Hugging Face's logo

OptimalScale
/
gpt2-inst-tuning

Text Generation
Transformers
PyTorch
gpt2
Generated from Trainer
text-generation-inference
Inference Endpoints
Model card Files Community
diao commited on
Commit
71d904e
•
1 Parent(s): ecff4d7

Upload 13 files

Browse files
Files changed (13) hide show
  1. README.md +55 -1
  2. all_results.json +8 -0
  3. config.json +39 -0
  4. generation_config.json +6 -0
  5. merges.txt +0 -0
  6. pytorch_model.bin +3 -0
  7. special_tokens_map.json +5 -0
  8. tokenizer.json +0 -0
  9. tokenizer_config.json +9 -0
  10. train_results.json +8 -0
  11. trainer_state.json +4495 -0
  12. training_args.bin +3 -0
  13. vocab.json +0 -0
README.md CHANGED
@@ -1,3 +1,57 @@
1
  ---
2
- license: apache-2.0
 
 
 
 
 
 
 
3
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
+ license: mit
3
+ tags:
4
+ - generated_from_trainer
5
+ datasets:
6
+ - lmflow_instruction
7
+ model-index:
8
+ - name: 044_inst-tuning_model-gpt_num-epoch-5_init-lr-2e-5_bf-16_blocksize768
9
+ results: []
10
  ---
11
+
12
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
13
+ should probably proofread and complete it, then remove this comment. -->
14
+
15
+ # 044_inst-tuning_model-gpt_num-epoch-5_init-lr-2e-5_bf-16_blocksize768
16
+
17
+ This model is a fine-tuned version of [gpt2](https://huggingface.co/gpt2) on the lmflow_instruction dataset.
18
+
19
+ ## Model description
20
+
21
+ More information needed
22
+
23
+ ## Intended uses & limitations
24
+
25
+ More information needed
26
+
27
+ ## Training and evaluation data
28
+
29
+ More information needed
30
+
31
+ ## Training procedure
32
+
33
+ ### Training hyperparameters
34
+
35
+ The following hyperparameters were used during training:
36
+ - learning_rate: 2e-05
37
+ - train_batch_size: 64
38
+ - eval_batch_size: 1
39
+ - seed: 42
40
+ - distributed_type: multi-GPU
41
+ - num_devices: 8
42
+ - total_train_batch_size: 512
43
+ - total_eval_batch_size: 8
44
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
45
+ - lr_scheduler_type: linear
46
+ - num_epochs: 5.0
47
+
48
+ ### Training results
49
+
50
+
51
+
52
+ ### Framework versions
53
+
54
+ - Transformers 4.27.0.dev0
55
+ - Pytorch 2.0.0+cu117
56
+ - Datasets 2.10.1
57
+ - Tokenizers 0.13.2
all_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 5.0,
3
+ "train_loss": 2.1389746749161076,
4
+ "train_runtime": 581.9582,
5
+ "train_samples": 76238,
6
+ "train_samples_per_second": 655.013,
7
+ "train_steps_per_second": 1.28
8
+ }
config.json ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "gpt2",
3
+ "activation_function": "gelu_new",
4
+ "architectures": [
5
+ "GPT2LMHeadModel"
6
+ ],
7
+ "attn_pdrop": 0.1,
8
+ "bos_token_id": 50256,
9
+ "embd_pdrop": 0.1,
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": 768,
16
+ "n_head": 12,
17
+ "n_inner": null,
18
+ "n_layer": 12,
19
+ "n_positions": 1024,
20
+ "reorder_and_upcast_attn": false,
21
+ "resid_pdrop": 0.1,
22
+ "scale_attn_by_inverse_layer_idx": false,
23
+ "scale_attn_weights": true,
24
+ "summary_activation": null,
25
+ "summary_first_dropout": 0.1,
26
+ "summary_proj_to_labels": true,
27
+ "summary_type": "cls_index",
28
+ "summary_use_proj": true,
29
+ "task_specific_params": {
30
+ "text-generation": {
31
+ "do_sample": true,
32
+ "max_length": 50
33
+ }
34
+ },
35
+ "torch_dtype": "bfloat16",
36
+ "transformers_version": "4.27.0.dev0",
37
+ "use_cache": true,
38
+ "vocab_size": 50257
39
+ }
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": "4.27.0.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 sha256:6cc52aa0bade0aa0f087380d2ec4e6f03914250caf89b5d3e004a19fbefcb3a5
3
+ size 261488764
special_tokens_map.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<|endoftext|>",
3
+ "eos_token": "<|endoftext|>",
4
+ "unk_token": "<|endoftext|>"
5
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "bos_token": "<|endoftext|>",
4
+ "eos_token": "<|endoftext|>",
5
+ "model_max_length": 1024,
6
+ "special_tokens_map_file": null,
7
+ "tokenizer_class": "GPT2Tokenizer",
8
+ "unk_token": "<|endoftext|>"
9
+ }
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 5.0,
3
+ "train_loss": 2.1389746749161076,
4
+ "train_runtime": 581.9582,
5
+ "train_samples": 76238,
6
+ "train_samples_per_second": 655.013,
7
+ "train_steps_per_second": 1.28
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,4495 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 5.0,
5
+ "global_step": 745,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.01,
12
+ "learning_rate": 1.997315436241611e-05,
13
+ "loss": 3.1953,
14
+ "step": 1
15
+ },
16
+ {
17
+ "epoch": 0.01,
18
+ "learning_rate": 1.9946308724832216e-05,
19
+ "loss": 3.1953,
20
+ "step": 2
21
+ },
22
+ {
23
+ "epoch": 0.02,
24
+ "learning_rate": 1.9919463087248324e-05,
25
+ "loss": 3.1523,
26
+ "step": 3
27
+ },
28
+ {
29
+ "epoch": 0.03,
30
+ "learning_rate": 1.989261744966443e-05,
31
+ "loss": 3.1562,
32
+ "step": 4
33
+ },
34
+ {
35
+ "epoch": 0.03,
36
+ "learning_rate": 1.986577181208054e-05,
37
+ "loss": 3.043,
38
+ "step": 5
39
+ },
40
+ {
41
+ "epoch": 0.04,
42
+ "learning_rate": 1.9838926174496646e-05,
43
+ "loss": 3.0391,
44
+ "step": 6
45
+ },
46
+ {
47
+ "epoch": 0.05,
48
+ "learning_rate": 1.9812080536912754e-05,
49
+ "loss": 3.0312,
50
+ "step": 7
51
+ },
52
+ {
53
+ "epoch": 0.05,
54
+ "learning_rate": 1.978523489932886e-05,
55
+ "loss": 2.9473,
56
+ "step": 8
57
+ },
58
+ {
59
+ "epoch": 0.06,
60
+ "learning_rate": 1.975838926174497e-05,
61
+ "loss": 2.9316,
62
+ "step": 9
63
+ },
64
+ {
65
+ "epoch": 0.07,
66
+ "learning_rate": 1.9731543624161076e-05,
67
+ "loss": 2.9316,
68
+ "step": 10
69
+ },
70
+ {
71
+ "epoch": 0.07,
72
+ "learning_rate": 1.9704697986577183e-05,
73
+ "loss": 2.9062,
74
+ "step": 11
75
+ },
76
+ {
77
+ "epoch": 0.08,
78
+ "learning_rate": 1.967785234899329e-05,
79
+ "loss": 2.8984,
80
+ "step": 12
81
+ },
82
+ {
83
+ "epoch": 0.09,
84
+ "learning_rate": 1.9651006711409398e-05,
85
+ "loss": 2.8652,
86
+ "step": 13
87
+ },
88
+ {
89
+ "epoch": 0.09,
90
+ "learning_rate": 1.9624161073825505e-05,
91
+ "loss": 2.8125,
92
+ "step": 14
93
+ },
94
+ {
95
+ "epoch": 0.1,
96
+ "learning_rate": 1.9597315436241613e-05,
97
+ "loss": 2.7734,
98
+ "step": 15
99
+ },
100
+ {
101
+ "epoch": 0.11,
102
+ "learning_rate": 1.957046979865772e-05,
103
+ "loss": 2.7773,
104
+ "step": 16
105
+ },
106
+ {
107
+ "epoch": 0.11,
108
+ "learning_rate": 1.9543624161073828e-05,
109
+ "loss": 2.75,
110
+ "step": 17
111
+ },
112
+ {
113
+ "epoch": 0.12,
114
+ "learning_rate": 1.9516778523489935e-05,
115
+ "loss": 2.7695,
116
+ "step": 18
117
+ },
118
+ {
119
+ "epoch": 0.13,
120
+ "learning_rate": 1.9489932885906042e-05,
121
+ "loss": 2.7285,
122
+ "step": 19
123
+ },
124
+ {
125
+ "epoch": 0.13,
126
+ "learning_rate": 1.946308724832215e-05,
127
+ "loss": 2.7148,
128
+ "step": 20
129
+ },
130
+ {
131
+ "epoch": 0.14,
132
+ "learning_rate": 1.9436241610738257e-05,
133
+ "loss": 2.7031,
134
+ "step": 21
135
+ },
136
+ {
137
+ "epoch": 0.15,
138
+ "learning_rate": 1.9409395973154365e-05,
139
+ "loss": 2.6816,
140
+ "step": 22
141
+ },
142
+ {
143
+ "epoch": 0.15,
144
+ "learning_rate": 1.9382550335570472e-05,
145
+ "loss": 2.6738,
146
+ "step": 23
147
+ },
148
+ {
149
+ "epoch": 0.16,
150
+ "learning_rate": 1.935570469798658e-05,
151
+ "loss": 2.6895,
152
+ "step": 24
153
+ },
154
+ {
155
+ "epoch": 0.17,
156
+ "learning_rate": 1.9328859060402687e-05,
157
+ "loss": 2.668,
158
+ "step": 25
159
+ },
160
+ {
161
+ "epoch": 0.17,
162
+ "learning_rate": 1.9302013422818794e-05,
163
+ "loss": 2.6699,
164
+ "step": 26
165
+ },
166
+ {
167
+ "epoch": 0.18,
168
+ "learning_rate": 1.9275167785234902e-05,
169
+ "loss": 2.625,
170
+ "step": 27
171
+ },
172
+ {
173
+ "epoch": 0.19,
174
+ "learning_rate": 1.924832214765101e-05,
175
+ "loss": 2.623,
176
+ "step": 28
177
+ },
178
+ {
179
+ "epoch": 0.19,
180
+ "learning_rate": 1.9221476510067117e-05,
181
+ "loss": 2.6055,
182
+ "step": 29
183
+ },
184
+ {
185
+ "epoch": 0.2,
186
+ "learning_rate": 1.919463087248322e-05,
187
+ "loss": 2.5723,
188
+ "step": 30
189
+ },
190
+ {
191
+ "epoch": 0.21,
192
+ "learning_rate": 1.916778523489933e-05,
193
+ "loss": 2.5762,
194
+ "step": 31
195
+ },
196
+ {
197
+ "epoch": 0.21,
198
+ "learning_rate": 1.914093959731544e-05,
199
+ "loss": 2.543,
200
+ "step": 32
201
+ },
202
+ {
203
+ "epoch": 0.22,
204
+ "learning_rate": 1.9114093959731546e-05,
205
+ "loss": 2.5742,
206
+ "step": 33
207
+ },
208
+ {
209
+ "epoch": 0.23,
210
+ "learning_rate": 1.9087248322147654e-05,
211
+ "loss": 2.5508,
212
+ "step": 34
213
+ },
214
+ {
215
+ "epoch": 0.23,
216
+ "learning_rate": 1.9060402684563758e-05,
217
+ "loss": 2.541,
218
+ "step": 35
219
+ },
220
+ {
221
+ "epoch": 0.24,
222
+ "learning_rate": 1.903355704697987e-05,
223
+ "loss": 2.5332,
224
+ "step": 36
225
+ },
226
+ {
227
+ "epoch": 0.25,
228
+ "learning_rate": 1.9006711409395976e-05,
229
+ "loss": 2.5449,
230
+ "step": 37
231
+ },
232
+ {
233
+ "epoch": 0.26,
234
+ "learning_rate": 1.8979865771812083e-05,
235
+ "loss": 2.4922,
236
+ "step": 38
237
+ },
238
+ {
239
+ "epoch": 0.26,
240
+ "learning_rate": 1.895302013422819e-05,
241
+ "loss": 2.5156,
242
+ "step": 39
243
+ },
244
+ {
245
+ "epoch": 0.27,
246
+ "learning_rate": 1.8926174496644295e-05,
247
+ "loss": 2.502,
248
+ "step": 40
249
+ },
250
+ {
251
+ "epoch": 0.28,
252
+ "learning_rate": 1.8899328859060406e-05,
253
+ "loss": 2.4941,
254
+ "step": 41
255
+ },
256
+ {
257
+ "epoch": 0.28,
258
+ "learning_rate": 1.8872483221476513e-05,
259
+ "loss": 2.4785,
260
+ "step": 42
261
+ },
262
+ {
263
+ "epoch": 0.29,
264
+ "learning_rate": 1.884563758389262e-05,
265
+ "loss": 2.4805,
266
+ "step": 43
267
+ },
268
+ {
269
+ "epoch": 0.3,
270
+ "learning_rate": 1.8818791946308724e-05,
271
+ "loss": 2.4805,
272
+ "step": 44
273
+ },
274
+ {
275
+ "epoch": 0.3,
276
+ "learning_rate": 1.8791946308724832e-05,
277
+ "loss": 2.4668,
278
+ "step": 45
279
+ },
280
+ {
281
+ "epoch": 0.31,
282
+ "learning_rate": 1.8765100671140943e-05,
283
+ "loss": 2.4805,
284
+ "step": 46
285
+ },
286
+ {
287
+ "epoch": 0.32,
288
+ "learning_rate": 1.873825503355705e-05,
289
+ "loss": 2.4531,
290
+ "step": 47
291
+ },
292
+ {
293
+ "epoch": 0.32,
294
+ "learning_rate": 1.8711409395973157e-05,
295
+ "loss": 2.4375,
296
+ "step": 48
297
+ },
298
+ {
299
+ "epoch": 0.33,
300
+ "learning_rate": 1.868456375838926e-05,
301
+ "loss": 2.4434,
302
+ "step": 49
303
+ },
304
+ {
305
+ "epoch": 0.34,
306
+ "learning_rate": 1.865771812080537e-05,
307
+ "loss": 2.4258,
308
+ "step": 50
309
+ },
310
+ {
311
+ "epoch": 0.34,
312
+ "learning_rate": 1.863087248322148e-05,
313
+ "loss": 2.4453,
314
+ "step": 51
315
+ },
316
+ {
317
+ "epoch": 0.35,
318
+ "learning_rate": 1.8604026845637587e-05,
319
+ "loss": 2.4297,
320
+ "step": 52
321
+ },
322
+ {
323
+ "epoch": 0.36,
324
+ "learning_rate": 1.857718120805369e-05,
325
+ "loss": 2.4316,
326
+ "step": 53
327
+ },
328
+ {
329
+ "epoch": 0.36,
330
+ "learning_rate": 1.85503355704698e-05,
331
+ "loss": 2.4121,
332
+ "step": 54
333
+ },
334
+ {
335
+ "epoch": 0.37,
336
+ "learning_rate": 1.8523489932885906e-05,
337
+ "loss": 2.4258,
338
+ "step": 55
339
+ },
340
+ {
341
+ "epoch": 0.38,
342
+ "learning_rate": 1.8496644295302017e-05,
343
+ "loss": 2.4121,
344
+ "step": 56
345
+ },
346
+ {
347
+ "epoch": 0.38,
348
+ "learning_rate": 1.8469798657718124e-05,
349
+ "loss": 2.3984,
350
+ "step": 57
351
+ },
352
+ {
353
+ "epoch": 0.39,
354
+ "learning_rate": 1.8442953020134228e-05,
355
+ "loss": 2.4082,
356
+ "step": 58
357
+ },
358
+ {
359
+ "epoch": 0.4,
360
+ "learning_rate": 1.8416107382550336e-05,
361
+ "loss": 2.3945,
362
+ "step": 59
363
+ },
364
+ {
365
+ "epoch": 0.4,
366
+ "learning_rate": 1.8389261744966443e-05,
367
+ "loss": 2.4277,
368
+ "step": 60
369
+ },
370
+ {
371
+ "epoch": 0.41,
372
+ "learning_rate": 1.8362416107382554e-05,
373
+ "loss": 2.3809,
374
+ "step": 61
375
+ },
376
+ {
377
+ "epoch": 0.42,
378
+ "learning_rate": 1.833557046979866e-05,
379
+ "loss": 2.3945,
380
+ "step": 62
381
+ },
382
+ {
383
+ "epoch": 0.42,
384
+ "learning_rate": 1.8308724832214765e-05,
385
+ "loss": 2.3691,
386
+ "step": 63
387
+ },
388
+ {
389
+ "epoch": 0.43,
390
+ "learning_rate": 1.8281879194630873e-05,
391
+ "loss": 2.3809,
392
+ "step": 64
393
+ },
394
+ {
395
+ "epoch": 0.44,
396
+ "learning_rate": 1.825503355704698e-05,
397
+ "loss": 2.375,
398
+ "step": 65
399
+ },
400
+ {
401
+ "epoch": 0.44,
402
+ "learning_rate": 1.822818791946309e-05,
403
+ "loss": 2.3711,
404
+ "step": 66
405
+ },
406
+ {
407
+ "epoch": 0.45,
408
+ "learning_rate": 1.8201342281879195e-05,
409
+ "loss": 2.3711,
410
+ "step": 67
411
+ },
412
+ {
413
+ "epoch": 0.46,
414
+ "learning_rate": 1.8174496644295302e-05,
415
+ "loss": 2.3711,
416
+ "step": 68
417
+ },
418
+ {
419
+ "epoch": 0.46,
420
+ "learning_rate": 1.814765100671141e-05,
421
+ "loss": 2.377,
422
+ "step": 69
423
+ },
424
+ {
425
+ "epoch": 0.47,
426
+ "learning_rate": 1.8120805369127517e-05,
427
+ "loss": 2.3535,
428
+ "step": 70
429
+ },
430
+ {
431
+ "epoch": 0.48,
432
+ "learning_rate": 1.8093959731543628e-05,
433
+ "loss": 2.3359,
434
+ "step": 71
435
+ },
436
+ {
437
+ "epoch": 0.48,
438
+ "learning_rate": 1.8067114093959732e-05,
439
+ "loss": 2.3613,
440
+ "step": 72
441
+ },
442
+ {
443
+ "epoch": 0.49,
444
+ "learning_rate": 1.804026845637584e-05,
445
+ "loss": 2.3574,
446
+ "step": 73
447
+ },
448
+ {
449
+ "epoch": 0.5,
450
+ "learning_rate": 1.8013422818791947e-05,
451
+ "loss": 2.3516,
452
+ "step": 74
453
+ },
454
+ {
455
+ "epoch": 0.5,
456
+ "learning_rate": 1.7986577181208054e-05,
457
+ "loss": 2.3594,
458
+ "step": 75
459
+ },
460
+ {
461
+ "epoch": 0.51,
462
+ "learning_rate": 1.795973154362416e-05,
463
+ "loss": 2.3359,
464
+ "step": 76
465
+ },
466
+ {
467
+ "epoch": 0.52,
468
+ "learning_rate": 1.793288590604027e-05,
469
+ "loss": 2.3438,
470
+ "step": 77
471
+ },
472
+ {
473
+ "epoch": 0.52,
474
+ "learning_rate": 1.7906040268456376e-05,
475
+ "loss": 2.3379,
476
+ "step": 78
477
+ },
478
+ {
479
+ "epoch": 0.53,
480
+ "learning_rate": 1.7879194630872484e-05,
481
+ "loss": 2.3359,
482
+ "step": 79
483
+ },
484
+ {
485
+ "epoch": 0.54,
486
+ "learning_rate": 1.7852348993288595e-05,
487
+ "loss": 2.3496,
488
+ "step": 80
489
+ },
490
+ {
491
+ "epoch": 0.54,
492
+ "learning_rate": 1.78255033557047e-05,
493
+ "loss": 2.3281,
494
+ "step": 81
495
+ },
496
+ {
497
+ "epoch": 0.55,
498
+ "learning_rate": 1.7798657718120806e-05,
499
+ "loss": 2.3359,
500
+ "step": 82
501
+ },
502
+ {
503
+ "epoch": 0.56,
504
+ "learning_rate": 1.7771812080536913e-05,
505
+ "loss": 2.3457,
506
+ "step": 83
507
+ },
508
+ {
509
+ "epoch": 0.56,
510
+ "learning_rate": 1.774496644295302e-05,
511
+ "loss": 2.3223,
512
+ "step": 84
513
+ },
514
+ {
515
+ "epoch": 0.57,
516
+ "learning_rate": 1.771812080536913e-05,
517
+ "loss": 2.3438,
518
+ "step": 85
519
+ },
520
+ {
521
+ "epoch": 0.58,
522
+ "learning_rate": 1.7691275167785236e-05,
523
+ "loss": 2.3242,
524
+ "step": 86
525
+ },
526
+ {
527
+ "epoch": 0.58,
528
+ "learning_rate": 1.7664429530201343e-05,
529
+ "loss": 2.3379,
530
+ "step": 87
531
+ },
532
+ {
533
+ "epoch": 0.59,
534
+ "learning_rate": 1.763758389261745e-05,
535
+ "loss": 2.3438,
536
+ "step": 88
537
+ },
538
+ {
539
+ "epoch": 0.6,
540
+ "learning_rate": 1.7610738255033558e-05,
541
+ "loss": 2.3145,
542
+ "step": 89
543
+ },
544
+ {
545
+ "epoch": 0.6,
546
+ "learning_rate": 1.7583892617449665e-05,
547
+ "loss": 2.2949,
548
+ "step": 90
549
+ },
550
+ {
551
+ "epoch": 0.61,
552
+ "learning_rate": 1.7557046979865773e-05,
553
+ "loss": 2.3066,
554
+ "step": 91
555
+ },
556
+ {
557
+ "epoch": 0.62,
558
+ "learning_rate": 1.753020134228188e-05,
559
+ "loss": 2.3027,
560
+ "step": 92
561
+ },
562
+ {
563
+ "epoch": 0.62,
564
+ "learning_rate": 1.7503355704697988e-05,
565
+ "loss": 2.2949,
566
+ "step": 93
567
+ },
568
+ {
569
+ "epoch": 0.63,
570
+ "learning_rate": 1.7476510067114095e-05,
571
+ "loss": 2.3086,
572
+ "step": 94
573
+ },
574
+ {
575
+ "epoch": 0.64,
576
+ "learning_rate": 1.7449664429530202e-05,
577
+ "loss": 2.3086,
578
+ "step": 95
579
+ },
580
+ {
581
+ "epoch": 0.64,
582
+ "learning_rate": 1.742281879194631e-05,
583
+ "loss": 2.3066,
584
+ "step": 96
585
+ },
586
+ {
587
+ "epoch": 0.65,
588
+ "learning_rate": 1.7395973154362417e-05,
589
+ "loss": 2.2793,
590
+ "step": 97
591
+ },
592
+ {
593
+ "epoch": 0.66,
594
+ "learning_rate": 1.7369127516778525e-05,
595
+ "loss": 2.3008,
596
+ "step": 98
597
+ },
598
+ {
599
+ "epoch": 0.66,
600
+ "learning_rate": 1.7342281879194632e-05,
601
+ "loss": 2.2949,
602
+ "step": 99
603
+ },
604
+ {
605
+ "epoch": 0.67,
606
+ "learning_rate": 1.731543624161074e-05,
607
+ "loss": 2.3086,
608
+ "step": 100
609
+ },
610
+ {
611
+ "epoch": 0.68,
612
+ "learning_rate": 1.7288590604026847e-05,
613
+ "loss": 2.2988,
614
+ "step": 101
615
+ },
616
+ {
617
+ "epoch": 0.68,
618
+ "learning_rate": 1.7261744966442954e-05,
619
+ "loss": 2.2715,
620
+ "step": 102
621
+ },
622
+ {
623
+ "epoch": 0.69,
624
+ "learning_rate": 1.7234899328859062e-05,
625
+ "loss": 2.2949,
626
+ "step": 103
627
+ },
628
+ {
629
+ "epoch": 0.7,
630
+ "learning_rate": 1.720805369127517e-05,
631
+ "loss": 2.2793,
632
+ "step": 104
633
+ },
634
+ {
635
+ "epoch": 0.7,
636
+ "learning_rate": 1.7181208053691277e-05,
637
+ "loss": 2.2637,
638
+ "step": 105
639
+ },
640
+ {
641
+ "epoch": 0.71,
642
+ "learning_rate": 1.7154362416107384e-05,
643
+ "loss": 2.2773,
644
+ "step": 106
645
+ },
646
+ {
647
+ "epoch": 0.72,
648
+ "learning_rate": 1.712751677852349e-05,
649
+ "loss": 2.293,
650
+ "step": 107
651
+ },
652
+ {
653
+ "epoch": 0.72,
654
+ "learning_rate": 1.71006711409396e-05,
655
+ "loss": 2.2793,
656
+ "step": 108
657
+ },
658
+ {
659
+ "epoch": 0.73,
660
+ "learning_rate": 1.7073825503355706e-05,
661
+ "loss": 2.2773,
662
+ "step": 109
663
+ },
664
+ {
665
+ "epoch": 0.74,
666
+ "learning_rate": 1.7046979865771814e-05,
667
+ "loss": 2.3027,
668
+ "step": 110
669
+ },
670
+ {
671
+ "epoch": 0.74,
672
+ "learning_rate": 1.702013422818792e-05,
673
+ "loss": 2.2676,
674
+ "step": 111
675
+ },
676
+ {
677
+ "epoch": 0.75,
678
+ "learning_rate": 1.699328859060403e-05,
679
+ "loss": 2.2852,
680
+ "step": 112
681
+ },
682
+ {
683
+ "epoch": 0.76,
684
+ "learning_rate": 1.6966442953020136e-05,
685
+ "loss": 2.2715,
686
+ "step": 113
687
+ },
688
+ {
689
+ "epoch": 0.77,
690
+ "learning_rate": 1.6939597315436243e-05,
691
+ "loss": 2.2559,
692
+ "step": 114
693
+ },
694
+ {
695
+ "epoch": 0.77,
696
+ "learning_rate": 1.691275167785235e-05,
697
+ "loss": 2.2812,
698
+ "step": 115
699
+ },
700
+ {
701
+ "epoch": 0.78,
702
+ "learning_rate": 1.6885906040268458e-05,
703
+ "loss": 2.2676,
704
+ "step": 116
705
+ },
706
+ {
707
+ "epoch": 0.79,
708
+ "learning_rate": 1.6859060402684565e-05,
709
+ "loss": 2.2832,
710
+ "step": 117
711
+ },
712
+ {
713
+ "epoch": 0.79,
714
+ "learning_rate": 1.6832214765100673e-05,
715
+ "loss": 2.2617,
716
+ "step": 118
717
+ },
718
+ {
719
+ "epoch": 0.8,
720
+ "learning_rate": 1.680536912751678e-05,
721
+ "loss": 2.2656,
722
+ "step": 119
723
+ },
724
+ {
725
+ "epoch": 0.81,
726
+ "learning_rate": 1.6778523489932888e-05,
727
+ "loss": 2.2578,
728
+ "step": 120
729
+ },
730
+ {
731
+ "epoch": 0.81,
732
+ "learning_rate": 1.6751677852348995e-05,
733
+ "loss": 2.2383,
734
+ "step": 121
735
+ },
736
+ {
737
+ "epoch": 0.82,
738
+ "learning_rate": 1.6724832214765103e-05,
739
+ "loss": 2.2441,
740
+ "step": 122
741
+ },
742
+ {
743
+ "epoch": 0.83,
744
+ "learning_rate": 1.669798657718121e-05,
745
+ "loss": 2.2539,
746
+ "step": 123
747
+ },
748
+ {
749
+ "epoch": 0.83,
750
+ "learning_rate": 1.6671140939597317e-05,
751
+ "loss": 2.2441,
752
+ "step": 124
753
+ },
754
+ {
755
+ "epoch": 0.84,
756
+ "learning_rate": 1.6644295302013425e-05,
757
+ "loss": 2.2363,
758
+ "step": 125
759
+ },
760
+ {
761
+ "epoch": 0.85,
762
+ "learning_rate": 1.6617449664429532e-05,
763
+ "loss": 2.2539,
764
+ "step": 126
765
+ },
766
+ {
767
+ "epoch": 0.85,
768
+ "learning_rate": 1.659060402684564e-05,
769
+ "loss": 2.2539,
770
+ "step": 127
771
+ },
772
+ {
773
+ "epoch": 0.86,
774
+ "learning_rate": 1.6563758389261747e-05,
775
+ "loss": 2.248,
776
+ "step": 128
777
+ },
778
+ {
779
+ "epoch": 0.87,
780
+ "learning_rate": 1.6536912751677854e-05,
781
+ "loss": 2.2402,
782
+ "step": 129
783
+ },
784
+ {
785
+ "epoch": 0.87,
786
+ "learning_rate": 1.6510067114093962e-05,
787
+ "loss": 2.2578,
788
+ "step": 130
789
+ },
790
+ {
791
+ "epoch": 0.88,
792
+ "learning_rate": 1.648322147651007e-05,
793
+ "loss": 2.248,
794
+ "step": 131
795
+ },
796
+ {
797
+ "epoch": 0.89,
798
+ "learning_rate": 1.6456375838926177e-05,
799
+ "loss": 2.2793,
800
+ "step": 132
801
+ },
802
+ {
803
+ "epoch": 0.89,
804
+ "learning_rate": 1.6429530201342284e-05,
805
+ "loss": 2.2227,
806
+ "step": 133
807
+ },
808
+ {
809
+ "epoch": 0.9,
810
+ "learning_rate": 1.640268456375839e-05,
811
+ "loss": 2.2422,
812
+ "step": 134
813
+ },
814
+ {
815
+ "epoch": 0.91,
816
+ "learning_rate": 1.63758389261745e-05,
817
+ "loss": 2.2402,
818
+ "step": 135
819
+ },
820
+ {
821
+ "epoch": 0.91,
822
+ "learning_rate": 1.6348993288590603e-05,
823
+ "loss": 2.2461,
824
+ "step": 136
825
+ },
826
+ {
827
+ "epoch": 0.92,
828
+ "learning_rate": 1.6322147651006714e-05,
829
+ "loss": 2.2246,
830
+ "step": 137
831
+ },
832
+ {
833
+ "epoch": 0.93,
834
+ "learning_rate": 1.629530201342282e-05,
835
+ "loss": 2.2363,
836
+ "step": 138
837
+ },
838
+ {
839
+ "epoch": 0.93,
840
+ "learning_rate": 1.626845637583893e-05,
841
+ "loss": 2.2285,
842
+ "step": 139
843
+ },
844
+ {
845
+ "epoch": 0.94,
846
+ "learning_rate": 1.6241610738255036e-05,
847
+ "loss": 2.2461,
848
+ "step": 140
849
+ },
850
+ {
851
+ "epoch": 0.95,
852
+ "learning_rate": 1.621476510067114e-05,
853
+ "loss": 2.2129,
854
+ "step": 141
855
+ },
856
+ {
857
+ "epoch": 0.95,
858
+ "learning_rate": 1.618791946308725e-05,
859
+ "loss": 2.2246,
860
+ "step": 142
861
+ },
862
+ {
863
+ "epoch": 0.96,
864
+ "learning_rate": 1.6161073825503358e-05,
865
+ "loss": 2.2148,
866
+ "step": 143
867
+ },
868
+ {
869
+ "epoch": 0.97,
870
+ "learning_rate": 1.6134228187919466e-05,
871
+ "loss": 2.2148,
872
+ "step": 144
873
+ },
874
+ {
875
+ "epoch": 0.97,
876
+ "learning_rate": 1.610738255033557e-05,
877
+ "loss": 2.2207,
878
+ "step": 145
879
+ },
880
+ {
881
+ "epoch": 0.98,
882
+ "learning_rate": 1.6080536912751677e-05,
883
+ "loss": 2.1973,
884
+ "step": 146
885
+ },
886
+ {
887
+ "epoch": 0.99,
888
+ "learning_rate": 1.6053691275167788e-05,
889
+ "loss": 2.2109,
890
+ "step": 147
891
+ },
892
+ {
893
+ "epoch": 0.99,
894
+ "learning_rate": 1.6026845637583895e-05,
895
+ "loss": 2.2266,
896
+ "step": 148
897
+ },
898
+ {
899
+ "epoch": 1.0,
900
+ "learning_rate": 1.6000000000000003e-05,
901
+ "loss": 2.2227,
902
+ "step": 149
903
+ },
904
+ {
905
+ "epoch": 1.01,
906
+ "learning_rate": 1.5973154362416107e-05,
907
+ "loss": 2.209,
908
+ "step": 150
909
+ },
910
+ {
911
+ "epoch": 1.01,
912
+ "learning_rate": 1.5946308724832214e-05,
913
+ "loss": 2.207,
914
+ "step": 151
915
+ },
916
+ {
917
+ "epoch": 1.02,
918
+ "learning_rate": 1.5919463087248325e-05,
919
+ "loss": 2.1934,
920
+ "step": 152
921
+ },
922
+ {
923
+ "epoch": 1.03,
924
+ "learning_rate": 1.5892617449664432e-05,
925
+ "loss": 2.2129,
926
+ "step": 153
927
+ },
928
+ {
929
+ "epoch": 1.03,
930
+ "learning_rate": 1.586577181208054e-05,
931
+ "loss": 2.207,
932
+ "step": 154
933
+ },
934
+ {
935
+ "epoch": 1.04,
936
+ "learning_rate": 1.5838926174496644e-05,
937
+ "loss": 2.2324,
938
+ "step": 155
939
+ },
940
+ {
941
+ "epoch": 1.05,
942
+ "learning_rate": 1.581208053691275e-05,
943
+ "loss": 2.2031,
944
+ "step": 156
945
+ },
946
+ {
947
+ "epoch": 1.05,
948
+ "learning_rate": 1.5785234899328862e-05,
949
+ "loss": 2.2031,
950
+ "step": 157
951
+ },
952
+ {
953
+ "epoch": 1.06,
954
+ "learning_rate": 1.575838926174497e-05,
955
+ "loss": 2.1992,
956
+ "step": 158
957
+ },
958
+ {
959
+ "epoch": 1.07,
960
+ "learning_rate": 1.5731543624161073e-05,
961
+ "loss": 2.2031,
962
+ "step": 159
963
+ },
964
+ {
965
+ "epoch": 1.07,
966
+ "learning_rate": 1.570469798657718e-05,
967
+ "loss": 2.1836,
968
+ "step": 160
969
+ },
970
+ {
971
+ "epoch": 1.08,
972
+ "learning_rate": 1.5677852348993288e-05,
973
+ "loss": 2.2266,
974
+ "step": 161
975
+ },
976
+ {
977
+ "epoch": 1.09,
978
+ "learning_rate": 1.56510067114094e-05,
979
+ "loss": 2.2129,
980
+ "step": 162
981
+ },
982
+ {
983
+ "epoch": 1.09,
984
+ "learning_rate": 1.5624161073825506e-05,
985
+ "loss": 2.1973,
986
+ "step": 163
987
+ },
988
+ {
989
+ "epoch": 1.1,
990
+ "learning_rate": 1.559731543624161e-05,
991
+ "loss": 2.2109,
992
+ "step": 164
993
+ },
994
+ {
995
+ "epoch": 1.11,
996
+ "learning_rate": 1.5570469798657718e-05,
997
+ "loss": 2.1934,
998
+ "step": 165
999
+ },
1000
+ {
1001
+ "epoch": 1.11,
1002
+ "learning_rate": 1.554362416107383e-05,
1003
+ "loss": 2.209,
1004
+ "step": 166
1005
+ },
1006
+ {
1007
+ "epoch": 1.12,
1008
+ "learning_rate": 1.5516778523489936e-05,
1009
+ "loss": 2.1992,
1010
+ "step": 167
1011
+ },
1012
+ {
1013
+ "epoch": 1.13,
1014
+ "learning_rate": 1.548993288590604e-05,
1015
+ "loss": 2.1816,
1016
+ "step": 168
1017
+ },
1018
+ {
1019
+ "epoch": 1.13,
1020
+ "learning_rate": 1.5463087248322148e-05,
1021
+ "loss": 2.2109,
1022
+ "step": 169
1023
+ },
1024
+ {
1025
+ "epoch": 1.14,
1026
+ "learning_rate": 1.5436241610738255e-05,
1027
+ "loss": 2.1992,
1028
+ "step": 170
1029
+ },
1030
+ {
1031
+ "epoch": 1.15,
1032
+ "learning_rate": 1.5409395973154366e-05,
1033
+ "loss": 2.1699,
1034
+ "step": 171
1035
+ },
1036
+ {
1037
+ "epoch": 1.15,
1038
+ "learning_rate": 1.5382550335570473e-05,
1039
+ "loss": 2.1777,
1040
+ "step": 172
1041
+ },
1042
+ {
1043
+ "epoch": 1.16,
1044
+ "learning_rate": 1.5355704697986577e-05,
1045
+ "loss": 2.2148,
1046
+ "step": 173
1047
+ },
1048
+ {
1049
+ "epoch": 1.17,
1050
+ "learning_rate": 1.5328859060402685e-05,
1051
+ "loss": 2.1953,
1052
+ "step": 174
1053
+ },
1054
+ {
1055
+ "epoch": 1.17,
1056
+ "learning_rate": 1.5302013422818792e-05,
1057
+ "loss": 2.168,
1058
+ "step": 175
1059
+ },
1060
+ {
1061
+ "epoch": 1.18,
1062
+ "learning_rate": 1.5275167785234903e-05,
1063
+ "loss": 2.1797,
1064
+ "step": 176
1065
+ },
1066
+ {
1067
+ "epoch": 1.19,
1068
+ "learning_rate": 1.5248322147651009e-05,
1069
+ "loss": 2.2109,
1070
+ "step": 177
1071
+ },
1072
+ {
1073
+ "epoch": 1.19,
1074
+ "learning_rate": 1.5221476510067114e-05,
1075
+ "loss": 2.1895,
1076
+ "step": 178
1077
+ },
1078
+ {
1079
+ "epoch": 1.2,
1080
+ "learning_rate": 1.5194630872483223e-05,
1081
+ "loss": 2.1816,
1082
+ "step": 179
1083
+ },
1084
+ {
1085
+ "epoch": 1.21,
1086
+ "learning_rate": 1.516778523489933e-05,
1087
+ "loss": 2.1621,
1088
+ "step": 180
1089
+ },
1090
+ {
1091
+ "epoch": 1.21,
1092
+ "learning_rate": 1.5140939597315438e-05,
1093
+ "loss": 2.1816,
1094
+ "step": 181
1095
+ },
1096
+ {
1097
+ "epoch": 1.22,
1098
+ "learning_rate": 1.5114093959731544e-05,
1099
+ "loss": 2.1777,
1100
+ "step": 182
1101
+ },
1102
+ {
1103
+ "epoch": 1.23,
1104
+ "learning_rate": 1.5087248322147651e-05,
1105
+ "loss": 2.1777,
1106
+ "step": 183
1107
+ },
1108
+ {
1109
+ "epoch": 1.23,
1110
+ "learning_rate": 1.506040268456376e-05,
1111
+ "loss": 2.1738,
1112
+ "step": 184
1113
+ },
1114
+ {
1115
+ "epoch": 1.24,
1116
+ "learning_rate": 1.5033557046979868e-05,
1117
+ "loss": 2.1836,
1118
+ "step": 185
1119
+ },
1120
+ {
1121
+ "epoch": 1.25,
1122
+ "learning_rate": 1.5006711409395975e-05,
1123
+ "loss": 2.1797,
1124
+ "step": 186
1125
+ },
1126
+ {
1127
+ "epoch": 1.26,
1128
+ "learning_rate": 1.4979865771812081e-05,
1129
+ "loss": 2.1738,
1130
+ "step": 187
1131
+ },
1132
+ {
1133
+ "epoch": 1.26,
1134
+ "learning_rate": 1.4953020134228188e-05,
1135
+ "loss": 2.1914,
1136
+ "step": 188
1137
+ },
1138
+ {
1139
+ "epoch": 1.27,
1140
+ "learning_rate": 1.4926174496644297e-05,
1141
+ "loss": 2.1719,
1142
+ "step": 189
1143
+ },
1144
+ {
1145
+ "epoch": 1.28,
1146
+ "learning_rate": 1.4899328859060405e-05,
1147
+ "loss": 2.1777,
1148
+ "step": 190
1149
+ },
1150
+ {
1151
+ "epoch": 1.28,
1152
+ "learning_rate": 1.487248322147651e-05,
1153
+ "loss": 2.1504,
1154
+ "step": 191
1155
+ },
1156
+ {
1157
+ "epoch": 1.29,
1158
+ "learning_rate": 1.4845637583892618e-05,
1159
+ "loss": 2.1992,
1160
+ "step": 192
1161
+ },
1162
+ {
1163
+ "epoch": 1.3,
1164
+ "learning_rate": 1.4818791946308725e-05,
1165
+ "loss": 2.1855,
1166
+ "step": 193
1167
+ },
1168
+ {
1169
+ "epoch": 1.3,
1170
+ "learning_rate": 1.4791946308724835e-05,
1171
+ "loss": 2.1777,
1172
+ "step": 194
1173
+ },
1174
+ {
1175
+ "epoch": 1.31,
1176
+ "learning_rate": 1.4765100671140942e-05,
1177
+ "loss": 2.1504,
1178
+ "step": 195
1179
+ },
1180
+ {
1181
+ "epoch": 1.32,
1182
+ "learning_rate": 1.4738255033557048e-05,
1183
+ "loss": 2.168,
1184
+ "step": 196
1185
+ },
1186
+ {
1187
+ "epoch": 1.32,
1188
+ "learning_rate": 1.4711409395973155e-05,
1189
+ "loss": 2.168,
1190
+ "step": 197
1191
+ },
1192
+ {
1193
+ "epoch": 1.33,
1194
+ "learning_rate": 1.4684563758389262e-05,
1195
+ "loss": 2.1777,
1196
+ "step": 198
1197
+ },
1198
+ {
1199
+ "epoch": 1.34,
1200
+ "learning_rate": 1.4657718120805372e-05,
1201
+ "loss": 2.1621,
1202
+ "step": 199
1203
+ },
1204
+ {
1205
+ "epoch": 1.34,
1206
+ "learning_rate": 1.4630872483221479e-05,
1207
+ "loss": 2.1758,
1208
+ "step": 200
1209
+ },
1210
+ {
1211
+ "epoch": 1.35,
1212
+ "learning_rate": 1.4604026845637585e-05,
1213
+ "loss": 2.1719,
1214
+ "step": 201
1215
+ },
1216
+ {
1217
+ "epoch": 1.36,
1218
+ "learning_rate": 1.4577181208053692e-05,
1219
+ "loss": 2.1777,
1220
+ "step": 202
1221
+ },
1222
+ {
1223
+ "epoch": 1.36,
1224
+ "learning_rate": 1.45503355704698e-05,
1225
+ "loss": 2.1445,
1226
+ "step": 203
1227
+ },
1228
+ {
1229
+ "epoch": 1.37,
1230
+ "learning_rate": 1.4523489932885909e-05,
1231
+ "loss": 2.1562,
1232
+ "step": 204
1233
+ },
1234
+ {
1235
+ "epoch": 1.38,
1236
+ "learning_rate": 1.4496644295302014e-05,
1237
+ "loss": 2.1758,
1238
+ "step": 205
1239
+ },
1240
+ {
1241
+ "epoch": 1.38,
1242
+ "learning_rate": 1.4469798657718122e-05,
1243
+ "loss": 2.1738,
1244
+ "step": 206
1245
+ },
1246
+ {
1247
+ "epoch": 1.39,
1248
+ "learning_rate": 1.444295302013423e-05,
1249
+ "loss": 2.1562,
1250
+ "step": 207
1251
+ },
1252
+ {
1253
+ "epoch": 1.4,
1254
+ "learning_rate": 1.4416107382550337e-05,
1255
+ "loss": 2.1758,
1256
+ "step": 208
1257
+ },
1258
+ {
1259
+ "epoch": 1.4,
1260
+ "learning_rate": 1.4389261744966446e-05,
1261
+ "loss": 2.1621,
1262
+ "step": 209
1263
+ },
1264
+ {
1265
+ "epoch": 1.41,
1266
+ "learning_rate": 1.4362416107382551e-05,
1267
+ "loss": 2.166,
1268
+ "step": 210
1269
+ },
1270
+ {
1271
+ "epoch": 1.42,
1272
+ "learning_rate": 1.4335570469798659e-05,
1273
+ "loss": 2.1582,
1274
+ "step": 211
1275
+ },
1276
+ {
1277
+ "epoch": 1.42,
1278
+ "learning_rate": 1.4308724832214766e-05,
1279
+ "loss": 2.1582,
1280
+ "step": 212
1281
+ },
1282
+ {
1283
+ "epoch": 1.43,
1284
+ "learning_rate": 1.4281879194630874e-05,
1285
+ "loss": 2.1562,
1286
+ "step": 213
1287
+ },
1288
+ {
1289
+ "epoch": 1.44,
1290
+ "learning_rate": 1.425503355704698e-05,
1291
+ "loss": 2.1641,
1292
+ "step": 214
1293
+ },
1294
+ {
1295
+ "epoch": 1.44,
1296
+ "learning_rate": 1.4228187919463088e-05,
1297
+ "loss": 2.1777,
1298
+ "step": 215
1299
+ },
1300
+ {
1301
+ "epoch": 1.45,
1302
+ "learning_rate": 1.4201342281879196e-05,
1303
+ "loss": 2.1426,
1304
+ "step": 216
1305
+ },
1306
+ {
1307
+ "epoch": 1.46,
1308
+ "learning_rate": 1.4174496644295303e-05,
1309
+ "loss": 2.1504,
1310
+ "step": 217
1311
+ },
1312
+ {
1313
+ "epoch": 1.46,
1314
+ "learning_rate": 1.414765100671141e-05,
1315
+ "loss": 2.1641,
1316
+ "step": 218
1317
+ },
1318
+ {
1319
+ "epoch": 1.47,
1320
+ "learning_rate": 1.4120805369127516e-05,
1321
+ "loss": 2.1484,
1322
+ "step": 219
1323
+ },
1324
+ {
1325
+ "epoch": 1.48,
1326
+ "learning_rate": 1.4093959731543626e-05,
1327
+ "loss": 2.1484,
1328
+ "step": 220
1329
+ },
1330
+ {
1331
+ "epoch": 1.48,
1332
+ "learning_rate": 1.4067114093959733e-05,
1333
+ "loss": 2.1289,
1334
+ "step": 221
1335
+ },
1336
+ {
1337
+ "epoch": 1.49,
1338
+ "learning_rate": 1.404026845637584e-05,
1339
+ "loss": 2.1426,
1340
+ "step": 222
1341
+ },
1342
+ {
1343
+ "epoch": 1.5,
1344
+ "learning_rate": 1.4013422818791948e-05,
1345
+ "loss": 2.1367,
1346
+ "step": 223
1347
+ },
1348
+ {
1349
+ "epoch": 1.5,
1350
+ "learning_rate": 1.3986577181208053e-05,
1351
+ "loss": 2.168,
1352
+ "step": 224
1353
+ },
1354
+ {
1355
+ "epoch": 1.51,
1356
+ "learning_rate": 1.3959731543624163e-05,
1357
+ "loss": 2.123,
1358
+ "step": 225
1359
+ },
1360
+ {
1361
+ "epoch": 1.52,
1362
+ "learning_rate": 1.393288590604027e-05,
1363
+ "loss": 2.1484,
1364
+ "step": 226
1365
+ },
1366
+ {
1367
+ "epoch": 1.52,
1368
+ "learning_rate": 1.3906040268456377e-05,
1369
+ "loss": 2.1641,
1370
+ "step": 227
1371
+ },
1372
+ {
1373
+ "epoch": 1.53,
1374
+ "learning_rate": 1.3879194630872483e-05,
1375
+ "loss": 2.1367,
1376
+ "step": 228
1377
+ },
1378
+ {
1379
+ "epoch": 1.54,
1380
+ "learning_rate": 1.385234899328859e-05,
1381
+ "loss": 2.1543,
1382
+ "step": 229
1383
+ },
1384
+ {
1385
+ "epoch": 1.54,
1386
+ "learning_rate": 1.38255033557047e-05,
1387
+ "loss": 2.166,
1388
+ "step": 230
1389
+ },
1390
+ {
1391
+ "epoch": 1.55,
1392
+ "learning_rate": 1.3798657718120807e-05,
1393
+ "loss": 2.1094,
1394
+ "step": 231
1395
+ },
1396
+ {
1397
+ "epoch": 1.56,
1398
+ "learning_rate": 1.3771812080536914e-05,
1399
+ "loss": 2.1289,
1400
+ "step": 232
1401
+ },
1402
+ {
1403
+ "epoch": 1.56,
1404
+ "learning_rate": 1.374496644295302e-05,
1405
+ "loss": 2.1484,
1406
+ "step": 233
1407
+ },
1408
+ {
1409
+ "epoch": 1.57,
1410
+ "learning_rate": 1.3718120805369128e-05,
1411
+ "loss": 2.1348,
1412
+ "step": 234
1413
+ },
1414
+ {
1415
+ "epoch": 1.58,
1416
+ "learning_rate": 1.3691275167785237e-05,
1417
+ "loss": 2.125,
1418
+ "step": 235
1419
+ },
1420
+ {
1421
+ "epoch": 1.58,
1422
+ "learning_rate": 1.3664429530201344e-05,
1423
+ "loss": 2.1426,
1424
+ "step": 236
1425
+ },
1426
+ {
1427
+ "epoch": 1.59,
1428
+ "learning_rate": 1.363758389261745e-05,
1429
+ "loss": 2.1504,
1430
+ "step": 237
1431
+ },
1432
+ {
1433
+ "epoch": 1.6,
1434
+ "learning_rate": 1.3610738255033557e-05,
1435
+ "loss": 2.1328,
1436
+ "step": 238
1437
+ },
1438
+ {
1439
+ "epoch": 1.6,
1440
+ "learning_rate": 1.3583892617449665e-05,
1441
+ "loss": 2.1406,
1442
+ "step": 239
1443
+ },
1444
+ {
1445
+ "epoch": 1.61,
1446
+ "learning_rate": 1.3557046979865774e-05,
1447
+ "loss": 2.1504,
1448
+ "step": 240
1449
+ },
1450
+ {
1451
+ "epoch": 1.62,
1452
+ "learning_rate": 1.3530201342281881e-05,
1453
+ "loss": 2.127,
1454
+ "step": 241
1455
+ },
1456
+ {
1457
+ "epoch": 1.62,
1458
+ "learning_rate": 1.3503355704697987e-05,
1459
+ "loss": 2.1211,
1460
+ "step": 242
1461
+ },
1462
+ {
1463
+ "epoch": 1.63,
1464
+ "learning_rate": 1.3476510067114094e-05,
1465
+ "loss": 2.1367,
1466
+ "step": 243
1467
+ },
1468
+ {
1469
+ "epoch": 1.64,
1470
+ "learning_rate": 1.3449664429530202e-05,
1471
+ "loss": 2.1406,
1472
+ "step": 244
1473
+ },
1474
+ {
1475
+ "epoch": 1.64,
1476
+ "learning_rate": 1.342281879194631e-05,
1477
+ "loss": 2.125,
1478
+ "step": 245
1479
+ },
1480
+ {
1481
+ "epoch": 1.65,
1482
+ "learning_rate": 1.3395973154362418e-05,
1483
+ "loss": 2.1445,
1484
+ "step": 246
1485
+ },
1486
+ {
1487
+ "epoch": 1.66,
1488
+ "learning_rate": 1.3369127516778524e-05,
1489
+ "loss": 2.1328,
1490
+ "step": 247
1491
+ },
1492
+ {
1493
+ "epoch": 1.66,
1494
+ "learning_rate": 1.3342281879194631e-05,
1495
+ "loss": 2.1328,
1496
+ "step": 248
1497
+ },
1498
+ {
1499
+ "epoch": 1.67,
1500
+ "learning_rate": 1.331543624161074e-05,
1501
+ "loss": 2.127,
1502
+ "step": 249
1503
+ },
1504
+ {
1505
+ "epoch": 1.68,
1506
+ "learning_rate": 1.3288590604026848e-05,
1507
+ "loss": 2.1309,
1508
+ "step": 250
1509
+ },
1510
+ {
1511
+ "epoch": 1.68,
1512
+ "learning_rate": 1.3261744966442954e-05,
1513
+ "loss": 2.1426,
1514
+ "step": 251
1515
+ },
1516
+ {
1517
+ "epoch": 1.69,
1518
+ "learning_rate": 1.3234899328859061e-05,
1519
+ "loss": 2.1289,
1520
+ "step": 252
1521
+ },
1522
+ {
1523
+ "epoch": 1.7,
1524
+ "learning_rate": 1.3208053691275168e-05,
1525
+ "loss": 2.1387,
1526
+ "step": 253
1527
+ },
1528
+ {
1529
+ "epoch": 1.7,
1530
+ "learning_rate": 1.3181208053691278e-05,
1531
+ "loss": 2.1172,
1532
+ "step": 254
1533
+ },
1534
+ {
1535
+ "epoch": 1.71,
1536
+ "learning_rate": 1.3154362416107385e-05,
1537
+ "loss": 2.1309,
1538
+ "step": 255
1539
+ },
1540
+ {
1541
+ "epoch": 1.72,
1542
+ "learning_rate": 1.312751677852349e-05,
1543
+ "loss": 2.1309,
1544
+ "step": 256
1545
+ },
1546
+ {
1547
+ "epoch": 1.72,
1548
+ "learning_rate": 1.3100671140939598e-05,
1549
+ "loss": 2.1074,
1550
+ "step": 257
1551
+ },
1552
+ {
1553
+ "epoch": 1.73,
1554
+ "learning_rate": 1.3073825503355706e-05,
1555
+ "loss": 2.0977,
1556
+ "step": 258
1557
+ },
1558
+ {
1559
+ "epoch": 1.74,
1560
+ "learning_rate": 1.3046979865771815e-05,
1561
+ "loss": 2.1133,
1562
+ "step": 259
1563
+ },
1564
+ {
1565
+ "epoch": 1.74,
1566
+ "learning_rate": 1.302013422818792e-05,
1567
+ "loss": 2.123,
1568
+ "step": 260
1569
+ },
1570
+ {
1571
+ "epoch": 1.75,
1572
+ "learning_rate": 1.2993288590604028e-05,
1573
+ "loss": 2.127,
1574
+ "step": 261
1575
+ },
1576
+ {
1577
+ "epoch": 1.76,
1578
+ "learning_rate": 1.2966442953020135e-05,
1579
+ "loss": 2.1191,
1580
+ "step": 262
1581
+ },
1582
+ {
1583
+ "epoch": 1.77,
1584
+ "learning_rate": 1.2939597315436243e-05,
1585
+ "loss": 2.0938,
1586
+ "step": 263
1587
+ },
1588
+ {
1589
+ "epoch": 1.77,
1590
+ "learning_rate": 1.2912751677852352e-05,
1591
+ "loss": 2.1367,
1592
+ "step": 264
1593
+ },
1594
+ {
1595
+ "epoch": 1.78,
1596
+ "learning_rate": 1.2885906040268457e-05,
1597
+ "loss": 2.1094,
1598
+ "step": 265
1599
+ },
1600
+ {
1601
+ "epoch": 1.79,
1602
+ "learning_rate": 1.2859060402684565e-05,
1603
+ "loss": 2.1035,
1604
+ "step": 266
1605
+ },
1606
+ {
1607
+ "epoch": 1.79,
1608
+ "learning_rate": 1.2832214765100672e-05,
1609
+ "loss": 2.1074,
1610
+ "step": 267
1611
+ },
1612
+ {
1613
+ "epoch": 1.8,
1614
+ "learning_rate": 1.280536912751678e-05,
1615
+ "loss": 2.1348,
1616
+ "step": 268
1617
+ },
1618
+ {
1619
+ "epoch": 1.81,
1620
+ "learning_rate": 1.2778523489932889e-05,
1621
+ "loss": 2.1309,
1622
+ "step": 269
1623
+ },
1624
+ {
1625
+ "epoch": 1.81,
1626
+ "learning_rate": 1.2751677852348994e-05,
1627
+ "loss": 2.1055,
1628
+ "step": 270
1629
+ },
1630
+ {
1631
+ "epoch": 1.82,
1632
+ "learning_rate": 1.2724832214765102e-05,
1633
+ "loss": 2.125,
1634
+ "step": 271
1635
+ },
1636
+ {
1637
+ "epoch": 1.83,
1638
+ "learning_rate": 1.269798657718121e-05,
1639
+ "loss": 2.0957,
1640
+ "step": 272
1641
+ },
1642
+ {
1643
+ "epoch": 1.83,
1644
+ "learning_rate": 1.2671140939597317e-05,
1645
+ "loss": 2.1309,
1646
+ "step": 273
1647
+ },
1648
+ {
1649
+ "epoch": 1.84,
1650
+ "learning_rate": 1.2644295302013422e-05,
1651
+ "loss": 2.1191,
1652
+ "step": 274
1653
+ },
1654
+ {
1655
+ "epoch": 1.85,
1656
+ "learning_rate": 1.2617449664429532e-05,
1657
+ "loss": 2.1211,
1658
+ "step": 275
1659
+ },
1660
+ {
1661
+ "epoch": 1.85,
1662
+ "learning_rate": 1.2590604026845639e-05,
1663
+ "loss": 2.1348,
1664
+ "step": 276
1665
+ },
1666
+ {
1667
+ "epoch": 1.86,
1668
+ "learning_rate": 1.2563758389261746e-05,
1669
+ "loss": 2.1191,
1670
+ "step": 277
1671
+ },
1672
+ {
1673
+ "epoch": 1.87,
1674
+ "learning_rate": 1.2536912751677854e-05,
1675
+ "loss": 2.1035,
1676
+ "step": 278
1677
+ },
1678
+ {
1679
+ "epoch": 1.87,
1680
+ "learning_rate": 1.251006711409396e-05,
1681
+ "loss": 2.1094,
1682
+ "step": 279
1683
+ },
1684
+ {
1685
+ "epoch": 1.88,
1686
+ "learning_rate": 1.2483221476510069e-05,
1687
+ "loss": 2.1133,
1688
+ "step": 280
1689
+ },
1690
+ {
1691
+ "epoch": 1.89,
1692
+ "learning_rate": 1.2456375838926176e-05,
1693
+ "loss": 2.1191,
1694
+ "step": 281
1695
+ },
1696
+ {
1697
+ "epoch": 1.89,
1698
+ "learning_rate": 1.2429530201342283e-05,
1699
+ "loss": 2.082,
1700
+ "step": 282
1701
+ },
1702
+ {
1703
+ "epoch": 1.9,
1704
+ "learning_rate": 1.2402684563758389e-05,
1705
+ "loss": 2.1133,
1706
+ "step": 283
1707
+ },
1708
+ {
1709
+ "epoch": 1.91,
1710
+ "learning_rate": 1.2375838926174497e-05,
1711
+ "loss": 2.1113,
1712
+ "step": 284
1713
+ },
1714
+ {
1715
+ "epoch": 1.91,
1716
+ "learning_rate": 1.2348993288590606e-05,
1717
+ "loss": 2.0996,
1718
+ "step": 285
1719
+ },
1720
+ {
1721
+ "epoch": 1.92,
1722
+ "learning_rate": 1.2322147651006713e-05,
1723
+ "loss": 2.1094,
1724
+ "step": 286
1725
+ },
1726
+ {
1727
+ "epoch": 1.93,
1728
+ "learning_rate": 1.229530201342282e-05,
1729
+ "loss": 2.0918,
1730
+ "step": 287
1731
+ },
1732
+ {
1733
+ "epoch": 1.93,
1734
+ "learning_rate": 1.2268456375838926e-05,
1735
+ "loss": 2.0938,
1736
+ "step": 288
1737
+ },
1738
+ {
1739
+ "epoch": 1.94,
1740
+ "learning_rate": 1.2241610738255034e-05,
1741
+ "loss": 2.1152,
1742
+ "step": 289
1743
+ },
1744
+ {
1745
+ "epoch": 1.95,
1746
+ "learning_rate": 1.2214765100671143e-05,
1747
+ "loss": 2.1016,
1748
+ "step": 290
1749
+ },
1750
+ {
1751
+ "epoch": 1.95,
1752
+ "learning_rate": 1.218791946308725e-05,
1753
+ "loss": 2.1152,
1754
+ "step": 291
1755
+ },
1756
+ {
1757
+ "epoch": 1.96,
1758
+ "learning_rate": 1.2161073825503358e-05,
1759
+ "loss": 2.1016,
1760
+ "step": 292
1761
+ },
1762
+ {
1763
+ "epoch": 1.97,
1764
+ "learning_rate": 1.2134228187919463e-05,
1765
+ "loss": 2.1055,
1766
+ "step": 293
1767
+ },
1768
+ {
1769
+ "epoch": 1.97,
1770
+ "learning_rate": 1.210738255033557e-05,
1771
+ "loss": 2.0938,
1772
+ "step": 294
1773
+ },
1774
+ {
1775
+ "epoch": 1.98,
1776
+ "learning_rate": 1.208053691275168e-05,
1777
+ "loss": 2.1211,
1778
+ "step": 295
1779
+ },
1780
+ {
1781
+ "epoch": 1.99,
1782
+ "learning_rate": 1.2053691275167787e-05,
1783
+ "loss": 2.1016,
1784
+ "step": 296
1785
+ },
1786
+ {
1787
+ "epoch": 1.99,
1788
+ "learning_rate": 1.2026845637583893e-05,
1789
+ "loss": 2.1152,
1790
+ "step": 297
1791
+ },
1792
+ {
1793
+ "epoch": 2.0,
1794
+ "learning_rate": 1.2e-05,
1795
+ "loss": 2.0957,
1796
+ "step": 298
1797
+ },
1798
+ {
1799
+ "epoch": 2.01,
1800
+ "learning_rate": 1.1973154362416108e-05,
1801
+ "loss": 2.0938,
1802
+ "step": 299
1803
+ },
1804
+ {
1805
+ "epoch": 2.01,
1806
+ "learning_rate": 1.1946308724832217e-05,
1807
+ "loss": 2.0859,
1808
+ "step": 300
1809
+ },
1810
+ {
1811
+ "epoch": 2.02,
1812
+ "learning_rate": 1.1919463087248324e-05,
1813
+ "loss": 2.0879,
1814
+ "step": 301
1815
+ },
1816
+ {
1817
+ "epoch": 2.03,
1818
+ "learning_rate": 1.189261744966443e-05,
1819
+ "loss": 2.082,
1820
+ "step": 302
1821
+ },
1822
+ {
1823
+ "epoch": 2.03,
1824
+ "learning_rate": 1.1865771812080537e-05,
1825
+ "loss": 2.1074,
1826
+ "step": 303
1827
+ },
1828
+ {
1829
+ "epoch": 2.04,
1830
+ "learning_rate": 1.1838926174496645e-05,
1831
+ "loss": 2.1016,
1832
+ "step": 304
1833
+ },
1834
+ {
1835
+ "epoch": 2.05,
1836
+ "learning_rate": 1.1812080536912754e-05,
1837
+ "loss": 2.0957,
1838
+ "step": 305
1839
+ },
1840
+ {
1841
+ "epoch": 2.05,
1842
+ "learning_rate": 1.178523489932886e-05,
1843
+ "loss": 2.0684,
1844
+ "step": 306
1845
+ },
1846
+ {
1847
+ "epoch": 2.06,
1848
+ "learning_rate": 1.1758389261744967e-05,
1849
+ "loss": 2.0898,
1850
+ "step": 307
1851
+ },
1852
+ {
1853
+ "epoch": 2.07,
1854
+ "learning_rate": 1.1731543624161074e-05,
1855
+ "loss": 2.0977,
1856
+ "step": 308
1857
+ },
1858
+ {
1859
+ "epoch": 2.07,
1860
+ "learning_rate": 1.1704697986577182e-05,
1861
+ "loss": 2.1074,
1862
+ "step": 309
1863
+ },
1864
+ {
1865
+ "epoch": 2.08,
1866
+ "learning_rate": 1.1677852348993291e-05,
1867
+ "loss": 2.125,
1868
+ "step": 310
1869
+ },
1870
+ {
1871
+ "epoch": 2.09,
1872
+ "learning_rate": 1.1651006711409397e-05,
1873
+ "loss": 2.1016,
1874
+ "step": 311
1875
+ },
1876
+ {
1877
+ "epoch": 2.09,
1878
+ "learning_rate": 1.1624161073825504e-05,
1879
+ "loss": 2.1035,
1880
+ "step": 312
1881
+ },
1882
+ {
1883
+ "epoch": 2.1,
1884
+ "learning_rate": 1.1597315436241611e-05,
1885
+ "loss": 2.084,
1886
+ "step": 313
1887
+ },
1888
+ {
1889
+ "epoch": 2.11,
1890
+ "learning_rate": 1.1570469798657719e-05,
1891
+ "loss": 2.0762,
1892
+ "step": 314
1893
+ },
1894
+ {
1895
+ "epoch": 2.11,
1896
+ "learning_rate": 1.1543624161073828e-05,
1897
+ "loss": 2.1133,
1898
+ "step": 315
1899
+ },
1900
+ {
1901
+ "epoch": 2.12,
1902
+ "learning_rate": 1.1516778523489934e-05,
1903
+ "loss": 2.1016,
1904
+ "step": 316
1905
+ },
1906
+ {
1907
+ "epoch": 2.13,
1908
+ "learning_rate": 1.1489932885906041e-05,
1909
+ "loss": 2.0742,
1910
+ "step": 317
1911
+ },
1912
+ {
1913
+ "epoch": 2.13,
1914
+ "learning_rate": 1.1463087248322149e-05,
1915
+ "loss": 2.0771,
1916
+ "step": 318
1917
+ },
1918
+ {
1919
+ "epoch": 2.14,
1920
+ "learning_rate": 1.1436241610738256e-05,
1921
+ "loss": 2.0879,
1922
+ "step": 319
1923
+ },
1924
+ {
1925
+ "epoch": 2.15,
1926
+ "learning_rate": 1.1409395973154362e-05,
1927
+ "loss": 2.0957,
1928
+ "step": 320
1929
+ },
1930
+ {
1931
+ "epoch": 2.15,
1932
+ "learning_rate": 1.138255033557047e-05,
1933
+ "loss": 2.0879,
1934
+ "step": 321
1935
+ },
1936
+ {
1937
+ "epoch": 2.16,
1938
+ "learning_rate": 1.1355704697986578e-05,
1939
+ "loss": 2.082,
1940
+ "step": 322
1941
+ },
1942
+ {
1943
+ "epoch": 2.17,
1944
+ "learning_rate": 1.1328859060402686e-05,
1945
+ "loss": 2.0996,
1946
+ "step": 323
1947
+ },
1948
+ {
1949
+ "epoch": 2.17,
1950
+ "learning_rate": 1.1302013422818795e-05,
1951
+ "loss": 2.0801,
1952
+ "step": 324
1953
+ },
1954
+ {
1955
+ "epoch": 2.18,
1956
+ "learning_rate": 1.1275167785234899e-05,
1957
+ "loss": 2.1113,
1958
+ "step": 325
1959
+ },
1960
+ {
1961
+ "epoch": 2.19,
1962
+ "learning_rate": 1.1248322147651008e-05,
1963
+ "loss": 2.0996,
1964
+ "step": 326
1965
+ },
1966
+ {
1967
+ "epoch": 2.19,
1968
+ "learning_rate": 1.1221476510067115e-05,
1969
+ "loss": 2.0762,
1970
+ "step": 327
1971
+ },
1972
+ {
1973
+ "epoch": 2.2,
1974
+ "learning_rate": 1.1194630872483223e-05,
1975
+ "loss": 2.1094,
1976
+ "step": 328
1977
+ },
1978
+ {
1979
+ "epoch": 2.21,
1980
+ "learning_rate": 1.1167785234899328e-05,
1981
+ "loss": 2.082,
1982
+ "step": 329
1983
+ },
1984
+ {
1985
+ "epoch": 2.21,
1986
+ "learning_rate": 1.1140939597315436e-05,
1987
+ "loss": 2.0527,
1988
+ "step": 330
1989
+ },
1990
+ {
1991
+ "epoch": 2.22,
1992
+ "learning_rate": 1.1114093959731545e-05,
1993
+ "loss": 2.0859,
1994
+ "step": 331
1995
+ },
1996
+ {
1997
+ "epoch": 2.23,
1998
+ "learning_rate": 1.1087248322147652e-05,
1999
+ "loss": 2.0508,
2000
+ "step": 332
2001
+ },
2002
+ {
2003
+ "epoch": 2.23,
2004
+ "learning_rate": 1.106040268456376e-05,
2005
+ "loss": 2.0977,
2006
+ "step": 333
2007
+ },
2008
+ {
2009
+ "epoch": 2.24,
2010
+ "learning_rate": 1.1033557046979865e-05,
2011
+ "loss": 2.0889,
2012
+ "step": 334
2013
+ },
2014
+ {
2015
+ "epoch": 2.25,
2016
+ "learning_rate": 1.1006711409395975e-05,
2017
+ "loss": 2.082,
2018
+ "step": 335
2019
+ },
2020
+ {
2021
+ "epoch": 2.26,
2022
+ "learning_rate": 1.0979865771812082e-05,
2023
+ "loss": 2.0742,
2024
+ "step": 336
2025
+ },
2026
+ {
2027
+ "epoch": 2.26,
2028
+ "learning_rate": 1.095302013422819e-05,
2029
+ "loss": 2.0879,
2030
+ "step": 337
2031
+ },
2032
+ {
2033
+ "epoch": 2.27,
2034
+ "learning_rate": 1.0926174496644297e-05,
2035
+ "loss": 2.082,
2036
+ "step": 338
2037
+ },
2038
+ {
2039
+ "epoch": 2.28,
2040
+ "learning_rate": 1.0899328859060403e-05,
2041
+ "loss": 2.0879,
2042
+ "step": 339
2043
+ },
2044
+ {
2045
+ "epoch": 2.28,
2046
+ "learning_rate": 1.0872483221476512e-05,
2047
+ "loss": 2.082,
2048
+ "step": 340
2049
+ },
2050
+ {
2051
+ "epoch": 2.29,
2052
+ "learning_rate": 1.0845637583892619e-05,
2053
+ "loss": 2.082,
2054
+ "step": 341
2055
+ },
2056
+ {
2057
+ "epoch": 2.3,
2058
+ "learning_rate": 1.0818791946308726e-05,
2059
+ "loss": 2.0547,
2060
+ "step": 342
2061
+ },
2062
+ {
2063
+ "epoch": 2.3,
2064
+ "learning_rate": 1.0791946308724832e-05,
2065
+ "loss": 2.0645,
2066
+ "step": 343
2067
+ },
2068
+ {
2069
+ "epoch": 2.31,
2070
+ "learning_rate": 1.076510067114094e-05,
2071
+ "loss": 2.0703,
2072
+ "step": 344
2073
+ },
2074
+ {
2075
+ "epoch": 2.32,
2076
+ "learning_rate": 1.0738255033557049e-05,
2077
+ "loss": 2.084,
2078
+ "step": 345
2079
+ },
2080
+ {
2081
+ "epoch": 2.32,
2082
+ "learning_rate": 1.0711409395973156e-05,
2083
+ "loss": 2.0762,
2084
+ "step": 346
2085
+ },
2086
+ {
2087
+ "epoch": 2.33,
2088
+ "learning_rate": 1.0684563758389264e-05,
2089
+ "loss": 2.0723,
2090
+ "step": 347
2091
+ },
2092
+ {
2093
+ "epoch": 2.34,
2094
+ "learning_rate": 1.065771812080537e-05,
2095
+ "loss": 2.0723,
2096
+ "step": 348
2097
+ },
2098
+ {
2099
+ "epoch": 2.34,
2100
+ "learning_rate": 1.0630872483221477e-05,
2101
+ "loss": 2.0723,
2102
+ "step": 349
2103
+ },
2104
+ {
2105
+ "epoch": 2.35,
2106
+ "learning_rate": 1.0604026845637586e-05,
2107
+ "loss": 2.084,
2108
+ "step": 350
2109
+ },
2110
+ {
2111
+ "epoch": 2.36,
2112
+ "learning_rate": 1.0577181208053693e-05,
2113
+ "loss": 2.0781,
2114
+ "step": 351
2115
+ },
2116
+ {
2117
+ "epoch": 2.36,
2118
+ "learning_rate": 1.0550335570469799e-05,
2119
+ "loss": 2.0742,
2120
+ "step": 352
2121
+ },
2122
+ {
2123
+ "epoch": 2.37,
2124
+ "learning_rate": 1.0523489932885906e-05,
2125
+ "loss": 2.0869,
2126
+ "step": 353
2127
+ },
2128
+ {
2129
+ "epoch": 2.38,
2130
+ "learning_rate": 1.0496644295302014e-05,
2131
+ "loss": 2.0625,
2132
+ "step": 354
2133
+ },
2134
+ {
2135
+ "epoch": 2.38,
2136
+ "learning_rate": 1.0469798657718123e-05,
2137
+ "loss": 2.0859,
2138
+ "step": 355
2139
+ },
2140
+ {
2141
+ "epoch": 2.39,
2142
+ "learning_rate": 1.044295302013423e-05,
2143
+ "loss": 2.0801,
2144
+ "step": 356
2145
+ },
2146
+ {
2147
+ "epoch": 2.4,
2148
+ "learning_rate": 1.0416107382550336e-05,
2149
+ "loss": 2.0566,
2150
+ "step": 357
2151
+ },
2152
+ {
2153
+ "epoch": 2.4,
2154
+ "learning_rate": 1.0389261744966443e-05,
2155
+ "loss": 2.0859,
2156
+ "step": 358
2157
+ },
2158
+ {
2159
+ "epoch": 2.41,
2160
+ "learning_rate": 1.036241610738255e-05,
2161
+ "loss": 2.084,
2162
+ "step": 359
2163
+ },
2164
+ {
2165
+ "epoch": 2.42,
2166
+ "learning_rate": 1.033557046979866e-05,
2167
+ "loss": 2.0762,
2168
+ "step": 360
2169
+ },
2170
+ {
2171
+ "epoch": 2.42,
2172
+ "learning_rate": 1.0308724832214767e-05,
2173
+ "loss": 2.0703,
2174
+ "step": 361
2175
+ },
2176
+ {
2177
+ "epoch": 2.43,
2178
+ "learning_rate": 1.0281879194630873e-05,
2179
+ "loss": 2.0762,
2180
+ "step": 362
2181
+ },
2182
+ {
2183
+ "epoch": 2.44,
2184
+ "learning_rate": 1.025503355704698e-05,
2185
+ "loss": 2.0762,
2186
+ "step": 363
2187
+ },
2188
+ {
2189
+ "epoch": 2.44,
2190
+ "learning_rate": 1.0228187919463088e-05,
2191
+ "loss": 2.0664,
2192
+ "step": 364
2193
+ },
2194
+ {
2195
+ "epoch": 2.45,
2196
+ "learning_rate": 1.0201342281879197e-05,
2197
+ "loss": 2.0586,
2198
+ "step": 365
2199
+ },
2200
+ {
2201
+ "epoch": 2.46,
2202
+ "learning_rate": 1.0174496644295303e-05,
2203
+ "loss": 2.0664,
2204
+ "step": 366
2205
+ },
2206
+ {
2207
+ "epoch": 2.46,
2208
+ "learning_rate": 1.014765100671141e-05,
2209
+ "loss": 2.0684,
2210
+ "step": 367
2211
+ },
2212
+ {
2213
+ "epoch": 2.47,
2214
+ "learning_rate": 1.0120805369127517e-05,
2215
+ "loss": 2.0605,
2216
+ "step": 368
2217
+ },
2218
+ {
2219
+ "epoch": 2.48,
2220
+ "learning_rate": 1.0093959731543625e-05,
2221
+ "loss": 2.0625,
2222
+ "step": 369
2223
+ },
2224
+ {
2225
+ "epoch": 2.48,
2226
+ "learning_rate": 1.0067114093959734e-05,
2227
+ "loss": 2.0859,
2228
+ "step": 370
2229
+ },
2230
+ {
2231
+ "epoch": 2.49,
2232
+ "learning_rate": 1.004026845637584e-05,
2233
+ "loss": 2.0527,
2234
+ "step": 371
2235
+ },
2236
+ {
2237
+ "epoch": 2.5,
2238
+ "learning_rate": 1.0013422818791947e-05,
2239
+ "loss": 2.0703,
2240
+ "step": 372
2241
+ },
2242
+ {
2243
+ "epoch": 2.5,
2244
+ "learning_rate": 9.986577181208055e-06,
2245
+ "loss": 2.0684,
2246
+ "step": 373
2247
+ },
2248
+ {
2249
+ "epoch": 2.51,
2250
+ "learning_rate": 9.959731543624162e-06,
2251
+ "loss": 2.0488,
2252
+ "step": 374
2253
+ },
2254
+ {
2255
+ "epoch": 2.52,
2256
+ "learning_rate": 9.93288590604027e-06,
2257
+ "loss": 2.0762,
2258
+ "step": 375
2259
+ },
2260
+ {
2261
+ "epoch": 2.52,
2262
+ "learning_rate": 9.906040268456377e-06,
2263
+ "loss": 2.0508,
2264
+ "step": 376
2265
+ },
2266
+ {
2267
+ "epoch": 2.53,
2268
+ "learning_rate": 9.879194630872484e-06,
2269
+ "loss": 2.0605,
2270
+ "step": 377
2271
+ },
2272
+ {
2273
+ "epoch": 2.54,
2274
+ "learning_rate": 9.852348993288592e-06,
2275
+ "loss": 2.0508,
2276
+ "step": 378
2277
+ },
2278
+ {
2279
+ "epoch": 2.54,
2280
+ "learning_rate": 9.825503355704699e-06,
2281
+ "loss": 2.0508,
2282
+ "step": 379
2283
+ },
2284
+ {
2285
+ "epoch": 2.55,
2286
+ "learning_rate": 9.798657718120806e-06,
2287
+ "loss": 2.0664,
2288
+ "step": 380
2289
+ },
2290
+ {
2291
+ "epoch": 2.56,
2292
+ "learning_rate": 9.771812080536914e-06,
2293
+ "loss": 2.0664,
2294
+ "step": 381
2295
+ },
2296
+ {
2297
+ "epoch": 2.56,
2298
+ "learning_rate": 9.744966442953021e-06,
2299
+ "loss": 2.0859,
2300
+ "step": 382
2301
+ },
2302
+ {
2303
+ "epoch": 2.57,
2304
+ "learning_rate": 9.718120805369129e-06,
2305
+ "loss": 2.0859,
2306
+ "step": 383
2307
+ },
2308
+ {
2309
+ "epoch": 2.58,
2310
+ "learning_rate": 9.691275167785236e-06,
2311
+ "loss": 2.0566,
2312
+ "step": 384
2313
+ },
2314
+ {
2315
+ "epoch": 2.58,
2316
+ "learning_rate": 9.664429530201343e-06,
2317
+ "loss": 2.0703,
2318
+ "step": 385
2319
+ },
2320
+ {
2321
+ "epoch": 2.59,
2322
+ "learning_rate": 9.637583892617451e-06,
2323
+ "loss": 2.0645,
2324
+ "step": 386
2325
+ },
2326
+ {
2327
+ "epoch": 2.6,
2328
+ "learning_rate": 9.610738255033558e-06,
2329
+ "loss": 2.0645,
2330
+ "step": 387
2331
+ },
2332
+ {
2333
+ "epoch": 2.6,
2334
+ "learning_rate": 9.583892617449666e-06,
2335
+ "loss": 2.0566,
2336
+ "step": 388
2337
+ },
2338
+ {
2339
+ "epoch": 2.61,
2340
+ "learning_rate": 9.557046979865773e-06,
2341
+ "loss": 2.0605,
2342
+ "step": 389
2343
+ },
2344
+ {
2345
+ "epoch": 2.62,
2346
+ "learning_rate": 9.530201342281879e-06,
2347
+ "loss": 2.0625,
2348
+ "step": 390
2349
+ },
2350
+ {
2351
+ "epoch": 2.62,
2352
+ "learning_rate": 9.503355704697988e-06,
2353
+ "loss": 2.0684,
2354
+ "step": 391
2355
+ },
2356
+ {
2357
+ "epoch": 2.63,
2358
+ "learning_rate": 9.476510067114095e-06,
2359
+ "loss": 2.0566,
2360
+ "step": 392
2361
+ },
2362
+ {
2363
+ "epoch": 2.64,
2364
+ "learning_rate": 9.449664429530203e-06,
2365
+ "loss": 2.0547,
2366
+ "step": 393
2367
+ },
2368
+ {
2369
+ "epoch": 2.64,
2370
+ "learning_rate": 9.42281879194631e-06,
2371
+ "loss": 2.0547,
2372
+ "step": 394
2373
+ },
2374
+ {
2375
+ "epoch": 2.65,
2376
+ "learning_rate": 9.395973154362416e-06,
2377
+ "loss": 2.0742,
2378
+ "step": 395
2379
+ },
2380
+ {
2381
+ "epoch": 2.66,
2382
+ "learning_rate": 9.369127516778525e-06,
2383
+ "loss": 2.084,
2384
+ "step": 396
2385
+ },
2386
+ {
2387
+ "epoch": 2.66,
2388
+ "learning_rate": 9.34228187919463e-06,
2389
+ "loss": 2.0518,
2390
+ "step": 397
2391
+ },
2392
+ {
2393
+ "epoch": 2.67,
2394
+ "learning_rate": 9.31543624161074e-06,
2395
+ "loss": 2.0439,
2396
+ "step": 398
2397
+ },
2398
+ {
2399
+ "epoch": 2.68,
2400
+ "learning_rate": 9.288590604026846e-06,
2401
+ "loss": 2.0645,
2402
+ "step": 399
2403
+ },
2404
+ {
2405
+ "epoch": 2.68,
2406
+ "learning_rate": 9.261744966442953e-06,
2407
+ "loss": 2.0605,
2408
+ "step": 400
2409
+ },
2410
+ {
2411
+ "epoch": 2.69,
2412
+ "learning_rate": 9.234899328859062e-06,
2413
+ "loss": 2.0664,
2414
+ "step": 401
2415
+ },
2416
+ {
2417
+ "epoch": 2.7,
2418
+ "learning_rate": 9.208053691275168e-06,
2419
+ "loss": 2.0645,
2420
+ "step": 402
2421
+ },
2422
+ {
2423
+ "epoch": 2.7,
2424
+ "learning_rate": 9.181208053691277e-06,
2425
+ "loss": 2.0527,
2426
+ "step": 403
2427
+ },
2428
+ {
2429
+ "epoch": 2.71,
2430
+ "learning_rate": 9.154362416107383e-06,
2431
+ "loss": 2.0605,
2432
+ "step": 404
2433
+ },
2434
+ {
2435
+ "epoch": 2.72,
2436
+ "learning_rate": 9.12751677852349e-06,
2437
+ "loss": 2.0381,
2438
+ "step": 405
2439
+ },
2440
+ {
2441
+ "epoch": 2.72,
2442
+ "learning_rate": 9.100671140939597e-06,
2443
+ "loss": 2.0352,
2444
+ "step": 406
2445
+ },
2446
+ {
2447
+ "epoch": 2.73,
2448
+ "learning_rate": 9.073825503355705e-06,
2449
+ "loss": 2.0352,
2450
+ "step": 407
2451
+ },
2452
+ {
2453
+ "epoch": 2.74,
2454
+ "learning_rate": 9.046979865771814e-06,
2455
+ "loss": 2.0439,
2456
+ "step": 408
2457
+ },
2458
+ {
2459
+ "epoch": 2.74,
2460
+ "learning_rate": 9.02013422818792e-06,
2461
+ "loss": 2.0332,
2462
+ "step": 409
2463
+ },
2464
+ {
2465
+ "epoch": 2.75,
2466
+ "learning_rate": 8.993288590604027e-06,
2467
+ "loss": 2.0566,
2468
+ "step": 410
2469
+ },
2470
+ {
2471
+ "epoch": 2.76,
2472
+ "learning_rate": 8.966442953020134e-06,
2473
+ "loss": 2.0625,
2474
+ "step": 411
2475
+ },
2476
+ {
2477
+ "epoch": 2.77,
2478
+ "learning_rate": 8.939597315436242e-06,
2479
+ "loss": 2.0488,
2480
+ "step": 412
2481
+ },
2482
+ {
2483
+ "epoch": 2.77,
2484
+ "learning_rate": 8.91275167785235e-06,
2485
+ "loss": 2.0488,
2486
+ "step": 413
2487
+ },
2488
+ {
2489
+ "epoch": 2.78,
2490
+ "learning_rate": 8.885906040268457e-06,
2491
+ "loss": 2.0664,
2492
+ "step": 414
2493
+ },
2494
+ {
2495
+ "epoch": 2.79,
2496
+ "learning_rate": 8.859060402684566e-06,
2497
+ "loss": 2.0645,
2498
+ "step": 415
2499
+ },
2500
+ {
2501
+ "epoch": 2.79,
2502
+ "learning_rate": 8.832214765100672e-06,
2503
+ "loss": 2.0625,
2504
+ "step": 416
2505
+ },
2506
+ {
2507
+ "epoch": 2.8,
2508
+ "learning_rate": 8.805369127516779e-06,
2509
+ "loss": 2.0547,
2510
+ "step": 417
2511
+ },
2512
+ {
2513
+ "epoch": 2.81,
2514
+ "learning_rate": 8.778523489932886e-06,
2515
+ "loss": 2.0645,
2516
+ "step": 418
2517
+ },
2518
+ {
2519
+ "epoch": 2.81,
2520
+ "learning_rate": 8.751677852348994e-06,
2521
+ "loss": 2.0459,
2522
+ "step": 419
2523
+ },
2524
+ {
2525
+ "epoch": 2.82,
2526
+ "learning_rate": 8.724832214765101e-06,
2527
+ "loss": 2.043,
2528
+ "step": 420
2529
+ },
2530
+ {
2531
+ "epoch": 2.83,
2532
+ "learning_rate": 8.697986577181209e-06,
2533
+ "loss": 2.0645,
2534
+ "step": 421
2535
+ },
2536
+ {
2537
+ "epoch": 2.83,
2538
+ "learning_rate": 8.671140939597316e-06,
2539
+ "loss": 2.0605,
2540
+ "step": 422
2541
+ },
2542
+ {
2543
+ "epoch": 2.84,
2544
+ "learning_rate": 8.644295302013423e-06,
2545
+ "loss": 2.0684,
2546
+ "step": 423
2547
+ },
2548
+ {
2549
+ "epoch": 2.85,
2550
+ "learning_rate": 8.617449664429531e-06,
2551
+ "loss": 2.0508,
2552
+ "step": 424
2553
+ },
2554
+ {
2555
+ "epoch": 2.85,
2556
+ "learning_rate": 8.590604026845638e-06,
2557
+ "loss": 2.0703,
2558
+ "step": 425
2559
+ },
2560
+ {
2561
+ "epoch": 2.86,
2562
+ "learning_rate": 8.563758389261746e-06,
2563
+ "loss": 2.0488,
2564
+ "step": 426
2565
+ },
2566
+ {
2567
+ "epoch": 2.87,
2568
+ "learning_rate": 8.536912751677853e-06,
2569
+ "loss": 2.0381,
2570
+ "step": 427
2571
+ },
2572
+ {
2573
+ "epoch": 2.87,
2574
+ "learning_rate": 8.51006711409396e-06,
2575
+ "loss": 2.04,
2576
+ "step": 428
2577
+ },
2578
+ {
2579
+ "epoch": 2.88,
2580
+ "learning_rate": 8.483221476510068e-06,
2581
+ "loss": 2.0566,
2582
+ "step": 429
2583
+ },
2584
+ {
2585
+ "epoch": 2.89,
2586
+ "learning_rate": 8.456375838926175e-06,
2587
+ "loss": 2.0547,
2588
+ "step": 430
2589
+ },
2590
+ {
2591
+ "epoch": 2.89,
2592
+ "learning_rate": 8.429530201342283e-06,
2593
+ "loss": 2.0449,
2594
+ "step": 431
2595
+ },
2596
+ {
2597
+ "epoch": 2.9,
2598
+ "learning_rate": 8.40268456375839e-06,
2599
+ "loss": 2.0352,
2600
+ "step": 432
2601
+ },
2602
+ {
2603
+ "epoch": 2.91,
2604
+ "learning_rate": 8.375838926174498e-06,
2605
+ "loss": 2.0605,
2606
+ "step": 433
2607
+ },
2608
+ {
2609
+ "epoch": 2.91,
2610
+ "learning_rate": 8.348993288590605e-06,
2611
+ "loss": 2.0684,
2612
+ "step": 434
2613
+ },
2614
+ {
2615
+ "epoch": 2.92,
2616
+ "learning_rate": 8.322147651006712e-06,
2617
+ "loss": 2.0508,
2618
+ "step": 435
2619
+ },
2620
+ {
2621
+ "epoch": 2.93,
2622
+ "learning_rate": 8.29530201342282e-06,
2623
+ "loss": 2.0664,
2624
+ "step": 436
2625
+ },
2626
+ {
2627
+ "epoch": 2.93,
2628
+ "learning_rate": 8.268456375838927e-06,
2629
+ "loss": 2.04,
2630
+ "step": 437
2631
+ },
2632
+ {
2633
+ "epoch": 2.94,
2634
+ "learning_rate": 8.241610738255035e-06,
2635
+ "loss": 2.0371,
2636
+ "step": 438
2637
+ },
2638
+ {
2639
+ "epoch": 2.95,
2640
+ "learning_rate": 8.214765100671142e-06,
2641
+ "loss": 2.0361,
2642
+ "step": 439
2643
+ },
2644
+ {
2645
+ "epoch": 2.95,
2646
+ "learning_rate": 8.18791946308725e-06,
2647
+ "loss": 2.0645,
2648
+ "step": 440
2649
+ },
2650
+ {
2651
+ "epoch": 2.96,
2652
+ "learning_rate": 8.161073825503357e-06,
2653
+ "loss": 2.0527,
2654
+ "step": 441
2655
+ },
2656
+ {
2657
+ "epoch": 2.97,
2658
+ "learning_rate": 8.134228187919464e-06,
2659
+ "loss": 2.0518,
2660
+ "step": 442
2661
+ },
2662
+ {
2663
+ "epoch": 2.97,
2664
+ "learning_rate": 8.10738255033557e-06,
2665
+ "loss": 2.0508,
2666
+ "step": 443
2667
+ },
2668
+ {
2669
+ "epoch": 2.98,
2670
+ "learning_rate": 8.080536912751679e-06,
2671
+ "loss": 2.0566,
2672
+ "step": 444
2673
+ },
2674
+ {
2675
+ "epoch": 2.99,
2676
+ "learning_rate": 8.053691275167785e-06,
2677
+ "loss": 2.042,
2678
+ "step": 445
2679
+ },
2680
+ {
2681
+ "epoch": 2.99,
2682
+ "learning_rate": 8.026845637583894e-06,
2683
+ "loss": 2.0527,
2684
+ "step": 446
2685
+ },
2686
+ {
2687
+ "epoch": 3.0,
2688
+ "learning_rate": 8.000000000000001e-06,
2689
+ "loss": 2.0342,
2690
+ "step": 447
2691
+ },
2692
+ {
2693
+ "epoch": 3.01,
2694
+ "learning_rate": 7.973154362416107e-06,
2695
+ "loss": 2.0391,
2696
+ "step": 448
2697
+ },
2698
+ {
2699
+ "epoch": 3.01,
2700
+ "learning_rate": 7.946308724832216e-06,
2701
+ "loss": 2.0361,
2702
+ "step": 449
2703
+ },
2704
+ {
2705
+ "epoch": 3.02,
2706
+ "learning_rate": 7.919463087248322e-06,
2707
+ "loss": 2.0645,
2708
+ "step": 450
2709
+ },
2710
+ {
2711
+ "epoch": 3.03,
2712
+ "learning_rate": 7.892617449664431e-06,
2713
+ "loss": 2.041,
2714
+ "step": 451
2715
+ },
2716
+ {
2717
+ "epoch": 3.03,
2718
+ "learning_rate": 7.865771812080537e-06,
2719
+ "loss": 2.0312,
2720
+ "step": 452
2721
+ },
2722
+ {
2723
+ "epoch": 3.04,
2724
+ "learning_rate": 7.838926174496644e-06,
2725
+ "loss": 2.0371,
2726
+ "step": 453
2727
+ },
2728
+ {
2729
+ "epoch": 3.05,
2730
+ "learning_rate": 7.812080536912753e-06,
2731
+ "loss": 2.0566,
2732
+ "step": 454
2733
+ },
2734
+ {
2735
+ "epoch": 3.05,
2736
+ "learning_rate": 7.785234899328859e-06,
2737
+ "loss": 2.0273,
2738
+ "step": 455
2739
+ },
2740
+ {
2741
+ "epoch": 3.06,
2742
+ "learning_rate": 7.758389261744968e-06,
2743
+ "loss": 2.0352,
2744
+ "step": 456
2745
+ },
2746
+ {
2747
+ "epoch": 3.07,
2748
+ "learning_rate": 7.731543624161074e-06,
2749
+ "loss": 2.0361,
2750
+ "step": 457
2751
+ },
2752
+ {
2753
+ "epoch": 3.07,
2754
+ "learning_rate": 7.704697986577183e-06,
2755
+ "loss": 2.0146,
2756
+ "step": 458
2757
+ },
2758
+ {
2759
+ "epoch": 3.08,
2760
+ "learning_rate": 7.677852348993289e-06,
2761
+ "loss": 2.0547,
2762
+ "step": 459
2763
+ },
2764
+ {
2765
+ "epoch": 3.09,
2766
+ "learning_rate": 7.651006711409396e-06,
2767
+ "loss": 2.0439,
2768
+ "step": 460
2769
+ },
2770
+ {
2771
+ "epoch": 3.09,
2772
+ "learning_rate": 7.624161073825504e-06,
2773
+ "loss": 2.0361,
2774
+ "step": 461
2775
+ },
2776
+ {
2777
+ "epoch": 3.1,
2778
+ "learning_rate": 7.597315436241612e-06,
2779
+ "loss": 2.0527,
2780
+ "step": 462
2781
+ },
2782
+ {
2783
+ "epoch": 3.11,
2784
+ "learning_rate": 7.570469798657719e-06,
2785
+ "loss": 2.0508,
2786
+ "step": 463
2787
+ },
2788
+ {
2789
+ "epoch": 3.11,
2790
+ "learning_rate": 7.543624161073826e-06,
2791
+ "loss": 2.0547,
2792
+ "step": 464
2793
+ },
2794
+ {
2795
+ "epoch": 3.12,
2796
+ "learning_rate": 7.516778523489934e-06,
2797
+ "loss": 2.0566,
2798
+ "step": 465
2799
+ },
2800
+ {
2801
+ "epoch": 3.13,
2802
+ "learning_rate": 7.4899328859060405e-06,
2803
+ "loss": 2.0547,
2804
+ "step": 466
2805
+ },
2806
+ {
2807
+ "epoch": 3.13,
2808
+ "learning_rate": 7.463087248322149e-06,
2809
+ "loss": 2.0234,
2810
+ "step": 467
2811
+ },
2812
+ {
2813
+ "epoch": 3.14,
2814
+ "learning_rate": 7.436241610738255e-06,
2815
+ "loss": 2.043,
2816
+ "step": 468
2817
+ },
2818
+ {
2819
+ "epoch": 3.15,
2820
+ "learning_rate": 7.409395973154363e-06,
2821
+ "loss": 2.0371,
2822
+ "step": 469
2823
+ },
2824
+ {
2825
+ "epoch": 3.15,
2826
+ "learning_rate": 7.382550335570471e-06,
2827
+ "loss": 2.0352,
2828
+ "step": 470
2829
+ },
2830
+ {
2831
+ "epoch": 3.16,
2832
+ "learning_rate": 7.3557046979865775e-06,
2833
+ "loss": 2.043,
2834
+ "step": 471
2835
+ },
2836
+ {
2837
+ "epoch": 3.17,
2838
+ "learning_rate": 7.328859060402686e-06,
2839
+ "loss": 2.0537,
2840
+ "step": 472
2841
+ },
2842
+ {
2843
+ "epoch": 3.17,
2844
+ "learning_rate": 7.302013422818792e-06,
2845
+ "loss": 2.0449,
2846
+ "step": 473
2847
+ },
2848
+ {
2849
+ "epoch": 3.18,
2850
+ "learning_rate": 7.2751677852349e-06,
2851
+ "loss": 2.0273,
2852
+ "step": 474
2853
+ },
2854
+ {
2855
+ "epoch": 3.19,
2856
+ "learning_rate": 7.248322147651007e-06,
2857
+ "loss": 2.0322,
2858
+ "step": 475
2859
+ },
2860
+ {
2861
+ "epoch": 3.19,
2862
+ "learning_rate": 7.221476510067115e-06,
2863
+ "loss": 2.0381,
2864
+ "step": 476
2865
+ },
2866
+ {
2867
+ "epoch": 3.2,
2868
+ "learning_rate": 7.194630872483223e-06,
2869
+ "loss": 2.0381,
2870
+ "step": 477
2871
+ },
2872
+ {
2873
+ "epoch": 3.21,
2874
+ "learning_rate": 7.167785234899329e-06,
2875
+ "loss": 2.0439,
2876
+ "step": 478
2877
+ },
2878
+ {
2879
+ "epoch": 3.21,
2880
+ "learning_rate": 7.140939597315437e-06,
2881
+ "loss": 2.04,
2882
+ "step": 479
2883
+ },
2884
+ {
2885
+ "epoch": 3.22,
2886
+ "learning_rate": 7.114093959731544e-06,
2887
+ "loss": 2.0352,
2888
+ "step": 480
2889
+ },
2890
+ {
2891
+ "epoch": 3.23,
2892
+ "learning_rate": 7.087248322147652e-06,
2893
+ "loss": 2.04,
2894
+ "step": 481
2895
+ },
2896
+ {
2897
+ "epoch": 3.23,
2898
+ "learning_rate": 7.060402684563758e-06,
2899
+ "loss": 2.0117,
2900
+ "step": 482
2901
+ },
2902
+ {
2903
+ "epoch": 3.24,
2904
+ "learning_rate": 7.0335570469798665e-06,
2905
+ "loss": 2.0342,
2906
+ "step": 483
2907
+ },
2908
+ {
2909
+ "epoch": 3.25,
2910
+ "learning_rate": 7.006711409395974e-06,
2911
+ "loss": 2.0146,
2912
+ "step": 484
2913
+ },
2914
+ {
2915
+ "epoch": 3.26,
2916
+ "learning_rate": 6.979865771812081e-06,
2917
+ "loss": 2.0518,
2918
+ "step": 485
2919
+ },
2920
+ {
2921
+ "epoch": 3.26,
2922
+ "learning_rate": 6.953020134228189e-06,
2923
+ "loss": 2.0283,
2924
+ "step": 486
2925
+ },
2926
+ {
2927
+ "epoch": 3.27,
2928
+ "learning_rate": 6.926174496644295e-06,
2929
+ "loss": 2.0322,
2930
+ "step": 487
2931
+ },
2932
+ {
2933
+ "epoch": 3.28,
2934
+ "learning_rate": 6.8993288590604035e-06,
2935
+ "loss": 2.0508,
2936
+ "step": 488
2937
+ },
2938
+ {
2939
+ "epoch": 3.28,
2940
+ "learning_rate": 6.87248322147651e-06,
2941
+ "loss": 2.0264,
2942
+ "step": 489
2943
+ },
2944
+ {
2945
+ "epoch": 3.29,
2946
+ "learning_rate": 6.845637583892618e-06,
2947
+ "loss": 2.041,
2948
+ "step": 490
2949
+ },
2950
+ {
2951
+ "epoch": 3.3,
2952
+ "learning_rate": 6.818791946308725e-06,
2953
+ "loss": 2.0264,
2954
+ "step": 491
2955
+ },
2956
+ {
2957
+ "epoch": 3.3,
2958
+ "learning_rate": 6.791946308724832e-06,
2959
+ "loss": 2.0371,
2960
+ "step": 492
2961
+ },
2962
+ {
2963
+ "epoch": 3.31,
2964
+ "learning_rate": 6.765100671140941e-06,
2965
+ "loss": 2.0371,
2966
+ "step": 493
2967
+ },
2968
+ {
2969
+ "epoch": 3.32,
2970
+ "learning_rate": 6.738255033557047e-06,
2971
+ "loss": 2.042,
2972
+ "step": 494
2973
+ },
2974
+ {
2975
+ "epoch": 3.32,
2976
+ "learning_rate": 6.711409395973155e-06,
2977
+ "loss": 2.0107,
2978
+ "step": 495
2979
+ },
2980
+ {
2981
+ "epoch": 3.33,
2982
+ "learning_rate": 6.684563758389262e-06,
2983
+ "loss": 2.0176,
2984
+ "step": 496
2985
+ },
2986
+ {
2987
+ "epoch": 3.34,
2988
+ "learning_rate": 6.65771812080537e-06,
2989
+ "loss": 2.0273,
2990
+ "step": 497
2991
+ },
2992
+ {
2993
+ "epoch": 3.34,
2994
+ "learning_rate": 6.630872483221477e-06,
2995
+ "loss": 2.0469,
2996
+ "step": 498
2997
+ },
2998
+ {
2999
+ "epoch": 3.35,
3000
+ "learning_rate": 6.604026845637584e-06,
3001
+ "loss": 2.0205,
3002
+ "step": 499
3003
+ },
3004
+ {
3005
+ "epoch": 3.36,
3006
+ "learning_rate": 6.5771812080536925e-06,
3007
+ "loss": 2.0137,
3008
+ "step": 500
3009
+ },
3010
+ {
3011
+ "epoch": 3.36,
3012
+ "learning_rate": 6.550335570469799e-06,
3013
+ "loss": 2.0352,
3014
+ "step": 501
3015
+ },
3016
+ {
3017
+ "epoch": 3.37,
3018
+ "learning_rate": 6.523489932885907e-06,
3019
+ "loss": 2.0547,
3020
+ "step": 502
3021
+ },
3022
+ {
3023
+ "epoch": 3.38,
3024
+ "learning_rate": 6.496644295302014e-06,
3025
+ "loss": 2.0137,
3026
+ "step": 503
3027
+ },
3028
+ {
3029
+ "epoch": 3.38,
3030
+ "learning_rate": 6.469798657718121e-06,
3031
+ "loss": 2.0029,
3032
+ "step": 504
3033
+ },
3034
+ {
3035
+ "epoch": 3.39,
3036
+ "learning_rate": 6.442953020134229e-06,
3037
+ "loss": 2.0176,
3038
+ "step": 505
3039
+ },
3040
+ {
3041
+ "epoch": 3.4,
3042
+ "learning_rate": 6.416107382550336e-06,
3043
+ "loss": 2.0283,
3044
+ "step": 506
3045
+ },
3046
+ {
3047
+ "epoch": 3.4,
3048
+ "learning_rate": 6.389261744966444e-06,
3049
+ "loss": 2.0488,
3050
+ "step": 507
3051
+ },
3052
+ {
3053
+ "epoch": 3.41,
3054
+ "learning_rate": 6.362416107382551e-06,
3055
+ "loss": 2.0371,
3056
+ "step": 508
3057
+ },
3058
+ {
3059
+ "epoch": 3.42,
3060
+ "learning_rate": 6.335570469798658e-06,
3061
+ "loss": 2.0107,
3062
+ "step": 509
3063
+ },
3064
+ {
3065
+ "epoch": 3.42,
3066
+ "learning_rate": 6.308724832214766e-06,
3067
+ "loss": 2.0273,
3068
+ "step": 510
3069
+ },
3070
+ {
3071
+ "epoch": 3.43,
3072
+ "learning_rate": 6.281879194630873e-06,
3073
+ "loss": 2.0508,
3074
+ "step": 511
3075
+ },
3076
+ {
3077
+ "epoch": 3.44,
3078
+ "learning_rate": 6.25503355704698e-06,
3079
+ "loss": 2.0342,
3080
+ "step": 512
3081
+ },
3082
+ {
3083
+ "epoch": 3.44,
3084
+ "learning_rate": 6.228187919463088e-06,
3085
+ "loss": 2.0586,
3086
+ "step": 513
3087
+ },
3088
+ {
3089
+ "epoch": 3.45,
3090
+ "learning_rate": 6.2013422818791946e-06,
3091
+ "loss": 2.0107,
3092
+ "step": 514
3093
+ },
3094
+ {
3095
+ "epoch": 3.46,
3096
+ "learning_rate": 6.174496644295303e-06,
3097
+ "loss": 2.0547,
3098
+ "step": 515
3099
+ },
3100
+ {
3101
+ "epoch": 3.46,
3102
+ "learning_rate": 6.14765100671141e-06,
3103
+ "loss": 2.0273,
3104
+ "step": 516
3105
+ },
3106
+ {
3107
+ "epoch": 3.47,
3108
+ "learning_rate": 6.120805369127517e-06,
3109
+ "loss": 2.0371,
3110
+ "step": 517
3111
+ },
3112
+ {
3113
+ "epoch": 3.48,
3114
+ "learning_rate": 6.093959731543625e-06,
3115
+ "loss": 2.0098,
3116
+ "step": 518
3117
+ },
3118
+ {
3119
+ "epoch": 3.48,
3120
+ "learning_rate": 6.067114093959732e-06,
3121
+ "loss": 2.0293,
3122
+ "step": 519
3123
+ },
3124
+ {
3125
+ "epoch": 3.49,
3126
+ "learning_rate": 6.04026845637584e-06,
3127
+ "loss": 2.0186,
3128
+ "step": 520
3129
+ },
3130
+ {
3131
+ "epoch": 3.5,
3132
+ "learning_rate": 6.0134228187919464e-06,
3133
+ "loss": 2.0146,
3134
+ "step": 521
3135
+ },
3136
+ {
3137
+ "epoch": 3.5,
3138
+ "learning_rate": 5.986577181208054e-06,
3139
+ "loss": 2.0361,
3140
+ "step": 522
3141
+ },
3142
+ {
3143
+ "epoch": 3.51,
3144
+ "learning_rate": 5.959731543624162e-06,
3145
+ "loss": 2.0117,
3146
+ "step": 523
3147
+ },
3148
+ {
3149
+ "epoch": 3.52,
3150
+ "learning_rate": 5.932885906040269e-06,
3151
+ "loss": 2.0146,
3152
+ "step": 524
3153
+ },
3154
+ {
3155
+ "epoch": 3.52,
3156
+ "learning_rate": 5.906040268456377e-06,
3157
+ "loss": 2.0264,
3158
+ "step": 525
3159
+ },
3160
+ {
3161
+ "epoch": 3.53,
3162
+ "learning_rate": 5.8791946308724835e-06,
3163
+ "loss": 2.0264,
3164
+ "step": 526
3165
+ },
3166
+ {
3167
+ "epoch": 3.54,
3168
+ "learning_rate": 5.852348993288591e-06,
3169
+ "loss": 2.0439,
3170
+ "step": 527
3171
+ },
3172
+ {
3173
+ "epoch": 3.54,
3174
+ "learning_rate": 5.825503355704698e-06,
3175
+ "loss": 2.0283,
3176
+ "step": 528
3177
+ },
3178
+ {
3179
+ "epoch": 3.55,
3180
+ "learning_rate": 5.798657718120806e-06,
3181
+ "loss": 2.0264,
3182
+ "step": 529
3183
+ },
3184
+ {
3185
+ "epoch": 3.56,
3186
+ "learning_rate": 5.771812080536914e-06,
3187
+ "loss": 2.0273,
3188
+ "step": 530
3189
+ },
3190
+ {
3191
+ "epoch": 3.56,
3192
+ "learning_rate": 5.7449664429530206e-06,
3193
+ "loss": 2.0254,
3194
+ "step": 531
3195
+ },
3196
+ {
3197
+ "epoch": 3.57,
3198
+ "learning_rate": 5.718120805369128e-06,
3199
+ "loss": 2.0146,
3200
+ "step": 532
3201
+ },
3202
+ {
3203
+ "epoch": 3.58,
3204
+ "learning_rate": 5.691275167785235e-06,
3205
+ "loss": 2.0273,
3206
+ "step": 533
3207
+ },
3208
+ {
3209
+ "epoch": 3.58,
3210
+ "learning_rate": 5.664429530201343e-06,
3211
+ "loss": 2.0439,
3212
+ "step": 534
3213
+ },
3214
+ {
3215
+ "epoch": 3.59,
3216
+ "learning_rate": 5.637583892617449e-06,
3217
+ "loss": 2.0293,
3218
+ "step": 535
3219
+ },
3220
+ {
3221
+ "epoch": 3.6,
3222
+ "learning_rate": 5.610738255033558e-06,
3223
+ "loss": 2.0254,
3224
+ "step": 536
3225
+ },
3226
+ {
3227
+ "epoch": 3.6,
3228
+ "learning_rate": 5.583892617449664e-06,
3229
+ "loss": 2.0283,
3230
+ "step": 537
3231
+ },
3232
+ {
3233
+ "epoch": 3.61,
3234
+ "learning_rate": 5.5570469798657725e-06,
3235
+ "loss": 2.0088,
3236
+ "step": 538
3237
+ },
3238
+ {
3239
+ "epoch": 3.62,
3240
+ "learning_rate": 5.53020134228188e-06,
3241
+ "loss": 2.0166,
3242
+ "step": 539
3243
+ },
3244
+ {
3245
+ "epoch": 3.62,
3246
+ "learning_rate": 5.503355704697987e-06,
3247
+ "loss": 2.0283,
3248
+ "step": 540
3249
+ },
3250
+ {
3251
+ "epoch": 3.63,
3252
+ "learning_rate": 5.476510067114095e-06,
3253
+ "loss": 2.002,
3254
+ "step": 541
3255
+ },
3256
+ {
3257
+ "epoch": 3.64,
3258
+ "learning_rate": 5.449664429530201e-06,
3259
+ "loss": 2.0205,
3260
+ "step": 542
3261
+ },
3262
+ {
3263
+ "epoch": 3.64,
3264
+ "learning_rate": 5.4228187919463095e-06,
3265
+ "loss": 2.0234,
3266
+ "step": 543
3267
+ },
3268
+ {
3269
+ "epoch": 3.65,
3270
+ "learning_rate": 5.395973154362416e-06,
3271
+ "loss": 2.043,
3272
+ "step": 544
3273
+ },
3274
+ {
3275
+ "epoch": 3.66,
3276
+ "learning_rate": 5.369127516778524e-06,
3277
+ "loss": 2.0127,
3278
+ "step": 545
3279
+ },
3280
+ {
3281
+ "epoch": 3.66,
3282
+ "learning_rate": 5.342281879194632e-06,
3283
+ "loss": 2.041,
3284
+ "step": 546
3285
+ },
3286
+ {
3287
+ "epoch": 3.67,
3288
+ "learning_rate": 5.315436241610738e-06,
3289
+ "loss": 2.0322,
3290
+ "step": 547
3291
+ },
3292
+ {
3293
+ "epoch": 3.68,
3294
+ "learning_rate": 5.2885906040268466e-06,
3295
+ "loss": 2.0332,
3296
+ "step": 548
3297
+ },
3298
+ {
3299
+ "epoch": 3.68,
3300
+ "learning_rate": 5.261744966442953e-06,
3301
+ "loss": 2.0244,
3302
+ "step": 549
3303
+ },
3304
+ {
3305
+ "epoch": 3.69,
3306
+ "learning_rate": 5.234899328859061e-06,
3307
+ "loss": 2.0244,
3308
+ "step": 550
3309
+ },
3310
+ {
3311
+ "epoch": 3.7,
3312
+ "learning_rate": 5.208053691275168e-06,
3313
+ "loss": 2.0195,
3314
+ "step": 551
3315
+ },
3316
+ {
3317
+ "epoch": 3.7,
3318
+ "learning_rate": 5.181208053691275e-06,
3319
+ "loss": 2.0254,
3320
+ "step": 552
3321
+ },
3322
+ {
3323
+ "epoch": 3.71,
3324
+ "learning_rate": 5.154362416107384e-06,
3325
+ "loss": 2.04,
3326
+ "step": 553
3327
+ },
3328
+ {
3329
+ "epoch": 3.72,
3330
+ "learning_rate": 5.12751677852349e-06,
3331
+ "loss": 2.0244,
3332
+ "step": 554
3333
+ },
3334
+ {
3335
+ "epoch": 3.72,
3336
+ "learning_rate": 5.1006711409395985e-06,
3337
+ "loss": 1.9902,
3338
+ "step": 555
3339
+ },
3340
+ {
3341
+ "epoch": 3.73,
3342
+ "learning_rate": 5.073825503355705e-06,
3343
+ "loss": 2.0127,
3344
+ "step": 556
3345
+ },
3346
+ {
3347
+ "epoch": 3.74,
3348
+ "learning_rate": 5.0469798657718124e-06,
3349
+ "loss": 2.0039,
3350
+ "step": 557
3351
+ },
3352
+ {
3353
+ "epoch": 3.74,
3354
+ "learning_rate": 5.02013422818792e-06,
3355
+ "loss": 2.0225,
3356
+ "step": 558
3357
+ },
3358
+ {
3359
+ "epoch": 3.75,
3360
+ "learning_rate": 4.993288590604027e-06,
3361
+ "loss": 2.0176,
3362
+ "step": 559
3363
+ },
3364
+ {
3365
+ "epoch": 3.76,
3366
+ "learning_rate": 4.966442953020135e-06,
3367
+ "loss": 2.0107,
3368
+ "step": 560
3369
+ },
3370
+ {
3371
+ "epoch": 3.77,
3372
+ "learning_rate": 4.939597315436242e-06,
3373
+ "loss": 2.0273,
3374
+ "step": 561
3375
+ },
3376
+ {
3377
+ "epoch": 3.77,
3378
+ "learning_rate": 4.9127516778523495e-06,
3379
+ "loss": 2.0078,
3380
+ "step": 562
3381
+ },
3382
+ {
3383
+ "epoch": 3.78,
3384
+ "learning_rate": 4.885906040268457e-06,
3385
+ "loss": 2.0361,
3386
+ "step": 563
3387
+ },
3388
+ {
3389
+ "epoch": 3.79,
3390
+ "learning_rate": 4.859060402684564e-06,
3391
+ "loss": 2.0352,
3392
+ "step": 564
3393
+ },
3394
+ {
3395
+ "epoch": 3.79,
3396
+ "learning_rate": 4.832214765100672e-06,
3397
+ "loss": 2.0381,
3398
+ "step": 565
3399
+ },
3400
+ {
3401
+ "epoch": 3.8,
3402
+ "learning_rate": 4.805369127516779e-06,
3403
+ "loss": 2.0215,
3404
+ "step": 566
3405
+ },
3406
+ {
3407
+ "epoch": 3.81,
3408
+ "learning_rate": 4.7785234899328866e-06,
3409
+ "loss": 2.0195,
3410
+ "step": 567
3411
+ },
3412
+ {
3413
+ "epoch": 3.81,
3414
+ "learning_rate": 4.751677852348994e-06,
3415
+ "loss": 2.0107,
3416
+ "step": 568
3417
+ },
3418
+ {
3419
+ "epoch": 3.82,
3420
+ "learning_rate": 4.724832214765101e-06,
3421
+ "loss": 2.0039,
3422
+ "step": 569
3423
+ },
3424
+ {
3425
+ "epoch": 3.83,
3426
+ "learning_rate": 4.697986577181208e-06,
3427
+ "loss": 2.0186,
3428
+ "step": 570
3429
+ },
3430
+ {
3431
+ "epoch": 3.83,
3432
+ "learning_rate": 4.671140939597315e-06,
3433
+ "loss": 2.0234,
3434
+ "step": 571
3435
+ },
3436
+ {
3437
+ "epoch": 3.84,
3438
+ "learning_rate": 4.644295302013423e-06,
3439
+ "loss": 2.0234,
3440
+ "step": 572
3441
+ },
3442
+ {
3443
+ "epoch": 3.85,
3444
+ "learning_rate": 4.617449664429531e-06,
3445
+ "loss": 2.0039,
3446
+ "step": 573
3447
+ },
3448
+ {
3449
+ "epoch": 3.85,
3450
+ "learning_rate": 4.5906040268456384e-06,
3451
+ "loss": 2.0039,
3452
+ "step": 574
3453
+ },
3454
+ {
3455
+ "epoch": 3.86,
3456
+ "learning_rate": 4.563758389261745e-06,
3457
+ "loss": 2.0215,
3458
+ "step": 575
3459
+ },
3460
+ {
3461
+ "epoch": 3.87,
3462
+ "learning_rate": 4.536912751677852e-06,
3463
+ "loss": 2.0215,
3464
+ "step": 576
3465
+ },
3466
+ {
3467
+ "epoch": 3.87,
3468
+ "learning_rate": 4.51006711409396e-06,
3469
+ "loss": 2.0146,
3470
+ "step": 577
3471
+ },
3472
+ {
3473
+ "epoch": 3.88,
3474
+ "learning_rate": 4.483221476510067e-06,
3475
+ "loss": 2.0391,
3476
+ "step": 578
3477
+ },
3478
+ {
3479
+ "epoch": 3.89,
3480
+ "learning_rate": 4.456375838926175e-06,
3481
+ "loss": 2.0137,
3482
+ "step": 579
3483
+ },
3484
+ {
3485
+ "epoch": 3.89,
3486
+ "learning_rate": 4.429530201342283e-06,
3487
+ "loss": 2.0107,
3488
+ "step": 580
3489
+ },
3490
+ {
3491
+ "epoch": 3.9,
3492
+ "learning_rate": 4.4026845637583895e-06,
3493
+ "loss": 2.0176,
3494
+ "step": 581
3495
+ },
3496
+ {
3497
+ "epoch": 3.91,
3498
+ "learning_rate": 4.375838926174497e-06,
3499
+ "loss": 2.0273,
3500
+ "step": 582
3501
+ },
3502
+ {
3503
+ "epoch": 3.91,
3504
+ "learning_rate": 4.348993288590604e-06,
3505
+ "loss": 2.0098,
3506
+ "step": 583
3507
+ },
3508
+ {
3509
+ "epoch": 3.92,
3510
+ "learning_rate": 4.322147651006712e-06,
3511
+ "loss": 2.0215,
3512
+ "step": 584
3513
+ },
3514
+ {
3515
+ "epoch": 3.93,
3516
+ "learning_rate": 4.295302013422819e-06,
3517
+ "loss": 2.0107,
3518
+ "step": 585
3519
+ },
3520
+ {
3521
+ "epoch": 3.93,
3522
+ "learning_rate": 4.2684563758389265e-06,
3523
+ "loss": 2.0293,
3524
+ "step": 586
3525
+ },
3526
+ {
3527
+ "epoch": 3.94,
3528
+ "learning_rate": 4.241610738255034e-06,
3529
+ "loss": 2.0312,
3530
+ "step": 587
3531
+ },
3532
+ {
3533
+ "epoch": 3.95,
3534
+ "learning_rate": 4.214765100671141e-06,
3535
+ "loss": 2.0156,
3536
+ "step": 588
3537
+ },
3538
+ {
3539
+ "epoch": 3.95,
3540
+ "learning_rate": 4.187919463087249e-06,
3541
+ "loss": 1.9922,
3542
+ "step": 589
3543
+ },
3544
+ {
3545
+ "epoch": 3.96,
3546
+ "learning_rate": 4.161073825503356e-06,
3547
+ "loss": 2.0225,
3548
+ "step": 590
3549
+ },
3550
+ {
3551
+ "epoch": 3.97,
3552
+ "learning_rate": 4.134228187919464e-06,
3553
+ "loss": 2.0391,
3554
+ "step": 591
3555
+ },
3556
+ {
3557
+ "epoch": 3.97,
3558
+ "learning_rate": 4.107382550335571e-06,
3559
+ "loss": 2.0293,
3560
+ "step": 592
3561
+ },
3562
+ {
3563
+ "epoch": 3.98,
3564
+ "learning_rate": 4.0805369127516784e-06,
3565
+ "loss": 2.0166,
3566
+ "step": 593
3567
+ },
3568
+ {
3569
+ "epoch": 3.99,
3570
+ "learning_rate": 4.053691275167785e-06,
3571
+ "loss": 2.001,
3572
+ "step": 594
3573
+ },
3574
+ {
3575
+ "epoch": 3.99,
3576
+ "learning_rate": 4.026845637583892e-06,
3577
+ "loss": 2.0215,
3578
+ "step": 595
3579
+ },
3580
+ {
3581
+ "epoch": 4.0,
3582
+ "learning_rate": 4.000000000000001e-06,
3583
+ "loss": 2.0166,
3584
+ "step": 596
3585
+ },
3586
+ {
3587
+ "epoch": 4.01,
3588
+ "learning_rate": 3.973154362416108e-06,
3589
+ "loss": 2.0088,
3590
+ "step": 597
3591
+ },
3592
+ {
3593
+ "epoch": 4.01,
3594
+ "learning_rate": 3.9463087248322155e-06,
3595
+ "loss": 2.0156,
3596
+ "step": 598
3597
+ },
3598
+ {
3599
+ "epoch": 4.02,
3600
+ "learning_rate": 3.919463087248322e-06,
3601
+ "loss": 2.0332,
3602
+ "step": 599
3603
+ },
3604
+ {
3605
+ "epoch": 4.03,
3606
+ "learning_rate": 3.8926174496644295e-06,
3607
+ "loss": 1.9951,
3608
+ "step": 600
3609
+ },
3610
+ {
3611
+ "epoch": 4.03,
3612
+ "learning_rate": 3.865771812080537e-06,
3613
+ "loss": 2.0215,
3614
+ "step": 601
3615
+ },
3616
+ {
3617
+ "epoch": 4.04,
3618
+ "learning_rate": 3.838926174496644e-06,
3619
+ "loss": 2.0146,
3620
+ "step": 602
3621
+ },
3622
+ {
3623
+ "epoch": 4.05,
3624
+ "learning_rate": 3.812080536912752e-06,
3625
+ "loss": 2.0332,
3626
+ "step": 603
3627
+ },
3628
+ {
3629
+ "epoch": 4.05,
3630
+ "learning_rate": 3.7852348993288595e-06,
3631
+ "loss": 2.0449,
3632
+ "step": 604
3633
+ },
3634
+ {
3635
+ "epoch": 4.06,
3636
+ "learning_rate": 3.758389261744967e-06,
3637
+ "loss": 2.0225,
3638
+ "step": 605
3639
+ },
3640
+ {
3641
+ "epoch": 4.07,
3642
+ "learning_rate": 3.7315436241610744e-06,
3643
+ "loss": 2.001,
3644
+ "step": 606
3645
+ },
3646
+ {
3647
+ "epoch": 4.07,
3648
+ "learning_rate": 3.7046979865771814e-06,
3649
+ "loss": 2.0146,
3650
+ "step": 607
3651
+ },
3652
+ {
3653
+ "epoch": 4.08,
3654
+ "learning_rate": 3.6778523489932888e-06,
3655
+ "loss": 1.9932,
3656
+ "step": 608
3657
+ },
3658
+ {
3659
+ "epoch": 4.09,
3660
+ "learning_rate": 3.651006711409396e-06,
3661
+ "loss": 2.0029,
3662
+ "step": 609
3663
+ },
3664
+ {
3665
+ "epoch": 4.09,
3666
+ "learning_rate": 3.6241610738255036e-06,
3667
+ "loss": 2.0029,
3668
+ "step": 610
3669
+ },
3670
+ {
3671
+ "epoch": 4.1,
3672
+ "learning_rate": 3.5973154362416114e-06,
3673
+ "loss": 2.0186,
3674
+ "step": 611
3675
+ },
3676
+ {
3677
+ "epoch": 4.11,
3678
+ "learning_rate": 3.5704697986577184e-06,
3679
+ "loss": 2.0117,
3680
+ "step": 612
3681
+ },
3682
+ {
3683
+ "epoch": 4.11,
3684
+ "learning_rate": 3.543624161073826e-06,
3685
+ "loss": 2.0186,
3686
+ "step": 613
3687
+ },
3688
+ {
3689
+ "epoch": 4.12,
3690
+ "learning_rate": 3.5167785234899332e-06,
3691
+ "loss": 1.999,
3692
+ "step": 614
3693
+ },
3694
+ {
3695
+ "epoch": 4.13,
3696
+ "learning_rate": 3.4899328859060407e-06,
3697
+ "loss": 2.0312,
3698
+ "step": 615
3699
+ },
3700
+ {
3701
+ "epoch": 4.13,
3702
+ "learning_rate": 3.4630872483221476e-06,
3703
+ "loss": 2.0234,
3704
+ "step": 616
3705
+ },
3706
+ {
3707
+ "epoch": 4.14,
3708
+ "learning_rate": 3.436241610738255e-06,
3709
+ "loss": 2.0225,
3710
+ "step": 617
3711
+ },
3712
+ {
3713
+ "epoch": 4.15,
3714
+ "learning_rate": 3.4093959731543625e-06,
3715
+ "loss": 2.0322,
3716
+ "step": 618
3717
+ },
3718
+ {
3719
+ "epoch": 4.15,
3720
+ "learning_rate": 3.3825503355704703e-06,
3721
+ "loss": 2.0186,
3722
+ "step": 619
3723
+ },
3724
+ {
3725
+ "epoch": 4.16,
3726
+ "learning_rate": 3.3557046979865777e-06,
3727
+ "loss": 2.001,
3728
+ "step": 620
3729
+ },
3730
+ {
3731
+ "epoch": 4.17,
3732
+ "learning_rate": 3.328859060402685e-06,
3733
+ "loss": 1.9971,
3734
+ "step": 621
3735
+ },
3736
+ {
3737
+ "epoch": 4.17,
3738
+ "learning_rate": 3.302013422818792e-06,
3739
+ "loss": 2.0117,
3740
+ "step": 622
3741
+ },
3742
+ {
3743
+ "epoch": 4.18,
3744
+ "learning_rate": 3.2751677852348995e-06,
3745
+ "loss": 1.9941,
3746
+ "step": 623
3747
+ },
3748
+ {
3749
+ "epoch": 4.19,
3750
+ "learning_rate": 3.248322147651007e-06,
3751
+ "loss": 2.0127,
3752
+ "step": 624
3753
+ },
3754
+ {
3755
+ "epoch": 4.19,
3756
+ "learning_rate": 3.2214765100671143e-06,
3757
+ "loss": 1.9932,
3758
+ "step": 625
3759
+ },
3760
+ {
3761
+ "epoch": 4.2,
3762
+ "learning_rate": 3.194630872483222e-06,
3763
+ "loss": 1.9932,
3764
+ "step": 626
3765
+ },
3766
+ {
3767
+ "epoch": 4.21,
3768
+ "learning_rate": 3.167785234899329e-06,
3769
+ "loss": 2.0059,
3770
+ "step": 627
3771
+ },
3772
+ {
3773
+ "epoch": 4.21,
3774
+ "learning_rate": 3.1409395973154366e-06,
3775
+ "loss": 1.9932,
3776
+ "step": 628
3777
+ },
3778
+ {
3779
+ "epoch": 4.22,
3780
+ "learning_rate": 3.114093959731544e-06,
3781
+ "loss": 2.0293,
3782
+ "step": 629
3783
+ },
3784
+ {
3785
+ "epoch": 4.23,
3786
+ "learning_rate": 3.0872483221476514e-06,
3787
+ "loss": 2.0,
3788
+ "step": 630
3789
+ },
3790
+ {
3791
+ "epoch": 4.23,
3792
+ "learning_rate": 3.0604026845637584e-06,
3793
+ "loss": 2.0156,
3794
+ "step": 631
3795
+ },
3796
+ {
3797
+ "epoch": 4.24,
3798
+ "learning_rate": 3.033557046979866e-06,
3799
+ "loss": 2.0303,
3800
+ "step": 632
3801
+ },
3802
+ {
3803
+ "epoch": 4.25,
3804
+ "learning_rate": 3.0067114093959732e-06,
3805
+ "loss": 2.0186,
3806
+ "step": 633
3807
+ },
3808
+ {
3809
+ "epoch": 4.26,
3810
+ "learning_rate": 2.979865771812081e-06,
3811
+ "loss": 2.0068,
3812
+ "step": 634
3813
+ },
3814
+ {
3815
+ "epoch": 4.26,
3816
+ "learning_rate": 2.9530201342281885e-06,
3817
+ "loss": 2.0381,
3818
+ "step": 635
3819
+ },
3820
+ {
3821
+ "epoch": 4.27,
3822
+ "learning_rate": 2.9261744966442955e-06,
3823
+ "loss": 2.0088,
3824
+ "step": 636
3825
+ },
3826
+ {
3827
+ "epoch": 4.28,
3828
+ "learning_rate": 2.899328859060403e-06,
3829
+ "loss": 1.9971,
3830
+ "step": 637
3831
+ },
3832
+ {
3833
+ "epoch": 4.28,
3834
+ "learning_rate": 2.8724832214765103e-06,
3835
+ "loss": 2.0293,
3836
+ "step": 638
3837
+ },
3838
+ {
3839
+ "epoch": 4.29,
3840
+ "learning_rate": 2.8456375838926177e-06,
3841
+ "loss": 1.9961,
3842
+ "step": 639
3843
+ },
3844
+ {
3845
+ "epoch": 4.3,
3846
+ "learning_rate": 2.8187919463087247e-06,
3847
+ "loss": 2.0117,
3848
+ "step": 640
3849
+ },
3850
+ {
3851
+ "epoch": 4.3,
3852
+ "learning_rate": 2.791946308724832e-06,
3853
+ "loss": 2.0068,
3854
+ "step": 641
3855
+ },
3856
+ {
3857
+ "epoch": 4.31,
3858
+ "learning_rate": 2.76510067114094e-06,
3859
+ "loss": 2.002,
3860
+ "step": 642
3861
+ },
3862
+ {
3863
+ "epoch": 4.32,
3864
+ "learning_rate": 2.7382550335570473e-06,
3865
+ "loss": 2.001,
3866
+ "step": 643
3867
+ },
3868
+ {
3869
+ "epoch": 4.32,
3870
+ "learning_rate": 2.7114093959731548e-06,
3871
+ "loss": 1.998,
3872
+ "step": 644
3873
+ },
3874
+ {
3875
+ "epoch": 4.33,
3876
+ "learning_rate": 2.684563758389262e-06,
3877
+ "loss": 2.0156,
3878
+ "step": 645
3879
+ },
3880
+ {
3881
+ "epoch": 4.34,
3882
+ "learning_rate": 2.657718120805369e-06,
3883
+ "loss": 2.0322,
3884
+ "step": 646
3885
+ },
3886
+ {
3887
+ "epoch": 4.34,
3888
+ "learning_rate": 2.6308724832214766e-06,
3889
+ "loss": 2.0156,
3890
+ "step": 647
3891
+ },
3892
+ {
3893
+ "epoch": 4.35,
3894
+ "learning_rate": 2.604026845637584e-06,
3895
+ "loss": 2.001,
3896
+ "step": 648
3897
+ },
3898
+ {
3899
+ "epoch": 4.36,
3900
+ "learning_rate": 2.577181208053692e-06,
3901
+ "loss": 1.9922,
3902
+ "step": 649
3903
+ },
3904
+ {
3905
+ "epoch": 4.36,
3906
+ "learning_rate": 2.5503355704697992e-06,
3907
+ "loss": 1.9883,
3908
+ "step": 650
3909
+ },
3910
+ {
3911
+ "epoch": 4.37,
3912
+ "learning_rate": 2.5234899328859062e-06,
3913
+ "loss": 2.0176,
3914
+ "step": 651
3915
+ },
3916
+ {
3917
+ "epoch": 4.38,
3918
+ "learning_rate": 2.4966442953020136e-06,
3919
+ "loss": 2.0039,
3920
+ "step": 652
3921
+ },
3922
+ {
3923
+ "epoch": 4.38,
3924
+ "learning_rate": 2.469798657718121e-06,
3925
+ "loss": 1.9961,
3926
+ "step": 653
3927
+ },
3928
+ {
3929
+ "epoch": 4.39,
3930
+ "learning_rate": 2.4429530201342285e-06,
3931
+ "loss": 2.002,
3932
+ "step": 654
3933
+ },
3934
+ {
3935
+ "epoch": 4.4,
3936
+ "learning_rate": 2.416107382550336e-06,
3937
+ "loss": 2.0205,
3938
+ "step": 655
3939
+ },
3940
+ {
3941
+ "epoch": 4.4,
3942
+ "learning_rate": 2.3892617449664433e-06,
3943
+ "loss": 1.9971,
3944
+ "step": 656
3945
+ },
3946
+ {
3947
+ "epoch": 4.41,
3948
+ "learning_rate": 2.3624161073825507e-06,
3949
+ "loss": 2.0088,
3950
+ "step": 657
3951
+ },
3952
+ {
3953
+ "epoch": 4.42,
3954
+ "learning_rate": 2.3355704697986577e-06,
3955
+ "loss": 2.0254,
3956
+ "step": 658
3957
+ },
3958
+ {
3959
+ "epoch": 4.42,
3960
+ "learning_rate": 2.3087248322147655e-06,
3961
+ "loss": 2.0117,
3962
+ "step": 659
3963
+ },
3964
+ {
3965
+ "epoch": 4.43,
3966
+ "learning_rate": 2.2818791946308725e-06,
3967
+ "loss": 1.9961,
3968
+ "step": 660
3969
+ },
3970
+ {
3971
+ "epoch": 4.44,
3972
+ "learning_rate": 2.25503355704698e-06,
3973
+ "loss": 2.0059,
3974
+ "step": 661
3975
+ },
3976
+ {
3977
+ "epoch": 4.44,
3978
+ "learning_rate": 2.2281879194630873e-06,
3979
+ "loss": 2.0029,
3980
+ "step": 662
3981
+ },
3982
+ {
3983
+ "epoch": 4.45,
3984
+ "learning_rate": 2.2013422818791947e-06,
3985
+ "loss": 2.0098,
3986
+ "step": 663
3987
+ },
3988
+ {
3989
+ "epoch": 4.46,
3990
+ "learning_rate": 2.174496644295302e-06,
3991
+ "loss": 2.0254,
3992
+ "step": 664
3993
+ },
3994
+ {
3995
+ "epoch": 4.46,
3996
+ "learning_rate": 2.1476510067114096e-06,
3997
+ "loss": 2.0107,
3998
+ "step": 665
3999
+ },
4000
+ {
4001
+ "epoch": 4.47,
4002
+ "learning_rate": 2.120805369127517e-06,
4003
+ "loss": 2.0059,
4004
+ "step": 666
4005
+ },
4006
+ {
4007
+ "epoch": 4.48,
4008
+ "learning_rate": 2.0939597315436244e-06,
4009
+ "loss": 2.0107,
4010
+ "step": 667
4011
+ },
4012
+ {
4013
+ "epoch": 4.48,
4014
+ "learning_rate": 2.067114093959732e-06,
4015
+ "loss": 1.9922,
4016
+ "step": 668
4017
+ },
4018
+ {
4019
+ "epoch": 4.49,
4020
+ "learning_rate": 2.0402684563758392e-06,
4021
+ "loss": 2.0137,
4022
+ "step": 669
4023
+ },
4024
+ {
4025
+ "epoch": 4.5,
4026
+ "learning_rate": 2.013422818791946e-06,
4027
+ "loss": 1.9912,
4028
+ "step": 670
4029
+ },
4030
+ {
4031
+ "epoch": 4.5,
4032
+ "learning_rate": 1.986577181208054e-06,
4033
+ "loss": 2.0146,
4034
+ "step": 671
4035
+ },
4036
+ {
4037
+ "epoch": 4.51,
4038
+ "learning_rate": 1.959731543624161e-06,
4039
+ "loss": 2.0039,
4040
+ "step": 672
4041
+ },
4042
+ {
4043
+ "epoch": 4.52,
4044
+ "learning_rate": 1.9328859060402684e-06,
4045
+ "loss": 1.9961,
4046
+ "step": 673
4047
+ },
4048
+ {
4049
+ "epoch": 4.52,
4050
+ "learning_rate": 1.906040268456376e-06,
4051
+ "loss": 2.0254,
4052
+ "step": 674
4053
+ },
4054
+ {
4055
+ "epoch": 4.53,
4056
+ "learning_rate": 1.8791946308724835e-06,
4057
+ "loss": 2.0156,
4058
+ "step": 675
4059
+ },
4060
+ {
4061
+ "epoch": 4.54,
4062
+ "learning_rate": 1.8523489932885907e-06,
4063
+ "loss": 2.001,
4064
+ "step": 676
4065
+ },
4066
+ {
4067
+ "epoch": 4.54,
4068
+ "learning_rate": 1.825503355704698e-06,
4069
+ "loss": 2.0029,
4070
+ "step": 677
4071
+ },
4072
+ {
4073
+ "epoch": 4.55,
4074
+ "learning_rate": 1.7986577181208057e-06,
4075
+ "loss": 2.0068,
4076
+ "step": 678
4077
+ },
4078
+ {
4079
+ "epoch": 4.56,
4080
+ "learning_rate": 1.771812080536913e-06,
4081
+ "loss": 2.0234,
4082
+ "step": 679
4083
+ },
4084
+ {
4085
+ "epoch": 4.56,
4086
+ "learning_rate": 1.7449664429530203e-06,
4087
+ "loss": 1.9824,
4088
+ "step": 680
4089
+ },
4090
+ {
4091
+ "epoch": 4.57,
4092
+ "learning_rate": 1.7181208053691275e-06,
4093
+ "loss": 2.0049,
4094
+ "step": 681
4095
+ },
4096
+ {
4097
+ "epoch": 4.58,
4098
+ "learning_rate": 1.6912751677852351e-06,
4099
+ "loss": 2.0029,
4100
+ "step": 682
4101
+ },
4102
+ {
4103
+ "epoch": 4.58,
4104
+ "learning_rate": 1.6644295302013426e-06,
4105
+ "loss": 2.0205,
4106
+ "step": 683
4107
+ },
4108
+ {
4109
+ "epoch": 4.59,
4110
+ "learning_rate": 1.6375838926174498e-06,
4111
+ "loss": 2.0029,
4112
+ "step": 684
4113
+ },
4114
+ {
4115
+ "epoch": 4.6,
4116
+ "learning_rate": 1.6107382550335572e-06,
4117
+ "loss": 2.0166,
4118
+ "step": 685
4119
+ },
4120
+ {
4121
+ "epoch": 4.6,
4122
+ "learning_rate": 1.5838926174496646e-06,
4123
+ "loss": 2.0059,
4124
+ "step": 686
4125
+ },
4126
+ {
4127
+ "epoch": 4.61,
4128
+ "learning_rate": 1.557046979865772e-06,
4129
+ "loss": 2.0186,
4130
+ "step": 687
4131
+ },
4132
+ {
4133
+ "epoch": 4.62,
4134
+ "learning_rate": 1.5302013422818792e-06,
4135
+ "loss": 2.0088,
4136
+ "step": 688
4137
+ },
4138
+ {
4139
+ "epoch": 4.62,
4140
+ "learning_rate": 1.5033557046979866e-06,
4141
+ "loss": 1.9961,
4142
+ "step": 689
4143
+ },
4144
+ {
4145
+ "epoch": 4.63,
4146
+ "learning_rate": 1.4765100671140942e-06,
4147
+ "loss": 2.0107,
4148
+ "step": 690
4149
+ },
4150
+ {
4151
+ "epoch": 4.64,
4152
+ "learning_rate": 1.4496644295302014e-06,
4153
+ "loss": 2.0234,
4154
+ "step": 691
4155
+ },
4156
+ {
4157
+ "epoch": 4.64,
4158
+ "learning_rate": 1.4228187919463088e-06,
4159
+ "loss": 2.0098,
4160
+ "step": 692
4161
+ },
4162
+ {
4163
+ "epoch": 4.65,
4164
+ "learning_rate": 1.395973154362416e-06,
4165
+ "loss": 2.0234,
4166
+ "step": 693
4167
+ },
4168
+ {
4169
+ "epoch": 4.66,
4170
+ "learning_rate": 1.3691275167785237e-06,
4171
+ "loss": 2.0254,
4172
+ "step": 694
4173
+ },
4174
+ {
4175
+ "epoch": 4.66,
4176
+ "learning_rate": 1.342281879194631e-06,
4177
+ "loss": 2.0029,
4178
+ "step": 695
4179
+ },
4180
+ {
4181
+ "epoch": 4.67,
4182
+ "learning_rate": 1.3154362416107383e-06,
4183
+ "loss": 2.0146,
4184
+ "step": 696
4185
+ },
4186
+ {
4187
+ "epoch": 4.68,
4188
+ "learning_rate": 1.288590604026846e-06,
4189
+ "loss": 2.0,
4190
+ "step": 697
4191
+ },
4192
+ {
4193
+ "epoch": 4.68,
4194
+ "learning_rate": 1.2617449664429531e-06,
4195
+ "loss": 2.0098,
4196
+ "step": 698
4197
+ },
4198
+ {
4199
+ "epoch": 4.69,
4200
+ "learning_rate": 1.2348993288590605e-06,
4201
+ "loss": 2.0293,
4202
+ "step": 699
4203
+ },
4204
+ {
4205
+ "epoch": 4.7,
4206
+ "learning_rate": 1.208053691275168e-06,
4207
+ "loss": 2.0059,
4208
+ "step": 700
4209
+ },
4210
+ {
4211
+ "epoch": 4.7,
4212
+ "learning_rate": 1.1812080536912753e-06,
4213
+ "loss": 1.9971,
4214
+ "step": 701
4215
+ },
4216
+ {
4217
+ "epoch": 4.71,
4218
+ "learning_rate": 1.1543624161073828e-06,
4219
+ "loss": 2.0303,
4220
+ "step": 702
4221
+ },
4222
+ {
4223
+ "epoch": 4.72,
4224
+ "learning_rate": 1.12751677852349e-06,
4225
+ "loss": 2.0088,
4226
+ "step": 703
4227
+ },
4228
+ {
4229
+ "epoch": 4.72,
4230
+ "learning_rate": 1.1006711409395974e-06,
4231
+ "loss": 2.0068,
4232
+ "step": 704
4233
+ },
4234
+ {
4235
+ "epoch": 4.73,
4236
+ "learning_rate": 1.0738255033557048e-06,
4237
+ "loss": 1.9951,
4238
+ "step": 705
4239
+ },
4240
+ {
4241
+ "epoch": 4.74,
4242
+ "learning_rate": 1.0469798657718122e-06,
4243
+ "loss": 2.0264,
4244
+ "step": 706
4245
+ },
4246
+ {
4247
+ "epoch": 4.74,
4248
+ "learning_rate": 1.0201342281879196e-06,
4249
+ "loss": 1.9951,
4250
+ "step": 707
4251
+ },
4252
+ {
4253
+ "epoch": 4.75,
4254
+ "learning_rate": 9.93288590604027e-07,
4255
+ "loss": 1.999,
4256
+ "step": 708
4257
+ },
4258
+ {
4259
+ "epoch": 4.76,
4260
+ "learning_rate": 9.664429530201342e-07,
4261
+ "loss": 2.0,
4262
+ "step": 709
4263
+ },
4264
+ {
4265
+ "epoch": 4.77,
4266
+ "learning_rate": 9.395973154362417e-07,
4267
+ "loss": 2.002,
4268
+ "step": 710
4269
+ },
4270
+ {
4271
+ "epoch": 4.77,
4272
+ "learning_rate": 9.12751677852349e-07,
4273
+ "loss": 2.0156,
4274
+ "step": 711
4275
+ },
4276
+ {
4277
+ "epoch": 4.78,
4278
+ "learning_rate": 8.859060402684565e-07,
4279
+ "loss": 2.0117,
4280
+ "step": 712
4281
+ },
4282
+ {
4283
+ "epoch": 4.79,
4284
+ "learning_rate": 8.590604026845638e-07,
4285
+ "loss": 1.9902,
4286
+ "step": 713
4287
+ },
4288
+ {
4289
+ "epoch": 4.79,
4290
+ "learning_rate": 8.322147651006713e-07,
4291
+ "loss": 1.9932,
4292
+ "step": 714
4293
+ },
4294
+ {
4295
+ "epoch": 4.8,
4296
+ "learning_rate": 8.053691275167786e-07,
4297
+ "loss": 2.001,
4298
+ "step": 715
4299
+ },
4300
+ {
4301
+ "epoch": 4.81,
4302
+ "learning_rate": 7.78523489932886e-07,
4303
+ "loss": 2.0059,
4304
+ "step": 716
4305
+ },
4306
+ {
4307
+ "epoch": 4.81,
4308
+ "learning_rate": 7.516778523489933e-07,
4309
+ "loss": 2.0078,
4310
+ "step": 717
4311
+ },
4312
+ {
4313
+ "epoch": 4.82,
4314
+ "learning_rate": 7.248322147651007e-07,
4315
+ "loss": 2.0049,
4316
+ "step": 718
4317
+ },
4318
+ {
4319
+ "epoch": 4.83,
4320
+ "learning_rate": 6.97986577181208e-07,
4321
+ "loss": 2.002,
4322
+ "step": 719
4323
+ },
4324
+ {
4325
+ "epoch": 4.83,
4326
+ "learning_rate": 6.711409395973155e-07,
4327
+ "loss": 2.0117,
4328
+ "step": 720
4329
+ },
4330
+ {
4331
+ "epoch": 4.84,
4332
+ "learning_rate": 6.44295302013423e-07,
4333
+ "loss": 1.9932,
4334
+ "step": 721
4335
+ },
4336
+ {
4337
+ "epoch": 4.85,
4338
+ "learning_rate": 6.174496644295303e-07,
4339
+ "loss": 1.998,
4340
+ "step": 722
4341
+ },
4342
+ {
4343
+ "epoch": 4.85,
4344
+ "learning_rate": 5.906040268456377e-07,
4345
+ "loss": 2.0146,
4346
+ "step": 723
4347
+ },
4348
+ {
4349
+ "epoch": 4.86,
4350
+ "learning_rate": 5.63758389261745e-07,
4351
+ "loss": 2.0049,
4352
+ "step": 724
4353
+ },
4354
+ {
4355
+ "epoch": 4.87,
4356
+ "learning_rate": 5.369127516778524e-07,
4357
+ "loss": 1.999,
4358
+ "step": 725
4359
+ },
4360
+ {
4361
+ "epoch": 4.87,
4362
+ "learning_rate": 5.100671140939598e-07,
4363
+ "loss": 2.0176,
4364
+ "step": 726
4365
+ },
4366
+ {
4367
+ "epoch": 4.88,
4368
+ "learning_rate": 4.832214765100671e-07,
4369
+ "loss": 2.0098,
4370
+ "step": 727
4371
+ },
4372
+ {
4373
+ "epoch": 4.89,
4374
+ "learning_rate": 4.563758389261745e-07,
4375
+ "loss": 2.0068,
4376
+ "step": 728
4377
+ },
4378
+ {
4379
+ "epoch": 4.89,
4380
+ "learning_rate": 4.295302013422819e-07,
4381
+ "loss": 2.0303,
4382
+ "step": 729
4383
+ },
4384
+ {
4385
+ "epoch": 4.9,
4386
+ "learning_rate": 4.026845637583893e-07,
4387
+ "loss": 2.0137,
4388
+ "step": 730
4389
+ },
4390
+ {
4391
+ "epoch": 4.91,
4392
+ "learning_rate": 3.7583892617449665e-07,
4393
+ "loss": 2.0098,
4394
+ "step": 731
4395
+ },
4396
+ {
4397
+ "epoch": 4.91,
4398
+ "learning_rate": 3.48993288590604e-07,
4399
+ "loss": 2.0244,
4400
+ "step": 732
4401
+ },
4402
+ {
4403
+ "epoch": 4.92,
4404
+ "learning_rate": 3.221476510067115e-07,
4405
+ "loss": 1.9922,
4406
+ "step": 733
4407
+ },
4408
+ {
4409
+ "epoch": 4.93,
4410
+ "learning_rate": 2.9530201342281884e-07,
4411
+ "loss": 1.998,
4412
+ "step": 734
4413
+ },
4414
+ {
4415
+ "epoch": 4.93,
4416
+ "learning_rate": 2.684563758389262e-07,
4417
+ "loss": 1.9961,
4418
+ "step": 735
4419
+ },
4420
+ {
4421
+ "epoch": 4.94,
4422
+ "learning_rate": 2.4161073825503355e-07,
4423
+ "loss": 1.9951,
4424
+ "step": 736
4425
+ },
4426
+ {
4427
+ "epoch": 4.95,
4428
+ "learning_rate": 2.1476510067114094e-07,
4429
+ "loss": 2.0117,
4430
+ "step": 737
4431
+ },
4432
+ {
4433
+ "epoch": 4.95,
4434
+ "learning_rate": 1.8791946308724833e-07,
4435
+ "loss": 2.0049,
4436
+ "step": 738
4437
+ },
4438
+ {
4439
+ "epoch": 4.96,
4440
+ "learning_rate": 1.6107382550335574e-07,
4441
+ "loss": 2.0146,
4442
+ "step": 739
4443
+ },
4444
+ {
4445
+ "epoch": 4.97,
4446
+ "learning_rate": 1.342281879194631e-07,
4447
+ "loss": 2.0303,
4448
+ "step": 740
4449
+ },
4450
+ {
4451
+ "epoch": 4.97,
4452
+ "learning_rate": 1.0738255033557047e-07,
4453
+ "loss": 2.0078,
4454
+ "step": 741
4455
+ },
4456
+ {
4457
+ "epoch": 4.98,
4458
+ "learning_rate": 8.053691275167787e-08,
4459
+ "loss": 2.0039,
4460
+ "step": 742
4461
+ },
4462
+ {
4463
+ "epoch": 4.99,
4464
+ "learning_rate": 5.3691275167785235e-08,
4465
+ "loss": 1.998,
4466
+ "step": 743
4467
+ },
4468
+ {
4469
+ "epoch": 4.99,
4470
+ "learning_rate": 2.6845637583892618e-08,
4471
+ "loss": 2.0186,
4472
+ "step": 744
4473
+ },
4474
+ {
4475
+ "epoch": 5.0,
4476
+ "learning_rate": 0.0,
4477
+ "loss": 1.9941,
4478
+ "step": 745
4479
+ },
4480
+ {
4481
+ "epoch": 5.0,
4482
+ "step": 745,
4483
+ "total_flos": 1.4940678387662848e+17,
4484
+ "train_loss": 2.1389746749161076,
4485
+ "train_runtime": 581.9582,
4486
+ "train_samples_per_second": 655.013,
4487
+ "train_steps_per_second": 1.28
4488
+ }
4489
+ ],
4490
+ "max_steps": 745,
4491
+ "num_train_epochs": 5,
4492
+ "total_flos": 1.4940678387662848e+17,
4493
+ "trial_name": null,
4494
+ "trial_params": null
4495
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6a75187b4f9d34bdc0969a7fb8b56df2299492431e0d943e943261a229431d7d
3
+ size 4795
vocab.json ADDED
The diff for this file is too large to render. See raw diff