Hugging Face's logo

ce-lery
/
dolly-japanese-gpt-1b-clone

Text Generation
Transformers
PyTorch
Safetensors
Japanese
gpt2
japanese
gpt
lm
nlp
Inference Endpoints
text-generation-inference
Model card Files Community
ce-lery commited on
Commit
5cb8212
•
1 Parent(s): 51b903b

style: 学習済みモデルを追加

Browse files
Files changed (12) hide show
  1. .gitattributes +2 -0
  2. output/README.md +51 -0
  3. output/all_results.json +8 -0
  4. output/config.json +34 -0
  5. output/generation_config.json +6 -0
  6. output/model.safetensors +3 -0
  7. output/special_tokens_map.json +9 -0
  8. output/spiece.model +3 -0
  9. output/tokenizer_config.json +17 -0
  10. output/train_results.json +8 -0
  11. output/trainer_state.json +4345 -0
  12. output/training_args.bin +0 -0
.gitattributes ADDED
@@ -0,0 +1,2 @@
 
 
 
1
+ output/model.safetensors filter=lfs diff=lfs merge=lfs -text
2
+ output/spiece.model filter=lfs diff=lfs merge=lfs -text
output/README.md ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: mit
3
+ tags:
4
+ - generated_from_trainer
5
+ model-index:
6
+ - name: output
7
+ results: []
8
+ ---
9
+
10
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
11
+ should probably proofread and complete it, then remove this comment. -->
12
+
13
+ # output
14
+
15
+ This model is a fine-tuned version of [rinna/japanese-gpt-1b](https://huggingface.co/rinna/japanese-gpt-1b) on an unknown dataset.
16
+
17
+ ## Model description
18
+
19
+ More information needed
20
+
21
+ ## Intended uses & limitations
22
+
23
+ More information needed
24
+
25
+ ## Training and evaluation data
26
+
27
+ More information needed
28
+
29
+ ## Training procedure
30
+
31
+ ### Training hyperparameters
32
+
33
+ The following hyperparameters were used during training:
34
+ - learning_rate: 1e-07
35
+ - train_batch_size: 2
36
+ - eval_batch_size: 8
37
+ - seed: 42
38
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
39
+ - lr_scheduler_type: constant
40
+ - num_epochs: 18.0
41
+
42
+ ### Training results
43
+
44
+
45
+
46
+ ### Framework versions
47
+
48
+ - Transformers 4.29.0.dev0
49
+ - Pytorch 2.1.0.dev20230503+cu117
50
+ - Datasets 2.12.0
51
+ - Tokenizers 0.13.3
output/all_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 18.0,
3
+ "train_loss": 3.0161131058520043,
4
+ "train_runtime": 38619.075,
5
+ "train_samples": 3040,
6
+ "train_samples_per_second": 1.417,
7
+ "train_steps_per_second": 0.708
8
+ }
output/config.json ADDED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "rinna/japanese-gpt-1b",
3
+ "activation_function": "gelu_fast",
4
+ "architectures": [
5
+ "GPT2LMHeadModel"
6
+ ],
7
+ "attn_pdrop": 0.1,
8
+ "bos_token_id": 2,
9
+ "embd_pdrop": 0.1,
10
+ "eos_token_id": 3,
11
+ "gradient_checkpointing": false,
12
+ "initializer_range": 0.02,
13
+ "layer_norm_epsilon": 1e-05,
14
+ "model_type": "gpt2",
15
+ "n_ctx": 1024,
16
+ "n_embd": 2048,
17
+ "n_head": 16,
18
+ "n_inner": 8192,
19
+ "n_layer": 24,
20
+ "n_positions": 1024,
21
+ "reorder_and_upcast_attn": false,
22
+ "resid_pdrop": 0.1,
23
+ "scale_attn_by_inverse_layer_idx": false,
24
+ "scale_attn_weights": true,
25
+ "summary_activation": null,
26
+ "summary_first_dropout": 0.1,
27
+ "summary_proj_to_labels": true,
28
+ "summary_type": "cls_index",
29
+ "summary_use_proj": true,
30
+ "torch_dtype": "float32",
31
+ "transformers_version": "4.29.0.dev0",
32
+ "use_cache": true,
33
+ "vocab_size": 44928
34
+ }
output/generation_config.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 2,
4
+ "eos_token_id": 3,
5
+ "transformers_version": "4.29.0.dev0"
6
+ }
output/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:47d98db14aee20d28f52982c02e371a0a5301ab39840807b4940c4c497d37d3d
3
+ size 5236050824
output/special_tokens_map.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "cls_token": "[CLS]",
4
+ "eos_token": "</s>",
5
+ "mask_token": "[MASK]",
6
+ "pad_token": "[PAD]",
7
+ "sep_token": "[SEP]",
8
+ "unk_token": "<unk>"
9
+ }
output/spiece.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9dbbd4ddbe43941051ed35fd44ff0d9d1c00ed345f7fd4d1969df174110f0609
3
+ size 1044749
output/tokenizer_config.json ADDED
@@ -0,0 +1,17 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [],
3
+ "bos_token": "<s>",
4
+ "clean_up_tokenization_spaces": true,
5
+ "cls_token": "[CLS]",
6
+ "do_lower_case": false,
7
+ "eos_token": "</s>",
8
+ "extra_ids": 0,
9
+ "mask_token": "[MASK]",
10
+ "model_max_length": 1000000000000000019884624838656,
11
+ "pad_token": "[PAD]",
12
+ "sep_token": "[SEP]",
13
+ "sp_model_kwargs": {},
14
+ "tokenizer_class": "T5Tokenizer",
15
+ "unk_token": "<unk>",
16
+ "use_fast": true
17
+ }
output/train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 18.0,
3
+ "train_loss": 3.0161131058520043,
4
+ "train_runtime": 38619.075,
5
+ "train_samples": 3040,
6
+ "train_samples_per_second": 1.417,
7
+ "train_steps_per_second": 0.708
8
+ }
output/trainer_state.json ADDED
@@ -0,0 +1,4345 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 18.0,
5
+ "global_step": 27360,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.03,
12
+ "learning_rate": 1e-07,
13
+ "loss": 6.0775,
14
+ "step": 38
15
+ },
16
+ {
17
+ "epoch": 0.05,
18
+ "learning_rate": 1e-07,
19
+ "loss": 6.0462,
20
+ "step": 76
21
+ },
22
+ {
23
+ "epoch": 0.07,
24
+ "learning_rate": 1e-07,
25
+ "loss": 6.0014,
26
+ "step": 114
27
+ },
28
+ {
29
+ "epoch": 0.1,
30
+ "learning_rate": 1e-07,
31
+ "loss": 5.9744,
32
+ "step": 152
33
+ },
34
+ {
35
+ "epoch": 0.12,
36
+ "learning_rate": 1e-07,
37
+ "loss": 5.886,
38
+ "step": 190
39
+ },
40
+ {
41
+ "epoch": 0.15,
42
+ "learning_rate": 1e-07,
43
+ "loss": 5.8752,
44
+ "step": 228
45
+ },
46
+ {
47
+ "epoch": 0.17,
48
+ "learning_rate": 1e-07,
49
+ "loss": 5.8586,
50
+ "step": 266
51
+ },
52
+ {
53
+ "epoch": 0.2,
54
+ "learning_rate": 1e-07,
55
+ "loss": 5.7714,
56
+ "step": 304
57
+ },
58
+ {
59
+ "epoch": 0.23,
60
+ "learning_rate": 1e-07,
61
+ "loss": 5.7406,
62
+ "step": 342
63
+ },
64
+ {
65
+ "epoch": 0.25,
66
+ "learning_rate": 1e-07,
67
+ "loss": 5.7096,
68
+ "step": 380
69
+ },
70
+ {
71
+ "epoch": 0.28,
72
+ "learning_rate": 1e-07,
73
+ "loss": 5.6967,
74
+ "step": 418
75
+ },
76
+ {
77
+ "epoch": 0.3,
78
+ "learning_rate": 1e-07,
79
+ "loss": 5.6248,
80
+ "step": 456
81
+ },
82
+ {
83
+ "epoch": 0.33,
84
+ "learning_rate": 1e-07,
85
+ "loss": 5.5632,
86
+ "step": 494
87
+ },
88
+ {
89
+ "epoch": 0.35,
90
+ "learning_rate": 1e-07,
91
+ "loss": 5.5463,
92
+ "step": 532
93
+ },
94
+ {
95
+ "epoch": 0.38,
96
+ "learning_rate": 1e-07,
97
+ "loss": 5.518,
98
+ "step": 570
99
+ },
100
+ {
101
+ "epoch": 0.4,
102
+ "learning_rate": 1e-07,
103
+ "loss": 5.4135,
104
+ "step": 608
105
+ },
106
+ {
107
+ "epoch": 0.42,
108
+ "learning_rate": 1e-07,
109
+ "loss": 5.4107,
110
+ "step": 646
111
+ },
112
+ {
113
+ "epoch": 0.45,
114
+ "learning_rate": 1e-07,
115
+ "loss": 5.3979,
116
+ "step": 684
117
+ },
118
+ {
119
+ "epoch": 0.47,
120
+ "learning_rate": 1e-07,
121
+ "loss": 5.3563,
122
+ "step": 722
123
+ },
124
+ {
125
+ "epoch": 0.5,
126
+ "learning_rate": 1e-07,
127
+ "loss": 5.3835,
128
+ "step": 760
129
+ },
130
+ {
131
+ "epoch": 0.53,
132
+ "learning_rate": 1e-07,
133
+ "loss": 5.3005,
134
+ "step": 798
135
+ },
136
+ {
137
+ "epoch": 0.55,
138
+ "learning_rate": 1e-07,
139
+ "loss": 5.2324,
140
+ "step": 836
141
+ },
142
+ {
143
+ "epoch": 0.57,
144
+ "learning_rate": 1e-07,
145
+ "loss": 5.2447,
146
+ "step": 874
147
+ },
148
+ {
149
+ "epoch": 0.6,
150
+ "learning_rate": 1e-07,
151
+ "loss": 5.1957,
152
+ "step": 912
153
+ },
154
+ {
155
+ "epoch": 0.62,
156
+ "learning_rate": 1e-07,
157
+ "loss": 5.1413,
158
+ "step": 950
159
+ },
160
+ {
161
+ "epoch": 0.65,
162
+ "learning_rate": 1e-07,
163
+ "loss": 5.1094,
164
+ "step": 988
165
+ },
166
+ {
167
+ "epoch": 0.68,
168
+ "learning_rate": 1e-07,
169
+ "loss": 5.0822,
170
+ "step": 1026
171
+ },
172
+ {
173
+ "epoch": 0.7,
174
+ "learning_rate": 1e-07,
175
+ "loss": 5.0826,
176
+ "step": 1064
177
+ },
178
+ {
179
+ "epoch": 0.72,
180
+ "learning_rate": 1e-07,
181
+ "loss": 5.0354,
182
+ "step": 1102
183
+ },
184
+ {
185
+ "epoch": 0.75,
186
+ "learning_rate": 1e-07,
187
+ "loss": 4.9899,
188
+ "step": 1140
189
+ },
190
+ {
191
+ "epoch": 0.78,
192
+ "learning_rate": 1e-07,
193
+ "loss": 5.0207,
194
+ "step": 1178
195
+ },
196
+ {
197
+ "epoch": 0.8,
198
+ "learning_rate": 1e-07,
199
+ "loss": 4.9249,
200
+ "step": 1216
201
+ },
202
+ {
203
+ "epoch": 0.82,
204
+ "learning_rate": 1e-07,
205
+ "loss": 4.9205,
206
+ "step": 1254
207
+ },
208
+ {
209
+ "epoch": 0.85,
210
+ "learning_rate": 1e-07,
211
+ "loss": 4.8866,
212
+ "step": 1292
213
+ },
214
+ {
215
+ "epoch": 0.88,
216
+ "learning_rate": 1e-07,
217
+ "loss": 4.9054,
218
+ "step": 1330
219
+ },
220
+ {
221
+ "epoch": 0.9,
222
+ "learning_rate": 1e-07,
223
+ "loss": 4.9223,
224
+ "step": 1368
225
+ },
226
+ {
227
+ "epoch": 0.93,
228
+ "learning_rate": 1e-07,
229
+ "loss": 4.8158,
230
+ "step": 1406
231
+ },
232
+ {
233
+ "epoch": 0.95,
234
+ "learning_rate": 1e-07,
235
+ "loss": 4.842,
236
+ "step": 1444
237
+ },
238
+ {
239
+ "epoch": 0.97,
240
+ "learning_rate": 1e-07,
241
+ "loss": 4.7924,
242
+ "step": 1482
243
+ },
244
+ {
245
+ "epoch": 1.0,
246
+ "learning_rate": 1e-07,
247
+ "loss": 4.7735,
248
+ "step": 1520
249
+ },
250
+ {
251
+ "epoch": 1.02,
252
+ "learning_rate": 1e-07,
253
+ "loss": 4.803,
254
+ "step": 1558
255
+ },
256
+ {
257
+ "epoch": 1.05,
258
+ "learning_rate": 1e-07,
259
+ "loss": 4.7616,
260
+ "step": 1596
261
+ },
262
+ {
263
+ "epoch": 1.07,
264
+ "learning_rate": 1e-07,
265
+ "loss": 4.7375,
266
+ "step": 1634
267
+ },
268
+ {
269
+ "epoch": 1.1,
270
+ "learning_rate": 1e-07,
271
+ "loss": 4.7046,
272
+ "step": 1672
273
+ },
274
+ {
275
+ "epoch": 1.12,
276
+ "learning_rate": 1e-07,
277
+ "loss": 4.6259,
278
+ "step": 1710
279
+ },
280
+ {
281
+ "epoch": 1.15,
282
+ "learning_rate": 1e-07,
283
+ "loss": 4.6591,
284
+ "step": 1748
285
+ },
286
+ {
287
+ "epoch": 1.18,
288
+ "learning_rate": 1e-07,
289
+ "loss": 4.6459,
290
+ "step": 1786
291
+ },
292
+ {
293
+ "epoch": 1.2,
294
+ "learning_rate": 1e-07,
295
+ "loss": 4.5775,
296
+ "step": 1824
297
+ },
298
+ {
299
+ "epoch": 1.23,
300
+ "learning_rate": 1e-07,
301
+ "loss": 4.5543,
302
+ "step": 1862
303
+ },
304
+ {
305
+ "epoch": 1.25,
306
+ "learning_rate": 1e-07,
307
+ "loss": 4.6039,
308
+ "step": 1900
309
+ },
310
+ {
311
+ "epoch": 1.27,
312
+ "learning_rate": 1e-07,
313
+ "loss": 4.5725,
314
+ "step": 1938
315
+ },
316
+ {
317
+ "epoch": 1.3,
318
+ "learning_rate": 1e-07,
319
+ "loss": 4.55,
320
+ "step": 1976
321
+ },
322
+ {
323
+ "epoch": 1.32,
324
+ "learning_rate": 1e-07,
325
+ "loss": 4.4842,
326
+ "step": 2014
327
+ },
328
+ {
329
+ "epoch": 1.35,
330
+ "learning_rate": 1e-07,
331
+ "loss": 4.4519,
332
+ "step": 2052
333
+ },
334
+ {
335
+ "epoch": 1.38,
336
+ "learning_rate": 1e-07,
337
+ "loss": 4.4428,
338
+ "step": 2090
339
+ },
340
+ {
341
+ "epoch": 1.4,
342
+ "learning_rate": 1e-07,
343
+ "loss": 4.4975,
344
+ "step": 2128
345
+ },
346
+ {
347
+ "epoch": 1.43,
348
+ "learning_rate": 1e-07,
349
+ "loss": 4.4482,
350
+ "step": 2166
351
+ },
352
+ {
353
+ "epoch": 1.45,
354
+ "learning_rate": 1e-07,
355
+ "loss": 4.439,
356
+ "step": 2204
357
+ },
358
+ {
359
+ "epoch": 1.48,
360
+ "learning_rate": 1e-07,
361
+ "loss": 4.4258,
362
+ "step": 2242
363
+ },
364
+ {
365
+ "epoch": 1.5,
366
+ "learning_rate": 1e-07,
367
+ "loss": 4.4161,
368
+ "step": 2280
369
+ },
370
+ {
371
+ "epoch": 1.52,
372
+ "learning_rate": 1e-07,
373
+ "loss": 4.4324,
374
+ "step": 2318
375
+ },
376
+ {
377
+ "epoch": 1.55,
378
+ "learning_rate": 1e-07,
379
+ "loss": 4.3676,
380
+ "step": 2356
381
+ },
382
+ {
383
+ "epoch": 1.57,
384
+ "learning_rate": 1e-07,
385
+ "loss": 4.35,
386
+ "step": 2394
387
+ },
388
+ {
389
+ "epoch": 1.6,
390
+ "learning_rate": 1e-07,
391
+ "loss": 4.3224,
392
+ "step": 2432
393
+ },
394
+ {
395
+ "epoch": 1.62,
396
+ "learning_rate": 1e-07,
397
+ "loss": 4.3432,
398
+ "step": 2470
399
+ },
400
+ {
401
+ "epoch": 1.65,
402
+ "learning_rate": 1e-07,
403
+ "loss": 4.2577,
404
+ "step": 2508
405
+ },
406
+ {
407
+ "epoch": 1.68,
408
+ "learning_rate": 1e-07,
409
+ "loss": 4.385,
410
+ "step": 2546
411
+ },
412
+ {
413
+ "epoch": 1.7,
414
+ "learning_rate": 1e-07,
415
+ "loss": 4.2489,
416
+ "step": 2584
417
+ },
418
+ {
419
+ "epoch": 1.73,
420
+ "learning_rate": 1e-07,
421
+ "loss": 4.2986,
422
+ "step": 2622
423
+ },
424
+ {
425
+ "epoch": 1.75,
426
+ "learning_rate": 1e-07,
427
+ "loss": 4.2423,
428
+ "step": 2660
429
+ },
430
+ {
431
+ "epoch": 1.77,
432
+ "learning_rate": 1e-07,
433
+ "loss": 4.2712,
434
+ "step": 2698
435
+ },
436
+ {
437
+ "epoch": 1.8,
438
+ "learning_rate": 1e-07,
439
+ "loss": 4.2693,
440
+ "step": 2736
441
+ },
442
+ {
443
+ "epoch": 1.82,
444
+ "learning_rate": 1e-07,
445
+ "loss": 4.1746,
446
+ "step": 2774
447
+ },
448
+ {
449
+ "epoch": 1.85,
450
+ "learning_rate": 1e-07,
451
+ "loss": 4.1345,
452
+ "step": 2812
453
+ },
454
+ {
455
+ "epoch": 1.88,
456
+ "learning_rate": 1e-07,
457
+ "loss": 4.0853,
458
+ "step": 2850
459
+ },
460
+ {
461
+ "epoch": 1.9,
462
+ "learning_rate": 1e-07,
463
+ "loss": 4.1348,
464
+ "step": 2888
465
+ },
466
+ {
467
+ "epoch": 1.93,
468
+ "learning_rate": 1e-07,
469
+ "loss": 4.1303,
470
+ "step": 2926
471
+ },
472
+ {
473
+ "epoch": 1.95,
474
+ "learning_rate": 1e-07,
475
+ "loss": 4.0622,
476
+ "step": 2964
477
+ },
478
+ {
479
+ "epoch": 1.98,
480
+ "learning_rate": 1e-07,
481
+ "loss": 4.0644,
482
+ "step": 3002
483
+ },
484
+ {
485
+ "epoch": 2.0,
486
+ "learning_rate": 1e-07,
487
+ "loss": 4.0578,
488
+ "step": 3040
489
+ },
490
+ {
491
+ "epoch": 2.02,
492
+ "learning_rate": 1e-07,
493
+ "loss": 3.9785,
494
+ "step": 3078
495
+ },
496
+ {
497
+ "epoch": 2.05,
498
+ "learning_rate": 1e-07,
499
+ "loss": 4.0323,
500
+ "step": 3116
501
+ },
502
+ {
503
+ "epoch": 2.08,
504
+ "learning_rate": 1e-07,
505
+ "loss": 4.0035,
506
+ "step": 3154
507
+ },
508
+ {
509
+ "epoch": 2.1,
510
+ "learning_rate": 1e-07,
511
+ "loss": 3.9697,
512
+ "step": 3192
513
+ },
514
+ {
515
+ "epoch": 2.12,
516
+ "learning_rate": 1e-07,
517
+ "loss": 3.9225,
518
+ "step": 3230
519
+ },
520
+ {
521
+ "epoch": 2.15,
522
+ "learning_rate": 1e-07,
523
+ "loss": 3.9822,
524
+ "step": 3268
525
+ },
526
+ {
527
+ "epoch": 2.17,
528
+ "learning_rate": 1e-07,
529
+ "loss": 3.9636,
530
+ "step": 3306
531
+ },
532
+ {
533
+ "epoch": 2.2,
534
+ "learning_rate": 1e-07,
535
+ "loss": 3.9574,
536
+ "step": 3344
537
+ },
538
+ {
539
+ "epoch": 2.23,
540
+ "learning_rate": 1e-07,
541
+ "loss": 3.8573,
542
+ "step": 3382
543
+ },
544
+ {
545
+ "epoch": 2.25,
546
+ "learning_rate": 1e-07,
547
+ "loss": 4.0107,
548
+ "step": 3420
549
+ },
550
+ {
551
+ "epoch": 2.27,
552
+ "learning_rate": 1e-07,
553
+ "loss": 3.8145,
554
+ "step": 3458
555
+ },
556
+ {
557
+ "epoch": 2.3,
558
+ "learning_rate": 1e-07,
559
+ "loss": 3.8424,
560
+ "step": 3496
561
+ },
562
+ {
563
+ "epoch": 2.33,
564
+ "learning_rate": 1e-07,
565
+ "loss": 3.8641,
566
+ "step": 3534
567
+ },
568
+ {
569
+ "epoch": 2.35,
570
+ "learning_rate": 1e-07,
571
+ "loss": 3.8606,
572
+ "step": 3572
573
+ },
574
+ {
575
+ "epoch": 2.38,
576
+ "learning_rate": 1e-07,
577
+ "loss": 3.7903,
578
+ "step": 3610
579
+ },
580
+ {
581
+ "epoch": 2.4,
582
+ "learning_rate": 1e-07,
583
+ "loss": 3.7378,
584
+ "step": 3648
585
+ },
586
+ {
587
+ "epoch": 2.42,
588
+ "learning_rate": 1e-07,
589
+ "loss": 3.7713,
590
+ "step": 3686
591
+ },
592
+ {
593
+ "epoch": 2.45,
594
+ "learning_rate": 1e-07,
595
+ "loss": 3.834,
596
+ "step": 3724
597
+ },
598
+ {
599
+ "epoch": 2.48,
600
+ "learning_rate": 1e-07,
601
+ "loss": 3.7785,
602
+ "step": 3762
603
+ },
604
+ {
605
+ "epoch": 2.5,
606
+ "learning_rate": 1e-07,
607
+ "loss": 3.8277,
608
+ "step": 3800
609
+ },
610
+ {
611
+ "epoch": 2.52,
612
+ "learning_rate": 1e-07,
613
+ "loss": 3.7287,
614
+ "step": 3838
615
+ },
616
+ {
617
+ "epoch": 2.55,
618
+ "learning_rate": 1e-07,
619
+ "loss": 3.7123,
620
+ "step": 3876
621
+ },
622
+ {
623
+ "epoch": 2.58,
624
+ "learning_rate": 1e-07,
625
+ "loss": 3.7185,
626
+ "step": 3914
627
+ },
628
+ {
629
+ "epoch": 2.6,
630
+ "learning_rate": 1e-07,
631
+ "loss": 3.6936,
632
+ "step": 3952
633
+ },
634
+ {
635
+ "epoch": 2.62,
636
+ "learning_rate": 1e-07,
637
+ "loss": 3.7462,
638
+ "step": 3990
639
+ },
640
+ {
641
+ "epoch": 2.65,
642
+ "learning_rate": 1e-07,
643
+ "loss": 3.6844,
644
+ "step": 4028
645
+ },
646
+ {
647
+ "epoch": 2.67,
648
+ "learning_rate": 1e-07,
649
+ "loss": 3.6709,
650
+ "step": 4066
651
+ },
652
+ {
653
+ "epoch": 2.7,
654
+ "learning_rate": 1e-07,
655
+ "loss": 3.6508,
656
+ "step": 4104
657
+ },
658
+ {
659
+ "epoch": 2.73,
660
+ "learning_rate": 1e-07,
661
+ "loss": 3.6247,
662
+ "step": 4142
663
+ },
664
+ {
665
+ "epoch": 2.75,
666
+ "learning_rate": 1e-07,
667
+ "loss": 3.6459,
668
+ "step": 4180
669
+ },
670
+ {
671
+ "epoch": 2.77,
672
+ "learning_rate": 1e-07,
673
+ "loss": 3.6169,
674
+ "step": 4218
675
+ },
676
+ {
677
+ "epoch": 2.8,
678
+ "learning_rate": 1e-07,
679
+ "loss": 3.6944,
680
+ "step": 4256
681
+ },
682
+ {
683
+ "epoch": 2.83,
684
+ "learning_rate": 1e-07,
685
+ "loss": 3.6458,
686
+ "step": 4294
687
+ },
688
+ {
689
+ "epoch": 2.85,
690
+ "learning_rate": 1e-07,
691
+ "loss": 3.5598,
692
+ "step": 4332
693
+ },
694
+ {
695
+ "epoch": 2.88,
696
+ "learning_rate": 1e-07,
697
+ "loss": 3.5221,
698
+ "step": 4370
699
+ },
700
+ {
701
+ "epoch": 2.9,
702
+ "learning_rate": 1e-07,
703
+ "loss": 3.6462,
704
+ "step": 4408
705
+ },
706
+ {
707
+ "epoch": 2.92,
708
+ "learning_rate": 1e-07,
709
+ "loss": 3.6155,
710
+ "step": 4446
711
+ },
712
+ {
713
+ "epoch": 2.95,
714
+ "learning_rate": 1e-07,
715
+ "loss": 3.5649,
716
+ "step": 4484
717
+ },
718
+ {
719
+ "epoch": 2.98,
720
+ "learning_rate": 1e-07,
721
+ "loss": 3.5088,
722
+ "step": 4522
723
+ },
724
+ {
725
+ "epoch": 3.0,
726
+ "learning_rate": 1e-07,
727
+ "loss": 3.6479,
728
+ "step": 4560
729
+ },
730
+ {
731
+ "epoch": 3.02,
732
+ "learning_rate": 1e-07,
733
+ "loss": 3.5495,
734
+ "step": 4598
735
+ },
736
+ {
737
+ "epoch": 3.05,
738
+ "learning_rate": 1e-07,
739
+ "loss": 3.4842,
740
+ "step": 4636
741
+ },
742
+ {
743
+ "epoch": 3.08,
744
+ "learning_rate": 1e-07,
745
+ "loss": 3.6004,
746
+ "step": 4674
747
+ },
748
+ {
749
+ "epoch": 3.1,
750
+ "learning_rate": 1e-07,
751
+ "loss": 3.4871,
752
+ "step": 4712
753
+ },
754
+ {
755
+ "epoch": 3.12,
756
+ "learning_rate": 1e-07,
757
+ "loss": 3.5237,
758
+ "step": 4750
759
+ },
760
+ {
761
+ "epoch": 3.15,
762
+ "learning_rate": 1e-07,
763
+ "loss": 3.5153,
764
+ "step": 4788
765
+ },
766
+ {
767
+ "epoch": 3.17,
768
+ "learning_rate": 1e-07,
769
+ "loss": 3.4592,
770
+ "step": 4826
771
+ },
772
+ {
773
+ "epoch": 3.2,
774
+ "learning_rate": 1e-07,
775
+ "loss": 3.5256,
776
+ "step": 4864
777
+ },
778
+ {
779
+ "epoch": 3.23,
780
+ "learning_rate": 1e-07,
781
+ "loss": 3.4191,
782
+ "step": 4902
783
+ },
784
+ {
785
+ "epoch": 3.25,
786
+ "learning_rate": 1e-07,
787
+ "loss": 3.4433,
788
+ "step": 4940
789
+ },
790
+ {
791
+ "epoch": 3.27,
792
+ "learning_rate": 1e-07,
793
+ "loss": 3.51,
794
+ "step": 4978
795
+ },
796
+ {
797
+ "epoch": 3.3,
798
+ "learning_rate": 1e-07,
799
+ "loss": 3.4299,
800
+ "step": 5016
801
+ },
802
+ {
803
+ "epoch": 3.33,
804
+ "learning_rate": 1e-07,
805
+ "loss": 3.4375,
806
+ "step": 5054
807
+ },
808
+ {
809
+ "epoch": 3.35,
810
+ "learning_rate": 1e-07,
811
+ "loss": 3.4766,
812
+ "step": 5092
813
+ },
814
+ {
815
+ "epoch": 3.38,
816
+ "learning_rate": 1e-07,
817
+ "loss": 3.4243,
818
+ "step": 5130
819
+ },
820
+ {
821
+ "epoch": 3.4,
822
+ "learning_rate": 1e-07,
823
+ "loss": 3.4521,
824
+ "step": 5168
825
+ },
826
+ {
827
+ "epoch": 3.42,
828
+ "learning_rate": 1e-07,
829
+ "loss": 3.4048,
830
+ "step": 5206
831
+ },
832
+ {
833
+ "epoch": 3.45,
834
+ "learning_rate": 1e-07,
835
+ "loss": 3.4903,
836
+ "step": 5244
837
+ },
838
+ {
839
+ "epoch": 3.48,
840
+ "learning_rate": 1e-07,
841
+ "loss": 3.3147,
842
+ "step": 5282
843
+ },
844
+ {
845
+ "epoch": 3.5,
846
+ "learning_rate": 1e-07,
847
+ "loss": 3.4223,
848
+ "step": 5320
849
+ },
850
+ {
851
+ "epoch": 3.52,
852
+ "learning_rate": 1e-07,
853
+ "loss": 3.3592,
854
+ "step": 5358
855
+ },
856
+ {
857
+ "epoch": 3.55,
858
+ "learning_rate": 1e-07,
859
+ "loss": 3.4287,
860
+ "step": 5396
861
+ },
862
+ {
863
+ "epoch": 3.58,
864
+ "learning_rate": 1e-07,
865
+ "loss": 3.2888,
866
+ "step": 5434
867
+ },
868
+ {
869
+ "epoch": 3.6,
870
+ "learning_rate": 1e-07,
871
+ "loss": 3.352,
872
+ "step": 5472
873
+ },
874
+ {
875
+ "epoch": 3.62,
876
+ "learning_rate": 1e-07,
877
+ "loss": 3.352,
878
+ "step": 5510
879
+ },
880
+ {
881
+ "epoch": 3.65,
882
+ "learning_rate": 1e-07,
883
+ "loss": 3.4211,
884
+ "step": 5548
885
+ },
886
+ {
887
+ "epoch": 3.67,
888
+ "learning_rate": 1e-07,
889
+ "loss": 3.4113,
890
+ "step": 5586
891
+ },
892
+ {
893
+ "epoch": 3.7,
894
+ "learning_rate": 1e-07,
895
+ "loss": 3.2818,
896
+ "step": 5624
897
+ },
898
+ {
899
+ "epoch": 3.73,
900
+ "learning_rate": 1e-07,
901
+ "loss": 3.368,
902
+ "step": 5662
903
+ },
904
+ {
905
+ "epoch": 3.75,
906
+ "learning_rate": 1e-07,
907
+ "loss": 3.3881,
908
+ "step": 5700
909
+ },
910
+ {
911
+ "epoch": 3.77,
912
+ "learning_rate": 1e-07,
913
+ "loss": 3.3362,
914
+ "step": 5738
915
+ },
916
+ {
917
+ "epoch": 3.8,
918
+ "learning_rate": 1e-07,
919
+ "loss": 3.3529,
920
+ "step": 5776
921
+ },
922
+ {
923
+ "epoch": 3.83,
924
+ "learning_rate": 1e-07,
925
+ "loss": 3.2348,
926
+ "step": 5814
927
+ },
928
+ {
929
+ "epoch": 3.85,
930
+ "learning_rate": 1e-07,
931
+ "loss": 3.3256,
932
+ "step": 5852
933
+ },
934
+ {
935
+ "epoch": 3.88,
936
+ "learning_rate": 1e-07,
937
+ "loss": 3.2392,
938
+ "step": 5890
939
+ },
940
+ {
941
+ "epoch": 3.9,
942
+ "learning_rate": 1e-07,
943
+ "loss": 3.2072,
944
+ "step": 5928
945
+ },
946
+ {
947
+ "epoch": 3.92,
948
+ "learning_rate": 1e-07,
949
+ "loss": 3.2969,
950
+ "step": 5966
951
+ },
952
+ {
953
+ "epoch": 3.95,
954
+ "learning_rate": 1e-07,
955
+ "loss": 3.2984,
956
+ "step": 6004
957
+ },
958
+ {
959
+ "epoch": 3.98,
960
+ "learning_rate": 1e-07,
961
+ "loss": 3.2254,
962
+ "step": 6042
963
+ },
964
+ {
965
+ "epoch": 4.0,
966
+ "learning_rate": 1e-07,
967
+ "loss": 3.3567,
968
+ "step": 6080
969
+ },
970
+ {
971
+ "epoch": 4.03,
972
+ "learning_rate": 1e-07,
973
+ "loss": 3.2742,
974
+ "step": 6118
975
+ },
976
+ {
977
+ "epoch": 4.05,
978
+ "learning_rate": 1e-07,
979
+ "loss": 3.3057,
980
+ "step": 6156
981
+ },
982
+ {
983
+ "epoch": 4.08,
984
+ "learning_rate": 1e-07,
985
+ "loss": 3.1336,
986
+ "step": 6194
987
+ },
988
+ {
989
+ "epoch": 4.1,
990
+ "learning_rate": 1e-07,
991
+ "loss": 3.2602,
992
+ "step": 6232
993
+ },
994
+ {
995
+ "epoch": 4.12,
996
+ "learning_rate": 1e-07,
997
+ "loss": 3.3183,
998
+ "step": 6270
999
+ },
1000
+ {
1001
+ "epoch": 4.15,
1002
+ "learning_rate": 1e-07,
1003
+ "loss": 3.2419,
1004
+ "step": 6308
1005
+ },
1006
+ {
1007
+ "epoch": 4.17,
1008
+ "learning_rate": 1e-07,
1009
+ "loss": 3.2496,
1010
+ "step": 6346
1011
+ },
1012
+ {
1013
+ "epoch": 4.2,
1014
+ "learning_rate": 1e-07,
1015
+ "loss": 3.2705,
1016
+ "step": 6384
1017
+ },
1018
+ {
1019
+ "epoch": 4.22,
1020
+ "learning_rate": 1e-07,
1021
+ "loss": 3.1646,
1022
+ "step": 6422
1023
+ },
1024
+ {
1025
+ "epoch": 4.25,
1026
+ "learning_rate": 1e-07,
1027
+ "loss": 3.1935,
1028
+ "step": 6460
1029
+ },
1030
+ {
1031
+ "epoch": 4.28,
1032
+ "learning_rate": 1e-07,
1033
+ "loss": 3.2376,
1034
+ "step": 6498
1035
+ },
1036
+ {
1037
+ "epoch": 4.3,
1038
+ "learning_rate": 1e-07,
1039
+ "loss": 3.1907,
1040
+ "step": 6536
1041
+ },
1042
+ {
1043
+ "epoch": 4.33,
1044
+ "learning_rate": 1e-07,
1045
+ "loss": 3.1895,
1046
+ "step": 6574
1047
+ },
1048
+ {
1049
+ "epoch": 4.35,
1050
+ "learning_rate": 1e-07,
1051
+ "loss": 3.2795,
1052
+ "step": 6612
1053
+ },
1054
+ {
1055
+ "epoch": 4.38,
1056
+ "learning_rate": 1e-07,
1057
+ "loss": 3.1587,
1058
+ "step": 6650
1059
+ },
1060
+ {
1061
+ "epoch": 4.4,
1062
+ "learning_rate": 1e-07,
1063
+ "loss": 3.2195,
1064
+ "step": 6688
1065
+ },
1066
+ {
1067
+ "epoch": 4.42,
1068
+ "learning_rate": 1e-07,
1069
+ "loss": 3.2117,
1070
+ "step": 6726
1071
+ },
1072
+ {
1073
+ "epoch": 4.45,
1074
+ "learning_rate": 1e-07,
1075
+ "loss": 3.1271,
1076
+ "step": 6764
1077
+ },
1078
+ {
1079
+ "epoch": 4.47,
1080
+ "learning_rate": 1e-07,
1081
+ "loss": 3.1652,
1082
+ "step": 6802
1083
+ },
1084
+ {
1085
+ "epoch": 4.5,
1086
+ "learning_rate": 1e-07,
1087
+ "loss": 3.1776,
1088
+ "step": 6840
1089
+ },
1090
+ {
1091
+ "epoch": 4.53,
1092
+ "learning_rate": 1e-07,
1093
+ "loss": 3.0865,
1094
+ "step": 6878
1095
+ },
1096
+ {
1097
+ "epoch": 4.55,
1098
+ "learning_rate": 1e-07,
1099
+ "loss": 3.1704,
1100
+ "step": 6916
1101
+ },
1102
+ {
1103
+ "epoch": 4.58,
1104
+ "learning_rate": 1e-07,
1105
+ "loss": 3.1358,
1106
+ "step": 6954
1107
+ },
1108
+ {
1109
+ "epoch": 4.6,
1110
+ "learning_rate": 1e-07,
1111
+ "loss": 3.1571,
1112
+ "step": 6992
1113
+ },
1114
+ {
1115
+ "epoch": 4.62,
1116
+ "learning_rate": 1e-07,
1117
+ "loss": 3.1921,
1118
+ "step": 7030
1119
+ },
1120
+ {
1121
+ "epoch": 4.65,
1122
+ "learning_rate": 1e-07,
1123
+ "loss": 3.1816,
1124
+ "step": 7068
1125
+ },
1126
+ {
1127
+ "epoch": 4.67,
1128
+ "learning_rate": 1e-07,
1129
+ "loss": 3.0454,
1130
+ "step": 7106
1131
+ },
1132
+ {
1133
+ "epoch": 4.7,
1134
+ "learning_rate": 1e-07,
1135
+ "loss": 3.0481,
1136
+ "step": 7144
1137
+ },
1138
+ {
1139
+ "epoch": 4.72,
1140
+ "learning_rate": 1e-07,
1141
+ "loss": 3.1092,
1142
+ "step": 7182
1143
+ },
1144
+ {
1145
+ "epoch": 4.75,
1146
+ "learning_rate": 1e-07,
1147
+ "loss": 3.1149,
1148
+ "step": 7220
1149
+ },
1150
+ {
1151
+ "epoch": 4.78,
1152
+ "learning_rate": 1e-07,
1153
+ "loss": 3.1654,
1154
+ "step": 7258
1155
+ },
1156
+ {
1157
+ "epoch": 4.8,
1158
+ "learning_rate": 1e-07,
1159
+ "loss": 3.1102,
1160
+ "step": 7296
1161
+ },
1162
+ {
1163
+ "epoch": 4.83,
1164
+ "learning_rate": 1e-07,
1165
+ "loss": 3.1573,
1166
+ "step": 7334
1167
+ },
1168
+ {
1169
+ "epoch": 4.85,
1170
+ "learning_rate": 1e-07,
1171
+ "loss": 3.0639,
1172
+ "step": 7372
1173
+ },
1174
+ {
1175
+ "epoch": 4.88,
1176
+ "learning_rate": 1e-07,
1177
+ "loss": 3.0567,
1178
+ "step": 7410
1179
+ },
1180
+ {
1181
+ "epoch": 4.9,
1182
+ "learning_rate": 1e-07,
1183
+ "loss": 3.0979,
1184
+ "step": 7448
1185
+ },
1186
+ {
1187
+ "epoch": 4.92,
1188
+ "learning_rate": 1e-07,
1189
+ "loss": 3.0217,
1190
+ "step": 7486
1191
+ },
1192
+ {
1193
+ "epoch": 4.95,
1194
+ "learning_rate": 1e-07,
1195
+ "loss": 3.0613,
1196
+ "step": 7524
1197
+ },
1198
+ {
1199
+ "epoch": 4.97,
1200
+ "learning_rate": 1e-07,
1201
+ "loss": 3.0275,
1202
+ "step": 7562
1203
+ },
1204
+ {
1205
+ "epoch": 5.0,
1206
+ "learning_rate": 1e-07,
1207
+ "loss": 3.0743,
1208
+ "step": 7600
1209
+ },
1210
+ {
1211
+ "epoch": 5.03,
1212
+ "learning_rate": 1e-07,
1213
+ "loss": 3.0545,
1214
+ "step": 7638
1215
+ },
1216
+ {
1217
+ "epoch": 5.05,
1218
+ "learning_rate": 1e-07,
1219
+ "loss": 3.0066,
1220
+ "step": 7676
1221
+ },
1222
+ {
1223
+ "epoch": 5.08,
1224
+ "learning_rate": 1e-07,
1225
+ "loss": 3.0473,
1226
+ "step": 7714
1227
+ },
1228
+ {
1229
+ "epoch": 5.1,
1230
+ "learning_rate": 1e-07,
1231
+ "loss": 3.0846,
1232
+ "step": 7752
1233
+ },
1234
+ {
1235
+ "epoch": 5.12,
1236
+ "learning_rate": 1e-07,
1237
+ "loss": 3.1315,
1238
+ "step": 7790
1239
+ },
1240
+ {
1241
+ "epoch": 5.15,
1242
+ "learning_rate": 1e-07,
1243
+ "loss": 2.9579,
1244
+ "step": 7828
1245
+ },
1246
+ {
1247
+ "epoch": 5.17,
1248
+ "learning_rate": 1e-07,
1249
+ "loss": 3.0408,
1250
+ "step": 7866
1251
+ },
1252
+ {
1253
+ "epoch": 5.2,
1254
+ "learning_rate": 1e-07,
1255
+ "loss": 3.0525,
1256
+ "step": 7904
1257
+ },
1258
+ {
1259
+ "epoch": 5.22,
1260
+ "learning_rate": 1e-07,
1261
+ "loss": 3.0084,
1262
+ "step": 7942
1263
+ },
1264
+ {
1265
+ "epoch": 5.25,
1266
+ "learning_rate": 1e-07,
1267
+ "loss": 3.0704,
1268
+ "step": 7980
1269
+ },
1270
+ {
1271
+ "epoch": 5.28,
1272
+ "learning_rate": 1e-07,
1273
+ "loss": 3.0312,
1274
+ "step": 8018
1275
+ },
1276
+ {
1277
+ "epoch": 5.3,
1278
+ "learning_rate": 1e-07,
1279
+ "loss": 2.9437,
1280
+ "step": 8056
1281
+ },
1282
+ {
1283
+ "epoch": 5.33,
1284
+ "learning_rate": 1e-07,
1285
+ "loss": 3.0565,
1286
+ "step": 8094
1287
+ },
1288
+ {
1289
+ "epoch": 5.35,
1290
+ "learning_rate": 1e-07,
1291
+ "loss": 2.9435,
1292
+ "step": 8132
1293
+ },
1294
+ {
1295
+ "epoch": 5.38,
1296
+ "learning_rate": 1e-07,
1297
+ "loss": 2.9414,
1298
+ "step": 8170
1299
+ },
1300
+ {
1301
+ "epoch": 5.4,
1302
+ "learning_rate": 1e-07,
1303
+ "loss": 3.0381,
1304
+ "step": 8208
1305
+ },
1306
+ {
1307
+ "epoch": 5.42,
1308
+ "learning_rate": 1e-07,
1309
+ "loss": 3.0162,
1310
+ "step": 8246
1311
+ },
1312
+ {
1313
+ "epoch": 5.45,
1314
+ "learning_rate": 1e-07,
1315
+ "loss": 3.0205,
1316
+ "step": 8284
1317
+ },
1318
+ {
1319
+ "epoch": 5.47,
1320
+ "learning_rate": 1e-07,
1321
+ "loss": 2.9968,
1322
+ "step": 8322
1323
+ },
1324
+ {
1325
+ "epoch": 5.5,
1326
+ "learning_rate": 1e-07,
1327
+ "loss": 3.0021,
1328
+ "step": 8360
1329
+ },
1330
+ {
1331
+ "epoch": 5.53,
1332
+ "learning_rate": 1e-07,
1333
+ "loss": 2.9997,
1334
+ "step": 8398
1335
+ },
1336
+ {
1337
+ "epoch": 5.55,
1338
+ "learning_rate": 1e-07,
1339
+ "loss": 3.0112,
1340
+ "step": 8436
1341
+ },
1342
+ {
1343
+ "epoch": 5.58,
1344
+ "learning_rate": 1e-07,
1345
+ "loss": 3.0385,
1346
+ "step": 8474
1347
+ },
1348
+ {
1349
+ "epoch": 5.6,
1350
+ "learning_rate": 1e-07,
1351
+ "loss": 2.9613,
1352
+ "step": 8512
1353
+ },
1354
+ {
1355
+ "epoch": 5.62,
1356
+ "learning_rate": 1e-07,
1357
+ "loss": 2.9484,
1358
+ "step": 8550
1359
+ },
1360
+ {
1361
+ "epoch": 5.65,
1362
+ "learning_rate": 1e-07,
1363
+ "loss": 2.979,
1364
+ "step": 8588
1365
+ },
1366
+ {
1367
+ "epoch": 5.67,
1368
+ "learning_rate": 1e-07,
1369
+ "loss": 2.9796,
1370
+ "step": 8626
1371
+ },
1372
+ {
1373
+ "epoch": 5.7,
1374
+ "learning_rate": 1e-07,
1375
+ "loss": 2.9801,
1376
+ "step": 8664
1377
+ },
1378
+ {
1379
+ "epoch": 5.72,
1380
+ "learning_rate": 1e-07,
1381
+ "loss": 3.0399,
1382
+ "step": 8702
1383
+ },
1384
+ {
1385
+ "epoch": 5.75,
1386
+ "learning_rate": 1e-07,
1387
+ "loss": 2.9223,
1388
+ "step": 8740
1389
+ },
1390
+ {
1391
+ "epoch": 5.78,
1392
+ "learning_rate": 1e-07,
1393
+ "loss": 3.0202,
1394
+ "step": 8778
1395
+ },
1396
+ {
1397
+ "epoch": 5.8,
1398
+ "learning_rate": 1e-07,
1399
+ "loss": 2.9057,
1400
+ "step": 8816
1401
+ },
1402
+ {
1403
+ "epoch": 5.83,
1404
+ "learning_rate": 1e-07,
1405
+ "loss": 2.9556,
1406
+ "step": 8854
1407
+ },
1408
+ {
1409
+ "epoch": 5.85,
1410
+ "learning_rate": 1e-07,
1411
+ "loss": 2.9582,
1412
+ "step": 8892
1413
+ },
1414
+ {
1415
+ "epoch": 5.88,
1416
+ "learning_rate": 1e-07,
1417
+ "loss": 2.8448,
1418
+ "step": 8930
1419
+ },
1420
+ {
1421
+ "epoch": 5.9,
1422
+ "learning_rate": 1e-07,
1423
+ "loss": 3.0643,
1424
+ "step": 8968
1425
+ },
1426
+ {
1427
+ "epoch": 5.92,
1428
+ "learning_rate": 1e-07,
1429
+ "loss": 2.8477,
1430
+ "step": 9006
1431
+ },
1432
+ {
1433
+ "epoch": 5.95,
1434
+ "learning_rate": 1e-07,
1435
+ "loss": 2.9684,
1436
+ "step": 9044
1437
+ },
1438
+ {
1439
+ "epoch": 5.97,
1440
+ "learning_rate": 1e-07,
1441
+ "loss": 2.9086,
1442
+ "step": 9082
1443
+ },
1444
+ {
1445
+ "epoch": 6.0,
1446
+ "learning_rate": 1e-07,
1447
+ "loss": 2.8426,
1448
+ "step": 9120
1449
+ },
1450
+ {
1451
+ "epoch": 6.03,
1452
+ "learning_rate": 1e-07,
1453
+ "loss": 2.903,
1454
+ "step": 9158
1455
+ },
1456
+ {
1457
+ "epoch": 6.05,
1458
+ "learning_rate": 1e-07,
1459
+ "loss": 2.8398,
1460
+ "step": 9196
1461
+ },
1462
+ {
1463
+ "epoch": 6.08,
1464
+ "learning_rate": 1e-07,
1465
+ "loss": 2.8839,
1466
+ "step": 9234
1467
+ },
1468
+ {
1469
+ "epoch": 6.1,
1470
+ "learning_rate": 1e-07,
1471
+ "loss": 2.9396,
1472
+ "step": 9272
1473
+ },
1474
+ {
1475
+ "epoch": 6.12,
1476
+ "learning_rate": 1e-07,
1477
+ "loss": 2.87,
1478
+ "step": 9310
1479
+ },
1480
+ {
1481
+ "epoch": 6.15,
1482
+ "learning_rate": 1e-07,
1483
+ "loss": 2.8619,
1484
+ "step": 9348
1485
+ },
1486
+ {
1487
+ "epoch": 6.17,
1488
+ "learning_rate": 1e-07,
1489
+ "loss": 2.8699,
1490
+ "step": 9386
1491
+ },
1492
+ {
1493
+ "epoch": 6.2,
1494
+ "learning_rate": 1e-07,
1495
+ "loss": 2.9366,
1496
+ "step": 9424
1497
+ },
1498
+ {
1499
+ "epoch": 6.22,
1500
+ "learning_rate": 1e-07,
1501
+ "loss": 2.895,
1502
+ "step": 9462
1503
+ },
1504
+ {
1505
+ "epoch": 6.25,
1506
+ "learning_rate": 1e-07,
1507
+ "loss": 2.8928,
1508
+ "step": 9500
1509
+ },
1510
+ {
1511
+ "epoch": 6.28,
1512
+ "learning_rate": 1e-07,
1513
+ "loss": 2.889,
1514
+ "step": 9538
1515
+ },
1516
+ {
1517
+ "epoch": 6.3,
1518
+ "learning_rate": 1e-07,
1519
+ "loss": 2.9291,
1520
+ "step": 9576
1521
+ },
1522
+ {
1523
+ "epoch": 6.33,
1524
+ "learning_rate": 1e-07,
1525
+ "loss": 2.8722,
1526
+ "step": 9614
1527
+ },
1528
+ {
1529
+ "epoch": 6.35,
1530
+ "learning_rate": 1e-07,
1531
+ "loss": 2.8362,
1532
+ "step": 9652
1533
+ },
1534
+ {
1535
+ "epoch": 6.38,
1536
+ "learning_rate": 1e-07,
1537
+ "loss": 2.8519,
1538
+ "step": 9690
1539
+ },
1540
+ {
1541
+ "epoch": 6.4,
1542
+ "learning_rate": 1e-07,
1543
+ "loss": 2.8364,
1544
+ "step": 9728
1545
+ },
1546
+ {
1547
+ "epoch": 6.42,
1548
+ "learning_rate": 1e-07,
1549
+ "loss": 2.9237,
1550
+ "step": 9766
1551
+ },
1552
+ {
1553
+ "epoch": 6.45,
1554
+ "learning_rate": 1e-07,
1555
+ "loss": 2.876,
1556
+ "step": 9804
1557
+ },
1558
+ {
1559
+ "epoch": 6.47,
1560
+ "learning_rate": 1e-07,
1561
+ "loss": 2.8436,
1562
+ "step": 9842
1563
+ },
1564
+ {
1565
+ "epoch": 6.5,
1566
+ "learning_rate": 1e-07,
1567
+ "loss": 2.8657,
1568
+ "step": 9880
1569
+ },
1570
+ {
1571
+ "epoch": 6.53,
1572
+ "learning_rate": 1e-07,
1573
+ "loss": 2.8283,
1574
+ "step": 9918
1575
+ },
1576
+ {
1577
+ "epoch": 6.55,
1578
+ "learning_rate": 1e-07,
1579
+ "loss": 2.8939,
1580
+ "step": 9956
1581
+ },
1582
+ {
1583
+ "epoch": 6.58,
1584
+ "learning_rate": 1e-07,
1585
+ "loss": 2.8956,
1586
+ "step": 9994
1587
+ },
1588
+ {
1589
+ "epoch": 6.6,
1590
+ "learning_rate": 1e-07,
1591
+ "loss": 2.8361,
1592
+ "step": 10032
1593
+ },
1594
+ {
1595
+ "epoch": 6.62,
1596
+ "learning_rate": 1e-07,
1597
+ "loss": 2.8309,
1598
+ "step": 10070
1599
+ },
1600
+ {
1601
+ "epoch": 6.65,
1602
+ "learning_rate": 1e-07,
1603
+ "loss": 2.8439,
1604
+ "step": 10108
1605
+ },
1606
+ {
1607
+ "epoch": 6.67,
1608
+ "learning_rate": 1e-07,
1609
+ "loss": 2.9308,
1610
+ "step": 10146
1611
+ },
1612
+ {
1613
+ "epoch": 6.7,
1614
+ "learning_rate": 1e-07,
1615
+ "loss": 2.8025,
1616
+ "step": 10184
1617
+ },
1618
+ {
1619
+ "epoch": 6.72,
1620
+ "learning_rate": 1e-07,
1621
+ "loss": 2.8796,
1622
+ "step": 10222
1623
+ },
1624
+ {
1625
+ "epoch": 6.75,
1626
+ "learning_rate": 1e-07,
1627
+ "loss": 2.8225,
1628
+ "step": 10260
1629
+ },
1630
+ {
1631
+ "epoch": 6.78,
1632
+ "learning_rate": 1e-07,
1633
+ "loss": 2.9773,
1634
+ "step": 10298
1635
+ },
1636
+ {
1637
+ "epoch": 6.8,
1638
+ "learning_rate": 1e-07,
1639
+ "loss": 2.8718,
1640
+ "step": 10336
1641
+ },
1642
+ {
1643
+ "epoch": 6.83,
1644
+ "learning_rate": 1e-07,
1645
+ "loss": 2.8718,
1646
+ "step": 10374
1647
+ },
1648
+ {
1649
+ "epoch": 6.85,
1650
+ "learning_rate": 1e-07,
1651
+ "loss": 2.8839,
1652
+ "step": 10412
1653
+ },
1654
+ {
1655
+ "epoch": 6.88,
1656
+ "learning_rate": 1e-07,
1657
+ "loss": 2.8241,
1658
+ "step": 10450
1659
+ },
1660
+ {
1661
+ "epoch": 6.9,
1662
+ "learning_rate": 1e-07,
1663
+ "loss": 2.811,
1664
+ "step": 10488
1665
+ },
1666
+ {
1667
+ "epoch": 6.92,
1668
+ "learning_rate": 1e-07,
1669
+ "loss": 2.8386,
1670
+ "step": 10526
1671
+ },
1672
+ {
1673
+ "epoch": 6.95,
1674
+ "learning_rate": 1e-07,
1675
+ "loss": 2.8042,
1676
+ "step": 10564
1677
+ },
1678
+ {
1679
+ "epoch": 6.97,
1680
+ "learning_rate": 1e-07,
1681
+ "loss": 2.7817,
1682
+ "step": 10602
1683
+ },
1684
+ {
1685
+ "epoch": 7.0,
1686
+ "learning_rate": 1e-07,
1687
+ "loss": 2.776,
1688
+ "step": 10640
1689
+ },
1690
+ {
1691
+ "epoch": 7.03,
1692
+ "learning_rate": 1e-07,
1693
+ "loss": 2.8606,
1694
+ "step": 10678
1695
+ },
1696
+ {
1697
+ "epoch": 7.05,
1698
+ "learning_rate": 1e-07,
1699
+ "loss": 2.8196,
1700
+ "step": 10716
1701
+ },
1702
+ {
1703
+ "epoch": 7.08,
1704
+ "learning_rate": 1e-07,
1705
+ "loss": 2.9032,
1706
+ "step": 10754
1707
+ },
1708
+ {
1709
+ "epoch": 7.1,
1710
+ "learning_rate": 1e-07,
1711
+ "loss": 2.8276,
1712
+ "step": 10792
1713
+ },
1714
+ {
1715
+ "epoch": 7.12,
1716
+ "learning_rate": 1e-07,
1717
+ "loss": 2.7629,
1718
+ "step": 10830
1719
+ },
1720
+ {
1721
+ "epoch": 7.15,
1722
+ "learning_rate": 1e-07,
1723
+ "loss": 2.7876,
1724
+ "step": 10868
1725
+ },
1726
+ {
1727
+ "epoch": 7.17,
1728
+ "learning_rate": 1e-07,
1729
+ "loss": 2.8237,
1730
+ "step": 10906
1731
+ },
1732
+ {
1733
+ "epoch": 7.2,
1734
+ "learning_rate": 1e-07,
1735
+ "loss": 2.8782,
1736
+ "step": 10944
1737
+ },
1738
+ {
1739
+ "epoch": 7.22,
1740
+ "learning_rate": 1e-07,
1741
+ "loss": 2.7644,
1742
+ "step": 10982
1743
+ },
1744
+ {
1745
+ "epoch": 7.25,
1746
+ "learning_rate": 1e-07,
1747
+ "loss": 2.7422,
1748
+ "step": 11020
1749
+ },
1750
+ {
1751
+ "epoch": 7.28,
1752
+ "learning_rate": 1e-07,
1753
+ "loss": 2.8169,
1754
+ "step": 11058
1755
+ },
1756
+ {
1757
+ "epoch": 7.3,
1758
+ "learning_rate": 1e-07,
1759
+ "loss": 2.8212,
1760
+ "step": 11096
1761
+ },
1762
+ {
1763
+ "epoch": 7.33,
1764
+ "learning_rate": 1e-07,
1765
+ "loss": 2.6941,
1766
+ "step": 11134
1767
+ },
1768
+ {
1769
+ "epoch": 7.35,
1770
+ "learning_rate": 1e-07,
1771
+ "loss": 2.8517,
1772
+ "step": 11172
1773
+ },
1774
+ {
1775
+ "epoch": 7.38,
1776
+ "learning_rate": 1e-07,
1777
+ "loss": 2.8385,
1778
+ "step": 11210
1779
+ },
1780
+ {
1781
+ "epoch": 7.4,
1782
+ "learning_rate": 1e-07,
1783
+ "loss": 2.755,
1784
+ "step": 11248
1785
+ },
1786
+ {
1787
+ "epoch": 7.42,
1788
+ "learning_rate": 1e-07,
1789
+ "loss": 2.7692,
1790
+ "step": 11286
1791
+ },
1792
+ {
1793
+ "epoch": 7.45,
1794
+ "learning_rate": 1e-07,
1795
+ "loss": 2.8242,
1796
+ "step": 11324
1797
+ },
1798
+ {
1799
+ "epoch": 7.47,
1800
+ "learning_rate": 1e-07,
1801
+ "loss": 2.7609,
1802
+ "step": 11362
1803
+ },
1804
+ {
1805
+ "epoch": 7.5,
1806
+ "learning_rate": 1e-07,
1807
+ "loss": 2.7633,
1808
+ "step": 11400
1809
+ },
1810
+ {
1811
+ "epoch": 7.53,
1812
+ "learning_rate": 1e-07,
1813
+ "loss": 2.7764,
1814
+ "step": 11438
1815
+ },
1816
+ {
1817
+ "epoch": 7.55,
1818
+ "learning_rate": 1e-07,
1819
+ "loss": 2.7956,
1820
+ "step": 11476
1821
+ },
1822
+ {
1823
+ "epoch": 7.58,
1824
+ "learning_rate": 1e-07,
1825
+ "loss": 2.7179,
1826
+ "step": 11514
1827
+ },
1828
+ {
1829
+ "epoch": 7.6,
1830
+ "learning_rate": 1e-07,
1831
+ "loss": 2.7766,
1832
+ "step": 11552
1833
+ },
1834
+ {
1835
+ "epoch": 7.62,
1836
+ "learning_rate": 1e-07,
1837
+ "loss": 2.8152,
1838
+ "step": 11590
1839
+ },
1840
+ {
1841
+ "epoch": 7.65,
1842
+ "learning_rate": 1e-07,
1843
+ "loss": 2.7367,
1844
+ "step": 11628
1845
+ },
1846
+ {
1847
+ "epoch": 7.67,
1848
+ "learning_rate": 1e-07,
1849
+ "loss": 2.7899,
1850
+ "step": 11666
1851
+ },
1852
+ {
1853
+ "epoch": 7.7,
1854
+ "learning_rate": 1e-07,
1855
+ "loss": 2.8211,
1856
+ "step": 11704
1857
+ },
1858
+ {
1859
+ "epoch": 7.72,
1860
+ "learning_rate": 1e-07,
1861
+ "loss": 2.7512,
1862
+ "step": 11742
1863
+ },
1864
+ {
1865
+ "epoch": 7.75,
1866
+ "learning_rate": 1e-07,
1867
+ "loss": 2.7689,
1868
+ "step": 11780
1869
+ },
1870
+ {
1871
+ "epoch": 7.78,
1872
+ "learning_rate": 1e-07,
1873
+ "loss": 2.7889,
1874
+ "step": 11818
1875
+ },
1876
+ {
1877
+ "epoch": 7.8,
1878
+ "learning_rate": 1e-07,
1879
+ "loss": 2.7053,
1880
+ "step": 11856
1881
+ },
1882
+ {
1883
+ "epoch": 7.83,
1884
+ "learning_rate": 1e-07,
1885
+ "loss": 2.7996,
1886
+ "step": 11894
1887
+ },
1888
+ {
1889
+ "epoch": 7.85,
1890
+ "learning_rate": 1e-07,
1891
+ "loss": 2.6374,
1892
+ "step": 11932
1893
+ },
1894
+ {
1895
+ "epoch": 7.88,
1896
+ "learning_rate": 1e-07,
1897
+ "loss": 2.7144,
1898
+ "step": 11970
1899
+ },
1900
+ {
1901
+ "epoch": 7.9,
1902
+ "learning_rate": 1e-07,
1903
+ "loss": 2.6325,
1904
+ "step": 12008
1905
+ },
1906
+ {
1907
+ "epoch": 7.92,
1908
+ "learning_rate": 1e-07,
1909
+ "loss": 2.7357,
1910
+ "step": 12046
1911
+ },
1912
+ {
1913
+ "epoch": 7.95,
1914
+ "learning_rate": 1e-07,
1915
+ "loss": 2.7822,
1916
+ "step": 12084
1917
+ },
1918
+ {
1919
+ "epoch": 7.97,
1920
+ "learning_rate": 1e-07,
1921
+ "loss": 2.7798,
1922
+ "step": 12122
1923
+ },
1924
+ {
1925
+ "epoch": 8.0,
1926
+ "learning_rate": 1e-07,
1927
+ "loss": 2.7708,
1928
+ "step": 12160
1929
+ },
1930
+ {
1931
+ "epoch": 8.03,
1932
+ "learning_rate": 1e-07,
1933
+ "loss": 2.7695,
1934
+ "step": 12198
1935
+ },
1936
+ {
1937
+ "epoch": 8.05,
1938
+ "learning_rate": 1e-07,
1939
+ "loss": 2.7509,
1940
+ "step": 12236
1941
+ },
1942
+ {
1943
+ "epoch": 8.07,
1944
+ "learning_rate": 1e-07,
1945
+ "loss": 2.6471,
1946
+ "step": 12274
1947
+ },
1948
+ {
1949
+ "epoch": 8.1,
1950
+ "learning_rate": 1e-07,
1951
+ "loss": 2.6833,
1952
+ "step": 12312
1953
+ },
1954
+ {
1955
+ "epoch": 8.12,
1956
+ "learning_rate": 1e-07,
1957
+ "loss": 2.737,
1958
+ "step": 12350
1959
+ },
1960
+ {
1961
+ "epoch": 8.15,
1962
+ "learning_rate": 1e-07,
1963
+ "loss": 2.7074,
1964
+ "step": 12388
1965
+ },
1966
+ {
1967
+ "epoch": 8.18,
1968
+ "learning_rate": 1e-07,
1969
+ "loss": 2.7347,
1970
+ "step": 12426
1971
+ },
1972
+ {
1973
+ "epoch": 8.2,
1974
+ "learning_rate": 1e-07,
1975
+ "loss": 2.6959,
1976
+ "step": 12464
1977
+ },
1978
+ {
1979
+ "epoch": 8.22,
1980
+ "learning_rate": 1e-07,
1981
+ "loss": 2.6782,
1982
+ "step": 12502
1983
+ },
1984
+ {
1985
+ "epoch": 8.25,
1986
+ "learning_rate": 1e-07,
1987
+ "loss": 2.727,
1988
+ "step": 12540
1989
+ },
1990
+ {
1991
+ "epoch": 8.28,
1992
+ "learning_rate": 1e-07,
1993
+ "loss": 2.6834,
1994
+ "step": 12578
1995
+ },
1996
+ {
1997
+ "epoch": 8.3,
1998
+ "learning_rate": 1e-07,
1999
+ "loss": 2.7919,
2000
+ "step": 12616
2001
+ },
2002
+ {
2003
+ "epoch": 8.32,
2004
+ "learning_rate": 1e-07,
2005
+ "loss": 2.7861,
2006
+ "step": 12654
2007
+ },
2008
+ {
2009
+ "epoch": 8.35,
2010
+ "learning_rate": 1e-07,
2011
+ "loss": 2.6642,
2012
+ "step": 12692
2013
+ },
2014
+ {
2015
+ "epoch": 8.38,
2016
+ "learning_rate": 1e-07,
2017
+ "loss": 2.7843,
2018
+ "step": 12730
2019
+ },
2020
+ {
2021
+ "epoch": 8.4,
2022
+ "learning_rate": 1e-07,
2023
+ "loss": 2.7456,
2024
+ "step": 12768
2025
+ },
2026
+ {
2027
+ "epoch": 8.43,
2028
+ "learning_rate": 1e-07,
2029
+ "loss": 2.749,
2030
+ "step": 12806
2031
+ },
2032
+ {
2033
+ "epoch": 8.45,
2034
+ "learning_rate": 1e-07,
2035
+ "loss": 2.6919,
2036
+ "step": 12844
2037
+ },
2038
+ {
2039
+ "epoch": 8.47,
2040
+ "learning_rate": 1e-07,
2041
+ "loss": 2.7122,
2042
+ "step": 12882
2043
+ },
2044
+ {
2045
+ "epoch": 8.5,
2046
+ "learning_rate": 1e-07,
2047
+ "loss": 2.6637,
2048
+ "step": 12920
2049
+ },
2050
+ {
2051
+ "epoch": 8.53,
2052
+ "learning_rate": 1e-07,
2053
+ "loss": 2.7101,
2054
+ "step": 12958
2055
+ },
2056
+ {
2057
+ "epoch": 8.55,
2058
+ "learning_rate": 1e-07,
2059
+ "loss": 2.716,
2060
+ "step": 12996
2061
+ },
2062
+ {
2063
+ "epoch": 8.57,
2064
+ "learning_rate": 1e-07,
2065
+ "loss": 2.698,
2066
+ "step": 13034
2067
+ },
2068
+ {
2069
+ "epoch": 8.6,
2070
+ "learning_rate": 1e-07,
2071
+ "loss": 2.7127,
2072
+ "step": 13072
2073
+ },
2074
+ {
2075
+ "epoch": 8.62,
2076
+ "learning_rate": 1e-07,
2077
+ "loss": 2.7368,
2078
+ "step": 13110
2079
+ },
2080
+ {
2081
+ "epoch": 8.65,
2082
+ "learning_rate": 1e-07,
2083
+ "loss": 2.6313,
2084
+ "step": 13148
2085
+ },
2086
+ {
2087
+ "epoch": 8.68,
2088
+ "learning_rate": 1e-07,
2089
+ "loss": 2.7304,
2090
+ "step": 13186
2091
+ },
2092
+ {
2093
+ "epoch": 8.7,
2094
+ "learning_rate": 1e-07,
2095
+ "loss": 2.7396,
2096
+ "step": 13224
2097
+ },
2098
+ {
2099
+ "epoch": 8.72,
2100
+ "learning_rate": 1e-07,
2101
+ "loss": 2.6746,
2102
+ "step": 13262
2103
+ },
2104
+ {
2105
+ "epoch": 8.75,
2106
+ "learning_rate": 1e-07,
2107
+ "loss": 2.6744,
2108
+ "step": 13300
2109
+ },
2110
+ {
2111
+ "epoch": 8.78,
2112
+ "learning_rate": 1e-07,
2113
+ "loss": 2.6228,
2114
+ "step": 13338
2115
+ },
2116
+ {
2117
+ "epoch": 8.8,
2118
+ "learning_rate": 1e-07,
2119
+ "loss": 2.7504,
2120
+ "step": 13376
2121
+ },
2122
+ {
2123
+ "epoch": 8.82,
2124
+ "learning_rate": 1e-07,
2125
+ "loss": 2.7281,
2126
+ "step": 13414
2127
+ },
2128
+ {
2129
+ "epoch": 8.85,
2130
+ "learning_rate": 1e-07,
2131
+ "loss": 2.7886,
2132
+ "step": 13452
2133
+ },
2134
+ {
2135
+ "epoch": 8.88,
2136
+ "learning_rate": 1e-07,
2137
+ "loss": 2.7505,
2138
+ "step": 13490
2139
+ },
2140
+ {
2141
+ "epoch": 8.9,
2142
+ "learning_rate": 1e-07,
2143
+ "loss": 2.7029,
2144
+ "step": 13528
2145
+ },
2146
+ {
2147
+ "epoch": 8.93,
2148
+ "learning_rate": 1e-07,
2149
+ "loss": 2.7687,
2150
+ "step": 13566
2151
+ },
2152
+ {
2153
+ "epoch": 8.95,
2154
+ "learning_rate": 1e-07,
2155
+ "loss": 2.6783,
2156
+ "step": 13604
2157
+ },
2158
+ {
2159
+ "epoch": 8.97,
2160
+ "learning_rate": 1e-07,
2161
+ "loss": 2.6507,
2162
+ "step": 13642
2163
+ },
2164
+ {
2165
+ "epoch": 9.0,
2166
+ "learning_rate": 1e-07,
2167
+ "loss": 2.673,
2168
+ "step": 13680
2169
+ },
2170
+ {
2171
+ "epoch": 9.03,
2172
+ "learning_rate": 1e-07,
2173
+ "loss": 2.6298,
2174
+ "step": 13718
2175
+ },
2176
+ {
2177
+ "epoch": 9.05,
2178
+ "learning_rate": 1e-07,
2179
+ "loss": 2.6612,
2180
+ "step": 13756
2181
+ },
2182
+ {
2183
+ "epoch": 9.07,
2184
+ "learning_rate": 1e-07,
2185
+ "loss": 2.7233,
2186
+ "step": 13794
2187
+ },
2188
+ {
2189
+ "epoch": 9.1,
2190
+ "learning_rate": 1e-07,
2191
+ "loss": 2.6386,
2192
+ "step": 13832
2193
+ },
2194
+ {
2195
+ "epoch": 9.12,
2196
+ "learning_rate": 1e-07,
2197
+ "loss": 2.6991,
2198
+ "step": 13870
2199
+ },
2200
+ {
2201
+ "epoch": 9.15,
2202
+ "learning_rate": 1e-07,
2203
+ "loss": 2.6629,
2204
+ "step": 13908
2205
+ },
2206
+ {
2207
+ "epoch": 9.18,
2208
+ "learning_rate": 1e-07,
2209
+ "loss": 2.6775,
2210
+ "step": 13946
2211
+ },
2212
+ {
2213
+ "epoch": 9.2,
2214
+ "learning_rate": 1e-07,
2215
+ "loss": 2.6804,
2216
+ "step": 13984
2217
+ },
2218
+ {
2219
+ "epoch": 9.22,
2220
+ "learning_rate": 1e-07,
2221
+ "loss": 2.6196,
2222
+ "step": 14022
2223
+ },
2224
+ {
2225
+ "epoch": 9.25,
2226
+ "learning_rate": 1e-07,
2227
+ "loss": 2.6489,
2228
+ "step": 14060
2229
+ },
2230
+ {
2231
+ "epoch": 9.28,
2232
+ "learning_rate": 1e-07,
2233
+ "loss": 2.7262,
2234
+ "step": 14098
2235
+ },
2236
+ {
2237
+ "epoch": 9.3,
2238
+ "learning_rate": 1e-07,
2239
+ "loss": 2.6845,
2240
+ "step": 14136
2241
+ },
2242
+ {
2243
+ "epoch": 9.32,
2244
+ "learning_rate": 1e-07,
2245
+ "loss": 2.6479,
2246
+ "step": 14174
2247
+ },
2248
+ {
2249
+ "epoch": 9.35,
2250
+ "learning_rate": 1e-07,
2251
+ "loss": 2.7273,
2252
+ "step": 14212
2253
+ },
2254
+ {
2255
+ "epoch": 9.38,
2256
+ "learning_rate": 1e-07,
2257
+ "loss": 2.6825,
2258
+ "step": 14250
2259
+ },
2260
+ {
2261
+ "epoch": 9.4,
2262
+ "learning_rate": 1e-07,
2263
+ "loss": 2.6207,
2264
+ "step": 14288
2265
+ },
2266
+ {
2267
+ "epoch": 9.43,
2268
+ "learning_rate": 1e-07,
2269
+ "loss": 2.6727,
2270
+ "step": 14326
2271
+ },
2272
+ {
2273
+ "epoch": 9.45,
2274
+ "learning_rate": 1e-07,
2275
+ "loss": 2.6411,
2276
+ "step": 14364
2277
+ },
2278
+ {
2279
+ "epoch": 9.47,
2280
+ "learning_rate": 1e-07,
2281
+ "loss": 2.7265,
2282
+ "step": 14402
2283
+ },
2284
+ {
2285
+ "epoch": 9.5,
2286
+ "learning_rate": 1e-07,
2287
+ "loss": 2.7001,
2288
+ "step": 14440
2289
+ },
2290
+ {
2291
+ "epoch": 9.53,
2292
+ "learning_rate": 1e-07,
2293
+ "loss": 2.699,
2294
+ "step": 14478
2295
+ },
2296
+ {
2297
+ "epoch": 9.55,
2298
+ "learning_rate": 1e-07,
2299
+ "loss": 2.612,
2300
+ "step": 14516
2301
+ },
2302
+ {
2303
+ "epoch": 9.57,
2304
+ "learning_rate": 1e-07,
2305
+ "loss": 2.6412,
2306
+ "step": 14554
2307
+ },
2308
+ {
2309
+ "epoch": 9.6,
2310
+ "learning_rate": 1e-07,
2311
+ "loss": 2.634,
2312
+ "step": 14592
2313
+ },
2314
+ {
2315
+ "epoch": 9.62,
2316
+ "learning_rate": 1e-07,
2317
+ "loss": 2.6596,
2318
+ "step": 14630
2319
+ },
2320
+ {
2321
+ "epoch": 9.65,
2322
+ "learning_rate": 1e-07,
2323
+ "loss": 2.702,
2324
+ "step": 14668
2325
+ },
2326
+ {
2327
+ "epoch": 9.68,
2328
+ "learning_rate": 1e-07,
2329
+ "loss": 2.692,
2330
+ "step": 14706
2331
+ },
2332
+ {
2333
+ "epoch": 9.7,
2334
+ "learning_rate": 1e-07,
2335
+ "loss": 2.6904,
2336
+ "step": 14744
2337
+ },
2338
+ {
2339
+ "epoch": 9.72,
2340
+ "learning_rate": 1e-07,
2341
+ "loss": 2.649,
2342
+ "step": 14782
2343
+ },
2344
+ {
2345
+ "epoch": 9.75,
2346
+ "learning_rate": 1e-07,
2347
+ "loss": 2.7208,
2348
+ "step": 14820
2349
+ },
2350
+ {
2351
+ "epoch": 9.78,
2352
+ "learning_rate": 1e-07,
2353
+ "loss": 2.6421,
2354
+ "step": 14858
2355
+ },
2356
+ {
2357
+ "epoch": 9.8,
2358
+ "learning_rate": 1e-07,
2359
+ "loss": 2.6062,
2360
+ "step": 14896
2361
+ },
2362
+ {
2363
+ "epoch": 9.82,
2364
+ "learning_rate": 1e-07,
2365
+ "loss": 2.6326,
2366
+ "step": 14934
2367
+ },
2368
+ {
2369
+ "epoch": 9.85,
2370
+ "learning_rate": 1e-07,
2371
+ "loss": 2.6574,
2372
+ "step": 14972
2373
+ },
2374
+ {
2375
+ "epoch": 9.88,
2376
+ "learning_rate": 1e-07,
2377
+ "loss": 2.6527,
2378
+ "step": 15010
2379
+ },
2380
+ {
2381
+ "epoch": 9.9,
2382
+ "learning_rate": 1e-07,
2383
+ "loss": 2.6796,
2384
+ "step": 15048
2385
+ },
2386
+ {
2387
+ "epoch": 9.93,
2388
+ "learning_rate": 1e-07,
2389
+ "loss": 2.543,
2390
+ "step": 15086
2391
+ },
2392
+ {
2393
+ "epoch": 9.95,
2394
+ "learning_rate": 1e-07,
2395
+ "loss": 2.6001,
2396
+ "step": 15124
2397
+ },
2398
+ {
2399
+ "epoch": 9.97,
2400
+ "learning_rate": 1e-07,
2401
+ "loss": 2.6147,
2402
+ "step": 15162
2403
+ },
2404
+ {
2405
+ "epoch": 10.0,
2406
+ "learning_rate": 1e-07,
2407
+ "loss": 2.6627,
2408
+ "step": 15200
2409
+ },
2410
+ {
2411
+ "epoch": 10.03,
2412
+ "learning_rate": 1e-07,
2413
+ "loss": 2.6809,
2414
+ "step": 15238
2415
+ },
2416
+ {
2417
+ "epoch": 10.05,
2418
+ "learning_rate": 1e-07,
2419
+ "loss": 2.6684,
2420
+ "step": 15276
2421
+ },
2422
+ {
2423
+ "epoch": 10.07,
2424
+ "learning_rate": 1e-07,
2425
+ "loss": 2.6266,
2426
+ "step": 15314
2427
+ },
2428
+ {
2429
+ "epoch": 10.1,
2430
+ "learning_rate": 1e-07,
2431
+ "loss": 2.6882,
2432
+ "step": 15352
2433
+ },
2434
+ {
2435
+ "epoch": 10.12,
2436
+ "learning_rate": 1e-07,
2437
+ "loss": 2.6337,
2438
+ "step": 15390
2439
+ },
2440
+ {
2441
+ "epoch": 10.15,
2442
+ "learning_rate": 1e-07,
2443
+ "loss": 2.6511,
2444
+ "step": 15428
2445
+ },
2446
+ {
2447
+ "epoch": 10.18,
2448
+ "learning_rate": 1e-07,
2449
+ "loss": 2.5565,
2450
+ "step": 15466
2451
+ },
2452
+ {
2453
+ "epoch": 10.2,
2454
+ "learning_rate": 1e-07,
2455
+ "loss": 2.6532,
2456
+ "step": 15504
2457
+ },
2458
+ {
2459
+ "epoch": 10.22,
2460
+ "learning_rate": 1e-07,
2461
+ "loss": 2.6808,
2462
+ "step": 15542
2463
+ },
2464
+ {
2465
+ "epoch": 10.25,
2466
+ "learning_rate": 1e-07,
2467
+ "loss": 2.6367,
2468
+ "step": 15580
2469
+ },
2470
+ {
2471
+ "epoch": 10.28,
2472
+ "learning_rate": 1e-07,
2473
+ "loss": 2.6816,
2474
+ "step": 15618
2475
+ },
2476
+ {
2477
+ "epoch": 10.3,
2478
+ "learning_rate": 1e-07,
2479
+ "loss": 2.5894,
2480
+ "step": 15656
2481
+ },
2482
+ {
2483
+ "epoch": 10.32,
2484
+ "learning_rate": 1e-07,
2485
+ "loss": 2.6045,
2486
+ "step": 15694
2487
+ },
2488
+ {
2489
+ "epoch": 10.35,
2490
+ "learning_rate": 1e-07,
2491
+ "loss": 2.6664,
2492
+ "step": 15732
2493
+ },
2494
+ {
2495
+ "epoch": 10.38,
2496
+ "learning_rate": 1e-07,
2497
+ "loss": 2.6359,
2498
+ "step": 15770
2499
+ },
2500
+ {
2501
+ "epoch": 10.4,
2502
+ "learning_rate": 1e-07,
2503
+ "loss": 2.595,
2504
+ "step": 15808
2505
+ },
2506
+ {
2507
+ "epoch": 10.43,
2508
+ "learning_rate": 1e-07,
2509
+ "loss": 2.7218,
2510
+ "step": 15846
2511
+ },
2512
+ {
2513
+ "epoch": 10.45,
2514
+ "learning_rate": 1e-07,
2515
+ "loss": 2.6015,
2516
+ "step": 15884
2517
+ },
2518
+ {
2519
+ "epoch": 10.47,
2520
+ "learning_rate": 1e-07,
2521
+ "loss": 2.6469,
2522
+ "step": 15922
2523
+ },
2524
+ {
2525
+ "epoch": 10.5,
2526
+ "learning_rate": 1e-07,
2527
+ "loss": 2.6632,
2528
+ "step": 15960
2529
+ },
2530
+ {
2531
+ "epoch": 10.53,
2532
+ "learning_rate": 1e-07,
2533
+ "loss": 2.6137,
2534
+ "step": 15998
2535
+ },
2536
+ {
2537
+ "epoch": 10.55,
2538
+ "learning_rate": 1e-07,
2539
+ "loss": 2.5723,
2540
+ "step": 16036
2541
+ },
2542
+ {
2543
+ "epoch": 10.57,
2544
+ "learning_rate": 1e-07,
2545
+ "loss": 2.5788,
2546
+ "step": 16074
2547
+ },
2548
+ {
2549
+ "epoch": 10.6,
2550
+ "learning_rate": 1e-07,
2551
+ "loss": 2.6213,
2552
+ "step": 16112
2553
+ },
2554
+ {
2555
+ "epoch": 10.62,
2556
+ "learning_rate": 1e-07,
2557
+ "loss": 2.6261,
2558
+ "step": 16150
2559
+ },
2560
+ {
2561
+ "epoch": 10.65,
2562
+ "learning_rate": 1e-07,
2563
+ "loss": 2.5937,
2564
+ "step": 16188
2565
+ },
2566
+ {
2567
+ "epoch": 10.68,
2568
+ "learning_rate": 1e-07,
2569
+ "loss": 2.5266,
2570
+ "step": 16226
2571
+ },
2572
+ {
2573
+ "epoch": 10.7,
2574
+ "learning_rate": 1e-07,
2575
+ "loss": 2.6844,
2576
+ "step": 16264
2577
+ },
2578
+ {
2579
+ "epoch": 10.72,
2580
+ "learning_rate": 1e-07,
2581
+ "loss": 2.5672,
2582
+ "step": 16302
2583
+ },
2584
+ {
2585
+ "epoch": 10.75,
2586
+ "learning_rate": 1e-07,
2587
+ "loss": 2.5905,
2588
+ "step": 16340
2589
+ },
2590
+ {
2591
+ "epoch": 10.78,
2592
+ "learning_rate": 1e-07,
2593
+ "loss": 2.5908,
2594
+ "step": 16378
2595
+ },
2596
+ {
2597
+ "epoch": 10.8,
2598
+ "learning_rate": 1e-07,
2599
+ "loss": 2.7097,
2600
+ "step": 16416
2601
+ },
2602
+ {
2603
+ "epoch": 10.82,
2604
+ "learning_rate": 1e-07,
2605
+ "loss": 2.6675,
2606
+ "step": 16454
2607
+ },
2608
+ {
2609
+ "epoch": 10.85,
2610
+ "learning_rate": 1e-07,
2611
+ "loss": 2.5913,
2612
+ "step": 16492
2613
+ },
2614
+ {
2615
+ "epoch": 10.88,
2616
+ "learning_rate": 1e-07,
2617
+ "loss": 2.6108,
2618
+ "step": 16530
2619
+ },
2620
+ {
2621
+ "epoch": 10.9,
2622
+ "learning_rate": 1e-07,
2623
+ "loss": 2.5413,
2624
+ "step": 16568
2625
+ },
2626
+ {
2627
+ "epoch": 10.93,
2628
+ "learning_rate": 1e-07,
2629
+ "loss": 2.6089,
2630
+ "step": 16606
2631
+ },
2632
+ {
2633
+ "epoch": 10.95,
2634
+ "learning_rate": 1e-07,
2635
+ "loss": 2.5889,
2636
+ "step": 16644
2637
+ },
2638
+ {
2639
+ "epoch": 10.97,
2640
+ "learning_rate": 1e-07,
2641
+ "loss": 2.5092,
2642
+ "step": 16682
2643
+ },
2644
+ {
2645
+ "epoch": 11.0,
2646
+ "learning_rate": 1e-07,
2647
+ "loss": 2.6164,
2648
+ "step": 16720
2649
+ },
2650
+ {
2651
+ "epoch": 11.03,
2652
+ "learning_rate": 1e-07,
2653
+ "loss": 2.6324,
2654
+ "step": 16758
2655
+ },
2656
+ {
2657
+ "epoch": 11.05,
2658
+ "learning_rate": 1e-07,
2659
+ "loss": 2.6398,
2660
+ "step": 16796
2661
+ },
2662
+ {
2663
+ "epoch": 11.07,
2664
+ "learning_rate": 1e-07,
2665
+ "loss": 2.6602,
2666
+ "step": 16834
2667
+ },
2668
+ {
2669
+ "epoch": 11.1,
2670
+ "learning_rate": 1e-07,
2671
+ "loss": 2.6362,
2672
+ "step": 16872
2673
+ },
2674
+ {
2675
+ "epoch": 11.12,
2676
+ "learning_rate": 1e-07,
2677
+ "loss": 2.574,
2678
+ "step": 16910
2679
+ },
2680
+ {
2681
+ "epoch": 11.15,
2682
+ "learning_rate": 1e-07,
2683
+ "loss": 2.5672,
2684
+ "step": 16948
2685
+ },
2686
+ {
2687
+ "epoch": 11.18,
2688
+ "learning_rate": 1e-07,
2689
+ "loss": 2.5726,
2690
+ "step": 16986
2691
+ },
2692
+ {
2693
+ "epoch": 11.2,
2694
+ "learning_rate": 1e-07,
2695
+ "loss": 2.6055,
2696
+ "step": 17024
2697
+ },
2698
+ {
2699
+ "epoch": 11.22,
2700
+ "learning_rate": 1e-07,
2701
+ "loss": 2.5427,
2702
+ "step": 17062
2703
+ },
2704
+ {
2705
+ "epoch": 11.25,
2706
+ "learning_rate": 1e-07,
2707
+ "loss": 2.5632,
2708
+ "step": 17100
2709
+ },
2710
+ {
2711
+ "epoch": 11.28,
2712
+ "learning_rate": 1e-07,
2713
+ "loss": 2.6167,
2714
+ "step": 17138
2715
+ },
2716
+ {
2717
+ "epoch": 11.3,
2718
+ "learning_rate": 1e-07,
2719
+ "loss": 2.584,
2720
+ "step": 17176
2721
+ },
2722
+ {
2723
+ "epoch": 11.32,
2724
+ "learning_rate": 1e-07,
2725
+ "loss": 2.6179,
2726
+ "step": 17214
2727
+ },
2728
+ {
2729
+ "epoch": 11.35,
2730
+ "learning_rate": 1e-07,
2731
+ "loss": 2.6232,
2732
+ "step": 17252
2733
+ },
2734
+ {
2735
+ "epoch": 11.38,
2736
+ "learning_rate": 1e-07,
2737
+ "loss": 2.6158,
2738
+ "step": 17290
2739
+ },
2740
+ {
2741
+ "epoch": 11.4,
2742
+ "learning_rate": 1e-07,
2743
+ "loss": 2.6115,
2744
+ "step": 17328
2745
+ },
2746
+ {
2747
+ "epoch": 11.43,
2748
+ "learning_rate": 1e-07,
2749
+ "loss": 2.5079,
2750
+ "step": 17366
2751
+ },
2752
+ {
2753
+ "epoch": 11.45,
2754
+ "learning_rate": 1e-07,
2755
+ "loss": 2.6525,
2756
+ "step": 17404
2757
+ },
2758
+ {
2759
+ "epoch": 11.47,
2760
+ "learning_rate": 1e-07,
2761
+ "loss": 2.5643,
2762
+ "step": 17442
2763
+ },
2764
+ {
2765
+ "epoch": 11.5,
2766
+ "learning_rate": 1e-07,
2767
+ "loss": 2.6215,
2768
+ "step": 17480
2769
+ },
2770
+ {
2771
+ "epoch": 11.53,
2772
+ "learning_rate": 1e-07,
2773
+ "loss": 2.5774,
2774
+ "step": 17518
2775
+ },
2776
+ {
2777
+ "epoch": 11.55,
2778
+ "learning_rate": 1e-07,
2779
+ "loss": 2.5696,
2780
+ "step": 17556
2781
+ },
2782
+ {
2783
+ "epoch": 11.57,
2784
+ "learning_rate": 1e-07,
2785
+ "loss": 2.5622,
2786
+ "step": 17594
2787
+ },
2788
+ {
2789
+ "epoch": 11.6,
2790
+ "learning_rate": 1e-07,
2791
+ "loss": 2.5572,
2792
+ "step": 17632
2793
+ },
2794
+ {
2795
+ "epoch": 11.62,
2796
+ "learning_rate": 1e-07,
2797
+ "loss": 2.5853,
2798
+ "step": 17670
2799
+ },
2800
+ {
2801
+ "epoch": 11.65,
2802
+ "learning_rate": 1e-07,
2803
+ "loss": 2.6487,
2804
+ "step": 17708
2805
+ },
2806
+ {
2807
+ "epoch": 11.68,
2808
+ "learning_rate": 1e-07,
2809
+ "loss": 2.5708,
2810
+ "step": 17746
2811
+ },
2812
+ {
2813
+ "epoch": 11.7,
2814
+ "learning_rate": 1e-07,
2815
+ "loss": 2.5757,
2816
+ "step": 17784
2817
+ },
2818
+ {
2819
+ "epoch": 11.72,
2820
+ "learning_rate": 1e-07,
2821
+ "loss": 2.5935,
2822
+ "step": 17822
2823
+ },
2824
+ {
2825
+ "epoch": 11.75,
2826
+ "learning_rate": 1e-07,
2827
+ "loss": 2.5212,
2828
+ "step": 17860
2829
+ },
2830
+ {
2831
+ "epoch": 11.78,
2832
+ "learning_rate": 1e-07,
2833
+ "loss": 2.6067,
2834
+ "step": 17898
2835
+ },
2836
+ {
2837
+ "epoch": 11.8,
2838
+ "learning_rate": 1e-07,
2839
+ "loss": 2.5712,
2840
+ "step": 17936
2841
+ },
2842
+ {
2843
+ "epoch": 11.82,
2844
+ "learning_rate": 1e-07,
2845
+ "loss": 2.5646,
2846
+ "step": 17974
2847
+ },
2848
+ {
2849
+ "epoch": 11.85,
2850
+ "learning_rate": 1e-07,
2851
+ "loss": 2.5789,
2852
+ "step": 18012
2853
+ },
2854
+ {
2855
+ "epoch": 11.88,
2856
+ "learning_rate": 1e-07,
2857
+ "loss": 2.6135,
2858
+ "step": 18050
2859
+ },
2860
+ {
2861
+ "epoch": 11.9,
2862
+ "learning_rate": 1e-07,
2863
+ "loss": 2.5586,
2864
+ "step": 18088
2865
+ },
2866
+ {
2867
+ "epoch": 11.93,
2868
+ "learning_rate": 1e-07,
2869
+ "loss": 2.5409,
2870
+ "step": 18126
2871
+ },
2872
+ {
2873
+ "epoch": 11.95,
2874
+ "learning_rate": 1e-07,
2875
+ "loss": 2.5607,
2876
+ "step": 18164
2877
+ },
2878
+ {
2879
+ "epoch": 11.97,
2880
+ "learning_rate": 1e-07,
2881
+ "loss": 2.5547,
2882
+ "step": 18202
2883
+ },
2884
+ {
2885
+ "epoch": 12.0,
2886
+ "learning_rate": 1e-07,
2887
+ "loss": 2.5457,
2888
+ "step": 18240
2889
+ },
2890
+ {
2891
+ "epoch": 12.03,
2892
+ "learning_rate": 1e-07,
2893
+ "loss": 2.5895,
2894
+ "step": 18278
2895
+ },
2896
+ {
2897
+ "epoch": 12.05,
2898
+ "learning_rate": 1e-07,
2899
+ "loss": 2.61,
2900
+ "step": 18316
2901
+ },
2902
+ {
2903
+ "epoch": 12.07,
2904
+ "learning_rate": 1e-07,
2905
+ "loss": 2.6503,
2906
+ "step": 18354
2907
+ },
2908
+ {
2909
+ "epoch": 12.1,
2910
+ "learning_rate": 1e-07,
2911
+ "loss": 2.5908,
2912
+ "step": 18392
2913
+ },
2914
+ {
2915
+ "epoch": 12.12,
2916
+ "learning_rate": 1e-07,
2917
+ "loss": 2.5728,
2918
+ "step": 18430
2919
+ },
2920
+ {
2921
+ "epoch": 12.15,
2922
+ "learning_rate": 1e-07,
2923
+ "loss": 2.5264,
2924
+ "step": 18468
2925
+ },
2926
+ {
2927
+ "epoch": 12.18,
2928
+ "learning_rate": 1e-07,
2929
+ "loss": 2.52,
2930
+ "step": 18506
2931
+ },
2932
+ {
2933
+ "epoch": 12.2,
2934
+ "learning_rate": 1e-07,
2935
+ "loss": 2.6832,
2936
+ "step": 18544
2937
+ },
2938
+ {
2939
+ "epoch": 12.22,
2940
+ "learning_rate": 1e-07,
2941
+ "loss": 2.6024,
2942
+ "step": 18582
2943
+ },
2944
+ {
2945
+ "epoch": 12.25,
2946
+ "learning_rate": 1e-07,
2947
+ "loss": 2.4557,
2948
+ "step": 18620
2949
+ },
2950
+ {
2951
+ "epoch": 12.28,
2952
+ "learning_rate": 1e-07,
2953
+ "loss": 2.6531,
2954
+ "step": 18658
2955
+ },
2956
+ {
2957
+ "epoch": 12.3,
2958
+ "learning_rate": 1e-07,
2959
+ "loss": 2.5052,
2960
+ "step": 18696
2961
+ },
2962
+ {
2963
+ "epoch": 12.32,
2964
+ "learning_rate": 1e-07,
2965
+ "loss": 2.4904,
2966
+ "step": 18734
2967
+ },
2968
+ {
2969
+ "epoch": 12.35,
2970
+ "learning_rate": 1e-07,
2971
+ "loss": 2.5336,
2972
+ "step": 18772
2973
+ },
2974
+ {
2975
+ "epoch": 12.38,
2976
+ "learning_rate": 1e-07,
2977
+ "loss": 2.5244,
2978
+ "step": 18810
2979
+ },
2980
+ {
2981
+ "epoch": 12.4,
2982
+ "learning_rate": 1e-07,
2983
+ "loss": 2.5641,
2984
+ "step": 18848
2985
+ },
2986
+ {
2987
+ "epoch": 12.43,
2988
+ "learning_rate": 1e-07,
2989
+ "loss": 2.5388,
2990
+ "step": 18886
2991
+ },
2992
+ {
2993
+ "epoch": 12.45,
2994
+ "learning_rate": 1e-07,
2995
+ "loss": 2.5346,
2996
+ "step": 18924
2997
+ },
2998
+ {
2999
+ "epoch": 12.47,
3000
+ "learning_rate": 1e-07,
3001
+ "loss": 2.5518,
3002
+ "step": 18962
3003
+ },
3004
+ {
3005
+ "epoch": 12.5,
3006
+ "learning_rate": 1e-07,
3007
+ "loss": 2.5202,
3008
+ "step": 19000
3009
+ },
3010
+ {
3011
+ "epoch": 12.53,
3012
+ "learning_rate": 1e-07,
3013
+ "loss": 2.5293,
3014
+ "step": 19038
3015
+ },
3016
+ {
3017
+ "epoch": 12.55,
3018
+ "learning_rate": 1e-07,
3019
+ "loss": 2.567,
3020
+ "step": 19076
3021
+ },
3022
+ {
3023
+ "epoch": 12.57,
3024
+ "learning_rate": 1e-07,
3025
+ "loss": 2.5357,
3026
+ "step": 19114
3027
+ },
3028
+ {
3029
+ "epoch": 12.6,
3030
+ "learning_rate": 1e-07,
3031
+ "loss": 2.5785,
3032
+ "step": 19152
3033
+ },
3034
+ {
3035
+ "epoch": 12.62,
3036
+ "learning_rate": 1e-07,
3037
+ "loss": 2.6327,
3038
+ "step": 19190
3039
+ },
3040
+ {
3041
+ "epoch": 12.65,
3042
+ "learning_rate": 1e-07,
3043
+ "loss": 2.6912,
3044
+ "step": 19228
3045
+ },
3046
+ {
3047
+ "epoch": 12.68,
3048
+ "learning_rate": 1e-07,
3049
+ "loss": 2.4824,
3050
+ "step": 19266
3051
+ },
3052
+ {
3053
+ "epoch": 12.7,
3054
+ "learning_rate": 1e-07,
3055
+ "loss": 2.5996,
3056
+ "step": 19304
3057
+ },
3058
+ {
3059
+ "epoch": 12.72,
3060
+ "learning_rate": 1e-07,
3061
+ "loss": 2.5246,
3062
+ "step": 19342
3063
+ },
3064
+ {
3065
+ "epoch": 12.75,
3066
+ "learning_rate": 1e-07,
3067
+ "loss": 2.5583,
3068
+ "step": 19380
3069
+ },
3070
+ {
3071
+ "epoch": 12.78,
3072
+ "learning_rate": 1e-07,
3073
+ "loss": 2.515,
3074
+ "step": 19418
3075
+ },
3076
+ {
3077
+ "epoch": 12.8,
3078
+ "learning_rate": 1e-07,
3079
+ "loss": 2.5677,
3080
+ "step": 19456
3081
+ },
3082
+ {
3083
+ "epoch": 12.82,
3084
+ "learning_rate": 1e-07,
3085
+ "loss": 2.5488,
3086
+ "step": 19494
3087
+ },
3088
+ {
3089
+ "epoch": 12.85,
3090
+ "learning_rate": 1e-07,
3091
+ "loss": 2.5562,
3092
+ "step": 19532
3093
+ },
3094
+ {
3095
+ "epoch": 12.88,
3096
+ "learning_rate": 1e-07,
3097
+ "loss": 2.544,
3098
+ "step": 19570
3099
+ },
3100
+ {
3101
+ "epoch": 12.9,
3102
+ "learning_rate": 1e-07,
3103
+ "loss": 2.5297,
3104
+ "step": 19608
3105
+ },
3106
+ {
3107
+ "epoch": 12.93,
3108
+ "learning_rate": 1e-07,
3109
+ "loss": 2.5091,
3110
+ "step": 19646
3111
+ },
3112
+ {
3113
+ "epoch": 12.95,
3114
+ "learning_rate": 1e-07,
3115
+ "loss": 2.5492,
3116
+ "step": 19684
3117
+ },
3118
+ {
3119
+ "epoch": 12.97,
3120
+ "learning_rate": 1e-07,
3121
+ "loss": 2.4553,
3122
+ "step": 19722
3123
+ },
3124
+ {
3125
+ "epoch": 13.0,
3126
+ "learning_rate": 1e-07,
3127
+ "loss": 2.5344,
3128
+ "step": 19760
3129
+ },
3130
+ {
3131
+ "epoch": 13.03,
3132
+ "learning_rate": 1e-07,
3133
+ "loss": 2.4807,
3134
+ "step": 19798
3135
+ },
3136
+ {
3137
+ "epoch": 13.05,
3138
+ "learning_rate": 1e-07,
3139
+ "loss": 2.4998,
3140
+ "step": 19836
3141
+ },
3142
+ {
3143
+ "epoch": 13.07,
3144
+ "learning_rate": 1e-07,
3145
+ "loss": 2.5033,
3146
+ "step": 19874
3147
+ },
3148
+ {
3149
+ "epoch": 13.1,
3150
+ "learning_rate": 1e-07,
3151
+ "loss": 2.5731,
3152
+ "step": 19912
3153
+ },
3154
+ {
3155
+ "epoch": 13.12,
3156
+ "learning_rate": 1e-07,
3157
+ "loss": 2.5362,
3158
+ "step": 19950
3159
+ },
3160
+ {
3161
+ "epoch": 13.15,
3162
+ "learning_rate": 1e-07,
3163
+ "loss": 2.5897,
3164
+ "step": 19988
3165
+ },
3166
+ {
3167
+ "epoch": 13.18,
3168
+ "learning_rate": 1e-07,
3169
+ "loss": 2.5419,
3170
+ "step": 20026
3171
+ },
3172
+ {
3173
+ "epoch": 13.2,
3174
+ "learning_rate": 1e-07,
3175
+ "loss": 2.5616,
3176
+ "step": 20064
3177
+ },
3178
+ {
3179
+ "epoch": 13.22,
3180
+ "learning_rate": 1e-07,
3181
+ "loss": 2.4845,
3182
+ "step": 20102
3183
+ },
3184
+ {
3185
+ "epoch": 13.25,
3186
+ "learning_rate": 1e-07,
3187
+ "loss": 2.5642,
3188
+ "step": 20140
3189
+ },
3190
+ {
3191
+ "epoch": 13.28,
3192
+ "learning_rate": 1e-07,
3193
+ "loss": 2.4496,
3194
+ "step": 20178
3195
+ },
3196
+ {
3197
+ "epoch": 13.3,
3198
+ "learning_rate": 1e-07,
3199
+ "loss": 2.5136,
3200
+ "step": 20216
3201
+ },
3202
+ {
3203
+ "epoch": 13.32,
3204
+ "learning_rate": 1e-07,
3205
+ "loss": 2.5947,
3206
+ "step": 20254
3207
+ },
3208
+ {
3209
+ "epoch": 13.35,
3210
+ "learning_rate": 1e-07,
3211
+ "loss": 2.5272,
3212
+ "step": 20292
3213
+ },
3214
+ {
3215
+ "epoch": 13.38,
3216
+ "learning_rate": 1e-07,
3217
+ "loss": 2.5801,
3218
+ "step": 20330
3219
+ },
3220
+ {
3221
+ "epoch": 13.4,
3222
+ "learning_rate": 1e-07,
3223
+ "loss": 2.4692,
3224
+ "step": 20368
3225
+ },
3226
+ {
3227
+ "epoch": 13.43,
3228
+ "learning_rate": 1e-07,
3229
+ "loss": 2.4849,
3230
+ "step": 20406
3231
+ },
3232
+ {
3233
+ "epoch": 13.45,
3234
+ "learning_rate": 1e-07,
3235
+ "loss": 2.5942,
3236
+ "step": 20444
3237
+ },
3238
+ {
3239
+ "epoch": 13.47,
3240
+ "learning_rate": 1e-07,
3241
+ "loss": 2.5341,
3242
+ "step": 20482
3243
+ },
3244
+ {
3245
+ "epoch": 13.5,
3246
+ "learning_rate": 1e-07,
3247
+ "loss": 2.4762,
3248
+ "step": 20520
3249
+ },
3250
+ {
3251
+ "epoch": 13.53,
3252
+ "learning_rate": 1e-07,
3253
+ "loss": 2.5744,
3254
+ "step": 20558
3255
+ },
3256
+ {
3257
+ "epoch": 13.55,
3258
+ "learning_rate": 1e-07,
3259
+ "loss": 2.5383,
3260
+ "step": 20596
3261
+ },
3262
+ {
3263
+ "epoch": 13.57,
3264
+ "learning_rate": 1e-07,
3265
+ "loss": 2.5769,
3266
+ "step": 20634
3267
+ },
3268
+ {
3269
+ "epoch": 13.6,
3270
+ "learning_rate": 1e-07,
3271
+ "loss": 2.5196,
3272
+ "step": 20672
3273
+ },
3274
+ {
3275
+ "epoch": 13.62,
3276
+ "learning_rate": 1e-07,
3277
+ "loss": 2.6015,
3278
+ "step": 20710
3279
+ },
3280
+ {
3281
+ "epoch": 13.65,
3282
+ "learning_rate": 1e-07,
3283
+ "loss": 2.5733,
3284
+ "step": 20748
3285
+ },
3286
+ {
3287
+ "epoch": 13.68,
3288
+ "learning_rate": 1e-07,
3289
+ "loss": 2.4808,
3290
+ "step": 20786
3291
+ },
3292
+ {
3293
+ "epoch": 13.7,
3294
+ "learning_rate": 1e-07,
3295
+ "loss": 2.4772,
3296
+ "step": 20824
3297
+ },
3298
+ {
3299
+ "epoch": 13.72,
3300
+ "learning_rate": 1e-07,
3301
+ "loss": 2.5212,
3302
+ "step": 20862
3303
+ },
3304
+ {
3305
+ "epoch": 13.75,
3306
+ "learning_rate": 1e-07,
3307
+ "loss": 2.5066,
3308
+ "step": 20900
3309
+ },
3310
+ {
3311
+ "epoch": 13.78,
3312
+ "learning_rate": 1e-07,
3313
+ "loss": 2.542,
3314
+ "step": 20938
3315
+ },
3316
+ {
3317
+ "epoch": 13.8,
3318
+ "learning_rate": 1e-07,
3319
+ "loss": 2.5422,
3320
+ "step": 20976
3321
+ },
3322
+ {
3323
+ "epoch": 13.82,
3324
+ "learning_rate": 1e-07,
3325
+ "loss": 2.5284,
3326
+ "step": 21014
3327
+ },
3328
+ {
3329
+ "epoch": 13.85,
3330
+ "learning_rate": 1e-07,
3331
+ "loss": 2.5472,
3332
+ "step": 21052
3333
+ },
3334
+ {
3335
+ "epoch": 13.88,
3336
+ "learning_rate": 1e-07,
3337
+ "loss": 2.6212,
3338
+ "step": 21090
3339
+ },
3340
+ {
3341
+ "epoch": 13.9,
3342
+ "learning_rate": 1e-07,
3343
+ "loss": 2.4717,
3344
+ "step": 21128
3345
+ },
3346
+ {
3347
+ "epoch": 13.93,
3348
+ "learning_rate": 1e-07,
3349
+ "loss": 2.5001,
3350
+ "step": 21166
3351
+ },
3352
+ {
3353
+ "epoch": 13.95,
3354
+ "learning_rate": 1e-07,
3355
+ "loss": 2.4892,
3356
+ "step": 21204
3357
+ },
3358
+ {
3359
+ "epoch": 13.97,
3360
+ "learning_rate": 1e-07,
3361
+ "loss": 2.4833,
3362
+ "step": 21242
3363
+ },
3364
+ {
3365
+ "epoch": 14.0,
3366
+ "learning_rate": 1e-07,
3367
+ "loss": 2.5363,
3368
+ "step": 21280
3369
+ },
3370
+ {
3371
+ "epoch": 14.03,
3372
+ "learning_rate": 1e-07,
3373
+ "loss": 2.5563,
3374
+ "step": 21318
3375
+ },
3376
+ {
3377
+ "epoch": 14.05,
3378
+ "learning_rate": 1e-07,
3379
+ "loss": 2.4392,
3380
+ "step": 21356
3381
+ },
3382
+ {
3383
+ "epoch": 14.07,
3384
+ "learning_rate": 1e-07,
3385
+ "loss": 2.4836,
3386
+ "step": 21394
3387
+ },
3388
+ {
3389
+ "epoch": 14.1,
3390
+ "learning_rate": 1e-07,
3391
+ "loss": 2.5705,
3392
+ "step": 21432
3393
+ },
3394
+ {
3395
+ "epoch": 14.12,
3396
+ "learning_rate": 1e-07,
3397
+ "loss": 2.494,
3398
+ "step": 21470
3399
+ },
3400
+ {
3401
+ "epoch": 14.15,
3402
+ "learning_rate": 1e-07,
3403
+ "loss": 2.5226,
3404
+ "step": 21508
3405
+ },
3406
+ {
3407
+ "epoch": 14.18,
3408
+ "learning_rate": 1e-07,
3409
+ "loss": 2.5026,
3410
+ "step": 21546
3411
+ },
3412
+ {
3413
+ "epoch": 14.2,
3414
+ "learning_rate": 1e-07,
3415
+ "loss": 2.4934,
3416
+ "step": 21584
3417
+ },
3418
+ {
3419
+ "epoch": 14.22,
3420
+ "learning_rate": 1e-07,
3421
+ "loss": 2.5057,
3422
+ "step": 21622
3423
+ },
3424
+ {
3425
+ "epoch": 14.25,
3426
+ "learning_rate": 1e-07,
3427
+ "loss": 2.4886,
3428
+ "step": 21660
3429
+ },
3430
+ {
3431
+ "epoch": 14.28,
3432
+ "learning_rate": 1e-07,
3433
+ "loss": 2.4816,
3434
+ "step": 21698
3435
+ },
3436
+ {
3437
+ "epoch": 14.3,
3438
+ "learning_rate": 1e-07,
3439
+ "loss": 2.4714,
3440
+ "step": 21736
3441
+ },
3442
+ {
3443
+ "epoch": 14.32,
3444
+ "learning_rate": 1e-07,
3445
+ "loss": 2.5459,
3446
+ "step": 21774
3447
+ },
3448
+ {
3449
+ "epoch": 14.35,
3450
+ "learning_rate": 1e-07,
3451
+ "loss": 2.549,
3452
+ "step": 21812
3453
+ },
3454
+ {
3455
+ "epoch": 14.38,
3456
+ "learning_rate": 1e-07,
3457
+ "loss": 2.4623,
3458
+ "step": 21850
3459
+ },
3460
+ {
3461
+ "epoch": 14.4,
3462
+ "learning_rate": 1e-07,
3463
+ "loss": 2.48,
3464
+ "step": 21888
3465
+ },
3466
+ {
3467
+ "epoch": 14.43,
3468
+ "learning_rate": 1e-07,
3469
+ "loss": 2.5529,
3470
+ "step": 21926
3471
+ },
3472
+ {
3473
+ "epoch": 14.45,
3474
+ "learning_rate": 1e-07,
3475
+ "loss": 2.5045,
3476
+ "step": 21964
3477
+ },
3478
+ {
3479
+ "epoch": 14.47,
3480
+ "learning_rate": 1e-07,
3481
+ "loss": 2.5599,
3482
+ "step": 22002
3483
+ },
3484
+ {
3485
+ "epoch": 14.5,
3486
+ "learning_rate": 1e-07,
3487
+ "loss": 2.4862,
3488
+ "step": 22040
3489
+ },
3490
+ {
3491
+ "epoch": 14.53,
3492
+ "learning_rate": 1e-07,
3493
+ "loss": 2.5145,
3494
+ "step": 22078
3495
+ },
3496
+ {
3497
+ "epoch": 14.55,
3498
+ "learning_rate": 1e-07,
3499
+ "loss": 2.4944,
3500
+ "step": 22116
3501
+ },
3502
+ {
3503
+ "epoch": 14.57,
3504
+ "learning_rate": 1e-07,
3505
+ "loss": 2.4581,
3506
+ "step": 22154
3507
+ },
3508
+ {
3509
+ "epoch": 14.6,
3510
+ "learning_rate": 1e-07,
3511
+ "loss": 2.552,
3512
+ "step": 22192
3513
+ },
3514
+ {
3515
+ "epoch": 14.62,
3516
+ "learning_rate": 1e-07,
3517
+ "loss": 2.5751,
3518
+ "step": 22230
3519
+ },
3520
+ {
3521
+ "epoch": 14.65,
3522
+ "learning_rate": 1e-07,
3523
+ "loss": 2.5383,
3524
+ "step": 22268
3525
+ },
3526
+ {
3527
+ "epoch": 14.68,
3528
+ "learning_rate": 1e-07,
3529
+ "loss": 2.5273,
3530
+ "step": 22306
3531
+ },
3532
+ {
3533
+ "epoch": 14.7,
3534
+ "learning_rate": 1e-07,
3535
+ "loss": 2.5118,
3536
+ "step": 22344
3537
+ },
3538
+ {
3539
+ "epoch": 14.72,
3540
+ "learning_rate": 1e-07,
3541
+ "loss": 2.5706,
3542
+ "step": 22382
3543
+ },
3544
+ {
3545
+ "epoch": 14.75,
3546
+ "learning_rate": 1e-07,
3547
+ "loss": 2.4765,
3548
+ "step": 22420
3549
+ },
3550
+ {
3551
+ "epoch": 14.78,
3552
+ "learning_rate": 1e-07,
3553
+ "loss": 2.4875,
3554
+ "step": 22458
3555
+ },
3556
+ {
3557
+ "epoch": 14.8,
3558
+ "learning_rate": 1e-07,
3559
+ "loss": 2.5111,
3560
+ "step": 22496
3561
+ },
3562
+ {
3563
+ "epoch": 14.82,
3564
+ "learning_rate": 1e-07,
3565
+ "loss": 2.4927,
3566
+ "step": 22534
3567
+ },
3568
+ {
3569
+ "epoch": 14.85,
3570
+ "learning_rate": 1e-07,
3571
+ "loss": 2.4156,
3572
+ "step": 22572
3573
+ },
3574
+ {
3575
+ "epoch": 14.88,
3576
+ "learning_rate": 1e-07,
3577
+ "loss": 2.5199,
3578
+ "step": 22610
3579
+ },
3580
+ {
3581
+ "epoch": 14.9,
3582
+ "learning_rate": 1e-07,
3583
+ "loss": 2.4974,
3584
+ "step": 22648
3585
+ },
3586
+ {
3587
+ "epoch": 14.93,
3588
+ "learning_rate": 1e-07,
3589
+ "loss": 2.5177,
3590
+ "step": 22686
3591
+ },
3592
+ {
3593
+ "epoch": 14.95,
3594
+ "learning_rate": 1e-07,
3595
+ "loss": 2.4438,
3596
+ "step": 22724
3597
+ },
3598
+ {
3599
+ "epoch": 14.97,
3600
+ "learning_rate": 1e-07,
3601
+ "loss": 2.4829,
3602
+ "step": 22762
3603
+ },
3604
+ {
3605
+ "epoch": 15.0,
3606
+ "learning_rate": 1e-07,
3607
+ "loss": 2.544,
3608
+ "step": 22800
3609
+ },
3610
+ {
3611
+ "epoch": 15.03,
3612
+ "learning_rate": 1e-07,
3613
+ "loss": 2.5036,
3614
+ "step": 22838
3615
+ },
3616
+ {
3617
+ "epoch": 15.05,
3618
+ "learning_rate": 1e-07,
3619
+ "loss": 2.4167,
3620
+ "step": 22876
3621
+ },
3622
+ {
3623
+ "epoch": 15.07,
3624
+ "learning_rate": 1e-07,
3625
+ "loss": 2.5017,
3626
+ "step": 22914
3627
+ },
3628
+ {
3629
+ "epoch": 15.1,
3630
+ "learning_rate": 1e-07,
3631
+ "loss": 2.4957,
3632
+ "step": 22952
3633
+ },
3634
+ {
3635
+ "epoch": 15.12,
3636
+ "learning_rate": 1e-07,
3637
+ "loss": 2.4176,
3638
+ "step": 22990
3639
+ },
3640
+ {
3641
+ "epoch": 15.15,
3642
+ "learning_rate": 1e-07,
3643
+ "loss": 2.535,
3644
+ "step": 23028
3645
+ },
3646
+ {
3647
+ "epoch": 15.18,
3648
+ "learning_rate": 1e-07,
3649
+ "loss": 2.5306,
3650
+ "step": 23066
3651
+ },
3652
+ {
3653
+ "epoch": 15.2,
3654
+ "learning_rate": 1e-07,
3655
+ "loss": 2.4867,
3656
+ "step": 23104
3657
+ },
3658
+ {
3659
+ "epoch": 15.22,
3660
+ "learning_rate": 1e-07,
3661
+ "loss": 2.5182,
3662
+ "step": 23142
3663
+ },
3664
+ {
3665
+ "epoch": 15.25,
3666
+ "learning_rate": 1e-07,
3667
+ "loss": 2.4443,
3668
+ "step": 23180
3669
+ },
3670
+ {
3671
+ "epoch": 15.28,
3672
+ "learning_rate": 1e-07,
3673
+ "loss": 2.4416,
3674
+ "step": 23218
3675
+ },
3676
+ {
3677
+ "epoch": 15.3,
3678
+ "learning_rate": 1e-07,
3679
+ "loss": 2.4665,
3680
+ "step": 23256
3681
+ },
3682
+ {
3683
+ "epoch": 15.32,
3684
+ "learning_rate": 1e-07,
3685
+ "loss": 2.4292,
3686
+ "step": 23294
3687
+ },
3688
+ {
3689
+ "epoch": 15.35,
3690
+ "learning_rate": 1e-07,
3691
+ "loss": 2.5607,
3692
+ "step": 23332
3693
+ },
3694
+ {
3695
+ "epoch": 15.38,
3696
+ "learning_rate": 1e-07,
3697
+ "loss": 2.4473,
3698
+ "step": 23370
3699
+ },
3700
+ {
3701
+ "epoch": 15.4,
3702
+ "learning_rate": 1e-07,
3703
+ "loss": 2.522,
3704
+ "step": 23408
3705
+ },
3706
+ {
3707
+ "epoch": 15.43,
3708
+ "learning_rate": 1e-07,
3709
+ "loss": 2.5633,
3710
+ "step": 23446
3711
+ },
3712
+ {
3713
+ "epoch": 15.45,
3714
+ "learning_rate": 1e-07,
3715
+ "loss": 2.4707,
3716
+ "step": 23484
3717
+ },
3718
+ {
3719
+ "epoch": 15.47,
3720
+ "learning_rate": 1e-07,
3721
+ "loss": 2.5822,
3722
+ "step": 23522
3723
+ },
3724
+ {
3725
+ "epoch": 15.5,
3726
+ "learning_rate": 1e-07,
3727
+ "loss": 2.4354,
3728
+ "step": 23560
3729
+ },
3730
+ {
3731
+ "epoch": 15.53,
3732
+ "learning_rate": 1e-07,
3733
+ "loss": 2.337,
3734
+ "step": 23598
3735
+ },
3736
+ {
3737
+ "epoch": 15.55,
3738
+ "learning_rate": 1e-07,
3739
+ "loss": 2.4321,
3740
+ "step": 23636
3741
+ },
3742
+ {
3743
+ "epoch": 15.57,
3744
+ "learning_rate": 1e-07,
3745
+ "loss": 2.453,
3746
+ "step": 23674
3747
+ },
3748
+ {
3749
+ "epoch": 15.6,
3750
+ "learning_rate": 1e-07,
3751
+ "loss": 2.4727,
3752
+ "step": 23712
3753
+ },
3754
+ {
3755
+ "epoch": 15.62,
3756
+ "learning_rate": 1e-07,
3757
+ "loss": 2.5775,
3758
+ "step": 23750
3759
+ },
3760
+ {
3761
+ "epoch": 15.65,
3762
+ "learning_rate": 1e-07,
3763
+ "loss": 2.4485,
3764
+ "step": 23788
3765
+ },
3766
+ {
3767
+ "epoch": 15.68,
3768
+ "learning_rate": 1e-07,
3769
+ "loss": 2.4905,
3770
+ "step": 23826
3771
+ },
3772
+ {
3773
+ "epoch": 15.7,
3774
+ "learning_rate": 1e-07,
3775
+ "loss": 2.486,
3776
+ "step": 23864
3777
+ },
3778
+ {
3779
+ "epoch": 15.72,
3780
+ "learning_rate": 1e-07,
3781
+ "loss": 2.4907,
3782
+ "step": 23902
3783
+ },
3784
+ {
3785
+ "epoch": 15.75,
3786
+ "learning_rate": 1e-07,
3787
+ "loss": 2.4584,
3788
+ "step": 23940
3789
+ },
3790
+ {
3791
+ "epoch": 15.78,
3792
+ "learning_rate": 1e-07,
3793
+ "loss": 2.5427,
3794
+ "step": 23978
3795
+ },
3796
+ {
3797
+ "epoch": 15.8,
3798
+ "learning_rate": 1e-07,
3799
+ "loss": 2.5468,
3800
+ "step": 24016
3801
+ },
3802
+ {
3803
+ "epoch": 15.82,
3804
+ "learning_rate": 1e-07,
3805
+ "loss": 2.5179,
3806
+ "step": 24054
3807
+ },
3808
+ {
3809
+ "epoch": 15.85,
3810
+ "learning_rate": 1e-07,
3811
+ "loss": 2.4527,
3812
+ "step": 24092
3813
+ },
3814
+ {
3815
+ "epoch": 15.88,
3816
+ "learning_rate": 1e-07,
3817
+ "loss": 2.4912,
3818
+ "step": 24130
3819
+ },
3820
+ {
3821
+ "epoch": 15.9,
3822
+ "learning_rate": 1e-07,
3823
+ "loss": 2.5246,
3824
+ "step": 24168
3825
+ },
3826
+ {
3827
+ "epoch": 15.93,
3828
+ "learning_rate": 1e-07,
3829
+ "loss": 2.4518,
3830
+ "step": 24206
3831
+ },
3832
+ {
3833
+ "epoch": 15.95,
3834
+ "learning_rate": 1e-07,
3835
+ "loss": 2.4702,
3836
+ "step": 24244
3837
+ },
3838
+ {
3839
+ "epoch": 15.97,
3840
+ "learning_rate": 1e-07,
3841
+ "loss": 2.4807,
3842
+ "step": 24282
3843
+ },
3844
+ {
3845
+ "epoch": 16.0,
3846
+ "learning_rate": 1e-07,
3847
+ "loss": 2.5016,
3848
+ "step": 24320
3849
+ },
3850
+ {
3851
+ "epoch": 16.02,
3852
+ "learning_rate": 1e-07,
3853
+ "loss": 2.4547,
3854
+ "step": 24358
3855
+ },
3856
+ {
3857
+ "epoch": 16.05,
3858
+ "learning_rate": 1e-07,
3859
+ "loss": 2.5046,
3860
+ "step": 24396
3861
+ },
3862
+ {
3863
+ "epoch": 16.07,
3864
+ "learning_rate": 1e-07,
3865
+ "loss": 2.4804,
3866
+ "step": 24434
3867
+ },
3868
+ {
3869
+ "epoch": 16.1,
3870
+ "learning_rate": 1e-07,
3871
+ "loss": 2.439,
3872
+ "step": 24472
3873
+ },
3874
+ {
3875
+ "epoch": 16.12,
3876
+ "learning_rate": 1e-07,
3877
+ "loss": 2.4343,
3878
+ "step": 24510
3879
+ },
3880
+ {
3881
+ "epoch": 16.15,
3882
+ "learning_rate": 1e-07,
3883
+ "loss": 2.5387,
3884
+ "step": 24548
3885
+ },
3886
+ {
3887
+ "epoch": 16.18,
3888
+ "learning_rate": 1e-07,
3889
+ "loss": 2.4459,
3890
+ "step": 24586
3891
+ },
3892
+ {
3893
+ "epoch": 16.2,
3894
+ "learning_rate": 1e-07,
3895
+ "loss": 2.4423,
3896
+ "step": 24624
3897
+ },
3898
+ {
3899
+ "epoch": 16.23,
3900
+ "learning_rate": 1e-07,
3901
+ "loss": 2.5521,
3902
+ "step": 24662
3903
+ },
3904
+ {
3905
+ "epoch": 16.25,
3906
+ "learning_rate": 1e-07,
3907
+ "loss": 2.5029,
3908
+ "step": 24700
3909
+ },
3910
+ {
3911
+ "epoch": 16.27,
3912
+ "learning_rate": 1e-07,
3913
+ "loss": 2.5005,
3914
+ "step": 24738
3915
+ },
3916
+ {
3917
+ "epoch": 16.3,
3918
+ "learning_rate": 1e-07,
3919
+ "loss": 2.4821,
3920
+ "step": 24776
3921
+ },
3922
+ {
3923
+ "epoch": 16.32,
3924
+ "learning_rate": 1e-07,
3925
+ "loss": 2.4868,
3926
+ "step": 24814
3927
+ },
3928
+ {
3929
+ "epoch": 16.35,
3930
+ "learning_rate": 1e-07,
3931
+ "loss": 2.4993,
3932
+ "step": 24852
3933
+ },
3934
+ {
3935
+ "epoch": 16.38,
3936
+ "learning_rate": 1e-07,
3937
+ "loss": 2.4132,
3938
+ "step": 24890
3939
+ },
3940
+ {
3941
+ "epoch": 16.4,
3942
+ "learning_rate": 1e-07,
3943
+ "loss": 2.4347,
3944
+ "step": 24928
3945
+ },
3946
+ {
3947
+ "epoch": 16.43,
3948
+ "learning_rate": 1e-07,
3949
+ "loss": 2.4667,
3950
+ "step": 24966
3951
+ },
3952
+ {
3953
+ "epoch": 16.45,
3954
+ "learning_rate": 1e-07,
3955
+ "loss": 2.4252,
3956
+ "step": 25004
3957
+ },
3958
+ {
3959
+ "epoch": 16.48,
3960
+ "learning_rate": 1e-07,
3961
+ "loss": 2.4267,
3962
+ "step": 25042
3963
+ },
3964
+ {
3965
+ "epoch": 16.5,
3966
+ "learning_rate": 1e-07,
3967
+ "loss": 2.4384,
3968
+ "step": 25080
3969
+ },
3970
+ {
3971
+ "epoch": 16.52,
3972
+ "learning_rate": 1e-07,
3973
+ "loss": 2.5273,
3974
+ "step": 25118
3975
+ },
3976
+ {
3977
+ "epoch": 16.55,
3978
+ "learning_rate": 1e-07,
3979
+ "loss": 2.5036,
3980
+ "step": 25156
3981
+ },
3982
+ {
3983
+ "epoch": 16.57,
3984
+ "learning_rate": 1e-07,
3985
+ "loss": 2.5159,
3986
+ "step": 25194
3987
+ },
3988
+ {
3989
+ "epoch": 16.6,
3990
+ "learning_rate": 1e-07,
3991
+ "loss": 2.4323,
3992
+ "step": 25232
3993
+ },
3994
+ {
3995
+ "epoch": 16.62,
3996
+ "learning_rate": 1e-07,
3997
+ "loss": 2.3885,
3998
+ "step": 25270
3999
+ },
4000
+ {
4001
+ "epoch": 16.65,
4002
+ "learning_rate": 1e-07,
4003
+ "loss": 2.4609,
4004
+ "step": 25308
4005
+ },
4006
+ {
4007
+ "epoch": 16.68,
4008
+ "learning_rate": 1e-07,
4009
+ "loss": 2.4227,
4010
+ "step": 25346
4011
+ },
4012
+ {
4013
+ "epoch": 16.7,
4014
+ "learning_rate": 1e-07,
4015
+ "loss": 2.5008,
4016
+ "step": 25384
4017
+ },
4018
+ {
4019
+ "epoch": 16.73,
4020
+ "learning_rate": 1e-07,
4021
+ "loss": 2.4119,
4022
+ "step": 25422
4023
+ },
4024
+ {
4025
+ "epoch": 16.75,
4026
+ "learning_rate": 1e-07,
4027
+ "loss": 2.4854,
4028
+ "step": 25460
4029
+ },
4030
+ {
4031
+ "epoch": 16.77,
4032
+ "learning_rate": 1e-07,
4033
+ "loss": 2.4073,
4034
+ "step": 25498
4035
+ },
4036
+ {
4037
+ "epoch": 16.8,
4038
+ "learning_rate": 1e-07,
4039
+ "loss": 2.4732,
4040
+ "step": 25536
4041
+ },
4042
+ {
4043
+ "epoch": 16.82,
4044
+ "learning_rate": 1e-07,
4045
+ "loss": 2.5109,
4046
+ "step": 25574
4047
+ },
4048
+ {
4049
+ "epoch": 16.85,
4050
+ "learning_rate": 1e-07,
4051
+ "loss": 2.4875,
4052
+ "step": 25612
4053
+ },
4054
+ {
4055
+ "epoch": 16.88,
4056
+ "learning_rate": 1e-07,
4057
+ "loss": 2.4114,
4058
+ "step": 25650
4059
+ },
4060
+ {
4061
+ "epoch": 16.9,
4062
+ "learning_rate": 1e-07,
4063
+ "loss": 2.5228,
4064
+ "step": 25688
4065
+ },
4066
+ {
4067
+ "epoch": 16.93,
4068
+ "learning_rate": 1e-07,
4069
+ "loss": 2.467,
4070
+ "step": 25726
4071
+ },
4072
+ {
4073
+ "epoch": 16.95,
4074
+ "learning_rate": 1e-07,
4075
+ "loss": 2.4497,
4076
+ "step": 25764
4077
+ },
4078
+ {
4079
+ "epoch": 16.98,
4080
+ "learning_rate": 1e-07,
4081
+ "loss": 2.4725,
4082
+ "step": 25802
4083
+ },
4084
+ {
4085
+ "epoch": 17.0,
4086
+ "learning_rate": 1e-07,
4087
+ "loss": 2.4368,
4088
+ "step": 25840
4089
+ },
4090
+ {
4091
+ "epoch": 17.02,
4092
+ "learning_rate": 1e-07,
4093
+ "loss": 2.4846,
4094
+ "step": 25878
4095
+ },
4096
+ {
4097
+ "epoch": 17.05,
4098
+ "learning_rate": 1e-07,
4099
+ "loss": 2.3977,
4100
+ "step": 25916
4101
+ },
4102
+ {
4103
+ "epoch": 17.07,
4104
+ "learning_rate": 1e-07,
4105
+ "loss": 2.4389,
4106
+ "step": 25954
4107
+ },
4108
+ {
4109
+ "epoch": 17.1,
4110
+ "learning_rate": 1e-07,
4111
+ "loss": 2.4865,
4112
+ "step": 25992
4113
+ },
4114
+ {
4115
+ "epoch": 17.12,
4116
+ "learning_rate": 1e-07,
4117
+ "loss": 2.3666,
4118
+ "step": 26030
4119
+ },
4120
+ {
4121
+ "epoch": 17.15,
4122
+ "learning_rate": 1e-07,
4123
+ "loss": 2.5045,
4124
+ "step": 26068
4125
+ },
4126
+ {
4127
+ "epoch": 17.18,
4128
+ "learning_rate": 1e-07,
4129
+ "loss": 2.4607,
4130
+ "step": 26106
4131
+ },
4132
+ {
4133
+ "epoch": 17.2,
4134
+ "learning_rate": 1e-07,
4135
+ "loss": 2.5089,
4136
+ "step": 26144
4137
+ },
4138
+ {
4139
+ "epoch": 17.23,
4140
+ "learning_rate": 1e-07,
4141
+ "loss": 2.4729,
4142
+ "step": 26182
4143
+ },
4144
+ {
4145
+ "epoch": 17.25,
4146
+ "learning_rate": 1e-07,
4147
+ "loss": 2.4228,
4148
+ "step": 26220
4149
+ },
4150
+ {
4151
+ "epoch": 17.27,
4152
+ "learning_rate": 1e-07,
4153
+ "loss": 2.4237,
4154
+ "step": 26258
4155
+ },
4156
+ {
4157
+ "epoch": 17.3,
4158
+ "learning_rate": 1e-07,
4159
+ "loss": 2.4593,
4160
+ "step": 26296
4161
+ },
4162
+ {
4163
+ "epoch": 17.32,
4164
+ "learning_rate": 1e-07,
4165
+ "loss": 2.4044,
4166
+ "step": 26334
4167
+ },
4168
+ {
4169
+ "epoch": 17.35,
4170
+ "learning_rate": 1e-07,
4171
+ "loss": 2.5488,
4172
+ "step": 26372
4173
+ },
4174
+ {
4175
+ "epoch": 17.38,
4176
+ "learning_rate": 1e-07,
4177
+ "loss": 2.3409,
4178
+ "step": 26410
4179
+ },
4180
+ {
4181
+ "epoch": 17.4,
4182
+ "learning_rate": 1e-07,
4183
+ "loss": 2.3963,
4184
+ "step": 26448
4185
+ },
4186
+ {
4187
+ "epoch": 17.43,
4188
+ "learning_rate": 1e-07,
4189
+ "loss": 2.4953,
4190
+ "step": 26486
4191
+ },
4192
+ {
4193
+ "epoch": 17.45,
4194
+ "learning_rate": 1e-07,
4195
+ "loss": 2.4604,
4196
+ "step": 26524
4197
+ },
4198
+ {
4199
+ "epoch": 17.48,
4200
+ "learning_rate": 1e-07,
4201
+ "loss": 2.4211,
4202
+ "step": 26562
4203
+ },
4204
+ {
4205
+ "epoch": 17.5,
4206
+ "learning_rate": 1e-07,
4207
+ "loss": 2.4791,
4208
+ "step": 26600
4209
+ },
4210
+ {
4211
+ "epoch": 17.52,
4212
+ "learning_rate": 1e-07,
4213
+ "loss": 2.442,
4214
+ "step": 26638
4215
+ },
4216
+ {
4217
+ "epoch": 17.55,
4218
+ "learning_rate": 1e-07,
4219
+ "loss": 2.4315,
4220
+ "step": 26676
4221
+ },
4222
+ {
4223
+ "epoch": 17.57,
4224
+ "learning_rate": 1e-07,
4225
+ "loss": 2.4432,
4226
+ "step": 26714
4227
+ },
4228
+ {
4229
+ "epoch": 17.6,
4230
+ "learning_rate": 1e-07,
4231
+ "loss": 2.4599,
4232
+ "step": 26752
4233
+ },
4234
+ {
4235
+ "epoch": 17.62,
4236
+ "learning_rate": 1e-07,
4237
+ "loss": 2.5794,
4238
+ "step": 26790
4239
+ },
4240
+ {
4241
+ "epoch": 17.65,
4242
+ "learning_rate": 1e-07,
4243
+ "loss": 2.4263,
4244
+ "step": 26828
4245
+ },
4246
+ {
4247
+ "epoch": 17.68,
4248
+ "learning_rate": 1e-07,
4249
+ "loss": 2.4788,
4250
+ "step": 26866
4251
+ },
4252
+ {
4253
+ "epoch": 17.7,
4254
+ "learning_rate": 1e-07,
4255
+ "loss": 2.4702,
4256
+ "step": 26904
4257
+ },
4258
+ {
4259
+ "epoch": 17.73,
4260
+ "learning_rate": 1e-07,
4261
+ "loss": 2.4099,
4262
+ "step": 26942
4263
+ },
4264
+ {
4265
+ "epoch": 17.75,
4266
+ "learning_rate": 1e-07,
4267
+ "loss": 2.3706,
4268
+ "step": 26980
4269
+ },
4270
+ {
4271
+ "epoch": 17.77,
4272
+ "learning_rate": 1e-07,
4273
+ "loss": 2.3648,
4274
+ "step": 27018
4275
+ },
4276
+ {
4277
+ "epoch": 17.8,
4278
+ "learning_rate": 1e-07,
4279
+ "loss": 2.4577,
4280
+ "step": 27056
4281
+ },
4282
+ {
4283
+ "epoch": 17.82,
4284
+ "learning_rate": 1e-07,
4285
+ "loss": 2.454,
4286
+ "step": 27094
4287
+ },
4288
+ {
4289
+ "epoch": 17.85,
4290
+ "learning_rate": 1e-07,
4291
+ "loss": 2.4376,
4292
+ "step": 27132
4293
+ },
4294
+ {
4295
+ "epoch": 17.88,
4296
+ "learning_rate": 1e-07,
4297
+ "loss": 2.4718,
4298
+ "step": 27170
4299
+ },
4300
+ {
4301
+ "epoch": 17.9,
4302
+ "learning_rate": 1e-07,
4303
+ "loss": 2.4469,
4304
+ "step": 27208
4305
+ },
4306
+ {
4307
+ "epoch": 17.93,
4308
+ "learning_rate": 1e-07,
4309
+ "loss": 2.4622,
4310
+ "step": 27246
4311
+ },
4312
+ {
4313
+ "epoch": 17.95,
4314
+ "learning_rate": 1e-07,
4315
+ "loss": 2.4444,
4316
+ "step": 27284
4317
+ },
4318
+ {
4319
+ "epoch": 17.98,
4320
+ "learning_rate": 1e-07,
4321
+ "loss": 2.4595,
4322
+ "step": 27322
4323
+ },
4324
+ {
4325
+ "epoch": 18.0,
4326
+ "learning_rate": 1e-07,
4327
+ "loss": 2.4624,
4328
+ "step": 27360
4329
+ },
4330
+ {
4331
+ "epoch": 18.0,
4332
+ "step": 27360,
4333
+ "total_flos": 4.063318154359603e+17,
4334
+ "train_loss": 3.0161131058520043,
4335
+ "train_runtime": 38619.075,
4336
+ "train_samples_per_second": 1.417,
4337
+ "train_steps_per_second": 0.708
4338
+ }
4339
+ ],
4340
+ "max_steps": 27360,
4341
+ "num_train_epochs": 18,
4342
+ "total_flos": 4.063318154359603e+17,
4343
+ "trial_name": null,
4344
+ "trial_params": null
4345
+ }
output/training_args.bin ADDED
Binary file (4.03 kB). View file