
+ "epoch": 0.53,
1621
+ "learning_rate": 5.314105096799684e-05,
1622
+ "loss": 0.6474,
1623
+ "step": 2690
1624
+ },
1625
+ {
1626
+ "epoch": 0.53,
1627
+ "learning_rate": 5.3338601343342556e-05,
1628
+ "loss": 0.6291,
1629
+ "step": 2700
1630
+ },
1631
+ {
1632
+ "epoch": 0.54,
1633
+ "learning_rate": 5.353615171868827e-05,
1634
+ "loss": 0.634,
1635
+ "step": 2710
1636
+ },
1637
+ {
1638
+ "epoch": 0.54,
1639
+ "learning_rate": 5.3733702094033975e-05,
1640
+ "loss": 0.67,
1641
+ "step": 2720
1642
+ },
1643
+ {
1644
+ "epoch": 0.54,
1645
+ "learning_rate": 5.393125246937969e-05,
1646
+ "loss": 0.6784,
1647
+ "step": 2730
1648
+ },
1649
+ {
1650
+ "epoch": 0.54,
1651
+ "learning_rate": 5.4128802844725414e-05,
1652
+ "loss": 0.6826,
1653
+ "step": 2740
1654
+ },
1655
+ {
1656
+ "epoch": 0.54,
1657
+ "learning_rate": 5.432635322007113e-05,
1658
+ "loss": 0.6485,
1659
+ "step": 2750
1660
+ },
1661
+ {
1662
+ "epoch": 0.55,
1663
+ "learning_rate": 5.452390359541684e-05,
1664
+ "loss": 0.6364,
1665
+ "step": 2760
1666
+ },
1667
+ {
1668
+ "epoch": 0.55,
1669
+ "learning_rate": 5.4721453970762546e-05,
1670
+ "loss": 0.6196,
1671
+ "step": 2770
1672
+ },
1673
+ {
1674
+ "epoch": 0.55,
1675
+ "learning_rate": 5.491900434610826e-05,
1676
+ "loss": 0.6812,
1677
+ "step": 2780
1678
+ },
1679
+ {
1680
+ "epoch": 0.55,
1681
+ "learning_rate": 5.511655472145397e-05,
1682
+ "loss": 0.6404,
1683
+ "step": 2790
1684
+ },
1685
+ {
1686
+ "epoch": 0.55,
1687
+ "learning_rate": 5.5314105096799684e-05,
1688
+ "loss": 0.6725,
1689
+ "step": 2800
1690
+ },
1691
+ {
1692
+ "epoch": 0.56,
1693
+ "learning_rate": 5.55116554721454e-05,
1694
+ "loss": 0.6811,
1695
+ "step": 2810
1696
+ },
1697
+ {
1698
+ "epoch": 0.56,
1699
+ "learning_rate": 5.570920584749111e-05,
1700
+ "loss": 0.6282,
1701
+ "step": 2820
1702
+ },
1703
+ {
1704
+ "epoch": 0.56,
1705
+ "learning_rate": 5.590675622283683e-05,
1706
+ "loss": 0.6505,
1707
+ "step": 2830
1708
+ },
1709
+ {
1710
+ "epoch": 0.56,
1711
+ "learning_rate": 5.610430659818254e-05,
1712
+ "loss": 0.6352,
1713
+ "step": 2840
1714
+ },
1715
+ {
1716
+ "epoch": 0.56,
1717
+ "learning_rate": 5.6301856973528255e-05,
1718
+ "loss": 0.6836,
1719
+ "step": 2850
1720
+ },
1721
+ {
1722
+ "epoch": 0.56,
1723
+ "learning_rate": 5.649940734887397e-05,
1724
+ "loss": 0.5995,
1725
+ "step": 2860
1726
+ },
1727
+ {
1728
+ "epoch": 0.57,
1729
+ "learning_rate": 5.669695772421968e-05,
1730
+ "loss": 0.6773,
1731
+ "step": 2870
1732
+ },
1733
+ {
1734
+ "epoch": 0.57,
1735
+ "learning_rate": 5.689450809956539e-05,
1736
+ "loss": 0.6272,
1737
+ "step": 2880
1738
+ },
1739
+ {
1740
+ "epoch": 0.57,
1741
+ "learning_rate": 5.7092058474911106e-05,
1742
+ "loss": 0.681,
1743
+ "step": 2890
1744
+ },
1745
+ {
1746
+ "epoch": 0.57,
1747
+ "learning_rate": 5.728960885025681e-05,
1748
+ "loss": 0.6307,
1749
+ "step": 2900
1750
+ },
1751
+ {
1752
+ "epoch": 0.57,
1753
+ "learning_rate": 5.7487159225602525e-05,
1754
+ "loss": 0.6643,
1755
+ "step": 2910
1756
+ },
1757
+ {
1758
+ "epoch": 0.58,
1759
+ "learning_rate": 5.768470960094824e-05,
1760
+ "loss": 0.6184,
1761
+ "step": 2920
1762
+ },
1763
+ {
1764
+ "epoch": 0.58,
1765
+ "learning_rate": 5.7882259976293964e-05,
1766
+ "loss": 0.6572,
1767
+ "step": 2930
1768
+ },
1769
+ {
1770
+ "epoch": 0.58,
1771
+ "learning_rate": 5.807981035163968e-05,
1772
+ "loss": 0.6898,
1773
+ "step": 2940
1774
+ },
1775
+ {
1776
+ "epoch": 0.58,
1777
+ "learning_rate": 5.827736072698538e-05,
1778
+ "loss": 0.6442,
1779
+ "step": 2950
1780
+ },
1781
+ {
1782
+ "epoch": 0.58,
1783
+ "learning_rate": 5.8474911102331095e-05,
1784
+ "loss": 0.6782,
1785
+ "step": 2960
1786
+ },
1787
+ {
1788
+ "epoch": 0.59,
1789
+ "learning_rate": 5.867246147767681e-05,
1790
+ "loss": 0.6749,
1791
+ "step": 2970
1792
+ },
1793
+ {
1794
+ "epoch": 0.59,
1795
+ "learning_rate": 5.887001185302252e-05,
1796
+ "loss": 0.6217,
1797
+ "step": 2980
1798
+ },
1799
+ {
1800
+ "epoch": 0.59,
1801
+ "learning_rate": 5.9067562228368234e-05,
1802
+ "loss": 0.6644,
1803
+ "step": 2990
1804
+ },
1805
+ {
1806
+ "epoch": 0.59,
1807
+ "learning_rate": 5.926511260371395e-05,
1808
+ "loss": 0.6527,
1809
+ "step": 3000
1810
+ },
1811
+ {
1812
+ "epoch": 0.59,
1813
+ "learning_rate": 5.946266297905966e-05,
1814
+ "loss": 0.6961,
1815
+ "step": 3010
1816
+ },
1817
+ {
1818
+ "epoch": 0.6,
1819
+ "learning_rate": 5.966021335440538e-05,
1820
+ "loss": 0.6514,
1821
+ "step": 3020
1822
+ },
1823
+ {
1824
+ "epoch": 0.6,
1825
+ "learning_rate": 5.985776372975109e-05,
1826
+ "loss": 0.6066,
1827
+ "step": 3030
1828
+ },
1829
+ {
1830
+ "epoch": 0.6,
1831
+ "learning_rate": 6.0055314105096805e-05,
1832
+ "loss": 0.673,
1833
+ "step": 3040
1834
+ },
1835
+ {
1836
+ "epoch": 0.6,
1837
+ "learning_rate": 6.025286448044252e-05,
1838
+ "loss": 0.6685,
1839
+ "step": 3050
1840
+ },
1841
+ {
1842
+ "epoch": 0.6,
1843
+ "learning_rate": 6.045041485578823e-05,
1844
+ "loss": 0.6274,
1845
+ "step": 3060
1846
+ },
1847
+ {
1848
+ "epoch": 0.61,
1849
+ "learning_rate": 6.064796523113394e-05,
1850
+ "loss": 0.7091,
1851
+ "step": 3070
1852
+ },
1853
+ {
1854
+ "epoch": 0.61,
1855
+ "learning_rate": 6.0845515606479656e-05,
1856
+ "loss": 0.6791,
1857
+ "step": 3080
1858
+ },
1859
+ {
1860
+ "epoch": 0.61,
1861
+ "learning_rate": 6.104306598182537e-05,
1862
+ "loss": 0.6348,
1863
+ "step": 3090
1864
+ },
1865
+ {
1866
+ "epoch": 0.61,
1867
+ "learning_rate": 6.124061635717107e-05,
1868
+ "loss": 0.6043,
1869
+ "step": 3100
1870
+ },
1871
+ {
1872
+ "epoch": 0.61,
1873
+ "learning_rate": 6.14381667325168e-05,
1874
+ "loss": 0.6991,
1875
+ "step": 3110
1876
+ },
1877
+ {
1878
+ "epoch": 0.62,
1879
+ "learning_rate": 6.163571710786251e-05,
1880
+ "loss": 0.6613,
1881
+ "step": 3120
1882
+ },
1883
+ {
1884
+ "epoch": 0.62,
1885
+ "learning_rate": 6.183326748320822e-05,
1886
+ "loss": 0.6498,
1887
+ "step": 3130
1888
+ },
1889
+ {
1890
+ "epoch": 0.62,
1891
+ "learning_rate": 6.203081785855394e-05,
1892
+ "loss": 0.6819,
1893
+ "step": 3140
1894
+ },
1895
+ {
1896
+ "epoch": 0.62,
1897
+ "learning_rate": 6.222836823389965e-05,
1898
+ "loss": 0.6316,
1899
+ "step": 3150
1900
+ },
1901
+ {
1902
+ "epoch": 0.62,
1903
+ "learning_rate": 6.242591860924536e-05,
1904
+ "loss": 0.6672,
1905
+ "step": 3160
1906
+ },
1907
+ {
1908
+ "epoch": 0.63,
1909
+ "learning_rate": 6.262346898459107e-05,
1910
+ "loss": 0.6602,
1911
+ "step": 3170
1912
+ },
1913
+ {
1914
+ "epoch": 0.63,
1915
+ "learning_rate": 6.282101935993678e-05,
1916
+ "loss": 0.6463,
1917
+ "step": 3180
1918
+ },
1919
+ {
1920
+ "epoch": 0.63,
1921
+ "learning_rate": 6.30185697352825e-05,
1922
+ "loss": 0.614,
1923
+ "step": 3190
1924
+ },
1925
+ {
1926
+ "epoch": 0.63,
1927
+ "learning_rate": 6.321612011062822e-05,
1928
+ "loss": 0.6689,
1929
+ "step": 3200
1930
+ },
1931
+ {
1932
+ "epoch": 0.63,
1933
+ "learning_rate": 6.341367048597394e-05,
1934
+ "loss": 0.6224,
1935
+ "step": 3210
1936
+ },
1937
+ {
1938
+ "epoch": 0.64,
1939
+ "learning_rate": 6.361122086131964e-05,
1940
+ "loss": 0.5965,
1941
+ "step": 3220
1942
+ },
1943
+ {
1944
+ "epoch": 0.64,
1945
+ "learning_rate": 6.380877123666536e-05,
1946
+ "loss": 0.6219,
1947
+ "step": 3230
1948
+ },
1949
+ {
1950
+ "epoch": 0.64,
1951
+ "learning_rate": 6.400632161201107e-05,
1952
+ "loss": 0.6675,
1953
+ "step": 3240
1954
+ },
1955
+ {
1956
+ "epoch": 0.64,
1957
+ "learning_rate": 6.420387198735677e-05,
1958
+ "loss": 0.7074,
1959
+ "step": 3250
1960
+ },
1961
+ {
1962
+ "epoch": 0.64,
1963
+ "learning_rate": 6.440142236270249e-05,
1964
+ "loss": 0.6304,
1965
+ "step": 3260
1966
+ },
1967
+ {
1968
+ "epoch": 0.65,
1969
+ "learning_rate": 6.45989727380482e-05,
1970
+ "loss": 0.6729,
1971
+ "step": 3270
1972
+ },
1973
+ {
1974
+ "epoch": 0.65,
1975
+ "learning_rate": 6.479652311339392e-05,
1976
+ "loss": 0.6177,
1977
+ "step": 3280
1978
+ },
1979
+ {
1980
+ "epoch": 0.65,
1981
+ "learning_rate": 6.499407348873962e-05,
1982
+ "loss": 0.6709,
1983
+ "step": 3290
1984
+ },
1985
+ {
1986
+ "epoch": 0.65,
1987
+ "learning_rate": 6.519162386408534e-05,
1988
+ "loss": 0.7026,
1989
+ "step": 3300
1990
+ },
1991
+ {
1992
+ "epoch": 0.65,
1993
+ "learning_rate": 6.538917423943106e-05,
1994
+ "loss": 0.6838,
1995
+ "step": 3310
1996
+ },
1997
+ {
1998
+ "epoch": 0.66,
1999
+ "learning_rate": 6.558672461477677e-05,
2000
+ "loss": 0.6294,
2001
+ "step": 3320
2002
+ },
2003
+ {
2004
+ "epoch": 0.66,
2005
+ "learning_rate": 6.578427499012249e-05,
2006
+ "loss": 0.6447,
2007
+ "step": 3330
2008
+ },
2009
+ {
2010
+ "epoch": 0.66,
2011
+ "learning_rate": 6.59818253654682e-05,
2012
+ "loss": 0.592,
2013
+ "step": 3340
2014
+ },
2015
+ {
2016
+ "epoch": 0.66,
2017
+ "learning_rate": 6.617937574081391e-05,
2018
+ "loss": 0.6698,
2019
+ "step": 3350
2020
+ },
2021
+ {
2022
+ "epoch": 0.66,
2023
+ "learning_rate": 6.637692611615962e-05,
2024
+ "loss": 0.6528,
2025
+ "step": 3360
2026
+ },
2027
+ {
2028
+ "epoch": 0.67,
2029
+ "learning_rate": 6.657447649150533e-05,
2030
+ "loss": 0.6679,
2031
+ "step": 3370
2032
+ },
2033
+ {
2034
+ "epoch": 0.67,
2035
+ "learning_rate": 6.677202686685105e-05,
2036
+ "loss": 0.6379,
2037
+ "step": 3380
2038
+ },
2039
+ {
2040
+ "epoch": 0.67,
2041
+ "learning_rate": 6.696957724219677e-05,
2042
+ "loss": 0.6213,
2043
+ "step": 3390
2044
+ },
2045
+ {
2046
+ "epoch": 0.67,
2047
+ "learning_rate": 6.716712761754249e-05,
2048
+ "loss": 0.6696,
2049
+ "step": 3400
2050
+ },
2051
+ {
2052
+ "epoch": 0.67,
2053
+ "learning_rate": 6.736467799288819e-05,
2054
+ "loss": 0.6174,
2055
+ "step": 3410
2056
+ },
2057
+ {
2058
+ "epoch": 0.68,
2059
+ "learning_rate": 6.75622283682339e-05,
2060
+ "loss": 0.6211,
2061
+ "step": 3420
2062
+ },
2063
+ {
2064
+ "epoch": 0.68,
2065
+ "learning_rate": 6.775977874357962e-05,
2066
+ "loss": 0.6016,
2067
+ "step": 3430
2068
+ },
2069
+ {
2070
+ "epoch": 0.68,
2071
+ "learning_rate": 6.795732911892532e-05,
2072
+ "loss": 0.6407,
2073
+ "step": 3440
2074
+ },
2075
+ {
2076
+ "epoch": 0.68,
2077
+ "learning_rate": 6.815487949427104e-05,
2078
+ "loss": 0.6349,
2079
+ "step": 3450
2080
+ },
2081
+ {
2082
+ "epoch": 0.68,
2083
+ "learning_rate": 6.835242986961675e-05,
2084
+ "loss": 0.6478,
2085
+ "step": 3460
2086
+ },
2087
+ {
2088
+ "epoch": 0.69,
2089
+ "learning_rate": 6.854998024496247e-05,
2090
+ "loss": 0.6733,
2091
+ "step": 3470
2092
+ },
2093
+ {
2094
+ "epoch": 0.69,
2095
+ "learning_rate": 6.874753062030819e-05,
2096
+ "loss": 0.6037,
2097
+ "step": 3480
2098
+ },
2099
+ {
2100
+ "epoch": 0.69,
2101
+ "learning_rate": 6.89450809956539e-05,
2102
+ "loss": 0.6621,
2103
+ "step": 3490
2104
+ },
2105
+ {
2106
+ "epoch": 0.69,
2107
+ "learning_rate": 6.914263137099961e-05,
2108
+ "loss": 0.6122,
2109
+ "step": 3500
2110
+ },
2111
+ {
2112
+ "epoch": 0.69,
2113
+ "learning_rate": 6.934018174634532e-05,
2114
+ "loss": 0.594,
2115
+ "step": 3510
2116
+ },
2117
+ {
2118
+ "epoch": 0.7,
2119
+ "learning_rate": 6.953773212169104e-05,
2120
+ "loss": 0.6445,
2121
+ "step": 3520
2122
+ },
2123
+ {
2124
+ "epoch": 0.7,
2125
+ "learning_rate": 6.973528249703675e-05,
2126
+ "loss": 0.6363,
2127
+ "step": 3530
2128
+ },
2129
+ {
2130
+ "epoch": 0.7,
2131
+ "learning_rate": 6.993283287238246e-05,
2132
+ "loss": 0.635,
2133
+ "step": 3540
2134
+ },
2135
+ {
2136
+ "epoch": 0.7,
2137
+ "learning_rate": 7.013038324772817e-05,
2138
+ "loss": 0.6506,
2139
+ "step": 3550
2140
+ },
2141
+ {
2142
+ "epoch": 0.7,
2143
+ "learning_rate": 7.032793362307388e-05,
2144
+ "loss": 0.6546,
2145
+ "step": 3560
2146
+ },
2147
+ {
2148
+ "epoch": 0.71,
2149
+ "learning_rate": 7.052548399841961e-05,
2150
+ "loss": 0.6486,
2151
+ "step": 3570
2152
+ },
2153
+ {
2154
+ "epoch": 0.71,
2155
+ "learning_rate": 7.072303437376532e-05,
2156
+ "loss": 0.6543,
2157
+ "step": 3580
2158
+ },
2159
+ {
2160
+ "epoch": 0.71,
2161
+ "learning_rate": 7.092058474911104e-05,
2162
+ "loss": 0.6799,
2163
+ "step": 3590
2164
+ },
2165
+ {
2166
+ "epoch": 0.71,
2167
+ "learning_rate": 7.111813512445674e-05,
2168
+ "loss": 0.665,
2169
+ "step": 3600
2170
+ },
2171
+ {
2172
+ "epoch": 0.71,
2173
+ "learning_rate": 7.131568549980245e-05,
2174
+ "loss": 0.6728,
2175
+ "step": 3610
2176
+ },
2177
+ {
2178
+ "epoch": 0.72,
2179
+ "learning_rate": 7.151323587514817e-05,
2180
+ "loss": 0.6574,
2181
+ "step": 3620
2182
+ },
2183
+ {
2184
+ "epoch": 0.72,
2185
+ "learning_rate": 7.171078625049387e-05,
2186
+ "loss": 0.6618,
2187
+ "step": 3630
2188
+ },
2189
+ {
2190
+ "epoch": 0.72,
2191
+ "learning_rate": 7.190833662583959e-05,
2192
+ "loss": 0.5915,
2193
+ "step": 3640
2194
+ },
2195
+ {
2196
+ "epoch": 0.72,
2197
+ "learning_rate": 7.21058870011853e-05,
2198
+ "loss": 0.5632,
2199
+ "step": 3650
2200
+ },
2201
+ {
2202
+ "epoch": 0.72,
2203
+ "learning_rate": 7.230343737653102e-05,
2204
+ "loss": 0.6802,
2205
+ "step": 3660
2206
+ },
2207
+ {
2208
+ "epoch": 0.72,
2209
+ "learning_rate": 7.250098775187674e-05,
2210
+ "loss": 0.6749,
2211
+ "step": 3670
2212
+ },
2213
+ {
2214
+ "epoch": 0.73,
2215
+ "learning_rate": 7.269853812722244e-05,
2216
+ "loss": 0.665,
2217
+ "step": 3680
2218
+ },
2219
+ {
2220
+ "epoch": 0.73,
2221
+ "learning_rate": 7.289608850256816e-05,
2222
+ "loss": 0.7057,
2223
+ "step": 3690
2224
+ },
2225
+ {
2226
+ "epoch": 0.73,
2227
+ "learning_rate": 7.309363887791387e-05,
2228
+ "loss": 0.6284,
2229
+ "step": 3700
2230
+ },
2231
+ {
2232
+ "epoch": 0.73,
2233
+ "learning_rate": 7.329118925325959e-05,
2234
+ "loss": 0.6445,
2235
+ "step": 3710
2236
+ },
2237
+ {
2238
+ "epoch": 0.73,
2239
+ "learning_rate": 7.34887396286053e-05,
2240
+ "loss": 0.6352,
2241
+ "step": 3720
2242
+ },
2243
+ {
2244
+ "epoch": 0.74,
2245
+ "learning_rate": 7.368629000395101e-05,
2246
+ "loss": 0.6989,
2247
+ "step": 3730
2248
+ },
2249
+ {
2250
+ "epoch": 0.74,
2251
+ "learning_rate": 7.388384037929672e-05,
2252
+ "loss": 0.6579,
2253
+ "step": 3740
2254
+ },
2255
+ {
2256
+ "epoch": 0.74,
2257
+ "learning_rate": 7.408139075464243e-05,
2258
+ "loss": 0.6215,
2259
+ "step": 3750
2260
+ },
2261
+ {
2262
+ "epoch": 0.74,
2263
+ "learning_rate": 7.427894112998816e-05,
2264
+ "loss": 0.5917,
2265
+ "step": 3760
2266
+ },
2267
+ {
2268
+ "epoch": 0.74,
2269
+ "learning_rate": 7.447649150533387e-05,
2270
+ "loss": 0.6685,
2271
+ "step": 3770
2272
+ },
2273
+ {
2274
+ "epoch": 0.75,
2275
+ "learning_rate": 7.467404188067959e-05,
2276
+ "loss": 0.6314,
2277
+ "step": 3780
2278
+ },
2279
+ {
2280
+ "epoch": 0.75,
2281
+ "learning_rate": 7.487159225602529e-05,
2282
+ "loss": 0.6184,
2283
+ "step": 3790
2284
+ },
2285
+ {
2286
+ "epoch": 0.75,
2287
+ "learning_rate": 7.5069142631371e-05,
2288
+ "loss": 0.6256,
2289
+ "step": 3800
2290
+ },
2291
+ {
2292
+ "epoch": 0.75,
2293
+ "learning_rate": 7.526669300671672e-05,
2294
+ "loss": 0.6055,
2295
+ "step": 3810
2296
+ },
2297
+ {
2298
+ "epoch": 0.75,
2299
+ "learning_rate": 7.546424338206242e-05,
2300
+ "loss": 0.6527,
2301
+ "step": 3820
2302
+ },
2303
+ {
2304
+ "epoch": 0.76,
2305
+ "learning_rate": 7.566179375740814e-05,
2306
+ "loss": 0.6632,
2307
+ "step": 3830
2308
+ },
2309
+ {
2310
+ "epoch": 0.76,
2311
+ "learning_rate": 7.585934413275385e-05,
2312
+ "loss": 0.5831,
2313
+ "step": 3840
2314
+ },
2315
+ {
2316
+ "epoch": 0.76,
2317
+ "learning_rate": 7.605689450809957e-05,
2318
+ "loss": 0.6703,
2319
+ "step": 3850
2320
+ },
2321
+ {
2322
+ "epoch": 0.76,
2323
+ "learning_rate": 7.625444488344529e-05,
2324
+ "loss": 0.6218,
2325
+ "step": 3860
2326
+ },
2327
+ {
2328
+ "epoch": 0.76,
2329
+ "learning_rate": 7.6451995258791e-05,
2330
+ "loss": 0.6795,
2331
+ "step": 3870
2332
+ },
2333
+ {
2334
+ "epoch": 0.77,
2335
+ "learning_rate": 7.664954563413671e-05,
2336
+ "loss": 0.6779,
2337
+ "step": 3880
2338
+ },
2339
+ {
2340
+ "epoch": 0.77,
2341
+ "learning_rate": 7.684709600948242e-05,
2342
+ "loss": 0.6174,
2343
+ "step": 3890
2344
+ },
2345
+ {
2346
+ "epoch": 0.77,
2347
+ "learning_rate": 7.704464638482814e-05,
2348
+ "loss": 0.6547,
2349
+ "step": 3900
2350
+ },
2351
+ {
2352
+ "epoch": 0.77,
2353
+ "learning_rate": 7.724219676017384e-05,
2354
+ "loss": 0.6003,
2355
+ "step": 3910
2356
+ },
2357
+ {
2358
+ "epoch": 0.77,
2359
+ "learning_rate": 7.743974713551956e-05,
2360
+ "loss": 0.6768,
2361
+ "step": 3920
2362
+ },
2363
+ {
2364
+ "epoch": 0.78,
2365
+ "learning_rate": 7.763729751086527e-05,
2366
+ "loss": 0.6117,
2367
+ "step": 3930
2368
+ },
2369
+ {
2370
+ "epoch": 0.78,
2371
+ "learning_rate": 7.783484788621098e-05,
2372
+ "loss": 0.618,
2373
+ "step": 3940
2374
+ },
2375
+ {
2376
+ "epoch": 0.78,
2377
+ "learning_rate": 7.803239826155671e-05,
2378
+ "loss": 0.6749,
2379
+ "step": 3950
2380
+ },
2381
+ {
2382
+ "epoch": 0.78,
2383
+ "learning_rate": 7.822994863690242e-05,
2384
+ "loss": 0.6358,
2385
+ "step": 3960
2386
+ },
2387
+ {
2388
+ "epoch": 0.78,
2389
+ "learning_rate": 7.842749901224814e-05,
2390
+ "loss": 0.6954,
2391
+ "step": 3970
2392
+ },
2393
+ {
2394
+ "epoch": 0.79,
2395
+ "learning_rate": 7.862504938759384e-05,
2396
+ "loss": 0.5934,
2397
+ "step": 3980
2398
+ },
2399
+ {
2400
+ "epoch": 0.79,
2401
+ "learning_rate": 7.882259976293955e-05,
2402
+ "loss": 0.6526,
2403
+ "step": 3990
2404
+ },
2405
+ {
2406
+ "epoch": 0.79,
2407
+ "learning_rate": 7.902015013828527e-05,
2408
+ "loss": 0.6171,
2409
+ "step": 4000
2410
+ },
2411
+ {
2412
+ "epoch": 0.79,
2413
+ "learning_rate": 7.921770051363097e-05,
2414
+ "loss": 0.6219,
2415
+ "step": 4010
2416
+ },
2417
+ {
2418
+ "epoch": 0.79,
2419
+ "learning_rate": 7.941525088897669e-05,
2420
+ "loss": 0.6421,
2421
+ "step": 4020
2422
+ },
2423
+ {
2424
+ "epoch": 0.8,
2425
+ "learning_rate": 7.96128012643224e-05,
2426
+ "loss": 0.6531,
2427
+ "step": 4030
2428
+ },
2429
+ {
2430
+ "epoch": 0.8,
2431
+ "learning_rate": 7.981035163966812e-05,
2432
+ "loss": 0.5801,
2433
+ "step": 4040
2434
+ },
2435
+ {
2436
+ "epoch": 0.8,
2437
+ "learning_rate": 8.000790201501384e-05,
2438
+ "loss": 0.63,
2439
+ "step": 4050
2440
+ },
2441
+ {
2442
+ "epoch": 0.8,
2443
+ "learning_rate": 8.020545239035954e-05,
2444
+ "loss": 0.6303,
2445
+ "step": 4060
2446
+ },
2447
+ {
2448
+ "epoch": 0.8,
2449
+ "learning_rate": 8.040300276570526e-05,
2450
+ "loss": 0.612,
2451
+ "step": 4070
2452
+ },
2453
+ {
2454
+ "epoch": 0.81,
2455
+ "learning_rate": 8.060055314105097e-05,
2456
+ "loss": 0.6031,
2457
+ "step": 4080
2458
+ },
2459
+ {
2460
+ "epoch": 0.81,
2461
+ "learning_rate": 8.079810351639669e-05,
2462
+ "loss": 0.6569,
2463
+ "step": 4090
2464
+ },
2465
+ {
2466
+ "epoch": 0.81,
2467
+ "learning_rate": 8.09956538917424e-05,
2468
+ "loss": 0.7017,
2469
+ "step": 4100
2470
+ },
2471
+ {
2472
+ "epoch": 0.81,
2473
+ "learning_rate": 8.119320426708811e-05,
2474
+ "loss": 0.6456,
2475
+ "step": 4110
2476
+ },
2477
+ {
2478
+ "epoch": 0.81,
2479
+ "learning_rate": 8.139075464243382e-05,
2480
+ "loss": 0.6752,
2481
+ "step": 4120
2482
+ },
2483
+ {
2484
+ "epoch": 0.82,
2485
+ "learning_rate": 8.158830501777954e-05,
2486
+ "loss": 0.5988,
2487
+ "step": 4130
2488
+ },
2489
+ {
2490
+ "epoch": 0.82,
2491
+ "learning_rate": 8.178585539312526e-05,
2492
+ "loss": 0.6224,
2493
+ "step": 4140
2494
+ },
2495
+ {
2496
+ "epoch": 0.82,
2497
+ "learning_rate": 8.198340576847097e-05,
2498
+ "loss": 0.6238,
2499
+ "step": 4150
2500
+ },
2501
+ {
2502
+ "epoch": 0.82,
2503
+ "learning_rate": 8.218095614381669e-05,
2504
+ "loss": 0.6953,
2505
+ "step": 4160
2506
+ },
2507
+ {
2508
+ "epoch": 0.82,
2509
+ "learning_rate": 8.237850651916239e-05,
2510
+ "loss": 0.675,
2511
+ "step": 4170
2512
+ },
2513
+ {
2514
+ "epoch": 0.83,
2515
+ "learning_rate": 8.25760568945081e-05,
2516
+ "loss": 0.6091,
2517
+ "step": 4180
2518
+ },
2519
+ {
2520
+ "epoch": 0.83,
2521
+ "learning_rate": 8.277360726985382e-05,
2522
+ "loss": 0.6527,
2523
+ "step": 4190
2524
+ },
2525
+ {
2526
+ "epoch": 0.83,
2527
+ "learning_rate": 8.297115764519952e-05,
2528
+ "loss": 0.6183,
2529
+ "step": 4200
2530
+ },
2531
+ {
2532
+ "epoch": 0.83,
2533
+ "learning_rate": 8.316870802054524e-05,
2534
+ "loss": 0.6101,
2535
+ "step": 4210
2536
+ },
2537
+ {
2538
+ "epoch": 0.83,
2539
+ "learning_rate": 8.336625839589096e-05,
2540
+ "loss": 0.6639,
2541
+ "step": 4220
2542
+ },
2543
+ {
2544
+ "epoch": 0.84,
2545
+ "learning_rate": 8.356380877123667e-05,
2546
+ "loss": 0.6226,
2547
+ "step": 4230
2548
+ },
2549
+ {
2550
+ "epoch": 0.84,
2551
+ "learning_rate": 8.376135914658239e-05,
2552
+ "loss": 0.6441,
2553
+ "step": 4240
2554
+ },
2555
+ {
2556
+ "epoch": 0.84,
2557
+ "learning_rate": 8.39589095219281e-05,
2558
+ "loss": 0.5824,
2559
+ "step": 4250
2560
+ },
2561
+ {
2562
+ "epoch": 0.84,
2563
+ "learning_rate": 8.415645989727381e-05,
2564
+ "loss": 0.6669,
2565
+ "step": 4260
2566
+ },
2567
+ {
2568
+ "epoch": 0.84,
2569
+ "learning_rate": 8.435401027261952e-05,
2570
+ "loss": 0.6821,
2571
+ "step": 4270
2572
+ },
2573
+ {
2574
+ "epoch": 0.85,
2575
+ "learning_rate": 8.455156064796524e-05,
2576
+ "loss": 0.6243,
2577
+ "step": 4280
2578
+ },
2579
+ {
2580
+ "epoch": 0.85,
2581
+ "learning_rate": 8.474911102331094e-05,
2582
+ "loss": 0.5964,
2583
+ "step": 4290
2584
+ },
2585
+ {
2586
+ "epoch": 0.85,
2587
+ "learning_rate": 8.494666139865665e-05,
2588
+ "loss": 0.5887,
2589
+ "step": 4300
2590
+ },
2591
+ {
2592
+ "epoch": 0.85,
2593
+ "learning_rate": 8.514421177400237e-05,
2594
+ "loss": 0.6496,
2595
+ "step": 4310
2596
+ },
2597
+ {
2598
+ "epoch": 0.85,
2599
+ "learning_rate": 8.534176214934809e-05,
2600
+ "loss": 0.6563,
2601
+ "step": 4320
2602
+ },
2603
+ {
2604
+ "epoch": 0.86,
2605
+ "learning_rate": 8.553931252469381e-05,
2606
+ "loss": 0.6282,
2607
+ "step": 4330
2608
+ },
2609
+ {
2610
+ "epoch": 0.86,
2611
+ "learning_rate": 8.573686290003952e-05,
2612
+ "loss": 0.6189,
2613
+ "step": 4340
2614
+ },
2615
+ {
2616
+ "epoch": 0.86,
2617
+ "learning_rate": 8.593441327538523e-05,
2618
+ "loss": 0.6376,
2619
+ "step": 4350
2620
+ },
2621
+ {
2622
+ "epoch": 0.86,
2623
+ "learning_rate": 8.613196365073094e-05,
2624
+ "loss": 0.5814,
2625
+ "step": 4360
2626
+ },
2627
+ {
2628
+ "epoch": 0.86,
2629
+ "learning_rate": 8.632951402607665e-05,
2630
+ "loss": 0.5461,
2631
+ "step": 4370
2632
+ },
2633
+ {
2634
+ "epoch": 0.87,
2635
+ "learning_rate": 8.652706440142237e-05,
2636
+ "loss": 0.6651,
2637
+ "step": 4380
2638
+ },
2639
+ {
2640
+ "epoch": 0.87,
2641
+ "learning_rate": 8.672461477676807e-05,
2642
+ "loss": 0.6614,
2643
+ "step": 4390
2644
+ },
2645
+ {
2646
+ "epoch": 0.87,
2647
+ "learning_rate": 8.692216515211379e-05,
2648
+ "loss": 0.604,
2649
+ "step": 4400
2650
+ },
2651
+ {
2652
+ "epoch": 0.87,
2653
+ "learning_rate": 8.711971552745951e-05,
2654
+ "loss": 0.6378,
2655
+ "step": 4410
2656
+ },
2657
+ {
2658
+ "epoch": 0.87,
2659
+ "learning_rate": 8.731726590280522e-05,
2660
+ "loss": 0.6211,
2661
+ "step": 4420
2662
+ },
2663
+ {
2664
+ "epoch": 0.88,
2665
+ "learning_rate": 8.751481627815094e-05,
2666
+ "loss": 0.5943,
2667
+ "step": 4430
2668
+ },
2669
+ {
2670
+ "epoch": 0.88,
2671
+ "learning_rate": 8.771236665349664e-05,
2672
+ "loss": 0.6493,
2673
+ "step": 4440
2674
+ },
2675
+ {
2676
+ "epoch": 0.88,
2677
+ "learning_rate": 8.790991702884236e-05,
2678
+ "loss": 0.643,
2679
+ "step": 4450
2680
+ },
2681
+ {
2682
+ "epoch": 0.88,
2683
+ "learning_rate": 8.810746740418807e-05,
2684
+ "loss": 0.6311,
2685
+ "step": 4460
2686
+ },
2687
+ {
2688
+ "epoch": 0.88,
2689
+ "learning_rate": 8.830501777953379e-05,
2690
+ "loss": 0.6606,
2691
+ "step": 4470
2692
+ },
2693
+ {
2694
+ "epoch": 0.88,
2695
+ "learning_rate": 8.85025681548795e-05,
2696
+ "loss": 0.615,
2697
+ "step": 4480
2698
+ },
2699
+ {
2700
+ "epoch": 0.89,
2701
+ "learning_rate": 8.87001185302252e-05,
2702
+ "loss": 0.6847,
2703
+ "step": 4490
2704
+ },
2705
+ {
2706
+ "epoch": 0.89,
2707
+ "learning_rate": 8.889766890557093e-05,
2708
+ "loss": 0.6207,
2709
+ "step": 4500
2710
+ },
2711
+ {
2712
+ "epoch": 0.89,
2713
+ "learning_rate": 8.909521928091664e-05,
2714
+ "loss": 0.6094,
2715
+ "step": 4510
2716
+ },
2717
+ {
2718
+ "epoch": 0.89,
2719
+ "learning_rate": 8.929276965626236e-05,
2720
+ "loss": 0.6294,
2721
+ "step": 4520
2722
+ },
2723
+ {
2724
+ "epoch": 0.89,
2725
+ "learning_rate": 8.949032003160807e-05,
2726
+ "loss": 0.644,
2727
+ "step": 4530
2728
+ },
2729
+ {
2730
+ "epoch": 0.9,
2731
+ "learning_rate": 8.968787040695377e-05,
2732
+ "loss": 0.6265,
2733
+ "step": 4540
2734
+ },
2735
+ {
2736
+ "epoch": 0.9,
2737
+ "learning_rate": 8.988542078229949e-05,
2738
+ "loss": 0.6348,
2739
+ "step": 4550
2740
+ },
2741
+ {
2742
+ "epoch": 0.9,
2743
+ "learning_rate": 9.00829711576452e-05,
2744
+ "loss": 0.5913,
2745
+ "step": 4560
2746
+ },
2747
+ {
2748
+ "epoch": 0.9,
2749
+ "learning_rate": 9.028052153299092e-05,
2750
+ "loss": 0.6377,
2751
+ "step": 4570
2752
+ },
2753
+ {
2754
+ "epoch": 0.9,
2755
+ "learning_rate": 9.047807190833662e-05,
2756
+ "loss": 0.5929,
2757
+ "step": 4580
2758
+ },
2759
+ {
2760
+ "epoch": 0.91,
2761
+ "learning_rate": 9.067562228368236e-05,
2762
+ "loss": 0.6708,
2763
+ "step": 4590
2764
+ },
2765
+ {
2766
+ "epoch": 0.91,
2767
+ "learning_rate": 9.087317265902806e-05,
2768
+ "loss": 0.6314,
2769
+ "step": 4600
2770
+ },
2771
+ {
2772
+ "epoch": 0.91,
2773
+ "learning_rate": 9.107072303437377e-05,
2774
+ "loss": 0.6853,
2775
+ "step": 4610
2776
+ },
2777
+ {
2778
+ "epoch": 0.91,
2779
+ "learning_rate": 9.126827340971949e-05,
2780
+ "loss": 0.6196,
2781
+ "step": 4620
2782
+ },
2783
+ {
2784
+ "epoch": 0.91,
2785
+ "learning_rate": 9.146582378506519e-05,
2786
+ "loss": 0.5957,
2787
+ "step": 4630
2788
+ },
2789
+ {
2790
+ "epoch": 0.92,
2791
+ "learning_rate": 9.166337416041091e-05,
2792
+ "loss": 0.5991,
2793
+ "step": 4640
2794
+ },
2795
+ {
2796
+ "epoch": 0.92,
2797
+ "learning_rate": 9.186092453575662e-05,
2798
+ "loss": 0.6602,
2799
+ "step": 4650
2800
+ },
2801
+ {
2802
+ "epoch": 0.92,
2803
+ "learning_rate": 9.205847491110234e-05,
2804
+ "loss": 0.6852,
2805
+ "step": 4660
2806
+ },
2807
+ {
2808
+ "epoch": 0.92,
2809
+ "learning_rate": 9.225602528644804e-05,
2810
+ "loss": 0.5723,
2811
+ "step": 4670
2812
+ },
2813
+ {
2814
+ "epoch": 0.92,
2815
+ "learning_rate": 9.245357566179375e-05,
2816
+ "loss": 0.6218,
2817
+ "step": 4680
2818
+ },
2819
+ {
2820
+ "epoch": 0.93,
2821
+ "learning_rate": 9.265112603713948e-05,
2822
+ "loss": 0.625,
2823
+ "step": 4690
2824
+ },
2825
+ {
2826
+ "epoch": 0.93,
2827
+ "learning_rate": 9.284867641248519e-05,
2828
+ "loss": 0.6035,
2829
+ "step": 4700
2830
+ },
2831
+ {
2832
+ "epoch": 0.93,
2833
+ "learning_rate": 9.304622678783091e-05,
2834
+ "loss": 0.6119,
2835
+ "step": 4710
2836
+ },
2837
+ {
2838
+ "epoch": 0.93,
2839
+ "learning_rate": 9.324377716317662e-05,
2840
+ "loss": 0.6284,
2841
+ "step": 4720
2842
+ },
2843
+ {
2844
+ "epoch": 0.93,
2845
+ "learning_rate": 9.344132753852232e-05,
2846
+ "loss": 0.5956,
2847
+ "step": 4730
2848
+ },
2849
+ {
2850
+ "epoch": 0.94,
2851
+ "learning_rate": 9.363887791386804e-05,
2852
+ "loss": 0.5773,
2853
+ "step": 4740
2854
+ },
2855
+ {
2856
+ "epoch": 0.94,
2857
+ "learning_rate": 9.383642828921375e-05,
2858
+ "loss": 0.6735,
2859
+ "step": 4750
2860
+ },
2861
+ {
2862
+ "epoch": 0.94,
2863
+ "learning_rate": 9.403397866455947e-05,
2864
+ "loss": 0.6019,
2865
+ "step": 4760
2866
+ },
2867
+ {
2868
+ "epoch": 0.94,
2869
+ "learning_rate": 9.423152903990517e-05,
2870
+ "loss": 0.6271,
2871
+ "step": 4770
2872
+ },
2873
+ {
2874
+ "epoch": 0.94,
2875
+ "learning_rate": 9.442907941525089e-05,
2876
+ "loss": 0.6201,
2877
+ "step": 4780
2878
+ },
2879
+ {
2880
+ "epoch": 0.95,
2881
+ "learning_rate": 9.462662979059661e-05,
2882
+ "loss": 0.6904,
2883
+ "step": 4790
2884
+ },
2885
+ {
2886
+ "epoch": 0.95,
2887
+ "learning_rate": 9.482418016594232e-05,
2888
+ "loss": 0.6405,
2889
+ "step": 4800
2890
+ },
2891
+ {
2892
+ "epoch": 0.95,
2893
+ "learning_rate": 9.502173054128804e-05,
2894
+ "loss": 0.5189,
2895
+ "step": 4810
2896
+ },
2897
+ {
2898
+ "epoch": 0.95,
2899
+ "learning_rate": 9.521928091663374e-05,
2900
+ "loss": 0.623,
2901
+ "step": 4820
2902
+ },
2903
+ {
2904
+ "epoch": 0.95,
2905
+ "learning_rate": 9.541683129197946e-05,
2906
+ "loss": 0.6278,
2907
+ "step": 4830
2908
+ },
2909
+ {
2910
+ "epoch": 0.96,
2911
+ "learning_rate": 9.561438166732517e-05,
2912
+ "loss": 0.6264,
2913
+ "step": 4840
2914
+ },
2915
+ {
2916
+ "epoch": 0.96,
2917
+ "learning_rate": 9.581193204267089e-05,
2918
+ "loss": 0.6407,
2919
+ "step": 4850
2920
+ },
2921
+ {
2922
+ "epoch": 0.96,
2923
+ "learning_rate": 9.60094824180166e-05,
2924
+ "loss": 0.6766,
2925
+ "step": 4860
2926
+ },
2927
+ {
2928
+ "epoch": 0.96,
2929
+ "learning_rate": 9.620703279336231e-05,
2930
+ "loss": 0.6492,
2931
+ "step": 4870
2932
+ },
2933
+ {
2934
+ "epoch": 0.96,
2935
+ "learning_rate": 9.640458316870803e-05,
2936
+ "loss": 0.6087,
2937
+ "step": 4880
2938
+ },
2939
+ {
2940
+ "epoch": 0.97,
2941
+ "learning_rate": 9.660213354405374e-05,
2942
+ "loss": 0.5838,
2943
+ "step": 4890
2944
+ },
2945
+ {
2946
+ "epoch": 0.97,
2947
+ "learning_rate": 9.679968391939946e-05,
2948
+ "loss": 0.6389,
2949
+ "step": 4900
2950
+ },
2951
+ {
2952
+ "epoch": 0.97,
2953
+ "learning_rate": 9.699723429474516e-05,
2954
+ "loss": 0.664,
2955
+ "step": 4910
2956
+ },
2957
+ {
2958
+ "epoch": 0.97,
2959
+ "learning_rate": 9.719478467009087e-05,
2960
+ "loss": 0.6164,
2961
+ "step": 4920
2962
+ },
2963
+ {
2964
+ "epoch": 0.97,
2965
+ "learning_rate": 9.739233504543659e-05,
2966
+ "loss": 0.6112,
2967
+ "step": 4930
2968
+ },
2969
+ {
2970
+ "epoch": 0.98,
2971
+ "learning_rate": 9.75898854207823e-05,
2972
+ "loss": 0.5928,
2973
+ "step": 4940
2974
+ },
2975
+ {
2976
+ "epoch": 0.98,
2977
+ "learning_rate": 9.778743579612802e-05,
2978
+ "loss": 0.6131,
2979
+ "step": 4950
2980
+ },
2981
+ {
2982
+ "epoch": 0.98,
2983
+ "learning_rate": 9.798498617147372e-05,
2984
+ "loss": 0.6221,
2985
+ "step": 4960
2986
+ },
2987
+ {
2988
+ "epoch": 0.98,
2989
+ "learning_rate": 9.818253654681944e-05,
2990
+ "loss": 0.6302,
2991
+ "step": 4970
2992
+ },
2993
+ {
2994
+ "epoch": 0.98,
2995
+ "learning_rate": 9.838008692216516e-05,
2996
+ "loss": 0.6185,
2997
+ "step": 4980
2998
+ },
2999
+ {
3000
+ "epoch": 0.99,
3001
+ "learning_rate": 9.857763729751087e-05,
3002
+ "loss": 0.6397,
3003
+ "step": 4990
3004
+ },
3005
+ {
3006
+ "epoch": 0.99,
3007
+ "learning_rate": 9.877518767285659e-05,
3008
+ "loss": 0.6381,
3009
+ "step": 5000
3010
+ },
3011
+ {
3012
+ "epoch": 0.99,
3013
+ "learning_rate": 9.897273804820229e-05,
3014
+ "loss": 0.6274,
3015
+ "step": 5010
3016
+ },
3017
+ {
3018
+ "epoch": 0.99,
3019
+ "learning_rate": 9.917028842354801e-05,
3020
+ "loss": 0.5895,
3021
+ "step": 5020
3022
+ },
3023
+ {
3024
+ "epoch": 0.99,
3025
+ "learning_rate": 9.936783879889372e-05,
3026
+ "loss": 0.6603,
3027
+ "step": 5030
3028
+ },
3029
+ {
3030
+ "epoch": 1.0,
3031
+ "learning_rate": 9.956538917423944e-05,
3032
+ "loss": 0.5812,
3033
+ "step": 5040
3034
+ },
3035
+ {
3036
+ "epoch": 1.0,
3037
+ "learning_rate": 9.976293954958514e-05,
3038
+ "loss": 0.6215,
3039
+ "step": 5050
3040
+ },
3041
+ {
3042
+ "epoch": 1.0,
3043
+ "learning_rate": 9.996048992493086e-05,
3044
+ "loss": 0.5478,
3045
+ "step": 5060
3046
+ },
3047
+ {
3048
+ "epoch": 1.0,
3049
+ "learning_rate": 9.996048992493086e-05,
3050
+ "loss": 0.5893,
3051
+ "step": 5070
3052
+ },
3053
+ {
3054
+ "epoch": 1.0,
3055
+ "learning_rate": 9.991110233109443e-05,
3056
+ "loss": 0.6172,
3057
+ "step": 5080
3058
+ },
3059
+ {
3060
+ "epoch": 1.01,
3061
+ "learning_rate": 9.986171473725801e-05,
3062
+ "loss": 0.609,
3063
+ "step": 5090
3064
+ },
3065
+ {
3066
+ "epoch": 1.01,
3067
+ "learning_rate": 9.981232714342158e-05,
3068
+ "loss": 0.5751,
3069
+ "step": 5100
3070
+ },
3071
+ {
3072
+ "epoch": 1.01,
3073
+ "learning_rate": 9.976293954958514e-05,
3074
+ "loss": 0.6062,
3075
+ "step": 5110
3076
+ },
3077
+ {
3078
+ "epoch": 1.01,
3079
+ "learning_rate": 9.971355195574872e-05,
3080
+ "loss": 0.5886,
3081
+ "step": 5120
3082
+ },
3083
+ {
3084
+ "epoch": 1.01,
3085
+ "learning_rate": 9.966416436191229e-05,
3086
+ "loss": 0.5286,
3087
+ "step": 5130
3088
+ },
3089
+ {
3090
+ "epoch": 1.02,
3091
+ "learning_rate": 9.961477676807587e-05,
3092
+ "loss": 0.611,
3093
+ "step": 5140
3094
+ },
3095
+ {
3096
+ "epoch": 1.02,
3097
+ "learning_rate": 9.956538917423944e-05,
3098
+ "loss": 0.5796,
3099
+ "step": 5150
3100
+ },
3101
+ {
3102
+ "epoch": 1.02,
3103
+ "learning_rate": 9.9516001580403e-05,
3104
+ "loss": 0.5771,
3105
+ "step": 5160
3106
+ },
3107
+ {
3108
+ "epoch": 1.02,
3109
+ "learning_rate": 9.946661398656658e-05,
3110
+ "loss": 0.5545,
3111
+ "step": 5170
3112
+ },
3113
+ {
3114
+ "epoch": 1.02,
3115
+ "learning_rate": 9.941722639273015e-05,
3116
+ "loss": 0.5793,
3117
+ "step": 5180
3118
+ },
3119
+ {
3120
+ "epoch": 1.03,
3121
+ "learning_rate": 9.936783879889372e-05,
3122
+ "loss": 0.6067,
3123
+ "step": 5190
3124
+ },
3125
+ {
3126
+ "epoch": 1.03,
3127
+ "learning_rate": 9.93184512050573e-05,
3128
+ "loss": 0.5685,
3129
+ "step": 5200
3130
+ },
3131
+ {
3132
+ "epoch": 1.03,
3133
+ "learning_rate": 9.926906361122087e-05,
3134
+ "loss": 0.5841,
3135
+ "step": 5210
3136
+ },
3137
+ {
3138
+ "epoch": 1.03,
3139
+ "learning_rate": 9.921967601738443e-05,
3140
+ "loss": 0.5445,
3141
+ "step": 5220
3142
+ },
3143
+ {
3144
+ "epoch": 1.03,
3145
+ "learning_rate": 9.917028842354801e-05,
3146
+ "loss": 0.5709,
3147
+ "step": 5230
3148
+ },
3149
+ {
3150
+ "epoch": 1.04,
3151
+ "learning_rate": 9.912090082971158e-05,
3152
+ "loss": 0.607,
3153
+ "step": 5240
3154
+ },
3155
+ {
3156
+ "epoch": 1.04,
3157
+ "learning_rate": 9.907151323587516e-05,
3158
+ "loss": 0.5702,
3159
+ "step": 5250
3160
+ },
3161
+ {
3162
+ "epoch": 1.04,
3163
+ "learning_rate": 9.902212564203873e-05,
3164
+ "loss": 0.5739,
3165
+ "step": 5260
3166
+ },
3167
+ {
3168
+ "epoch": 1.04,
3169
+ "learning_rate": 9.897273804820229e-05,
3170
+ "loss": 0.6375,
3171
+ "step": 5270
3172
+ },
3173
+ {
3174
+ "epoch": 1.04,
3175
+ "learning_rate": 9.892335045436587e-05,
3176
+ "loss": 0.6164,
3177
+ "step": 5280
3178
+ },
3179
+ {
3180
+ "epoch": 1.04,
3181
+ "learning_rate": 9.887396286052944e-05,
3182
+ "loss": 0.6189,
3183
+ "step": 5290
3184
+ },
3185
+ {
3186
+ "epoch": 1.05,
3187
+ "learning_rate": 9.8824575266693e-05,
3188
+ "loss": 0.5755,
3189
+ "step": 5300
3190
+ },
3191
+ {
3192
+ "epoch": 1.05,
3193
+ "learning_rate": 9.877518767285659e-05,
3194
+ "loss": 0.5973,
3195
+ "step": 5310
3196
+ },
3197
+ {
3198
+ "epoch": 1.05,
3199
+ "learning_rate": 9.872580007902015e-05,
3200
+ "loss": 0.6061,
3201
+ "step": 5320
3202
+ },
3203
+ {
3204
+ "epoch": 1.05,
3205
+ "learning_rate": 9.867641248518372e-05,
3206
+ "loss": 0.5847,
3207
+ "step": 5330
3208
+ },
3209
+ {
3210
+ "epoch": 1.05,
3211
+ "learning_rate": 9.86270248913473e-05,
3212
+ "loss": 0.5368,
3213
+ "step": 5340
3214
+ },
3215
+ {
3216
+ "epoch": 1.06,
3217
+ "learning_rate": 9.857763729751087e-05,
3218
+ "loss": 0.5204,
3219
+ "step": 5350
3220
+ },
3221
+ {
3222
+ "epoch": 1.06,
3223
+ "learning_rate": 9.852824970367445e-05,
3224
+ "loss": 0.5737,
3225
+ "step": 5360
3226
+ },
3227
+ {
3228
+ "epoch": 1.06,
3229
+ "learning_rate": 9.847886210983801e-05,
3230
+ "loss": 0.5824,
3231
+ "step": 5370
3232
+ },
3233
+ {
3234
+ "epoch": 1.06,
3235
+ "learning_rate": 9.842947451600158e-05,
3236
+ "loss": 0.5648,
3237
+ "step": 5380
3238
+ },
3239
+ {
3240
+ "epoch": 1.06,
3241
+ "learning_rate": 9.838008692216516e-05,
3242
+ "loss": 0.5426,
3243
+ "step": 5390
3244
+ },
3245
+ {
3246
+ "epoch": 1.07,
3247
+ "learning_rate": 9.833069932832873e-05,
3248
+ "loss": 0.6269,
3249
+ "step": 5400
3250
+ },
3251
+ {
3252
+ "epoch": 1.07,
3253
+ "learning_rate": 9.82813117344923e-05,
3254
+ "loss": 0.5373,
3255
+ "step": 5410
3256
+ },
3257
+ {
3258
+ "epoch": 1.07,
3259
+ "learning_rate": 9.823192414065587e-05,
3260
+ "loss": 0.5895,
3261
+ "step": 5420
3262
+ },
3263
+ {
3264
+ "epoch": 1.07,
3265
+ "learning_rate": 9.818253654681944e-05,
3266
+ "loss": 0.5642,
3267
+ "step": 5430
3268
+ },
3269
+ {
3270
+ "epoch": 1.07,
3271
+ "learning_rate": 9.813314895298301e-05,
3272
+ "loss": 0.5318,
3273
+ "step": 5440
3274
+ },
3275
+ {
3276
+ "epoch": 1.08,
3277
+ "learning_rate": 9.808376135914659e-05,
3278
+ "loss": 0.6078,
3279
+ "step": 5450
3280
+ },
3281
+ {
3282
+ "epoch": 1.08,
3283
+ "learning_rate": 9.803437376531016e-05,
3284
+ "loss": 0.5631,
3285
+ "step": 5460
3286
+ },
3287
+ {
3288
+ "epoch": 1.08,
3289
+ "learning_rate": 9.798498617147372e-05,
3290
+ "loss": 0.6024,
3291
+ "step": 5470
3292
+ },
3293
+ {
3294
+ "epoch": 1.08,
3295
+ "learning_rate": 9.79355985776373e-05,
3296
+ "loss": 0.6337,
3297
+ "step": 5480
3298
+ },
3299
+ {
3300
+ "epoch": 1.08,
3301
+ "learning_rate": 9.788621098380087e-05,
3302
+ "loss": 0.5616,
3303
+ "step": 5490
3304
+ },
3305
+ {
3306
+ "epoch": 1.09,
3307
+ "learning_rate": 9.783682338996445e-05,
3308
+ "loss": 0.5906,
3309
+ "step": 5500
3310
+ },
3311
+ {
3312
+ "epoch": 1.09,
3313
+ "learning_rate": 9.778743579612802e-05,
3314
+ "loss": 0.6046,
3315
+ "step": 5510
3316
+ },
3317
+ {
3318
+ "epoch": 1.09,
3319
+ "learning_rate": 9.773804820229158e-05,
3320
+ "loss": 0.5795,
3321
+ "step": 5520
3322
+ },
3323
+ {
3324
+ "epoch": 1.09,
3325
+ "learning_rate": 9.768866060845516e-05,
3326
+ "loss": 0.6262,
3327
+ "step": 5530
3328
+ },
3329
+ {
3330
+ "epoch": 1.09,
3331
+ "learning_rate": 9.763927301461874e-05,
3332
+ "loss": 0.6531,
3333
+ "step": 5540
3334
+ },
3335
+ {
3336
+ "epoch": 1.1,
3337
+ "learning_rate": 9.75898854207823e-05,
3338
+ "loss": 0.6423,
3339
+ "step": 5550
3340
+ },
3341
+ {
3342
+ "epoch": 1.1,
3343
+ "learning_rate": 9.754049782694588e-05,
3344
+ "loss": 0.6074,
3345
+ "step": 5560
3346
+ },
3347
+ {
3348
+ "epoch": 1.1,
3349
+ "learning_rate": 9.749111023310946e-05,
3350
+ "loss": 0.5798,
3351
+ "step": 5570
3352
+ },
3353
+ {
3354
+ "epoch": 1.1,
3355
+ "learning_rate": 9.744172263927301e-05,
3356
+ "loss": 0.5806,
3357
+ "step": 5580
3358
+ },
3359
+ {
3360
+ "epoch": 1.1,
3361
+ "learning_rate": 9.739233504543659e-05,
3362
+ "loss": 0.5655,
3363
+ "step": 5590
3364
+ },
3365
+ {
3366
+ "epoch": 1.11,
3367
+ "learning_rate": 9.734294745160016e-05,
3368
+ "loss": 0.523,
3369
+ "step": 5600
3370
+ },
3371
+ {
3372
+ "epoch": 1.11,
3373
+ "learning_rate": 9.729355985776374e-05,
3374
+ "loss": 0.566,
3375
+ "step": 5610
3376
+ },
3377
+ {
3378
+ "epoch": 1.11,
3379
+ "learning_rate": 9.72441722639273e-05,
3380
+ "loss": 0.5482,
3381
+ "step": 5620
3382
+ },
3383
+ {
3384
+ "epoch": 1.11,
3385
+ "learning_rate": 9.719478467009087e-05,
3386
+ "loss": 0.5899,
3387
+ "step": 5630
3388
+ },
3389
+ {
3390
+ "epoch": 1.11,
3391
+ "learning_rate": 9.714539707625445e-05,
3392
+ "loss": 0.5508,
3393
+ "step": 5640
3394
+ },
3395
+ {
3396
+ "epoch": 1.12,
3397
+ "learning_rate": 9.709600948241803e-05,
3398
+ "loss": 0.5866,
3399
+ "step": 5650
3400
+ },
3401
+ {
3402
+ "epoch": 1.12,
3403
+ "learning_rate": 9.704662188858158e-05,
3404
+ "loss": 0.5878,
3405
+ "step": 5660
3406
+ },
3407
+ {
3408
+ "epoch": 1.12,
3409
+ "learning_rate": 9.699723429474516e-05,
3410
+ "loss": 0.5563,
3411
+ "step": 5670
3412
+ },
3413
+ {
3414
+ "epoch": 1.12,
3415
+ "learning_rate": 9.694784670090875e-05,
3416
+ "loss": 0.6075,
3417
+ "step": 5680
3418
+ },
3419
+ {
3420
+ "epoch": 1.12,
3421
+ "learning_rate": 9.68984591070723e-05,
3422
+ "loss": 0.6032,
3423
+ "step": 5690
3424
+ },
3425
+ {
3426
+ "epoch": 1.13,
3427
+ "learning_rate": 9.684907151323588e-05,
3428
+ "loss": 0.5908,
3429
+ "step": 5700
3430
+ },
3431
+ {
3432
+ "epoch": 1.13,
3433
+ "learning_rate": 9.679968391939946e-05,
3434
+ "loss": 0.5863,
3435
+ "step": 5710
3436
+ },
3437
+ {
3438
+ "epoch": 1.13,
3439
+ "learning_rate": 9.675029632556303e-05,
3440
+ "loss": 0.5858,
3441
+ "step": 5720
3442
+ },
3443
+ {
3444
+ "epoch": 1.13,
3445
+ "learning_rate": 9.670090873172659e-05,
3446
+ "loss": 0.5929,
3447
+ "step": 5730
3448
+ },
3449
+ {
3450
+ "epoch": 1.13,
3451
+ "learning_rate": 9.665152113789017e-05,
3452
+ "loss": 0.5891,
3453
+ "step": 5740
3454
+ },
3455
+ {
3456
+ "epoch": 1.14,
3457
+ "learning_rate": 9.660213354405374e-05,
3458
+ "loss": 0.618,
3459
+ "step": 5750
3460
+ },
3461
+ {
3462
+ "epoch": 1.14,
3463
+ "learning_rate": 9.65527459502173e-05,
3464
+ "loss": 0.6104,
3465
+ "step": 5760
3466
+ },
3467
+ {
3468
+ "epoch": 1.14,
3469
+ "learning_rate": 9.650335835638089e-05,
3470
+ "loss": 0.588,
3471
+ "step": 5770
3472
+ },
3473
+ {
3474
+ "epoch": 1.14,
3475
+ "learning_rate": 9.645397076254445e-05,
3476
+ "loss": 0.6054,
3477
+ "step": 5780
3478
+ },
3479
+ {
3480
+ "epoch": 1.14,
3481
+ "learning_rate": 9.640458316870803e-05,
3482
+ "loss": 0.5846,
3483
+ "step": 5790
3484
+ },
3485
+ {
3486
+ "epoch": 1.15,
3487
+ "learning_rate": 9.635519557487159e-05,
3488
+ "loss": 0.6161,
3489
+ "step": 5800
3490
+ },
3491
+ {
3492
+ "epoch": 1.15,
3493
+ "learning_rate": 9.630580798103517e-05,
3494
+ "loss": 0.5995,
3495
+ "step": 5810
3496
+ },
3497
+ {
3498
+ "epoch": 1.15,
3499
+ "learning_rate": 9.625642038719875e-05,
3500
+ "loss": 0.5678,
3501
+ "step": 5820
3502
+ },
3503
+ {
3504
+ "epoch": 1.15,
3505
+ "learning_rate": 9.620703279336231e-05,
3506
+ "loss": 0.6255,
3507
+ "step": 5830
3508
+ },
3509
+ {
3510
+ "epoch": 1.15,
3511
+ "learning_rate": 9.615764519952588e-05,
3512
+ "loss": 0.5903,
3513
+ "step": 5840
3514
+ },
3515
+ {
3516
+ "epoch": 1.16,
3517
+ "learning_rate": 9.610825760568946e-05,
3518
+ "loss": 0.5584,
3519
+ "step": 5850
3520
+ },
3521
+ {
3522
+ "epoch": 1.16,
3523
+ "learning_rate": 9.605887001185303e-05,
3524
+ "loss": 0.5297,
3525
+ "step": 5860
3526
+ },
3527
+ {
3528
+ "epoch": 1.16,
3529
+ "learning_rate": 9.60094824180166e-05,
3530
+ "loss": 0.5777,
3531
+ "step": 5870
3532
+ },
3533
+ {
3534
+ "epoch": 1.16,
3535
+ "learning_rate": 9.596009482418017e-05,
3536
+ "loss": 0.6187,
3537
+ "step": 5880
3538
+ },
3539
+ {
3540
+ "epoch": 1.16,
3541
+ "learning_rate": 9.591070723034374e-05,
3542
+ "loss": 0.5935,
3543
+ "step": 5890
3544
+ },
3545
+ {
3546
+ "epoch": 1.17,
3547
+ "learning_rate": 9.586131963650732e-05,
3548
+ "loss": 0.5897,
3549
+ "step": 5900
3550
+ },
3551
+ {
3552
+ "epoch": 1.17,
3553
+ "learning_rate": 9.581193204267089e-05,
3554
+ "loss": 0.5581,
3555
+ "step": 5910
3556
+ },
3557
+ {
3558
+ "epoch": 1.17,
3559
+ "learning_rate": 9.576254444883445e-05,
3560
+ "loss": 0.6178,
3561
+ "step": 5920
3562
+ },
3563
+ {
3564
+ "epoch": 1.17,
3565
+ "learning_rate": 9.571315685499804e-05,
3566
+ "loss": 0.5879,
3567
+ "step": 5930
3568
+ },
3569
+ {
3570
+ "epoch": 1.17,
3571
+ "learning_rate": 9.56637692611616e-05,
3572
+ "loss": 0.6004,
3573
+ "step": 5940
3574
+ },
3575
+ {
3576
+ "epoch": 1.18,
3577
+ "learning_rate": 9.561438166732517e-05,
3578
+ "loss": 0.5726,
3579
+ "step": 5950
3580
+ },
3581
+ {
3582
+ "epoch": 1.18,
3583
+ "learning_rate": 9.556499407348875e-05,
3584
+ "loss": 0.5794,
3585
+ "step": 5960
3586
+ },
3587
+ {
3588
+ "epoch": 1.18,
3589
+ "learning_rate": 9.551560647965232e-05,
3590
+ "loss": 0.586,
3591
+ "step": 5970
3592
+ },
3593
+ {
3594
+ "epoch": 1.18,
3595
+ "learning_rate": 9.546621888581588e-05,
3596
+ "loss": 0.5863,
3597
+ "step": 5980
3598
+ },
3599
+ {
3600
+ "epoch": 1.18,
3601
+ "learning_rate": 9.541683129197946e-05,
3602
+ "loss": 0.6062,
3603
+ "step": 5990
3604
+ },
3605
+ {
3606
+ "epoch": 1.19,
3607
+ "learning_rate": 9.536744369814303e-05,
3608
+ "loss": 0.6103,
3609
+ "step": 6000
3610
+ },
3611
+ {
3612
+ "epoch": 1.19,
3613
+ "learning_rate": 9.531805610430661e-05,
3614
+ "loss": 0.6179,
3615
+ "step": 6010
3616
+ },
3617
+ {
3618
+ "epoch": 1.19,
3619
+ "learning_rate": 9.526866851047018e-05,
3620
+ "loss": 0.5564,
3621
+ "step": 6020
3622
+ },
3623
+ {
3624
+ "epoch": 1.19,
3625
+ "learning_rate": 9.521928091663374e-05,
3626
+ "loss": 0.5664,
3627
+ "step": 6030
3628
+ },
3629
+ {
3630
+ "epoch": 1.19,
3631
+ "learning_rate": 9.516989332279732e-05,
3632
+ "loss": 0.5681,
3633
+ "step": 6040
3634
+ },
3635
+ {
3636
+ "epoch": 1.2,
3637
+ "learning_rate": 9.512050572896089e-05,
3638
+ "loss": 0.5789,
3639
+ "step": 6050
3640
+ },
3641
+ {
3642
+ "epoch": 1.2,
3643
+ "learning_rate": 9.507111813512446e-05,
3644
+ "loss": 0.59,
3645
+ "step": 6060
3646
+ },
3647
+ {
3648
+ "epoch": 1.2,
3649
+ "learning_rate": 9.502173054128804e-05,
3650
+ "loss": 0.5384,
3651
+ "step": 6070
3652
+ },
3653
+ {
3654
+ "epoch": 1.2,
3655
+ "learning_rate": 9.49723429474516e-05,
3656
+ "loss": 0.5677,
3657
+ "step": 6080
3658
+ },
3659
+ {
3660
+ "epoch": 1.2,
3661
+ "learning_rate": 9.492295535361517e-05,
3662
+ "loss": 0.6119,
3663
+ "step": 6090
3664
+ },
3665
+ {
3666
+ "epoch": 1.2,
3667
+ "learning_rate": 9.487356775977875e-05,
3668
+ "loss": 0.5626,
3669
+ "step": 6100
3670
+ },
3671
+ {
3672
+ "epoch": 1.21,
3673
+ "learning_rate": 9.482418016594232e-05,
3674
+ "loss": 0.5925,
3675
+ "step": 6110
3676
+ },
3677
+ {
3678
+ "epoch": 1.21,
3679
+ "learning_rate": 9.477479257210588e-05,
3680
+ "loss": 0.5993,
3681
+ "step": 6120
3682
+ },
3683
+ {
3684
+ "epoch": 1.21,
3685
+ "learning_rate": 9.472540497826946e-05,
3686
+ "loss": 0.5264,
3687
+ "step": 6130
3688
+ },
3689
+ {
3690
+ "epoch": 1.21,
3691
+ "learning_rate": 9.467601738443303e-05,
3692
+ "loss": 0.5545,
3693
+ "step": 6140
3694
+ },
3695
+ {
3696
+ "epoch": 1.21,
3697
+ "learning_rate": 9.462662979059661e-05,
3698
+ "loss": 0.5822,
3699
+ "step": 6150
3700
+ },
3701
+ {
3702
+ "epoch": 1.22,
3703
+ "learning_rate": 9.457724219676018e-05,
3704
+ "loss": 0.5915,
3705
+ "step": 6160
3706
+ },
3707
+ {
3708
+ "epoch": 1.22,
3709
+ "learning_rate": 9.452785460292374e-05,
3710
+ "loss": 0.5689,
3711
+ "step": 6170
3712
+ },
3713
+ {
3714
+ "epoch": 1.22,
3715
+ "learning_rate": 9.447846700908733e-05,
3716
+ "loss": 0.5721,
3717
+ "step": 6180
3718
+ },
3719
+ {
3720
+ "epoch": 1.22,
3721
+ "learning_rate": 9.442907941525089e-05,
3722
+ "loss": 0.5692,
3723
+ "step": 6190
3724
+ },
3725
+ {
3726
+ "epoch": 1.22,
3727
+ "learning_rate": 9.437969182141446e-05,
3728
+ "loss": 0.5578,
3729
+ "step": 6200
3730
+ },
3731
+ {
3732
+ "epoch": 1.23,
3733
+ "learning_rate": 9.433030422757804e-05,
3734
+ "loss": 0.5678,
3735
+ "step": 6210
3736
+ },
3737
+ {
3738
+ "epoch": 1.23,
3739
+ "learning_rate": 9.42809166337416e-05,
3740
+ "loss": 0.5738,
3741
+ "step": 6220
3742
+ },
3743
+ {
3744
+ "epoch": 1.23,
3745
+ "learning_rate": 9.423152903990517e-05,
3746
+ "loss": 0.5563,
3747
+ "step": 6230
3748
+ },
3749
+ {
3750
+ "epoch": 1.23,
3751
+ "learning_rate": 9.418214144606875e-05,
3752
+ "loss": 0.5703,
3753
+ "step": 6240
3754
+ },
3755
+ {
3756
+ "epoch": 1.23,
3757
+ "learning_rate": 9.413275385223232e-05,
3758
+ "loss": 0.5612,
3759
+ "step": 6250
3760
+ },
3761
+ {
3762
+ "epoch": 1.24,
3763
+ "learning_rate": 9.40833662583959e-05,
3764
+ "loss": 0.5578,
3765
+ "step": 6260
3766
+ },
3767
+ {
3768
+ "epoch": 1.24,
3769
+ "learning_rate": 9.403397866455947e-05,
3770
+ "loss": 0.609,
3771
+ "step": 6270
3772
+ },
3773
+ {
3774
+ "epoch": 1.24,
3775
+ "learning_rate": 9.398459107072303e-05,
3776
+ "loss": 0.5784,
3777
+ "step": 6280
3778
+ },
3779
+ {
3780
+ "epoch": 1.24,
3781
+ "learning_rate": 9.393520347688661e-05,
3782
+ "loss": 0.5322,
3783
+ "step": 6290
3784
+ },
3785
+ {
3786
+ "epoch": 1.24,
3787
+ "learning_rate": 9.38858158830502e-05,
3788
+ "loss": 0.5788,
3789
+ "step": 6300
3790
+ },
3791
+ {
3792
+ "epoch": 1.25,
3793
+ "learning_rate": 9.383642828921375e-05,
3794
+ "loss": 0.5519,
3795
+ "step": 6310
3796
+ },
3797
+ {
3798
+ "epoch": 1.25,
3799
+ "learning_rate": 9.378704069537733e-05,
3800
+ "loss": 0.5735,
3801
+ "step": 6320
3802
+ },
3803
+ {
3804
+ "epoch": 1.25,
3805
+ "learning_rate": 9.373765310154091e-05,
3806
+ "loss": 0.5542,
3807
+ "step": 6330
3808
+ },
3809
+ {
3810
+ "epoch": 1.25,
3811
+ "learning_rate": 9.368826550770446e-05,
3812
+ "loss": 0.5588,
3813
+ "step": 6340
3814
+ },
3815
+ {
3816
+ "epoch": 1.25,
3817
+ "learning_rate": 9.363887791386804e-05,
3818
+ "loss": 0.6296,
3819
+ "step": 6350
3820
+ },
3821
+ {
3822
+ "epoch": 1.26,
3823
+ "learning_rate": 9.358949032003162e-05,
3824
+ "loss": 0.592,
3825
+ "step": 6360
3826
+ },
3827
+ {
3828
+ "epoch": 1.26,
3829
+ "learning_rate": 9.354010272619519e-05,
3830
+ "loss": 0.54,
3831
+ "step": 6370
3832
+ },
3833
+ {
3834
+ "epoch": 1.26,
3835
+ "learning_rate": 9.349071513235875e-05,
3836
+ "loss": 0.5527,
3837
+ "step": 6380
3838
+ },
3839
+ {
3840
+ "epoch": 1.26,
3841
+ "learning_rate": 9.344132753852232e-05,
3842
+ "loss": 0.5999,
3843
+ "step": 6390
3844
+ },
3845
+ {
3846
+ "epoch": 1.26,
3847
+ "learning_rate": 9.33919399446859e-05,
3848
+ "loss": 0.5323,
3849
+ "step": 6400
3850
+ },
3851
+ {
3852
+ "epoch": 1.27,
3853
+ "learning_rate": 9.334255235084947e-05,
3854
+ "loss": 0.5933,
3855
+ "step": 6410
3856
+ },
3857
+ {
3858
+ "epoch": 1.27,
3859
+ "learning_rate": 9.329316475701303e-05,
3860
+ "loss": 0.5564,
3861
+ "step": 6420
3862
+ },
3863
+ {
3864
+ "epoch": 1.27,
3865
+ "learning_rate": 9.324377716317662e-05,
3866
+ "loss": 0.6013,
3867
+ "step": 6430
3868
+ },
3869
+ {
3870
+ "epoch": 1.27,
3871
+ "learning_rate": 9.31943895693402e-05,
3872
+ "loss": 0.5646,
3873
+ "step": 6440
3874
+ },
3875
+ {
3876
+ "epoch": 1.27,
3877
+ "learning_rate": 9.314500197550375e-05,
3878
+ "loss": 0.5512,
3879
+ "step": 6450
3880
+ },
3881
+ {
3882
+ "epoch": 1.28,
3883
+ "learning_rate": 9.309561438166733e-05,
3884
+ "loss": 0.5529,
3885
+ "step": 6460
3886
+ },
3887
+ {
3888
+ "epoch": 1.28,
3889
+ "learning_rate": 9.304622678783091e-05,
3890
+ "loss": 0.6227,
3891
+ "step": 6470
3892
+ },
3893
+ {
3894
+ "epoch": 1.28,
3895
+ "learning_rate": 9.299683919399448e-05,
3896
+ "loss": 0.5761,
3897
+ "step": 6480
3898
+ },
3899
+ {
3900
+ "epoch": 1.28,
3901
+ "learning_rate": 9.294745160015804e-05,
3902
+ "loss": 0.5036,
3903
+ "step": 6490
3904
+ },
3905
+ {
3906
+ "epoch": 1.28,
3907
+ "learning_rate": 9.289806400632162e-05,
3908
+ "loss": 0.5148,
3909
+ "step": 6500
3910
+ },
3911
+ {
3912
+ "epoch": 1.29,
3913
+ "learning_rate": 9.284867641248519e-05,
3914
+ "loss": 0.5503,
3915
+ "step": 6510
3916
+ },
3917
+ {
3918
+ "epoch": 1.29,
3919
+ "learning_rate": 9.279928881864876e-05,
3920
+ "loss": 0.5882,
3921
+ "step": 6520
3922
+ },
3923
+ {
3924
+ "epoch": 1.29,
3925
+ "learning_rate": 9.274990122481234e-05,
3926
+ "loss": 0.5803,
3927
+ "step": 6530
3928
+ },
3929
+ {
3930
+ "epoch": 1.29,
3931
+ "learning_rate": 9.27005136309759e-05,
3932
+ "loss": 0.5741,
3933
+ "step": 6540
3934
+ },
3935
+ {
3936
+ "epoch": 1.29,
3937
+ "learning_rate": 9.265112603713948e-05,
3938
+ "loss": 0.6181,
3939
+ "step": 6550
3940
+ },
3941
+ {
3942
+ "epoch": 1.3,
3943
+ "learning_rate": 9.260173844330304e-05,
3944
+ "loss": 0.594,
3945
+ "step": 6560
3946
+ },
3947
+ {
3948
+ "epoch": 1.3,
3949
+ "learning_rate": 9.255235084946662e-05,
3950
+ "loss": 0.5033,
3951
+ "step": 6570
3952
+ },
3953
+ {
3954
+ "epoch": 1.3,
3955
+ "learning_rate": 9.25029632556302e-05,
3956
+ "loss": 0.6282,
3957
+ "step": 6580
3958
+ },
3959
+ {
3960
+ "epoch": 1.3,
3961
+ "learning_rate": 9.245357566179375e-05,
3962
+ "loss": 0.5376,
3963
+ "step": 6590
3964
+ },
3965
+ {
3966
+ "epoch": 1.3,
3967
+ "learning_rate": 9.240418806795733e-05,
3968
+ "loss": 0.5705,
3969
+ "step": 6600
3970
+ },
3971
+ {
3972
+ "epoch": 1.31,
3973
+ "learning_rate": 9.235480047412091e-05,
3974
+ "loss": 0.5539,
3975
+ "step": 6610
3976
+ },
3977
+ {
3978
+ "epoch": 1.31,
3979
+ "learning_rate": 9.230541288028448e-05,
3980
+ "loss": 0.5966,
3981
+ "step": 6620
3982
+ },
3983
+ {
3984
+ "epoch": 1.31,
3985
+ "learning_rate": 9.225602528644804e-05,
3986
+ "loss": 0.5605,
3987
+ "step": 6630
3988
+ },
3989
+ {
3990
+ "epoch": 1.31,
3991
+ "learning_rate": 9.220663769261162e-05,
3992
+ "loss": 0.6484,
3993
+ "step": 6640
3994
+ },
3995
+ {
3996
+ "epoch": 1.31,
3997
+ "learning_rate": 9.215725009877519e-05,
3998
+ "loss": 0.5461,
3999
+ "step": 6650
4000
+ },
4001
+ {
4002
+ "epoch": 1.32,
4003
+ "learning_rate": 9.210786250493877e-05,
4004
+ "loss": 0.6082,
4005
+ "step": 6660
4006
+ },
4007
+ {
4008
+ "epoch": 1.32,
4009
+ "learning_rate": 9.205847491110234e-05,
4010
+ "loss": 0.5517,
4011
+ "step": 6670
4012
+ },
4013
+ {
4014
+ "epoch": 1.32,
4015
+ "learning_rate": 9.20090873172659e-05,
4016
+ "loss": 0.5512,
4017
+ "step": 6680
4018
+ },
4019
+ {
4020
+ "epoch": 1.32,
4021
+ "learning_rate": 9.195969972342949e-05,
4022
+ "loss": 0.4835,
4023
+ "step": 6690
4024
+ },
4025
+ {
4026
+ "epoch": 1.32,
4027
+ "learning_rate": 9.191031212959305e-05,
4028
+ "loss": 0.5864,
4029
+ "step": 6700
4030
+ },
4031
+ {
4032
+ "epoch": 1.33,
4033
+ "learning_rate": 9.186092453575662e-05,
4034
+ "loss": 0.5513,
4035
+ "step": 6710
4036
+ },
4037
+ {
4038
+ "epoch": 1.33,
4039
+ "learning_rate": 9.18115369419202e-05,
4040
+ "loss": 0.5527,
4041
+ "step": 6720
4042
+ },
4043
+ {
4044
+ "epoch": 1.33,
4045
+ "learning_rate": 9.176214934808377e-05,
4046
+ "loss": 0.5251,
4047
+ "step": 6730
4048
+ },
4049
+ {
4050
+ "epoch": 1.33,
4051
+ "learning_rate": 9.171276175424733e-05,
4052
+ "loss": 0.5467,
4053
+ "step": 6740
4054
+ },
4055
+ {
4056
+ "epoch": 1.33,
4057
+ "learning_rate": 9.166337416041091e-05,
4058
+ "loss": 0.5539,
4059
+ "step": 6750
4060
+ },
4061
+ {
4062
+ "epoch": 1.34,
4063
+ "learning_rate": 9.161398656657448e-05,
4064
+ "loss": 0.5269,
4065
+ "step": 6760
4066
+ },
4067
+ {
4068
+ "epoch": 1.34,
4069
+ "learning_rate": 9.156459897273805e-05,
4070
+ "loss": 0.5198,
4071
+ "step": 6770
4072
+ },
4073
+ {
4074
+ "epoch": 1.34,
4075
+ "learning_rate": 9.151521137890163e-05,
4076
+ "loss": 0.5678,
4077
+ "step": 6780
4078
+ },
4079
+ {
4080
+ "epoch": 1.34,
4081
+ "learning_rate": 9.146582378506519e-05,
4082
+ "loss": 0.6333,
4083
+ "step": 6790
4084
+ },
4085
+ {
4086
+ "epoch": 1.34,
4087
+ "learning_rate": 9.141643619122877e-05,
4088
+ "loss": 0.5445,
4089
+ "step": 6800
4090
+ },
4091
+ {
4092
+ "epoch": 1.35,
4093
+ "learning_rate": 9.136704859739234e-05,
4094
+ "loss": 0.5594,
4095
+ "step": 6810
4096
+ },
4097
+ {
4098
+ "epoch": 1.35,
4099
+ "learning_rate": 9.131766100355591e-05,
4100
+ "loss": 0.5557,
4101
+ "step": 6820
4102
+ },
4103
+ {
4104
+ "epoch": 1.35,
4105
+ "learning_rate": 9.126827340971949e-05,
4106
+ "loss": 0.5797,
4107
+ "step": 6830
4108
+ },
4109
+ {
4110
+ "epoch": 1.35,
4111
+ "learning_rate": 9.121888581588305e-05,
4112
+ "loss": 0.5981,
4113
+ "step": 6840
4114
+ },
4115
+ {
4116
+ "epoch": 1.35,
4117
+ "learning_rate": 9.116949822204662e-05,
4118
+ "loss": 0.535,
4119
+ "step": 6850
4120
+ },
4121
+ {
4122
+ "epoch": 1.35,
4123
+ "learning_rate": 9.11201106282102e-05,
4124
+ "loss": 0.5275,
4125
+ "step": 6860
4126
+ },
4127
+ {
4128
+ "epoch": 1.36,
4129
+ "learning_rate": 9.107072303437377e-05,
4130
+ "loss": 0.5531,
4131
+ "step": 6870
4132
+ },
4133
+ {
4134
+ "epoch": 1.36,
4135
+ "learning_rate": 9.102133544053733e-05,
4136
+ "loss": 0.5669,
4137
+ "step": 6880
4138
+ },
4139
+ {
4140
+ "epoch": 1.36,
4141
+ "learning_rate": 9.097194784670091e-05,
4142
+ "loss": 0.5635,
4143
+ "step": 6890
4144
+ },
4145
+ {
4146
+ "epoch": 1.36,
4147
+ "learning_rate": 9.092256025286448e-05,
4148
+ "loss": 0.5911,
4149
+ "step": 6900
4150
+ },
4151
+ {
4152
+ "epoch": 1.36,
4153
+ "learning_rate": 9.087317265902806e-05,
4154
+ "loss": 0.5518,
4155
+ "step": 6910
4156
+ },
4157
+ {
4158
+ "epoch": 1.37,
4159
+ "learning_rate": 9.082378506519163e-05,
4160
+ "loss": 0.6044,
4161
+ "step": 6920
4162
+ },
4163
+ {
4164
+ "epoch": 1.37,
4165
+ "learning_rate": 9.07743974713552e-05,
4166
+ "loss": 0.5872,
4167
+ "step": 6930
4168
+ },
4169
+ {
4170
+ "epoch": 1.37,
4171
+ "learning_rate": 9.072500987751878e-05,
4172
+ "loss": 0.5214,
4173
+ "step": 6940
4174
+ },
4175
+ {
4176
+ "epoch": 1.37,
4177
+ "learning_rate": 9.067562228368236e-05,
4178
+ "loss": 0.5623,
4179
+ "step": 6950
4180
+ },
4181
+ {
4182
+ "epoch": 1.37,
4183
+ "learning_rate": 9.062623468984591e-05,
4184
+ "loss": 0.5604,
4185
+ "step": 6960
4186
+ },
4187
+ {
4188
+ "epoch": 1.38,
4189
+ "learning_rate": 9.057684709600949e-05,
4190
+ "loss": 0.5896,
4191
+ "step": 6970
4192
+ },
4193
+ {
4194
+ "epoch": 1.38,
4195
+ "learning_rate": 9.052745950217306e-05,
4196
+ "loss": 0.5493,
4197
+ "step": 6980
4198
+ },
4199
+ {
4200
+ "epoch": 1.38,
4201
+ "learning_rate": 9.047807190833662e-05,
4202
+ "loss": 0.4864,
4203
+ "step": 6990
4204
+ },
4205
+ {
4206
+ "epoch": 1.38,
4207
+ "learning_rate": 9.04286843145002e-05,
4208
+ "loss": 0.5621,
4209
+ "step": 7000
4210
+ }
4211
+ ],
4212
+ "logging_steps": 10,
4213
+ "max_steps": 25310,
4214
+ "num_train_epochs": 5,
4215
+ "save_steps": 7000,
4216
+ "total_flos": 1.8202371090127258e+17,
4217
+ "trial_name": null,
4218
+ "trial_params": null
4219
+ }
checkpoint-7000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6b861ebe1aabc6404f5654d57dd93aa3f72512ce343953af0f357d2575b8b9d4
3
+ size 4027
checkpoint-7000/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
config.json ADDED
@@ -0,0 +1,41 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "chronopt-research/vietnamese-gpt2-medium",
3
+ "activation_function": "gelu_new",
4
+ "architectures": [
5
+ "GPT2LMHeadModel"
6
+ ],
7
+ "attn_pdrop": 0.0,
8
+ "bos_token_id": 50256,
9
+ "embd_pdrop": 0.0,
10
+ "eos_token_id": 50256,
11
+ "initializer_range": 0.02,
12
+ "layer_norm_epsilon": 1e-05,
13
+ "model_type": "gpt2",
14
+ "n_ctx": 1024,
15
+ "n_embd": 1024,
16
+ "n_head": 16,
17
+ "n_inner": null,
18
+ "n_layer": 24,
19
+ "n_positions": 1024,
20
+ "n_special": 0,
21
+ "predict_special_tokens": true,
22
+ "reorder_and_upcast_attn": false,
23
+ "resid_pdrop": 0.0,
24
+ "scale_attn_by_inverse_layer_idx": false,
25
+ "scale_attn_weights": true,
26
+ "summary_activation": null,
27
+ "summary_first_dropout": 0.1,
28
+ "summary_proj_to_labels": true,
29
+ "summary_type": "cls_index",
30
+ "summary_use_proj": true,
31
+ "task_specific_params": {
32
+ "text-generation": {
33
+ "do_sample": true,
34
+ "max_length": 50
35
+ }
36
+ },
37
+ "torch_dtype": "float32",
38
+ "transformers_version": "4.33.0.dev0",
39
+ "use_cache": true,
40
+ "vocab_size": 50262
41
+ }
eval_results.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 5.0,
3
+ "eval_accuracy": 0.8983410461950713,
4
+ "eval_loss": 0.5218655467033386,
5
+ "eval_runtime": 103.6013,
6
+ "eval_samples": 8529,
7
+ "eval_samples_per_second": 82.325,
8
+ "eval_steps_per_second": 10.299,
9
+ "perplexity": 1.6851684796072373
10
+ }
generation_config.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 50256,
4
+ "eos_token_id": 50256,
5
+ "transformers_version": "4.33.0.dev0"
6
+ }
merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6277a304937564b3a561092253461b7c2617a66bae9aff83780608d1bf594622
3
+ size 1419408349