Hugging Face's logo

Weni
/
WeniGPT-Mistral-7B-instructBase

Text Generation
Transformers
PyTorch
mistral
Generated from Trainer
conversational
text-generation-inference
Inference Endpoints
Model card Files Community
1
Mel-Iza0 commited on
Commit
3e8ca30
•
1 Parent(s): 5bb664b

Training in progress, epoch 1, checkpoint

Browse files
Files changed (13) hide show
  1. checkpoint-8000/README.md +10 -0
  2. checkpoint-8000/adapter_config.json +22 -0
  3. checkpoint-8000/adapter_model.bin +3 -0
  4. checkpoint-8000/added_tokens.json +5 -0
  5. checkpoint-8000/optimizer.pt +3 -0
  6. checkpoint-8000/rng_state.pth +3 -0
  7. checkpoint-8000/scheduler.pt +3 -0
  8. checkpoint-8000/special_tokens_map.json +11 -0
  9. checkpoint-8000/tokenizer.json +0 -0
  10. checkpoint-8000/tokenizer.model +3 -0
  11. checkpoint-8000/tokenizer_config.json +45 -0
  12. checkpoint-8000/trainer_state.json +4819 -0
  13. checkpoint-8000/training_args.bin +3 -0
checkpoint-8000/README.md ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: peft
3
+ ---
4
+ ## Training procedure
5
+
6
+ ### Framework versions
7
+
8
+ - PEFT 0.5.0
9
+
10
+ - PEFT 0.5.0
checkpoint-8000/adapter_config.json ADDED
@@ -0,0 +1,22 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "auto_mapping": null,
3
+ "base_model_name_or_path": "mistralai/Mistral-7B-Instruct-v0.1",
4
+ "bias": "none",
5
+ "fan_in_fan_out": false,
6
+ "inference_mode": true,
7
+ "init_lora_weights": true,
8
+ "layers_pattern": null,
9
+ "layers_to_transform": null,
10
+ "lora_alpha": 16,
11
+ "lora_dropout": 0.1,
12
+ "modules_to_save": null,
13
+ "peft_type": "LORA",
14
+ "r": 16,
15
+ "revision": null,
16
+ "target_modules": [
17
+ "q_proj",
18
+ "k_proj",
19
+ "v_projo_proj"
20
+ ],
21
+ "task_type": "CAUSAL_LM"
22
+ }
checkpoint-8000/adapter_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fe5664142023400abd072e20e50dbbb0eab7a9bf03a173d59b34b3f17b4a9552
3
+ size 27308941
checkpoint-8000/added_tokens.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "</s>": 2,
3
+ "<s>": 1,
4
+ "<unk>": 0
5
+ }
checkpoint-8000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9b52719f2b07bba159e956656a9dab9e4bf30e3cd98c25a9f69d58f9158b4a36
3
+ size 54633541
checkpoint-8000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8e278c1dd0b29a83911f1eb3ab38283ebe672596453978025b20108c1a5924ed
3
+ size 14575
checkpoint-8000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1845e0aed5b14e6ceddffdbd894bbc4114507e4caa0126112bd24244d477dc12
3
+ size 627
checkpoint-8000/special_tokens_map.json ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<unk>",
4
+ "<s>",
5
+ "</s>"
6
+ ],
7
+ "bos_token": "<s>",
8
+ "eos_token": "</s>",
9
+ "pad_token": "</s>",
10
+ "unk_token": "<unk>"
11
+ }
checkpoint-8000/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-8000/tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dadfd56d766715c61d2ef780a525ab43b8e6da4de6865bda3d95fdef5e134055
3
+ size 493443
checkpoint-8000/tokenizer_config.json ADDED
@@ -0,0 +1,45 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<unk>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<s>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "</s>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ }
27
+ },
28
+ "additional_special_tokens": [
29
+ "<unk>",
30
+ "<s>",
31
+ "</s>"
32
+ ],
33
+ "bos_token": "<s>",
34
+ "chat_template": "{{ bos_token }}{% for message in messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% if message['role'] == 'user' %}{{ '[INST] ' + message['content'] + ' [/INST]' }}{% elif message['role'] == 'assistant' %}{{ message['content'] + eos_token + ' ' }}{% else %}{{ raise_exception('Only user and assistant roles are supported!') }}{% endif %}{% endfor %}",
35
+ "clean_up_tokenization_spaces": false,
36
+ "eos_token": "</s>",
37
+ "legacy": true,
38
+ "model_max_length": 1000000000000000019884624838656,
39
+ "pad_token": "</s>",
40
+ "sp_model_kwargs": {},
41
+ "spaces_between_special_tokens": false,
42
+ "tokenizer_class": "LlamaTokenizer",
43
+ "unk_token": "<unk>",
44
+ "use_default_system_prompt": true
45
+ }
checkpoint-8000/trainer_state.json ADDED
@@ -0,0 +1,4819 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 1.1006397468528581,
5
+ "eval_steps": 500,
6
+ "global_step": 8000,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.0,
13
+ "learning_rate": 1.6666666666666667e-05,
14
+ "loss": 0.9934,
15
+ "step": 10
16
+ },
17
+ {
18
+ "epoch": 0.0,
19
+ "learning_rate": 3.3333333333333335e-05,
20
+ "loss": 1.0101,
21
+ "step": 20
22
+ },
23
+ {
24
+ "epoch": 0.0,
25
+ "learning_rate": 5e-05,
26
+ "loss": 0.98,
27
+ "step": 30
28
+ },
29
+ {
30
+ "epoch": 0.01,
31
+ "learning_rate": 6.666666666666667e-05,
32
+ "loss": 0.962,
33
+ "step": 40
34
+ },
35
+ {
36
+ "epoch": 0.01,
37
+ "learning_rate": 8.333333333333334e-05,
38
+ "loss": 0.9046,
39
+ "step": 50
40
+ },
41
+ {
42
+ "epoch": 0.01,
43
+ "learning_rate": 0.0001,
44
+ "loss": 0.8624,
45
+ "step": 60
46
+ },
47
+ {
48
+ "epoch": 0.01,
49
+ "learning_rate": 0.00011666666666666668,
50
+ "loss": 0.8147,
51
+ "step": 70
52
+ },
53
+ {
54
+ "epoch": 0.01,
55
+ "learning_rate": 0.00013333333333333334,
56
+ "loss": 0.807,
57
+ "step": 80
58
+ },
59
+ {
60
+ "epoch": 0.01,
61
+ "learning_rate": 0.00015000000000000001,
62
+ "loss": 0.7814,
63
+ "step": 90
64
+ },
65
+ {
66
+ "epoch": 0.01,
67
+ "learning_rate": 0.0001666666666666667,
68
+ "loss": 0.7618,
69
+ "step": 100
70
+ },
71
+ {
72
+ "epoch": 0.02,
73
+ "learning_rate": 0.00018333333333333334,
74
+ "loss": 0.7724,
75
+ "step": 110
76
+ },
77
+ {
78
+ "epoch": 0.02,
79
+ "learning_rate": 0.0002,
80
+ "loss": 0.7593,
81
+ "step": 120
82
+ },
83
+ {
84
+ "epoch": 0.02,
85
+ "learning_rate": 0.00021666666666666666,
86
+ "loss": 0.7546,
87
+ "step": 130
88
+ },
89
+ {
90
+ "epoch": 0.02,
91
+ "learning_rate": 0.00023333333333333336,
92
+ "loss": 0.7347,
93
+ "step": 140
94
+ },
95
+ {
96
+ "epoch": 0.02,
97
+ "learning_rate": 0.00025,
98
+ "loss": 0.7227,
99
+ "step": 150
100
+ },
101
+ {
102
+ "epoch": 0.02,
103
+ "learning_rate": 0.0002666666666666667,
104
+ "loss": 0.7359,
105
+ "step": 160
106
+ },
107
+ {
108
+ "epoch": 0.02,
109
+ "learning_rate": 0.00028333333333333335,
110
+ "loss": 0.7309,
111
+ "step": 170
112
+ },
113
+ {
114
+ "epoch": 0.02,
115
+ "learning_rate": 0.00030000000000000003,
116
+ "loss": 0.7237,
117
+ "step": 180
118
+ },
119
+ {
120
+ "epoch": 0.03,
121
+ "learning_rate": 0.00031666666666666665,
122
+ "loss": 0.7288,
123
+ "step": 190
124
+ },
125
+ {
126
+ "epoch": 0.03,
127
+ "learning_rate": 0.0003333333333333334,
128
+ "loss": 0.7329,
129
+ "step": 200
130
+ },
131
+ {
132
+ "epoch": 0.03,
133
+ "learning_rate": 0.00035,
134
+ "loss": 0.7229,
135
+ "step": 210
136
+ },
137
+ {
138
+ "epoch": 0.03,
139
+ "learning_rate": 0.00036666666666666667,
140
+ "loss": 0.7089,
141
+ "step": 220
142
+ },
143
+ {
144
+ "epoch": 0.03,
145
+ "learning_rate": 0.00038333333333333334,
146
+ "loss": 0.7214,
147
+ "step": 230
148
+ },
149
+ {
150
+ "epoch": 0.03,
151
+ "learning_rate": 0.0004,
152
+ "loss": 0.7022,
153
+ "step": 240
154
+ },
155
+ {
156
+ "epoch": 0.03,
157
+ "learning_rate": 0.0004,
158
+ "loss": 0.7097,
159
+ "step": 250
160
+ },
161
+ {
162
+ "epoch": 0.04,
163
+ "learning_rate": 0.0004,
164
+ "loss": 0.7128,
165
+ "step": 260
166
+ },
167
+ {
168
+ "epoch": 0.04,
169
+ "learning_rate": 0.0004,
170
+ "loss": 0.6895,
171
+ "step": 270
172
+ },
173
+ {
174
+ "epoch": 0.04,
175
+ "learning_rate": 0.0004,
176
+ "loss": 0.7036,
177
+ "step": 280
178
+ },
179
+ {
180
+ "epoch": 0.04,
181
+ "learning_rate": 0.0004,
182
+ "loss": 0.6802,
183
+ "step": 290
184
+ },
185
+ {
186
+ "epoch": 0.04,
187
+ "learning_rate": 0.0004,
188
+ "loss": 0.7051,
189
+ "step": 300
190
+ },
191
+ {
192
+ "epoch": 0.04,
193
+ "learning_rate": 0.0004,
194
+ "loss": 0.6901,
195
+ "step": 310
196
+ },
197
+ {
198
+ "epoch": 0.04,
199
+ "learning_rate": 0.0004,
200
+ "loss": 0.6852,
201
+ "step": 320
202
+ },
203
+ {
204
+ "epoch": 0.05,
205
+ "learning_rate": 0.0004,
206
+ "loss": 0.6823,
207
+ "step": 330
208
+ },
209
+ {
210
+ "epoch": 0.05,
211
+ "learning_rate": 0.0004,
212
+ "loss": 0.6889,
213
+ "step": 340
214
+ },
215
+ {
216
+ "epoch": 0.05,
217
+ "learning_rate": 0.0004,
218
+ "loss": 0.6698,
219
+ "step": 350
220
+ },
221
+ {
222
+ "epoch": 0.05,
223
+ "learning_rate": 0.0004,
224
+ "loss": 0.6695,
225
+ "step": 360
226
+ },
227
+ {
228
+ "epoch": 0.05,
229
+ "learning_rate": 0.0004,
230
+ "loss": 0.681,
231
+ "step": 370
232
+ },
233
+ {
234
+ "epoch": 0.05,
235
+ "learning_rate": 0.0004,
236
+ "loss": 0.6878,
237
+ "step": 380
238
+ },
239
+ {
240
+ "epoch": 0.05,
241
+ "learning_rate": 0.0004,
242
+ "loss": 0.6718,
243
+ "step": 390
244
+ },
245
+ {
246
+ "epoch": 0.06,
247
+ "learning_rate": 0.0004,
248
+ "loss": 0.6505,
249
+ "step": 400
250
+ },
251
+ {
252
+ "epoch": 0.06,
253
+ "learning_rate": 0.0004,
254
+ "loss": 0.686,
255
+ "step": 410
256
+ },
257
+ {
258
+ "epoch": 0.06,
259
+ "learning_rate": 0.0004,
260
+ "loss": 0.6634,
261
+ "step": 420
262
+ },
263
+ {
264
+ "epoch": 0.06,
265
+ "learning_rate": 0.0004,
266
+ "loss": 0.6701,
267
+ "step": 430
268
+ },
269
+ {
270
+ "epoch": 0.06,
271
+ "learning_rate": 0.0004,
272
+ "loss": 0.6733,
273
+ "step": 440
274
+ },
275
+ {
276
+ "epoch": 0.06,
277
+ "learning_rate": 0.0004,
278
+ "loss": 0.6889,
279
+ "step": 450
280
+ },
281
+ {
282
+ "epoch": 0.06,
283
+ "learning_rate": 0.0004,
284
+ "loss": 0.6591,
285
+ "step": 460
286
+ },
287
+ {
288
+ "epoch": 0.06,
289
+ "learning_rate": 0.0004,
290
+ "loss": 0.6769,
291
+ "step": 470
292
+ },
293
+ {
294
+ "epoch": 0.07,
295
+ "learning_rate": 0.0004,
296
+ "loss": 0.6839,
297
+ "step": 480
298
+ },
299
+ {
300
+ "epoch": 0.07,
301
+ "learning_rate": 0.0004,
302
+ "loss": 0.6579,
303
+ "step": 490
304
+ },
305
+ {
306
+ "epoch": 0.07,
307
+ "learning_rate": 0.0004,
308
+ "loss": 0.6768,
309
+ "step": 500
310
+ },
311
+ {
312
+ "epoch": 0.07,
313
+ "learning_rate": 0.0004,
314
+ "loss": 0.6644,
315
+ "step": 510
316
+ },
317
+ {
318
+ "epoch": 0.07,
319
+ "learning_rate": 0.0004,
320
+ "loss": 0.6488,
321
+ "step": 520
322
+ },
323
+ {
324
+ "epoch": 0.07,
325
+ "learning_rate": 0.0004,
326
+ "loss": 0.6557,
327
+ "step": 530
328
+ },
329
+ {
330
+ "epoch": 0.07,
331
+ "learning_rate": 0.0004,
332
+ "loss": 0.6738,
333
+ "step": 540
334
+ },
335
+ {
336
+ "epoch": 0.08,
337
+ "learning_rate": 0.0004,
338
+ "loss": 0.6536,
339
+ "step": 550
340
+ },
341
+ {
342
+ "epoch": 0.08,
343
+ "learning_rate": 0.0004,
344
+ "loss": 0.6679,
345
+ "step": 560
346
+ },
347
+ {
348
+ "epoch": 0.08,
349
+ "learning_rate": 0.0004,
350
+ "loss": 0.6359,
351
+ "step": 570
352
+ },
353
+ {
354
+ "epoch": 0.08,
355
+ "learning_rate": 0.0004,
356
+ "loss": 0.6492,
357
+ "step": 580
358
+ },
359
+ {
360
+ "epoch": 0.08,
361
+ "learning_rate": 0.0004,
362
+ "loss": 0.6488,
363
+ "step": 590
364
+ },
365
+ {
366
+ "epoch": 0.08,
367
+ "learning_rate": 0.0004,
368
+ "loss": 0.6764,
369
+ "step": 600
370
+ },
371
+ {
372
+ "epoch": 0.08,
373
+ "learning_rate": 0.0004,
374
+ "loss": 0.6511,
375
+ "step": 610
376
+ },
377
+ {
378
+ "epoch": 0.09,
379
+ "learning_rate": 0.0004,
380
+ "loss": 0.6738,
381
+ "step": 620
382
+ },
383
+ {
384
+ "epoch": 0.09,
385
+ "learning_rate": 0.0004,
386
+ "loss": 0.6424,
387
+ "step": 630
388
+ },
389
+ {
390
+ "epoch": 0.09,
391
+ "learning_rate": 0.0004,
392
+ "loss": 0.6582,
393
+ "step": 640
394
+ },
395
+ {
396
+ "epoch": 0.09,
397
+ "learning_rate": 0.0004,
398
+ "loss": 0.6649,
399
+ "step": 650
400
+ },
401
+ {
402
+ "epoch": 0.09,
403
+ "learning_rate": 0.0004,
404
+ "loss": 0.6697,
405
+ "step": 660
406
+ },
407
+ {
408
+ "epoch": 0.09,
409
+ "learning_rate": 0.0004,
410
+ "loss": 0.6491,
411
+ "step": 670
412
+ },
413
+ {
414
+ "epoch": 0.09,
415
+ "learning_rate": 0.0004,
416
+ "loss": 0.665,
417
+ "step": 680
418
+ },
419
+ {
420
+ "epoch": 0.09,
421
+ "learning_rate": 0.0004,
422
+ "loss": 0.643,
423
+ "step": 690
424
+ },
425
+ {
426
+ "epoch": 0.1,
427
+ "learning_rate": 0.0004,
428
+ "loss": 0.656,
429
+ "step": 700
430
+ },
431
+ {
432
+ "epoch": 0.1,
433
+ "learning_rate": 0.0004,
434
+ "loss": 0.6469,
435
+ "step": 710
436
+ },
437
+ {
438
+ "epoch": 0.1,
439
+ "learning_rate": 0.0004,
440
+ "loss": 0.6543,
441
+ "step": 720
442
+ },
443
+ {
444
+ "epoch": 0.1,
445
+ "learning_rate": 0.0004,
446
+ "loss": 0.6673,
447
+ "step": 730
448
+ },
449
+ {
450
+ "epoch": 0.1,
451
+ "learning_rate": 0.0004,
452
+ "loss": 0.6515,
453
+ "step": 740
454
+ },
455
+ {
456
+ "epoch": 0.1,
457
+ "learning_rate": 0.0004,
458
+ "loss": 0.6462,
459
+ "step": 750
460
+ },
461
+ {
462
+ "epoch": 0.1,
463
+ "learning_rate": 0.0004,
464
+ "loss": 0.6559,
465
+ "step": 760
466
+ },
467
+ {
468
+ "epoch": 0.11,
469
+ "learning_rate": 0.0004,
470
+ "loss": 0.6439,
471
+ "step": 770
472
+ },
473
+ {
474
+ "epoch": 0.11,
475
+ "learning_rate": 0.0004,
476
+ "loss": 0.6417,
477
+ "step": 780
478
+ },
479
+ {
480
+ "epoch": 0.11,
481
+ "learning_rate": 0.0004,
482
+ "loss": 0.6525,
483
+ "step": 790
484
+ },
485
+ {
486
+ "epoch": 0.11,
487
+ "learning_rate": 0.0004,
488
+ "loss": 0.6406,
489
+ "step": 800
490
+ },
491
+ {
492
+ "epoch": 0.11,
493
+ "learning_rate": 0.0004,
494
+ "loss": 0.6524,
495
+ "step": 810
496
+ },
497
+ {
498
+ "epoch": 0.11,
499
+ "learning_rate": 0.0004,
500
+ "loss": 0.6399,
501
+ "step": 820
502
+ },
503
+ {
504
+ "epoch": 0.11,
505
+ "learning_rate": 0.0004,
506
+ "loss": 0.6609,
507
+ "step": 830
508
+ },
509
+ {
510
+ "epoch": 0.12,
511
+ "learning_rate": 0.0004,
512
+ "loss": 0.6179,
513
+ "step": 840
514
+ },
515
+ {
516
+ "epoch": 0.12,
517
+ "learning_rate": 0.0004,
518
+ "loss": 0.6528,
519
+ "step": 850
520
+ },
521
+ {
522
+ "epoch": 0.12,
523
+ "learning_rate": 0.0004,
524
+ "loss": 0.6528,
525
+ "step": 860
526
+ },
527
+ {
528
+ "epoch": 0.12,
529
+ "learning_rate": 0.0004,
530
+ "loss": 0.6432,
531
+ "step": 870
532
+ },
533
+ {
534
+ "epoch": 0.12,
535
+ "learning_rate": 0.0004,
536
+ "loss": 0.6584,
537
+ "step": 880
538
+ },
539
+ {
540
+ "epoch": 0.12,
541
+ "learning_rate": 0.0004,
542
+ "loss": 0.6601,
543
+ "step": 890
544
+ },
545
+ {
546
+ "epoch": 0.12,
547
+ "learning_rate": 0.0004,
548
+ "loss": 0.6547,
549
+ "step": 900
550
+ },
551
+ {
552
+ "epoch": 0.13,
553
+ "learning_rate": 0.0004,
554
+ "loss": 0.6497,
555
+ "step": 910
556
+ },
557
+ {
558
+ "epoch": 0.13,
559
+ "learning_rate": 0.0004,
560
+ "loss": 0.6312,
561
+ "step": 920
562
+ },
563
+ {
564
+ "epoch": 0.13,
565
+ "learning_rate": 0.0004,
566
+ "loss": 0.6408,
567
+ "step": 930
568
+ },
569
+ {
570
+ "epoch": 0.13,
571
+ "learning_rate": 0.0004,
572
+ "loss": 0.6441,
573
+ "step": 940
574
+ },
575
+ {
576
+ "epoch": 0.13,
577
+ "learning_rate": 0.0004,
578
+ "loss": 0.6168,
579
+ "step": 950
580
+ },
581
+ {
582
+ "epoch": 0.13,
583
+ "learning_rate": 0.0004,
584
+ "loss": 0.6247,
585
+ "step": 960
586
+ },
587
+ {
588
+ "epoch": 0.13,
589
+ "learning_rate": 0.0004,
590
+ "loss": 0.6398,
591
+ "step": 970
592
+ },
593
+ {
594
+ "epoch": 0.13,
595
+ "learning_rate": 0.0004,
596
+ "loss": 0.6227,
597
+ "step": 980
598
+ },
599
+ {
600
+ "epoch": 0.14,
601
+ "learning_rate": 0.0004,
602
+ "loss": 0.6254,
603
+ "step": 990
604
+ },
605
+ {
606
+ "epoch": 0.14,
607
+ "learning_rate": 0.0004,
608
+ "loss": 0.6468,
609
+ "step": 1000
610
+ },
611
+ {
612
+ "epoch": 0.14,
613
+ "learning_rate": 0.0004,
614
+ "loss": 0.6318,
615
+ "step": 1010
616
+ },
617
+ {
618
+ "epoch": 0.14,
619
+ "learning_rate": 0.0004,
620
+ "loss": 0.611,
621
+ "step": 1020
622
+ },
623
+ {
624
+ "epoch": 0.14,
625
+ "learning_rate": 0.0004,
626
+ "loss": 0.6401,
627
+ "step": 1030
628
+ },
629
+ {
630
+ "epoch": 0.14,
631
+ "learning_rate": 0.0004,
632
+ "loss": 0.6165,
633
+ "step": 1040
634
+ },
635
+ {
636
+ "epoch": 0.14,
637
+ "learning_rate": 0.0004,
638
+ "loss": 0.6352,
639
+ "step": 1050
640
+ },
641
+ {
642
+ "epoch": 0.15,
643
+ "learning_rate": 0.0004,
644
+ "loss": 0.6532,
645
+ "step": 1060
646
+ },
647
+ {
648
+ "epoch": 0.15,
649
+ "learning_rate": 0.0004,
650
+ "loss": 0.6129,
651
+ "step": 1070
652
+ },
653
+ {
654
+ "epoch": 0.15,
655
+ "learning_rate": 0.0004,
656
+ "loss": 0.6254,
657
+ "step": 1080
658
+ },
659
+ {
660
+ "epoch": 0.15,
661
+ "learning_rate": 0.0004,
662
+ "loss": 0.6179,
663
+ "step": 1090
664
+ },
665
+ {
666
+ "epoch": 0.15,
667
+ "learning_rate": 0.0004,
668
+ "loss": 0.6141,
669
+ "step": 1100
670
+ },
671
+ {
672
+ "epoch": 0.15,
673
+ "learning_rate": 0.0004,
674
+ "loss": 0.6215,
675
+ "step": 1110
676
+ },
677
+ {
678
+ "epoch": 0.15,
679
+ "learning_rate": 0.0004,
680
+ "loss": 0.6333,
681
+ "step": 1120
682
+ },
683
+ {
684
+ "epoch": 0.16,
685
+ "learning_rate": 0.0004,
686
+ "loss": 0.6254,
687
+ "step": 1130
688
+ },
689
+ {
690
+ "epoch": 0.16,
691
+ "learning_rate": 0.0004,
692
+ "loss": 0.6284,
693
+ "step": 1140
694
+ },
695
+ {
696
+ "epoch": 0.16,
697
+ "learning_rate": 0.0004,
698
+ "loss": 0.6287,
699
+ "step": 1150
700
+ },
701
+ {
702
+ "epoch": 0.16,
703
+ "learning_rate": 0.0004,
704
+ "loss": 0.6085,
705
+ "step": 1160
706
+ },
707
+ {
708
+ "epoch": 0.16,
709
+ "learning_rate": 0.0004,
710
+ "loss": 0.6288,
711
+ "step": 1170
712
+ },
713
+ {
714
+ "epoch": 0.16,
715
+ "learning_rate": 0.0004,
716
+ "loss": 0.6339,
717
+ "step": 1180
718
+ },
719
+ {
720
+ "epoch": 0.16,
721
+ "learning_rate": 0.0004,
722
+ "loss": 0.6256,
723
+ "step": 1190
724
+ },
725
+ {
726
+ "epoch": 0.17,
727
+ "learning_rate": 0.0004,
728
+ "loss": 0.6382,
729
+ "step": 1200
730
+ },
731
+ {
732
+ "epoch": 0.17,
733
+ "learning_rate": 0.0004,
734
+ "loss": 0.6284,
735
+ "step": 1210
736
+ },
737
+ {
738
+ "epoch": 0.17,
739
+ "learning_rate": 0.0004,
740
+ "loss": 0.6147,
741
+ "step": 1220
742
+ },
743
+ {
744
+ "epoch": 0.17,
745
+ "learning_rate": 0.0004,
746
+ "loss": 0.6215,
747
+ "step": 1230
748
+ },
749
+ {
750
+ "epoch": 0.17,
751
+ "learning_rate": 0.0004,
752
+ "loss": 0.6045,
753
+ "step": 1240
754
+ },
755
+ {
756
+ "epoch": 0.17,
757
+ "learning_rate": 0.0004,
758
+ "loss": 0.6203,
759
+ "step": 1250
760
+ },
761
+ {
762
+ "epoch": 0.17,
763
+ "learning_rate": 0.0004,
764
+ "loss": 0.5984,
765
+ "step": 1260
766
+ },
767
+ {
768
+ "epoch": 0.17,
769
+ "learning_rate": 0.0004,
770
+ "loss": 0.5862,
771
+ "step": 1270
772
+ },
773
+ {
774
+ "epoch": 0.18,
775
+ "learning_rate": 0.0004,
776
+ "loss": 0.6168,
777
+ "step": 1280
778
+ },
779
+ {
780
+ "epoch": 0.18,
781
+ "learning_rate": 0.0004,
782
+ "loss": 0.6334,
783
+ "step": 1290
784
+ },
785
+ {
786
+ "epoch": 0.18,
787
+ "learning_rate": 0.0004,
788
+ "loss": 0.5835,
789
+ "step": 1300
790
+ },
791
+ {
792
+ "epoch": 0.18,
793
+ "learning_rate": 0.0004,
794
+ "loss": 0.6111,
795
+ "step": 1310
796
+ },
797
+ {
798
+ "epoch": 0.18,
799
+ "learning_rate": 0.0004,
800
+ "loss": 0.6041,
801
+ "step": 1320
802
+ },
803
+ {
804
+ "epoch": 0.18,
805
+ "learning_rate": 0.0004,
806
+ "loss": 0.6144,
807
+ "step": 1330
808
+ },
809
+ {
810
+ "epoch": 0.18,
811
+ "learning_rate": 0.0004,
812
+ "loss": 0.6292,
813
+ "step": 1340
814
+ },
815
+ {
816
+ "epoch": 0.19,
817
+ "learning_rate": 0.0004,
818
+ "loss": 0.5969,
819
+ "step": 1350
820
+ },
821
+ {
822
+ "epoch": 0.19,
823
+ "learning_rate": 0.0004,
824
+ "loss": 0.6309,
825
+ "step": 1360
826
+ },
827
+ {
828
+ "epoch": 0.19,
829
+ "learning_rate": 0.0004,
830
+ "loss": 0.6183,
831
+ "step": 1370
832
+ },
833
+ {
834
+ "epoch": 0.19,
835
+ "learning_rate": 0.0004,
836
+ "loss": 0.6182,
837
+ "step": 1380
838
+ },
839
+ {
840
+ "epoch": 0.19,
841
+ "learning_rate": 0.0004,
842
+ "loss": 0.6181,
843
+ "step": 1390
844
+ },
845
+ {
846
+ "epoch": 0.19,
847
+ "learning_rate": 0.0004,
848
+ "loss": 0.6138,
849
+ "step": 1400
850
+ },
851
+ {
852
+ "epoch": 0.19,
853
+ "learning_rate": 0.0004,
854
+ "loss": 0.617,
855
+ "step": 1410
856
+ },
857
+ {
858
+ "epoch": 0.2,
859
+ "learning_rate": 0.0004,
860
+ "loss": 0.6379,
861
+ "step": 1420
862
+ },
863
+ {
864
+ "epoch": 0.2,
865
+ "learning_rate": 0.0004,
866
+ "loss": 0.6138,
867
+ "step": 1430
868
+ },
869
+ {
870
+ "epoch": 0.2,
871
+ "learning_rate": 0.0004,
872
+ "loss": 0.6153,
873
+ "step": 1440
874
+ },
875
+ {
876
+ "epoch": 0.2,
877
+ "learning_rate": 0.0004,
878
+ "loss": 0.6062,
879
+ "step": 1450
880
+ },
881
+ {
882
+ "epoch": 0.2,
883
+ "learning_rate": 0.0004,
884
+ "loss": 0.5951,
885
+ "step": 1460
886
+ },
887
+ {
888
+ "epoch": 0.2,
889
+ "learning_rate": 0.0004,
890
+ "loss": 0.6155,
891
+ "step": 1470
892
+ },
893
+ {
894
+ "epoch": 0.2,
895
+ "learning_rate": 0.0004,
896
+ "loss": 0.5928,
897
+ "step": 1480
898
+ },
899
+ {
900
+ "epoch": 0.2,
901
+ "learning_rate": 0.0004,
902
+ "loss": 0.6116,
903
+ "step": 1490
904
+ },
905
+ {
906
+ "epoch": 0.21,
907
+ "learning_rate": 0.0004,
908
+ "loss": 0.5976,
909
+ "step": 1500
910
+ },
911
+ {
912
+ "epoch": 0.21,
913
+ "learning_rate": 0.0004,
914
+ "loss": 0.6131,
915
+ "step": 1510
916
+ },
917
+ {
918
+ "epoch": 0.21,
919
+ "learning_rate": 0.0004,
920
+ "loss": 0.606,
921
+ "step": 1520
922
+ },
923
+ {
924
+ "epoch": 0.21,
925
+ "learning_rate": 0.0004,
926
+ "loss": 0.6052,
927
+ "step": 1530
928
+ },
929
+ {
930
+ "epoch": 0.21,
931
+ "learning_rate": 0.0004,
932
+ "loss": 0.6266,
933
+ "step": 1540
934
+ },
935
+ {
936
+ "epoch": 0.21,
937
+ "learning_rate": 0.0004,
938
+ "loss": 0.6089,
939
+ "step": 1550
940
+ },
941
+ {
942
+ "epoch": 0.21,
943
+ "learning_rate": 0.0004,
944
+ "loss": 0.6095,
945
+ "step": 1560
946
+ },
947
+ {
948
+ "epoch": 0.22,
949
+ "learning_rate": 0.0004,
950
+ "loss": 0.6151,
951
+ "step": 1570
952
+ },
953
+ {
954
+ "epoch": 0.22,
955
+ "learning_rate": 0.0004,
956
+ "loss": 0.6155,
957
+ "step": 1580
958
+ },
959
+ {
960
+ "epoch": 0.22,
961
+ "learning_rate": 0.0004,
962
+ "loss": 0.6056,
963
+ "step": 1590
964
+ },
965
+ {
966
+ "epoch": 0.22,
967
+ "learning_rate": 0.0004,
968
+ "loss": 0.599,
969
+ "step": 1600
970
+ },
971
+ {
972
+ "epoch": 0.22,
973
+ "learning_rate": 0.0004,
974
+ "loss": 0.6191,
975
+ "step": 1610
976
+ },
977
+ {
978
+ "epoch": 0.22,
979
+ "learning_rate": 0.0004,
980
+ "loss": 0.61,
981
+ "step": 1620
982
+ },
983
+ {
984
+ "epoch": 0.22,
985
+ "learning_rate": 0.0004,
986
+ "loss": 0.5986,
987
+ "step": 1630
988
+ },
989
+ {
990
+ "epoch": 0.23,
991
+ "learning_rate": 0.0004,
992
+ "loss": 0.6285,
993
+ "step": 1640
994
+ },
995
+ {
996
+ "epoch": 0.23,
997
+ "learning_rate": 0.0004,
998
+ "loss": 0.6076,
999
+ "step": 1650
1000
+ },
1001
+ {
1002
+ "epoch": 0.23,
1003
+ "learning_rate": 0.0004,
1004
+ "loss": 0.5945,
1005
+ "step": 1660
1006
+ },
1007
+ {
1008
+ "epoch": 0.23,
1009
+ "learning_rate": 0.0004,
1010
+ "loss": 0.6044,
1011
+ "step": 1670
1012
+ },
1013
+ {
1014
+ "epoch": 0.23,
1015
+ "learning_rate": 0.0004,
1016
+ "loss": 0.6068,
1017
+ "step": 1680
1018
+ },
1019
+ {
1020
+ "epoch": 0.23,
1021
+ "learning_rate": 0.0004,
1022
+ "loss": 0.6028,
1023
+ "step": 1690
1024
+ },
1025
+ {
1026
+ "epoch": 0.23,
1027
+ "learning_rate": 0.0004,
1028
+ "loss": 0.6155,
1029
+ "step": 1700
1030
+ },
1031
+ {
1032
+ "epoch": 0.24,
1033
+ "learning_rate": 0.0004,
1034
+ "loss": 0.5815,
1035
+ "step": 1710
1036
+ },
1037
+ {
1038
+ "epoch": 0.24,
1039
+ "learning_rate": 0.0004,
1040
+ "loss": 0.6384,
1041
+ "step": 1720
1042
+ },
1043
+ {
1044
+ "epoch": 0.24,
1045
+ "learning_rate": 0.0004,
1046
+ "loss": 0.6108,
1047
+ "step": 1730
1048
+ },
1049
+ {
1050
+ "epoch": 0.24,
1051
+ "learning_rate": 0.0004,
1052
+ "loss": 0.5913,
1053
+ "step": 1740
1054
+ },
1055
+ {
1056
+ "epoch": 0.24,
1057
+ "learning_rate": 0.0004,
1058
+ "loss": 0.6011,
1059
+ "step": 1750
1060
+ },
1061
+ {
1062
+ "epoch": 0.24,
1063
+ "learning_rate": 0.0004,
1064
+ "loss": 0.6169,
1065
+ "step": 1760
1066
+ },
1067
+ {
1068
+ "epoch": 0.24,
1069
+ "learning_rate": 0.0004,
1070
+ "loss": 0.5958,
1071
+ "step": 1770
1072
+ },
1073
+ {
1074
+ "epoch": 0.24,
1075
+ "learning_rate": 0.0004,
1076
+ "loss": 0.6034,
1077
+ "step": 1780
1078
+ },
1079
+ {
1080
+ "epoch": 0.25,
1081
+ "learning_rate": 0.0004,
1082
+ "loss": 0.6142,
1083
+ "step": 1790
1084
+ },
1085
+ {
1086
+ "epoch": 0.25,
1087
+ "learning_rate": 0.0004,
1088
+ "loss": 0.587,
1089
+ "step": 1800
1090
+ },
1091
+ {
1092
+ "epoch": 0.25,
1093
+ "learning_rate": 0.0004,
1094
+ "loss": 0.5915,
1095
+ "step": 1810
1096
+ },
1097
+ {
1098
+ "epoch": 0.25,
1099
+ "learning_rate": 0.0004,
1100
+ "loss": 0.5921,
1101
+ "step": 1820
1102
+ },
1103
+ {
1104
+ "epoch": 0.25,
1105
+ "learning_rate": 0.0004,
1106
+ "loss": 0.6091,
1107
+ "step": 1830
1108
+ },
1109
+ {
1110
+ "epoch": 0.25,
1111
+ "learning_rate": 0.0004,
1112
+ "loss": 0.6042,
1113
+ "step": 1840
1114
+ },
1115
+ {
1116
+ "epoch": 0.25,
1117
+ "learning_rate": 0.0004,
1118
+ "loss": 0.5896,
1119
+ "step": 1850
1120
+ },
1121
+ {
1122
+ "epoch": 0.26,
1123
+ "learning_rate": 0.0004,
1124
+ "loss": 0.6045,
1125
+ "step": 1860
1126
+ },
1127
+ {
1128
+ "epoch": 0.26,
1129
+ "learning_rate": 0.0004,
1130
+ "loss": 0.5877,
1131
+ "step": 1870
1132
+ },
1133
+ {
1134
+ "epoch": 0.26,
1135
+ "learning_rate": 0.0004,
1136
+ "loss": 0.5885,
1137
+ "step": 1880
1138
+ },
1139
+ {
1140
+ "epoch": 0.26,
1141
+ "learning_rate": 0.0004,
1142
+ "loss": 0.6048,
1143
+ "step": 1890
1144
+ },
1145
+ {
1146
+ "epoch": 0.26,
1147
+ "learning_rate": 0.0004,
1148
+ "loss": 0.6154,
1149
+ "step": 1900
1150
+ },
1151
+ {
1152
+ "epoch": 0.26,
1153
+ "learning_rate": 0.0004,
1154
+ "loss": 0.6036,
1155
+ "step": 1910
1156
+ },
1157
+ {
1158
+ "epoch": 0.26,
1159
+ "learning_rate": 0.0004,
1160
+ "loss": 0.6164,
1161
+ "step": 1920
1162
+ },
1163
+ {
1164
+ "epoch": 0.27,
1165
+ "learning_rate": 0.0004,
1166
+ "loss": 0.5781,
1167
+ "step": 1930
1168
+ },
1169
+ {
1170
+ "epoch": 0.27,
1171
+ "learning_rate": 0.0004,
1172
+ "loss": 0.5881,
1173
+ "step": 1940
1174
+ },
1175
+ {
1176
+ "epoch": 0.27,
1177
+ "learning_rate": 0.0004,
1178
+ "loss": 0.583,
1179
+ "step": 1950
1180
+ },
1181
+ {
1182
+ "epoch": 0.27,
1183
+ "learning_rate": 0.0004,
1184
+ "loss": 0.5912,
1185
+ "step": 1960
1186
+ },
1187
+ {
1188
+ "epoch": 0.27,
1189
+ "learning_rate": 0.0004,
1190
+ "loss": 0.5967,
1191
+ "step": 1970
1192
+ },
1193
+ {
1194
+ "epoch": 0.27,
1195
+ "learning_rate": 0.0004,
1196
+ "loss": 0.5733,
1197
+ "step": 1980
1198
+ },
1199
+ {
1200
+ "epoch": 0.27,
1201
+ "learning_rate": 0.0004,
1202
+ "loss": 0.6067,
1203
+ "step": 1990
1204
+ },
1205
+ {
1206
+ "epoch": 0.28,
1207
+ "learning_rate": 0.0004,
1208
+ "loss": 0.5786,
1209
+ "step": 2000
1210
+ },
1211
+ {
1212
+ "epoch": 0.28,
1213
+ "learning_rate": 0.0004,
1214
+ "loss": 0.5948,
1215
+ "step": 2010
1216
+ },
1217
+ {
1218
+ "epoch": 0.28,
1219
+ "learning_rate": 0.0004,
1220
+ "loss": 0.5933,
1221
+ "step": 2020
1222
+ },
1223
+ {
1224
+ "epoch": 0.28,
1225
+ "learning_rate": 0.0004,
1226
+ "loss": 0.5901,
1227
+ "step": 2030
1228
+ },
1229
+ {
1230
+ "epoch": 0.28,
1231
+ "learning_rate": 0.0004,
1232
+ "loss": 0.5929,
1233
+ "step": 2040
1234
+ },
1235
+ {
1236
+ "epoch": 0.28,
1237
+ "learning_rate": 0.0004,
1238
+ "loss": 0.5845,
1239
+ "step": 2050
1240
+ },
1241
+ {
1242
+ "epoch": 0.28,
1243
+ "learning_rate": 0.0004,
1244
+ "loss": 0.5843,
1245
+ "step": 2060
1246
+ },
1247
+ {
1248
+ "epoch": 0.28,
1249
+ "learning_rate": 0.0004,
1250
+ "loss": 0.5986,
1251
+ "step": 2070
1252
+ },
1253
+ {
1254
+ "epoch": 0.29,
1255
+ "learning_rate": 0.0004,
1256
+ "loss": 0.6027,
1257
+ "step": 2080
1258
+ },
1259
+ {
1260
+ "epoch": 0.29,
1261
+ "learning_rate": 0.0004,
1262
+ "loss": 0.5922,
1263
+ "step": 2090
1264
+ },
1265
+ {
1266
+ "epoch": 0.29,
1267
+ "learning_rate": 0.0004,
1268
+ "loss": 0.5832,
1269
+ "step": 2100
1270
+ },
1271
+ {
1272
+ "epoch": 0.29,
1273
+ "learning_rate": 0.0004,
1274
+ "loss": 0.5973,
1275
+ "step": 2110
1276
+ },
1277
+ {
1278
+ "epoch": 0.29,
1279
+ "learning_rate": 0.0004,
1280
+ "loss": 0.6133,
1281
+ "step": 2120
1282
+ },
1283
+ {
1284
+ "epoch": 0.29,
1285
+ "learning_rate": 0.0004,
1286
+ "loss": 0.591,
1287
+ "step": 2130
1288
+ },
1289
+ {
1290
+ "epoch": 0.29,
1291
+ "learning_rate": 0.0004,
1292
+ "loss": 0.5949,
1293
+ "step": 2140
1294
+ },
1295
+ {
1296
+ "epoch": 0.3,
1297
+ "learning_rate": 0.0004,
1298
+ "loss": 0.5804,
1299
+ "step": 2150
1300
+ },
1301
+ {
1302
+ "epoch": 0.3,
1303
+ "learning_rate": 0.0004,
1304
+ "loss": 0.5818,
1305
+ "step": 2160
1306
+ },
1307
+ {
1308
+ "epoch": 0.3,
1309
+ "learning_rate": 0.0004,
1310
+ "loss": 0.5997,
1311
+ "step": 2170
1312
+ },
1313
+ {
1314
+ "epoch": 0.3,
1315
+ "learning_rate": 0.0004,
1316
+ "loss": 0.5953,
1317
+ "step": 2180
1318
+ },
1319
+ {
1320
+ "epoch": 0.3,
1321
+ "learning_rate": 0.0004,
1322
+ "loss": 0.5913,
1323
+ "step": 2190
1324
+ },
1325
+ {
1326
+ "epoch": 0.3,
1327
+ "learning_rate": 0.0004,
1328
+ "loss": 0.5865,
1329
+ "step": 2200
1330
+ },
1331
+ {
1332
+ "epoch": 0.3,
1333
+ "learning_rate": 0.0004,
1334
+ "loss": 0.5891,
1335
+ "step": 2210
1336
+ },
1337
+ {
1338
+ "epoch": 0.31,
1339
+ "learning_rate": 0.0004,
1340
+ "loss": 0.584,
1341
+ "step": 2220
1342
+ },
1343
+ {
1344
+ "epoch": 0.31,
1345
+ "learning_rate": 0.0004,
1346
+ "loss": 0.5916,
1347
+ "step": 2230
1348
+ },
1349
+ {
1350
+ "epoch": 0.31,
1351
+ "learning_rate": 0.0004,
1352
+ "loss": 0.5828,
1353
+ "step": 2240
1354
+ },
1355
+ {
1356
+ "epoch": 0.31,
1357
+ "learning_rate": 0.0004,
1358
+ "loss": 0.581,
1359
+ "step": 2250
1360
+ },
1361
+ {
1362
+ "epoch": 0.31,
1363
+ "learning_rate": 0.0004,
1364
+ "loss": 0.5856,
1365
+ "step": 2260
1366
+ },
1367
+ {
1368
+ "epoch": 0.31,
1369
+ "learning_rate": 0.0004,
1370
+ "loss": 0.5822,
1371
+ "step": 2270
1372
+ },
1373
+ {
1374
+ "epoch": 0.31,
1375
+ "learning_rate": 0.0004,
1376
+ "loss": 0.5786,
1377
+ "step": 2280
1378
+ },
1379
+ {
1380
+ "epoch": 0.32,
1381
+ "learning_rate": 0.0004,
1382
+ "loss": 0.5882,
1383
+ "step": 2290
1384
+ },
1385
+ {
1386
+ "epoch": 0.32,
1387
+ "learning_rate": 0.0004,
1388
+ "loss": 0.5876,
1389
+ "step": 2300
1390
+ },
1391
+ {
1392
+ "epoch": 0.32,
1393
+ "learning_rate": 0.0004,
1394
+ "loss": 0.5954,
1395
+ "step": 2310
1396
+ },
1397
+ {
1398
+ "epoch": 0.32,
1399
+ "learning_rate": 0.0004,
1400
+ "loss": 0.5752,
1401
+ "step": 2320
1402
+ },
1403
+ {
1404
+ "epoch": 0.32,
1405
+ "learning_rate": 0.0004,
1406
+ "loss": 0.5977,
1407
+ "step": 2330
1408
+ },
1409
+ {
1410
+ "epoch": 0.32,
1411
+ "learning_rate": 0.0004,
1412
+ "loss": 0.5705,
1413
+ "step": 2340
1414
+ },
1415
+ {
1416
+ "epoch": 0.32,
1417
+ "learning_rate": 0.0004,
1418
+ "loss": 0.5805,
1419
+ "step": 2350
1420
+ },
1421
+ {
1422
+ "epoch": 0.32,
1423
+ "learning_rate": 0.0004,
1424
+ "loss": 0.5896,
1425
+ "step": 2360
1426
+ },
1427
+ {
1428
+ "epoch": 0.33,
1429
+ "learning_rate": 0.0004,
1430
+ "loss": 0.5804,
1431
+ "step": 2370
1432
+ },
1433
+ {
1434
+ "epoch": 0.33,
1435
+ "learning_rate": 0.0004,
1436
+ "loss": 0.5922,
1437
+ "step": 2380
1438
+ },
1439
+ {
1440
+ "epoch": 0.33,
1441
+ "learning_rate": 0.0004,
1442
+ "loss": 0.59,
1443
+ "step": 2390
1444
+ },
1445
+ {
1446
+ "epoch": 0.33,
1447
+ "learning_rate": 0.0004,
1448
+ "loss": 0.5868,
1449
+ "step": 2400
1450
+ },
1451
+ {
1452
+ "epoch": 0.33,
1453
+ "learning_rate": 0.0004,
1454
+ "loss": 0.5954,
1455
+ "step": 2410
1456
+ },
1457
+ {
1458
+ "epoch": 0.33,
1459
+ "learning_rate": 0.0004,
1460
+ "loss": 0.5729,
1461
+ "step": 2420
1462
+ },
1463
+ {
1464
+ "epoch": 0.33,
1465
+ "learning_rate": 0.0004,
1466
+ "loss": 0.5768,
1467
+ "step": 2430
1468
+ },
1469
+ {
1470
+ "epoch": 0.34,
1471
+ "learning_rate": 0.0004,
1472
+ "loss": 0.5826,
1473
+ "step": 2440
1474
+ },
1475
+ {
1476
+ "epoch": 0.34,
1477
+ "learning_rate": 0.0004,
1478
+ "loss": 0.5836,
1479
+ "step": 2450
1480
+ },
1481
+ {
1482
+ "epoch": 0.34,
1483
+ "learning_rate": 0.0004,
1484
+ "loss": 0.575,
1485
+ "step": 2460
1486
+ },
1487
+ {
1488
+ "epoch": 0.34,
1489
+ "learning_rate": 0.0004,
1490
+ "loss": 0.5957,
1491
+ "step": 2470
1492
+ },
1493
+ {
1494
+ "epoch": 0.34,
1495
+ "learning_rate": 0.0004,
1496
+ "loss": 0.5653,
1497
+ "step": 2480
1498
+ },
1499
+ {
1500
+ "epoch": 0.34,
1501
+ "learning_rate": 0.0004,
1502
+ "loss": 0.5697,
1503
+ "step": 2490
1504
+ },
1505
+ {
1506
+ "epoch": 0.34,
1507
+ "learning_rate": 0.0004,
1508
+ "loss": 0.5876,
1509
+ "step": 2500
1510
+ },
1511
+ {
1512
+ "epoch": 0.35,
1513
+ "learning_rate": 0.0004,
1514
+ "loss": 0.5703,
1515
+ "step": 2510
1516
+ },
1517
+ {
1518
+ "epoch": 0.35,
1519
+ "learning_rate": 0.0004,
1520
+ "loss": 0.5711,
1521
+ "step": 2520
1522
+ },
1523
+ {
1524
+ "epoch": 0.35,
1525
+ "learning_rate": 0.0004,
1526
+ "loss": 0.5714,
1527
+ "step": 2530
1528
+ },
1529
+ {
1530
+ "epoch": 0.35,
1531
+ "learning_rate": 0.0004,
1532
+ "loss": 0.5933,
1533
+ "step": 2540
1534
+ },
1535
+ {
1536
+ "epoch": 0.35,
1537
+ "learning_rate": 0.0004,
1538
+ "loss": 0.5651,
1539
+ "step": 2550
1540
+ },
1541
+ {
1542
+ "epoch": 0.35,
1543
+ "learning_rate": 0.0004,
1544
+ "loss": 0.5806,
1545
+ "step": 2560
1546
+ },
1547
+ {
1548
+ "epoch": 0.35,
1549
+ "learning_rate": 0.0004,
1550
+ "loss": 0.5819,
1551
+ "step": 2570
1552
+ },
1553
+ {
1554
+ "epoch": 0.35,
1555
+ "learning_rate": 0.0004,
1556
+ "loss": 0.5818,
1557
+ "step": 2580
1558
+ },
1559
+ {
1560
+ "epoch": 0.36,
1561
+ "learning_rate": 0.0004,
1562
+ "loss": 0.5686,
1563
+ "step": 2590
1564
+ },
1565
+ {
1566
+ "epoch": 0.36,
1567
+ "learning_rate": 0.0004,
1568
+ "loss": 0.5809,
1569
+ "step": 2600
1570
+ },
1571
+ {
1572
+ "epoch": 0.36,
1573
+ "learning_rate": 0.0004,
1574
+ "loss": 0.5801,
1575
+ "step": 2610
1576
+ },
1577
+ {
1578
+ "epoch": 0.36,
1579
+ "learning_rate": 0.0004,
1580
+ "loss": 0.5729,
1581
+ "step": 2620
1582
+ },
1583
+ {
1584
+ "epoch": 0.36,
1585
+ "learning_rate": 0.0004,
1586
+ "loss": 0.5673,
1587
+ "step": 2630
1588
+ },
1589
+ {
1590
+ "epoch": 0.36,
1591
+ "learning_rate": 0.0004,
1592
+ "loss": 0.5733,
1593
+ "step": 2640
1594
+ },
1595
+ {
1596
+ "epoch": 0.36,
1597
+ "learning_rate": 0.0004,
1598
+ "loss": 0.5652,
1599
+ "step": 2650
1600
+ },
1601
+ {
1602
+ "epoch": 0.37,
1603
+ "learning_rate": 0.0004,
1604
+ "loss": 0.5834,
1605
+ "step": 2660
1606
+ },
1607
+ {
1608
+ "epoch": 0.37,
1609
+ "learning_rate": 0.0004,
1610
+ "loss": 0.5689,
1611
+ "step": 2670
1612
+ },
1613
+ {
1614
+ "epoch": 0.37,
1615
+ "learning_rate": 0.0004,
1616
+ "loss": 0.5708,
1617
+ "step": 2680
1618
+ },
1619
+ {
1620
+ "epoch": 0.37,
1621
+ "learning_rate": 0.0004,
1622
+ "loss": 0.5689,
1623
+ "step": 2690
1624
+ },
1625
+ {
1626
+ "epoch": 0.37,
1627
+ "learning_rate": 0.0004,
1628
+ "loss": 0.5807,
1629
+ "step": 2700
1630
+ },
1631
+ {
1632
+ "epoch": 0.37,
1633
+ "learning_rate": 0.0004,
1634
+ "loss": 0.5909,
1635
+ "step": 2710
1636
+ },
1637
+ {
1638
+ "epoch": 0.37,
1639
+ "learning_rate": 0.0004,
1640
+ "loss": 0.5859,
1641
+ "step": 2720
1642
+ },
1643
+ {
1644
+ "epoch": 0.38,
1645
+ "learning_rate": 0.0004,
1646
+ "loss": 0.5694,
1647
+ "step": 2730
1648
+ },
1649
+ {
1650
+ "epoch": 0.38,
1651
+ "learning_rate": 0.0004,
1652
+ "loss": 0.5855,
1653
+ "step": 2740
1654
+ },
1655
+ {
1656
+ "epoch": 0.38,
1657
+ "learning_rate": 0.0004,
1658
+ "loss": 0.5655,
1659
+ "step": 2750
1660
+ },
1661
+ {
1662
+ "epoch": 0.38,
1663
+ "learning_rate": 0.0004,
1664
+ "loss": 0.5816,
1665
+ "step": 2760
1666
+ },
1667
+ {
1668
+ "epoch": 0.38,
1669
+ "learning_rate": 0.0004,
1670
+ "loss": 0.5745,
1671
+ "step": 2770
1672
+ },
1673
+ {
1674
+ "epoch": 0.38,
1675
+ "learning_rate": 0.0004,
1676
+ "loss": 0.5803,
1677
+ "step": 2780
1678
+ },
1679
+ {
1680
+ "epoch": 0.38,
1681
+ "learning_rate": 0.0004,
1682
+ "loss": 0.5718,
1683
+ "step": 2790
1684
+ },
1685
+ {
1686
+ "epoch": 0.39,
1687
+ "learning_rate": 0.0004,
1688
+ "loss": 0.5839,
1689
+ "step": 2800
1690
+ },
1691
+ {
1692
+ "epoch": 0.39,
1693
+ "learning_rate": 0.0004,
1694
+ "loss": 0.564,
1695
+ "step": 2810
1696
+ },
1697
+ {
1698
+ "epoch": 0.39,
1699
+ "learning_rate": 0.0004,
1700
+ "loss": 0.5592,
1701
+ "step": 2820
1702
+ },
1703
+ {
1704
+ "epoch": 0.39,
1705
+ "learning_rate": 0.0004,
1706
+ "loss": 0.5561,
1707
+ "step": 2830
1708
+ },
1709
+ {
1710
+ "epoch": 0.39,
1711
+ "learning_rate": 0.0004,
1712
+ "loss": 0.5673,
1713
+ "step": 2840
1714
+ },
1715
+ {
1716
+ "epoch": 0.39,
1717
+ "learning_rate": 0.0004,
1718
+ "loss": 0.5781,
1719
+ "step": 2850
1720
+ },
1721
+ {
1722
+ "epoch": 0.39,
1723
+ "learning_rate": 0.0004,
1724
+ "loss": 0.5787,
1725
+ "step": 2860
1726
+ },
1727
+ {
1728
+ "epoch": 0.39,
1729
+ "learning_rate": 0.0004,
1730
+ "loss": 0.5649,
1731
+ "step": 2870
1732
+ },
1733
+ {
1734
+ "epoch": 0.4,
1735
+ "learning_rate": 0.0004,
1736
+ "loss": 0.5706,
1737
+ "step": 2880
1738
+ },
1739
+ {
1740
+ "epoch": 0.4,
1741
+ "learning_rate": 0.0004,
1742
+ "loss": 0.5649,
1743
+ "step": 2890
1744
+ },
1745
+ {
1746
+ "epoch": 0.4,
1747
+ "learning_rate": 0.0004,
1748
+ "loss": 0.5743,
1749
+ "step": 2900
1750
+ },
1751
+ {
1752
+ "epoch": 0.4,
1753
+ "learning_rate": 0.0004,
1754
+ "loss": 0.5665,
1755
+ "step": 2910
1756
+ },
1757
+ {
1758
+ "epoch": 0.4,
1759
+ "learning_rate": 0.0004,
1760
+ "loss": 0.5653,
1761
+ "step": 2920
1762
+ },
1763
+ {
1764
+ "epoch": 0.4,
1765
+ "learning_rate": 0.0004,
1766
+ "loss": 0.5723,
1767
+ "step": 2930
1768
+ },
1769
+ {
1770
+ "epoch": 0.4,
1771
+ "learning_rate": 0.0004,
1772
+ "loss": 0.5728,
1773
+ "step": 2940
1774
+ },
1775
+ {
1776
+ "epoch": 0.41,
1777
+ "learning_rate": 0.0004,
1778
+ "loss": 0.5627,
1779
+ "step": 2950
1780
+ },
1781
+ {
1782
+ "epoch": 0.41,
1783
+ "learning_rate": 0.0004,
1784
+ "loss": 0.5582,
1785
+ "step": 2960
1786
+ },
1787
+ {
1788
+ "epoch": 0.41,
1789
+ "learning_rate": 0.0004,
1790
+ "loss": 0.5783,
1791
+ "step": 2970
1792
+ },
1793
+ {
1794
+ "epoch": 0.41,
1795
+ "learning_rate": 0.0004,
1796
+ "loss": 0.5683,
1797
+ "step": 2980
1798
+ },
1799
+ {
1800
+ "epoch": 0.41,
1801
+ "learning_rate": 0.0004,
1802
+ "loss": 0.571,
1803
+ "step": 2990
1804
+ },
1805
+ {
1806
+ "epoch": 0.41,
1807
+ "learning_rate": 0.0004,
1808
+ "loss": 0.5644,
1809
+ "step": 3000
1810
+ },
1811
+ {
1812
+ "epoch": 0.41,
1813
+ "learning_rate": 0.0004,
1814
+ "loss": 0.5693,
1815
+ "step": 3010
1816
+ },
1817
+ {
1818
+ "epoch": 0.42,
1819
+ "learning_rate": 0.0004,
1820
+ "loss": 0.5663,
1821
+ "step": 3020
1822
+ },
1823
+ {
1824
+ "epoch": 0.42,
1825
+ "learning_rate": 0.0004,
1826
+ "loss": 0.5728,
1827
+ "step": 3030
1828
+ },
1829
+ {
1830
+ "epoch": 0.42,
1831
+ "learning_rate": 0.0004,
1832
+ "loss": 0.5683,
1833
+ "step": 3040
1834
+ },
1835
+ {
1836
+ "epoch": 0.42,
1837
+ "learning_rate": 0.0004,
1838
+ "loss": 0.5829,
1839
+ "step": 3050
1840
+ },
1841
+ {
1842
+ "epoch": 0.42,
1843
+ "learning_rate": 0.0004,
1844
+ "loss": 0.5659,
1845
+ "step": 3060
1846
+ },
1847
+ {
1848
+ "epoch": 0.42,
1849
+ "learning_rate": 0.0004,
1850
+ "loss": 0.5676,
1851
+ "step": 3070
1852
+ },
1853
+ {
1854
+ "epoch": 0.42,
1855
+ "learning_rate": 0.0004,
1856
+ "loss": 0.5776,
1857
+ "step": 3080
1858
+ },
1859
+ {
1860
+ "epoch": 0.43,
1861
+ "learning_rate": 0.0004,
1862
+ "loss": 0.5473,
1863
+ "step": 3090
1864
+ },
1865
+ {
1866
+ "epoch": 0.43,
1867
+ "learning_rate": 0.0004,
1868
+ "loss": 0.5703,
1869
+ "step": 3100
1870
+ },
1871
+ {
1872
+ "epoch": 0.43,
1873
+ "learning_rate": 0.0004,
1874
+ "loss": 0.5412,
1875
+ "step": 3110
1876
+ },
1877
+ {
1878
+ "epoch": 0.43,
1879
+ "learning_rate": 0.0004,
1880
+ "loss": 0.5584,
1881
+ "step": 3120
1882
+ },
1883
+ {
1884
+ "epoch": 0.43,
1885
+ "learning_rate": 0.0004,
1886
+ "loss": 0.5571,
1887
+ "step": 3130
1888
+ },
1889
+ {
1890
+ "epoch": 0.43,
1891
+ "learning_rate": 0.0004,
1892
+ "loss": 0.556,
1893
+ "step": 3140
1894
+ },
1895
+ {
1896
+ "epoch": 0.43,
1897
+ "learning_rate": 0.0004,
1898
+ "loss": 0.5774,
1899
+ "step": 3150
1900
+ },
1901
+ {
1902
+ "epoch": 0.43,
1903
+ "learning_rate": 0.0004,
1904
+ "loss": 0.5523,
1905
+ "step": 3160
1906
+ },
1907
+ {
1908
+ "epoch": 0.44,
1909
+ "learning_rate": 0.0004,
1910
+ "loss": 0.5712,
1911
+ "step": 3170
1912
+ },
1913
+ {
1914
+ "epoch": 0.44,
1915
+ "learning_rate": 0.0004,
1916
+ "loss": 0.5693,
1917
+ "step": 3180
1918
+ },
1919
+ {
1920
+ "epoch": 0.44,
1921
+ "learning_rate": 0.0004,
1922
+ "loss": 0.5712,
1923
+ "step": 3190
1924
+ },
1925
+ {
1926
+ "epoch": 0.44,
1927
+ "learning_rate": 0.0004,
1928
+ "loss": 0.566,
1929
+ "step": 3200
1930
+ },
1931
+ {
1932
+ "epoch": 0.44,
1933
+ "learning_rate": 0.0004,
1934
+ "loss": 0.5677,
1935
+ "step": 3210
1936
+ },
1937
+ {
1938
+ "epoch": 0.44,
1939
+ "learning_rate": 0.0004,
1940
+ "loss": 0.5625,
1941
+ "step": 3220
1942
+ },
1943
+ {
1944
+ "epoch": 0.44,
1945
+ "learning_rate": 0.0004,
1946
+ "loss": 0.5561,
1947
+ "step": 3230
1948
+ },
1949
+ {
1950
+ "epoch": 0.45,
1951
+ "learning_rate": 0.0004,
1952
+ "loss": 0.5867,
1953
+ "step": 3240
1954
+ },
1955
+ {
1956
+ "epoch": 0.45,
1957
+ "learning_rate": 0.0004,
1958
+ "loss": 0.5406,
1959
+ "step": 3250
1960
+ },
1961
+ {
1962
+ "epoch": 0.45,
1963
+ "learning_rate": 0.0004,
1964
+ "loss": 0.5673,
1965
+ "step": 3260
1966
+ },
1967
+ {
1968
+ "epoch": 0.45,
1969
+ "learning_rate": 0.0004,
1970
+ "loss": 0.5676,
1971
+ "step": 3270
1972
+ },
1973
+ {
1974
+ "epoch": 0.45,
1975
+ "learning_rate": 0.0004,
1976
+ "loss": 0.555,
1977
+ "step": 3280
1978
+ },
1979
+ {
1980
+ "epoch": 0.45,
1981
+ "learning_rate": 0.0004,
1982
+ "loss": 0.5585,
1983
+ "step": 3290
1984
+ },
1985
+ {
1986
+ "epoch": 0.45,
1987
+ "learning_rate": 0.0004,
1988
+ "loss": 0.5795,
1989
+ "step": 3300
1990
+ },
1991
+ {
1992
+ "epoch": 0.46,
1993
+ "learning_rate": 0.0004,
1994
+ "loss": 0.5663,
1995
+ "step": 3310
1996
+ },
1997
+ {
1998
+ "epoch": 0.46,
1999
+ "learning_rate": 0.0004,
2000
+ "loss": 0.5717,
2001
+ "step": 3320
2002
+ },
2003
+ {
2004
+ "epoch": 0.46,
2005
+ "learning_rate": 0.0004,
2006
+ "loss": 0.5676,
2007
+ "step": 3330
2008
+ },
2009
+ {
2010
+ "epoch": 0.46,
2011
+ "learning_rate": 0.0004,
2012
+ "loss": 0.5592,
2013
+ "step": 3340
2014
+ },
2015
+ {
2016
+ "epoch": 0.46,
2017
+ "learning_rate": 0.0004,
2018
+ "loss": 0.5566,
2019
+ "step": 3350
2020
+ },
2021
+ {
2022
+ "epoch": 0.46,
2023
+ "learning_rate": 0.0004,
2024
+ "loss": 0.5694,
2025
+ "step": 3360
2026
+ },
2027
+ {
2028
+ "epoch": 0.46,
2029
+ "learning_rate": 0.0004,
2030
+ "loss": 0.5656,
2031
+ "step": 3370
2032
+ },
2033
+ {
2034
+ "epoch": 0.47,
2035
+ "learning_rate": 0.0004,
2036
+ "loss": 0.5751,
2037
+ "step": 3380
2038
+ },
2039
+ {
2040
+ "epoch": 0.47,
2041
+ "learning_rate": 0.0004,
2042
+ "loss": 0.5439,
2043
+ "step": 3390
2044
+ },
2045
+ {
2046
+ "epoch": 0.47,
2047
+ "learning_rate": 0.0004,
2048
+ "loss": 0.5595,
2049
+ "step": 3400
2050
+ },
2051
+ {
2052
+ "epoch": 0.47,
2053
+ "learning_rate": 0.0004,
2054
+ "loss": 0.5614,
2055
+ "step": 3410
2056
+ },
2057
+ {
2058
+ "epoch": 0.47,
2059
+ "learning_rate": 0.0004,
2060
+ "loss": 0.5281,
2061
+ "step": 3420
2062
+ },
2063
+ {
2064
+ "epoch": 0.47,
2065
+ "learning_rate": 0.0004,
2066
+ "loss": 0.5468,
2067
+ "step": 3430
2068
+ },
2069
+ {
2070
+ "epoch": 0.47,
2071
+ "learning_rate": 0.0004,
2072
+ "loss": 0.529,
2073
+ "step": 3440
2074
+ },
2075
+ {
2076
+ "epoch": 0.47,
2077
+ "learning_rate": 0.0004,
2078
+ "loss": 0.5733,
2079
+ "step": 3450
2080
+ },
2081
+ {
2082
+ "epoch": 0.48,
2083
+ "learning_rate": 0.0004,
2084
+ "loss": 0.5489,
2085
+ "step": 3460
2086
+ },
2087
+ {
2088
+ "epoch": 0.48,
2089
+ "learning_rate": 0.0004,
2090
+ "loss": 0.5494,
2091
+ "step": 3470
2092
+ },
2093
+ {
2094
+ "epoch": 0.48,
2095
+ "learning_rate": 0.0004,
2096
+ "loss": 0.5452,
2097
+ "step": 3480
2098
+ },
2099
+ {
2100
+ "epoch": 0.48,
2101
+ "learning_rate": 0.0004,
2102
+ "loss": 0.5643,
2103
+ "step": 3490
2104
+ },
2105
+ {
2106
+ "epoch": 0.48,
2107
+ "learning_rate": 0.0004,
2108
+ "loss": 0.5633,
2109
+ "step": 3500
2110
+ },
2111
+ {
2112
+ "epoch": 0.48,
2113
+ "learning_rate": 0.0004,
2114
+ "loss": 0.5599,
2115
+ "step": 3510
2116
+ },
2117
+ {
2118
+ "epoch": 0.48,
2119
+ "learning_rate": 0.0004,
2120
+ "loss": 0.5598,
2121
+ "step": 3520
2122
+ },
2123
+ {
2124
+ "epoch": 0.49,
2125
+ "learning_rate": 0.0004,
2126
+ "loss": 0.5383,
2127
+ "step": 3530
2128
+ },
2129
+ {
2130
+ "epoch": 0.49,
2131
+ "learning_rate": 0.0004,
2132
+ "loss": 0.5421,
2133
+ "step": 3540
2134
+ },
2135
+ {
2136
+ "epoch": 0.49,
2137
+ "learning_rate": 0.0004,
2138
+ "loss": 0.5691,
2139
+ "step": 3550
2140
+ },
2141
+ {
2142
+ "epoch": 0.49,
2143
+ "learning_rate": 0.0004,
2144
+ "loss": 0.5675,
2145
+ "step": 3560
2146
+ },
2147
+ {
2148
+ "epoch": 0.49,
2149
+ "learning_rate": 0.0004,
2150
+ "loss": 0.5517,
2151
+ "step": 3570
2152
+ },
2153
+ {
2154
+ "epoch": 0.49,
2155
+ "learning_rate": 0.0004,
2156
+ "loss": 0.5462,
2157
+ "step": 3580
2158
+ },
2159
+ {
2160
+ "epoch": 0.49,
2161
+ "learning_rate": 0.0004,
2162
+ "loss": 0.558,
2163
+ "step": 3590
2164
+ },
2165
+ {
2166
+ "epoch": 0.5,
2167
+ "learning_rate": 0.0004,
2168
+ "loss": 0.563,
2169
+ "step": 3600
2170
+ },
2171
+ {
2172
+ "epoch": 0.5,
2173
+ "learning_rate": 0.0004,
2174
+ "loss": 0.5406,
2175
+ "step": 3610
2176
+ },
2177
+ {
2178
+ "epoch": 0.5,
2179
+ "learning_rate": 0.0004,
2180
+ "loss": 0.5541,
2181
+ "step": 3620
2182
+ },
2183
+ {
2184
+ "epoch": 0.5,
2185
+ "learning_rate": 0.0004,
2186
+ "loss": 0.5613,
2187
+ "step": 3630
2188
+ },
2189
+ {
2190
+ "epoch": 0.5,
2191
+ "learning_rate": 0.0004,
2192
+ "loss": 0.5341,
2193
+ "step": 3640
2194
+ },
2195
+ {
2196
+ "epoch": 0.5,
2197
+ "learning_rate": 0.0004,
2198
+ "loss": 0.5778,
2199
+ "step": 3650
2200
+ },
2201
+ {
2202
+ "epoch": 0.5,
2203
+ "learning_rate": 0.0004,
2204
+ "loss": 0.5505,
2205
+ "step": 3660
2206
+ },
2207
+ {
2208
+ "epoch": 0.5,
2209
+ "learning_rate": 0.0004,
2210
+ "loss": 0.5547,
2211
+ "step": 3670
2212
+ },
2213
+ {
2214
+ "epoch": 0.51,
2215
+ "learning_rate": 0.0004,
2216
+ "loss": 0.5444,
2217
+ "step": 3680
2218
+ },
2219
+ {
2220
+ "epoch": 0.51,
2221
+ "learning_rate": 0.0004,
2222
+ "loss": 0.5398,
2223
+ "step": 3690
2224
+ },
2225
+ {
2226
+ "epoch": 0.51,
2227
+ "learning_rate": 0.0004,
2228
+ "loss": 0.5362,
2229
+ "step": 3700
2230
+ },
2231
+ {
2232
+ "epoch": 0.51,
2233
+ "learning_rate": 0.0004,
2234
+ "loss": 0.5711,
2235
+ "step": 3710
2236
+ },
2237
+ {
2238
+ "epoch": 0.51,
2239
+ "learning_rate": 0.0004,
2240
+ "loss": 0.5807,
2241
+ "step": 3720
2242
+ },
2243
+ {
2244
+ "epoch": 0.51,
2245
+ "learning_rate": 0.0004,
2246
+ "loss": 0.5608,
2247
+ "step": 3730
2248
+ },
2249
+ {
2250
+ "epoch": 0.51,
2251
+ "learning_rate": 0.0004,
2252
+ "loss": 0.5509,
2253
+ "step": 3740
2254
+ },
2255
+ {
2256
+ "epoch": 0.52,
2257
+ "learning_rate": 0.0004,
2258
+ "loss": 0.5554,
2259
+ "step": 3750
2260
+ },
2261
+ {
2262
+ "epoch": 0.52,
2263
+ "learning_rate": 0.0004,
2264
+ "loss": 0.5524,
2265
+ "step": 3760
2266
+ },
2267
+ {
2268
+ "epoch": 0.52,
2269
+ "learning_rate": 0.0004,
2270
+ "loss": 0.5323,
2271
+ "step": 3770
2272
+ },
2273
+ {
2274
+ "epoch": 0.52,
2275
+ "learning_rate": 0.0004,
2276
+ "loss": 0.552,
2277
+ "step": 3780
2278
+ },
2279
+ {
2280
+ "epoch": 0.52,
2281
+ "learning_rate": 0.0004,
2282
+ "loss": 0.5616,
2283
+ "step": 3790
2284
+ },
2285
+ {
2286
+ "epoch": 0.52,
2287
+ "learning_rate": 0.0004,
2288
+ "loss": 0.5462,
2289
+ "step": 3800
2290
+ },
2291
+ {
2292
+ "epoch": 0.52,
2293
+ "learning_rate": 0.0004,
2294
+ "loss": 0.5636,
2295
+ "step": 3810
2296
+ },
2297
+ {
2298
+ "epoch": 0.53,
2299
+ "learning_rate": 0.0004,
2300
+ "loss": 0.5556,
2301
+ "step": 3820
2302
+ },
2303
+ {
2304
+ "epoch": 0.53,
2305
+ "learning_rate": 0.0004,
2306
+ "loss": 0.5631,
2307
+ "step": 3830
2308
+ },
2309
+ {
2310
+ "epoch": 0.53,
2311
+ "learning_rate": 0.0004,
2312
+ "loss": 0.5554,
2313
+ "step": 3840
2314
+ },
2315
+ {
2316
+ "epoch": 0.53,
2317
+ "learning_rate": 0.0004,
2318
+ "loss": 0.5614,
2319
+ "step": 3850
2320
+ },
2321
+ {
2322
+ "epoch": 0.53,
2323
+ "learning_rate": 0.0004,
2324
+ "loss": 0.5507,
2325
+ "step": 3860
2326
+ },
2327
+ {
2328
+ "epoch": 0.53,
2329
+ "learning_rate": 0.0004,
2330
+ "loss": 0.576,
2331
+ "step": 3870
2332
+ },
2333
+ {
2334
+ "epoch": 0.53,
2335
+ "learning_rate": 0.0004,
2336
+ "loss": 0.5495,
2337
+ "step": 3880
2338
+ },
2339
+ {
2340
+ "epoch": 0.54,
2341
+ "learning_rate": 0.0004,
2342
+ "loss": 0.5673,
2343
+ "step": 3890
2344
+ },
2345
+ {
2346
+ "epoch": 0.54,
2347
+ "learning_rate": 0.0004,
2348
+ "loss": 0.5611,
2349
+ "step": 3900
2350
+ },
2351
+ {
2352
+ "epoch": 0.54,
2353
+ "learning_rate": 0.0004,
2354
+ "loss": 0.5412,
2355
+ "step": 3910
2356
+ },
2357
+ {
2358
+ "epoch": 0.54,
2359
+ "learning_rate": 0.0004,
2360
+ "loss": 0.5356,
2361
+ "step": 3920
2362
+ },
2363
+ {
2364
+ "epoch": 0.54,
2365
+ "learning_rate": 0.0004,
2366
+ "loss": 0.568,
2367
+ "step": 3930
2368
+ },
2369
+ {
2370
+ "epoch": 0.54,
2371
+ "learning_rate": 0.0004,
2372
+ "loss": 0.5221,
2373
+ "step": 3940
2374
+ },
2375
+ {
2376
+ "epoch": 0.54,
2377
+ "learning_rate": 0.0004,
2378
+ "loss": 0.5448,
2379
+ "step": 3950
2380
+ },
2381
+ {
2382
+ "epoch": 0.54,
2383
+ "learning_rate": 0.0004,
2384
+ "loss": 0.5453,
2385
+ "step": 3960
2386
+ },
2387
+ {
2388
+ "epoch": 0.55,
2389
+ "learning_rate": 0.0004,
2390
+ "loss": 0.5405,
2391
+ "step": 3970
2392
+ },
2393
+ {
2394
+ "epoch": 0.55,
2395
+ "learning_rate": 0.0004,
2396
+ "loss": 0.5473,
2397
+ "step": 3980
2398
+ },
2399
+ {
2400
+ "epoch": 0.55,
2401
+ "learning_rate": 0.0004,
2402
+ "loss": 0.5419,
2403
+ "step": 3990
2404
+ },
2405
+ {
2406
+ "epoch": 0.55,
2407
+ "learning_rate": 0.0004,
2408
+ "loss": 0.5503,
2409
+ "step": 4000
2410
+ },
2411
+ {
2412
+ "epoch": 0.55,
2413
+ "learning_rate": 0.0004,
2414
+ "loss": 0.5486,
2415
+ "step": 4010
2416
+ },
2417
+ {
2418
+ "epoch": 0.55,
2419
+ "learning_rate": 0.0004,
2420
+ "loss": 0.5464,
2421
+ "step": 4020
2422
+ },
2423
+ {
2424
+ "epoch": 0.55,
2425
+ "learning_rate": 0.0004,
2426
+ "loss": 0.5374,
2427
+ "step": 4030
2428
+ },
2429
+ {
2430
+ "epoch": 0.56,
2431
+ "learning_rate": 0.0004,
2432
+ "loss": 0.531,
2433
+ "step": 4040
2434
+ },
2435
+ {
2436
+ "epoch": 0.56,
2437
+ "learning_rate": 0.0004,
2438
+ "loss": 0.5624,
2439
+ "step": 4050
2440
+ },
2441
+ {
2442
+ "epoch": 0.56,
2443
+ "learning_rate": 0.0004,
2444
+ "loss": 0.5268,
2445
+ "step": 4060
2446
+ },
2447
+ {
2448
+ "epoch": 0.56,
2449
+ "learning_rate": 0.0004,
2450
+ "loss": 0.5612,
2451
+ "step": 4070
2452
+ },
2453
+ {
2454
+ "epoch": 0.56,
2455
+ "learning_rate": 0.0004,
2456
+ "loss": 0.5606,
2457
+ "step": 4080
2458
+ },
2459
+ {
2460
+ "epoch": 0.56,
2461
+ "learning_rate": 0.0004,
2462
+ "loss": 0.548,
2463
+ "step": 4090
2464
+ },
2465
+ {
2466
+ "epoch": 0.56,
2467
+ "learning_rate": 0.0004,
2468
+ "loss": 0.5552,
2469
+ "step": 4100
2470
+ },
2471
+ {
2472
+ "epoch": 0.57,
2473
+ "learning_rate": 0.0004,
2474
+ "loss": 0.5403,
2475
+ "step": 4110
2476
+ },
2477
+ {
2478
+ "epoch": 0.57,
2479
+ "learning_rate": 0.0004,
2480
+ "loss": 0.5516,
2481
+ "step": 4120
2482
+ },
2483
+ {
2484
+ "epoch": 0.57,
2485
+ "learning_rate": 0.0004,
2486
+ "loss": 0.542,
2487
+ "step": 4130
2488
+ },
2489
+ {
2490
+ "epoch": 0.57,
2491
+ "learning_rate": 0.0004,
2492
+ "loss": 0.5519,
2493
+ "step": 4140
2494
+ },
2495
+ {
2496
+ "epoch": 0.57,
2497
+ "learning_rate": 0.0004,
2498
+ "loss": 0.5438,
2499
+ "step": 4150
2500
+ },
2501
+ {
2502
+ "epoch": 0.57,
2503
+ "learning_rate": 0.0004,
2504
+ "loss": 0.5379,
2505
+ "step": 4160
2506
+ },
2507
+ {
2508
+ "epoch": 0.57,
2509
+ "learning_rate": 0.0004,
2510
+ "loss": 0.5441,
2511
+ "step": 4170
2512
+ },
2513
+ {
2514
+ "epoch": 0.58,
2515
+ "learning_rate": 0.0004,
2516
+ "loss": 0.5496,
2517
+ "step": 4180
2518
+ },
2519
+ {
2520
+ "epoch": 0.58,
2521
+ "learning_rate": 0.0004,
2522
+ "loss": 0.5462,
2523
+ "step": 4190
2524
+ },
2525
+ {
2526
+ "epoch": 0.58,
2527
+ "learning_rate": 0.0004,
2528
+ "loss": 0.5498,
2529
+ "step": 4200
2530
+ },
2531
+ {
2532
+ "epoch": 0.58,
2533
+ "learning_rate": 0.0004,
2534
+ "loss": 0.5462,
2535
+ "step": 4210
2536
+ },
2537
+ {
2538
+ "epoch": 0.58,
2539
+ "learning_rate": 0.0004,
2540
+ "loss": 0.5197,
2541
+ "step": 4220
2542
+ },
2543
+ {
2544
+ "epoch": 0.58,
2545
+ "learning_rate": 0.0004,
2546
+ "loss": 0.541,
2547
+ "step": 4230
2548
+ },
2549
+ {
2550
+ "epoch": 0.58,
2551
+ "learning_rate": 0.0004,
2552
+ "loss": 0.5453,
2553
+ "step": 4240
2554
+ },
2555
+ {
2556
+ "epoch": 0.58,
2557
+ "learning_rate": 0.0004,
2558
+ "loss": 0.5305,
2559
+ "step": 4250
2560
+ },
2561
+ {
2562
+ "epoch": 0.59,
2563
+ "learning_rate": 0.0004,
2564
+ "loss": 0.5402,
2565
+ "step": 4260
2566
+ },
2567
+ {
2568
+ "epoch": 0.59,
2569
+ "learning_rate": 0.0004,
2570
+ "loss": 0.533,
2571
+ "step": 4270
2572
+ },
2573
+ {
2574
+ "epoch": 0.59,
2575
+ "learning_rate": 0.0004,
2576
+ "loss": 0.5434,
2577
+ "step": 4280
2578
+ },
2579
+ {
2580
+ "epoch": 0.59,
2581
+ "learning_rate": 0.0004,
2582
+ "loss": 0.5453,
2583
+ "step": 4290
2584
+ },
2585
+ {
2586
+ "epoch": 0.59,
2587
+ "learning_rate": 0.0004,
2588
+ "loss": 0.5623,
2589
+ "step": 4300
2590
+ },
2591
+ {
2592
+ "epoch": 0.59,
2593
+ "learning_rate": 0.0004,
2594
+ "loss": 0.5416,
2595
+ "step": 4310
2596
+ },
2597
+ {
2598
+ "epoch": 0.59,
2599
+ "learning_rate": 0.0004,
2600
+ "loss": 0.5291,
2601
+ "step": 4320
2602
+ },
2603
+ {
2604
+ "epoch": 0.6,
2605
+ "learning_rate": 0.0004,
2606
+ "loss": 0.55,
2607
+ "step": 4330
2608
+ },
2609
+ {
2610
+ "epoch": 0.6,
2611
+ "learning_rate": 0.0004,
2612
+ "loss": 0.5325,
2613
+ "step": 4340
2614
+ },
2615
+ {
2616
+ "epoch": 0.6,
2617
+ "learning_rate": 0.0004,
2618
+ "loss": 0.5287,
2619
+ "step": 4350
2620
+ },
2621
+ {
2622
+ "epoch": 0.6,
2623
+ "learning_rate": 0.0004,
2624
+ "loss": 0.5207,
2625
+ "step": 4360
2626
+ },
2627
+ {
2628
+ "epoch": 0.6,
2629
+ "learning_rate": 0.0004,
2630
+ "loss": 0.5536,
2631
+ "step": 4370
2632
+ },
2633
+ {
2634
+ "epoch": 0.6,
2635
+ "learning_rate": 0.0004,
2636
+ "loss": 0.5495,
2637
+ "step": 4380
2638
+ },
2639
+ {
2640
+ "epoch": 0.6,
2641
+ "learning_rate": 0.0004,
2642
+ "loss": 0.5408,
2643
+ "step": 4390
2644
+ },
2645
+ {
2646
+ "epoch": 0.61,
2647
+ "learning_rate": 0.0004,
2648
+ "loss": 0.5391,
2649
+ "step": 4400
2650
+ },
2651
+ {
2652
+ "epoch": 0.61,
2653
+ "learning_rate": 0.0004,
2654
+ "loss": 0.5366,
2655
+ "step": 4410
2656
+ },
2657
+ {
2658
+ "epoch": 0.61,
2659
+ "learning_rate": 0.0004,
2660
+ "loss": 0.5355,
2661
+ "step": 4420
2662
+ },
2663
+ {
2664
+ "epoch": 0.61,
2665
+ "learning_rate": 0.0004,
2666
+ "loss": 0.5398,
2667
+ "step": 4430
2668
+ },
2669
+ {
2670
+ "epoch": 0.61,
2671
+ "learning_rate": 0.0004,
2672
+ "loss": 0.5378,
2673
+ "step": 4440
2674
+ },
2675
+ {
2676
+ "epoch": 0.61,
2677
+ "learning_rate": 0.0004,
2678
+ "loss": 0.5508,
2679
+ "step": 4450
2680
+ },
2681
+ {
2682
+ "epoch": 0.61,
2683
+ "learning_rate": 0.0004,
2684
+ "loss": 0.5188,
2685
+ "step": 4460
2686
+ },
2687
+ {
2688
+ "epoch": 0.61,
2689
+ "learning_rate": 0.0004,
2690
+ "loss": 0.539,
2691
+ "step": 4470
2692
+ },
2693
+ {
2694
+ "epoch": 0.62,
2695
+ "learning_rate": 0.0004,
2696
+ "loss": 0.5391,
2697
+ "step": 4480
2698
+ },
2699
+ {
2700
+ "epoch": 0.62,
2701
+ "learning_rate": 0.0004,
2702
+ "loss": 0.5352,
2703
+ "step": 4490
2704
+ },
2705
+ {
2706
+ "epoch": 0.62,
2707
+ "learning_rate": 0.0004,
2708
+ "loss": 0.55,
2709
+ "step": 4500
2710
+ },
2711
+ {
2712
+ "epoch": 0.62,
2713
+ "learning_rate": 0.0004,
2714
+ "loss": 0.5444,
2715
+ "step": 4510
2716
+ },
2717
+ {
2718
+ "epoch": 0.62,
2719
+ "learning_rate": 0.0004,
2720
+ "loss": 0.5212,
2721
+ "step": 4520
2722
+ },
2723
+ {
2724
+ "epoch": 0.62,
2725
+ "learning_rate": 0.0004,
2726
+ "loss": 0.5293,
2727
+ "step": 4530
2728
+ },
2729
+ {
2730
+ "epoch": 0.62,
2731
+ "learning_rate": 0.0004,
2732
+ "loss": 0.544,
2733
+ "step": 4540
2734
+ },
2735
+ {
2736
+ "epoch": 0.63,
2737
+ "learning_rate": 0.0004,
2738
+ "loss": 0.5524,
2739
+ "step": 4550
2740
+ },
2741
+ {
2742
+ "epoch": 0.63,
2743
+ "learning_rate": 0.0004,
2744
+ "loss": 0.539,
2745
+ "step": 4560
2746
+ },
2747
+ {
2748
+ "epoch": 0.63,
2749
+ "learning_rate": 0.0004,
2750
+ "loss": 0.533,
2751
+ "step": 4570
2752
+ },
2753
+ {
2754
+ "epoch": 0.63,
2755
+ "learning_rate": 0.0004,
2756
+ "loss": 0.5288,
2757
+ "step": 4580
2758
+ },
2759
+ {
2760
+ "epoch": 0.63,
2761
+ "learning_rate": 0.0004,
2762
+ "loss": 0.5078,
2763
+ "step": 4590
2764
+ },
2765
+ {
2766
+ "epoch": 0.63,
2767
+ "learning_rate": 0.0004,
2768
+ "loss": 0.5397,
2769
+ "step": 4600
2770
+ },
2771
+ {
2772
+ "epoch": 0.63,
2773
+ "learning_rate": 0.0004,
2774
+ "loss": 0.541,
2775
+ "step": 4610
2776
+ },
2777
+ {
2778
+ "epoch": 0.64,
2779
+ "learning_rate": 0.0004,
2780
+ "loss": 0.5461,
2781
+ "step": 4620
2782
+ },
2783
+ {
2784
+ "epoch": 0.64,
2785
+ "learning_rate": 0.0004,
2786
+ "loss": 0.5524,
2787
+ "step": 4630
2788
+ },
2789
+ {
2790
+ "epoch": 0.64,
2791
+ "learning_rate": 0.0004,
2792
+ "loss": 0.5428,
2793
+ "step": 4640
2794
+ },
2795
+ {
2796
+ "epoch": 0.64,
2797
+ "learning_rate": 0.0004,
2798
+ "loss": 0.5352,
2799
+ "step": 4650
2800
+ },
2801
+ {
2802
+ "epoch": 0.64,
2803
+ "learning_rate": 0.0004,
2804
+ "loss": 0.5264,
2805
+ "step": 4660
2806
+ },
2807
+ {
2808
+ "epoch": 0.64,
2809
+ "learning_rate": 0.0004,
2810
+ "loss": 0.5436,
2811
+ "step": 4670
2812
+ },
2813
+ {
2814
+ "epoch": 0.64,
2815
+ "learning_rate": 0.0004,
2816
+ "loss": 0.5402,
2817
+ "step": 4680
2818
+ },
2819
+ {
2820
+ "epoch": 0.65,
2821
+ "learning_rate": 0.0004,
2822
+ "loss": 0.5353,
2823
+ "step": 4690
2824
+ },
2825
+ {
2826
+ "epoch": 0.65,
2827
+ "learning_rate": 0.0004,
2828
+ "loss": 0.5155,
2829
+ "step": 4700
2830
+ },
2831
+ {
2832
+ "epoch": 0.65,
2833
+ "learning_rate": 0.0004,
2834
+ "loss": 0.5484,
2835
+ "step": 4710
2836
+ },
2837
+ {
2838
+ "epoch": 0.65,
2839
+ "learning_rate": 0.0004,
2840
+ "loss": 0.5634,
2841
+ "step": 4720
2842
+ },
2843
+ {
2844
+ "epoch": 0.65,
2845
+ "learning_rate": 0.0004,
2846
+ "loss": 0.5414,
2847
+ "step": 4730
2848
+ },
2849
+ {
2850
+ "epoch": 0.65,
2851
+ "learning_rate": 0.0004,
2852
+ "loss": 0.5566,
2853
+ "step": 4740
2854
+ },
2855
+ {
2856
+ "epoch": 0.65,
2857
+ "learning_rate": 0.0004,
2858
+ "loss": 0.5365,
2859
+ "step": 4750
2860
+ },
2861
+ {
2862
+ "epoch": 0.65,
2863
+ "learning_rate": 0.0004,
2864
+ "loss": 0.5352,
2865
+ "step": 4760
2866
+ },
2867
+ {
2868
+ "epoch": 0.66,
2869
+ "learning_rate": 0.0004,
2870
+ "loss": 0.5372,
2871
+ "step": 4770
2872
+ },
2873
+ {
2874
+ "epoch": 0.66,
2875
+ "learning_rate": 0.0004,
2876
+ "loss": 0.5374,
2877
+ "step": 4780
2878
+ },
2879
+ {
2880
+ "epoch": 0.66,
2881
+ "learning_rate": 0.0004,
2882
+ "loss": 0.5376,
2883
+ "step": 4790
2884
+ },
2885
+ {
2886
+ "epoch": 0.66,
2887
+ "learning_rate": 0.0004,
2888
+ "loss": 0.5338,
2889
+ "step": 4800
2890
+ },
2891
+ {
2892
+ "epoch": 0.66,
2893
+ "learning_rate": 0.0004,
2894
+ "loss": 0.5446,
2895
+ "step": 4810
2896
+ },
2897
+ {
2898
+ "epoch": 0.66,
2899
+ "learning_rate": 0.0004,
2900
+ "loss": 0.5394,
2901
+ "step": 4820
2902
+ },
2903
+ {
2904
+ "epoch": 0.66,
2905
+ "learning_rate": 0.0004,
2906
+ "loss": 0.5255,
2907
+ "step": 4830
2908
+ },
2909
+ {
2910
+ "epoch": 0.67,
2911
+ "learning_rate": 0.0004,
2912
+ "loss": 0.5203,
2913
+ "step": 4840
2914
+ },
2915
+ {
2916
+ "epoch": 0.67,
2917
+ "learning_rate": 0.0004,
2918
+ "loss": 0.5401,
2919
+ "step": 4850
2920
+ },
2921
+ {
2922
+ "epoch": 0.67,
2923
+ "learning_rate": 0.0004,
2924
+ "loss": 0.539,
2925
+ "step": 4860
2926
+ },
2927
+ {
2928
+ "epoch": 0.67,
2929
+ "learning_rate": 0.0004,
2930
+ "loss": 0.5341,
2931
+ "step": 4870
2932
+ },
2933
+ {
2934
+ "epoch": 0.67,
2935
+ "learning_rate": 0.0004,
2936
+ "loss": 0.5309,
2937
+ "step": 4880
2938
+ },
2939
+ {
2940
+ "epoch": 0.67,
2941
+ "learning_rate": 0.0004,
2942
+ "loss": 0.5272,
2943
+ "step": 4890
2944
+ },
2945
+ {
2946
+ "epoch": 0.67,
2947
+ "learning_rate": 0.0004,
2948
+ "loss": 0.5432,
2949
+ "step": 4900
2950
+ },
2951
+ {
2952
+ "epoch": 0.68,
2953
+ "learning_rate": 0.0004,
2954
+ "loss": 0.5432,
2955
+ "step": 4910
2956
+ },
2957
+ {
2958
+ "epoch": 0.68,
2959
+ "learning_rate": 0.0004,
2960
+ "loss": 0.5228,
2961
+ "step": 4920
2962
+ },
2963
+ {
2964
+ "epoch": 0.68,
2965
+ "learning_rate": 0.0004,
2966
+ "loss": 0.5361,
2967
+ "step": 4930
2968
+ },
2969
+ {
2970
+ "epoch": 0.68,
2971
+ "learning_rate": 0.0004,
2972
+ "loss": 0.5275,
2973
+ "step": 4940
2974
+ },
2975
+ {
2976
+ "epoch": 0.68,
2977
+ "learning_rate": 0.0004,
2978
+ "loss": 0.5315,
2979
+ "step": 4950
2980
+ },
2981
+ {
2982
+ "epoch": 0.68,
2983
+ "learning_rate": 0.0004,
2984
+ "loss": 0.5289,
2985
+ "step": 4960
2986
+ },
2987
+ {
2988
+ "epoch": 0.68,
2989
+ "learning_rate": 0.0004,
2990
+ "loss": 0.5473,
2991
+ "step": 4970
2992
+ },
2993
+ {
2994
+ "epoch": 0.69,
2995
+ "learning_rate": 0.0004,
2996
+ "loss": 0.5315,
2997
+ "step": 4980
2998
+ },
2999
+ {
3000
+ "epoch": 0.69,
3001
+ "learning_rate": 0.0004,
3002
+ "loss": 0.5326,
3003
+ "step": 4990
3004
+ },
3005
+ {
3006
+ "epoch": 0.69,
3007
+ "learning_rate": 0.0004,
3008
+ "loss": 0.5083,
3009
+ "step": 5000
3010
+ },
3011
+ {
3012
+ "epoch": 0.69,
3013
+ "learning_rate": 0.0004,
3014
+ "loss": 0.5256,
3015
+ "step": 5010
3016
+ },
3017
+ {
3018
+ "epoch": 0.69,
3019
+ "learning_rate": 0.0004,
3020
+ "loss": 0.5264,
3021
+ "step": 5020
3022
+ },
3023
+ {
3024
+ "epoch": 0.69,
3025
+ "learning_rate": 0.0004,
3026
+ "loss": 0.5248,
3027
+ "step": 5030
3028
+ },
3029
+ {
3030
+ "epoch": 0.69,
3031
+ "learning_rate": 0.0004,
3032
+ "loss": 0.5353,
3033
+ "step": 5040
3034
+ },
3035
+ {
3036
+ "epoch": 0.69,
3037
+ "learning_rate": 0.0004,
3038
+ "loss": 0.5302,
3039
+ "step": 5050
3040
+ },
3041
+ {
3042
+ "epoch": 0.7,
3043
+ "learning_rate": 0.0004,
3044
+ "loss": 0.5262,
3045
+ "step": 5060
3046
+ },
3047
+ {
3048
+ "epoch": 0.7,
3049
+ "learning_rate": 0.0004,
3050
+ "loss": 0.5296,
3051
+ "step": 5070
3052
+ },
3053
+ {
3054
+ "epoch": 0.7,
3055
+ "learning_rate": 0.0004,
3056
+ "loss": 0.5231,
3057
+ "step": 5080
3058
+ },
3059
+ {
3060
+ "epoch": 0.7,
3061
+ "learning_rate": 0.0004,
3062
+ "loss": 0.5276,
3063
+ "step": 5090
3064
+ },
3065
+ {
3066
+ "epoch": 0.7,
3067
+ "learning_rate": 0.0004,
3068
+ "loss": 0.53,
3069
+ "step": 5100
3070
+ },
3071
+ {
3072
+ "epoch": 0.7,
3073
+ "learning_rate": 0.0004,
3074
+ "loss": 0.537,
3075
+ "step": 5110
3076
+ },
3077
+ {
3078
+ "epoch": 0.7,
3079
+ "learning_rate": 0.0004,
3080
+ "loss": 0.5207,
3081
+ "step": 5120
3082
+ },
3083
+ {
3084
+ "epoch": 0.71,
3085
+ "learning_rate": 0.0004,
3086
+ "loss": 0.5055,
3087
+ "step": 5130
3088
+ },
3089
+ {
3090
+ "epoch": 0.71,
3091
+ "learning_rate": 0.0004,
3092
+ "loss": 0.5222,
3093
+ "step": 5140
3094
+ },
3095
+ {
3096
+ "epoch": 0.71,
3097
+ "learning_rate": 0.0004,
3098
+ "loss": 0.5062,
3099
+ "step": 5150
3100
+ },
3101
+ {
3102
+ "epoch": 0.71,
3103
+ "learning_rate": 0.0004,
3104
+ "loss": 0.5285,
3105
+ "step": 5160
3106
+ },
3107
+ {
3108
+ "epoch": 0.71,
3109
+ "learning_rate": 0.0004,
3110
+ "loss": 0.5419,
3111
+ "step": 5170
3112
+ },
3113
+ {
3114
+ "epoch": 0.71,
3115
+ "learning_rate": 0.0004,
3116
+ "loss": 0.5356,
3117
+ "step": 5180
3118
+ },
3119
+ {
3120
+ "epoch": 0.71,
3121
+ "learning_rate": 0.0004,
3122
+ "loss": 0.5237,
3123
+ "step": 5190
3124
+ },
3125
+ {
3126
+ "epoch": 0.72,
3127
+ "learning_rate": 0.0004,
3128
+ "loss": 0.5516,
3129
+ "step": 5200
3130
+ },
3131
+ {
3132
+ "epoch": 0.72,
3133
+ "learning_rate": 0.0004,
3134
+ "loss": 0.5375,
3135
+ "step": 5210
3136
+ },
3137
+ {
3138
+ "epoch": 0.72,
3139
+ "learning_rate": 0.0004,
3140
+ "loss": 0.5193,
3141
+ "step": 5220
3142
+ },
3143
+ {
3144
+ "epoch": 0.72,
3145
+ "learning_rate": 0.0004,
3146
+ "loss": 0.5165,
3147
+ "step": 5230
3148
+ },
3149
+ {
3150
+ "epoch": 0.72,
3151
+ "learning_rate": 0.0004,
3152
+ "loss": 0.5263,
3153
+ "step": 5240
3154
+ },
3155
+ {
3156
+ "epoch": 0.72,
3157
+ "learning_rate": 0.0004,
3158
+ "loss": 0.5101,
3159
+ "step": 5250
3160
+ },
3161
+ {
3162
+ "epoch": 0.72,
3163
+ "learning_rate": 0.0004,
3164
+ "loss": 0.5258,
3165
+ "step": 5260
3166
+ },
3167
+ {
3168
+ "epoch": 0.73,
3169
+ "learning_rate": 0.0004,
3170
+ "loss": 0.5206,
3171
+ "step": 5270
3172
+ },
3173
+ {
3174
+ "epoch": 0.73,
3175
+ "learning_rate": 0.0004,
3176
+ "loss": 0.5226,
3177
+ "step": 5280
3178
+ },
3179
+ {
3180
+ "epoch": 0.73,
3181
+ "learning_rate": 0.0004,
3182
+ "loss": 0.5244,
3183
+ "step": 5290
3184
+ },
3185
+ {
3186
+ "epoch": 0.73,
3187
+ "learning_rate": 0.0004,
3188
+ "loss": 0.5103,
3189
+ "step": 5300
3190
+ },
3191
+ {
3192
+ "epoch": 0.73,
3193
+ "learning_rate": 0.0004,
3194
+ "loss": 0.5365,
3195
+ "step": 5310
3196
+ },
3197
+ {
3198
+ "epoch": 0.73,
3199
+ "learning_rate": 0.0004,
3200
+ "loss": 0.5152,
3201
+ "step": 5320
3202
+ },
3203
+ {
3204
+ "epoch": 0.73,
3205
+ "learning_rate": 0.0004,
3206
+ "loss": 0.5316,
3207
+ "step": 5330
3208
+ },
3209
+ {
3210
+ "epoch": 0.73,
3211
+ "learning_rate": 0.0004,
3212
+ "loss": 0.5429,
3213
+ "step": 5340
3214
+ },
3215
+ {
3216
+ "epoch": 0.74,
3217
+ "learning_rate": 0.0004,
3218
+ "loss": 0.5204,
3219
+ "step": 5350
3220
+ },
3221
+ {
3222
+ "epoch": 0.74,
3223
+ "learning_rate": 0.0004,
3224
+ "loss": 0.5362,
3225
+ "step": 5360
3226
+ },
3227
+ {
3228
+ "epoch": 0.74,
3229
+ "learning_rate": 0.0004,
3230
+ "loss": 0.5296,
3231
+ "step": 5370
3232
+ },
3233
+ {
3234
+ "epoch": 0.74,
3235
+ "learning_rate": 0.0004,
3236
+ "loss": 0.5302,
3237
+ "step": 5380
3238
+ },
3239
+ {
3240
+ "epoch": 0.74,
3241
+ "learning_rate": 0.0004,
3242
+ "loss": 0.5276,
3243
+ "step": 5390
3244
+ },
3245
+ {
3246
+ "epoch": 0.74,
3247
+ "learning_rate": 0.0004,
3248
+ "loss": 0.5247,
3249
+ "step": 5400
3250
+ },
3251
+ {
3252
+ "epoch": 0.74,
3253
+ "learning_rate": 0.0004,
3254
+ "loss": 0.5216,
3255
+ "step": 5410
3256
+ },
3257
+ {
3258
+ "epoch": 0.75,
3259
+ "learning_rate": 0.0004,
3260
+ "loss": 0.5397,
3261
+ "step": 5420
3262
+ },
3263
+ {
3264
+ "epoch": 0.75,
3265
+ "learning_rate": 0.0004,
3266
+ "loss": 0.5137,
3267
+ "step": 5430
3268
+ },
3269
+ {
3270
+ "epoch": 0.75,
3271
+ "learning_rate": 0.0004,
3272
+ "loss": 0.5194,
3273
+ "step": 5440
3274
+ },
3275
+ {
3276
+ "epoch": 0.75,
3277
+ "learning_rate": 0.0004,
3278
+ "loss": 0.5208,
3279
+ "step": 5450
3280
+ },
3281
+ {
3282
+ "epoch": 0.75,
3283
+ "learning_rate": 0.0004,
3284
+ "loss": 0.506,
3285
+ "step": 5460
3286
+ },
3287
+ {
3288
+ "epoch": 0.75,
3289
+ "learning_rate": 0.0004,
3290
+ "loss": 0.5002,
3291
+ "step": 5470
3292
+ },
3293
+ {
3294
+ "epoch": 0.75,
3295
+ "learning_rate": 0.0004,
3296
+ "loss": 0.5337,
3297
+ "step": 5480
3298
+ },
3299
+ {
3300
+ "epoch": 0.76,
3301
+ "learning_rate": 0.0004,
3302
+ "loss": 0.5241,
3303
+ "step": 5490
3304
+ },
3305
+ {
3306
+ "epoch": 0.76,
3307
+ "learning_rate": 0.0004,
3308
+ "loss": 0.5517,
3309
+ "step": 5500
3310
+ },
3311
+ {
3312
+ "epoch": 0.76,
3313
+ "learning_rate": 0.0004,
3314
+ "loss": 0.5073,
3315
+ "step": 5510
3316
+ },
3317
+ {
3318
+ "epoch": 0.76,
3319
+ "learning_rate": 0.0004,
3320
+ "loss": 0.5424,
3321
+ "step": 5520
3322
+ },
3323
+ {
3324
+ "epoch": 0.76,
3325
+ "learning_rate": 0.0004,
3326
+ "loss": 0.5254,
3327
+ "step": 5530
3328
+ },
3329
+ {
3330
+ "epoch": 0.76,
3331
+ "learning_rate": 0.0004,
3332
+ "loss": 0.5248,
3333
+ "step": 5540
3334
+ },
3335
+ {
3336
+ "epoch": 0.76,
3337
+ "learning_rate": 0.0004,
3338
+ "loss": 0.5214,
3339
+ "step": 5550
3340
+ },
3341
+ {
3342
+ "epoch": 0.76,
3343
+ "learning_rate": 0.0004,
3344
+ "loss": 0.5372,
3345
+ "step": 5560
3346
+ },
3347
+ {
3348
+ "epoch": 0.77,
3349
+ "learning_rate": 0.0004,
3350
+ "loss": 0.5317,
3351
+ "step": 5570
3352
+ },
3353
+ {
3354
+ "epoch": 0.77,
3355
+ "learning_rate": 0.0004,
3356
+ "loss": 0.5149,
3357
+ "step": 5580
3358
+ },
3359
+ {
3360
+ "epoch": 0.77,
3361
+ "learning_rate": 0.0004,
3362
+ "loss": 0.5194,
3363
+ "step": 5590
3364
+ },
3365
+ {
3366
+ "epoch": 0.77,
3367
+ "learning_rate": 0.0004,
3368
+ "loss": 0.5223,
3369
+ "step": 5600
3370
+ },
3371
+ {
3372
+ "epoch": 0.77,
3373
+ "learning_rate": 0.0004,
3374
+ "loss": 0.5353,
3375
+ "step": 5610
3376
+ },
3377
+ {
3378
+ "epoch": 0.77,
3379
+ "learning_rate": 0.0004,
3380
+ "loss": 0.5325,
3381
+ "step": 5620
3382
+ },
3383
+ {
3384
+ "epoch": 0.77,
3385
+ "learning_rate": 0.0004,
3386
+ "loss": 0.5255,
3387
+ "step": 5630
3388
+ },
3389
+ {
3390
+ "epoch": 0.78,
3391
+ "learning_rate": 0.0004,
3392
+ "loss": 0.512,
3393
+ "step": 5640
3394
+ },
3395
+ {
3396
+ "epoch": 0.78,
3397
+ "learning_rate": 0.0004,
3398
+ "loss": 0.5189,
3399
+ "step": 5650
3400
+ },
3401
+ {
3402
+ "epoch": 0.78,
3403
+ "learning_rate": 0.0004,
3404
+ "loss": 0.5184,
3405
+ "step": 5660
3406
+ },
3407
+ {
3408
+ "epoch": 0.78,
3409
+ "learning_rate": 0.0004,
3410
+ "loss": 0.5409,
3411
+ "step": 5670
3412
+ },
3413
+ {
3414
+ "epoch": 0.78,
3415
+ "learning_rate": 0.0004,
3416
+ "loss": 0.5241,
3417
+ "step": 5680
3418
+ },
3419
+ {
3420
+ "epoch": 0.78,
3421
+ "learning_rate": 0.0004,
3422
+ "loss": 0.5438,
3423
+ "step": 5690
3424
+ },
3425
+ {
3426
+ "epoch": 0.78,
3427
+ "learning_rate": 0.0004,
3428
+ "loss": 0.5361,
3429
+ "step": 5700
3430
+ },
3431
+ {
3432
+ "epoch": 0.79,
3433
+ "learning_rate": 0.0004,
3434
+ "loss": 0.5368,
3435
+ "step": 5710
3436
+ },
3437
+ {
3438
+ "epoch": 0.79,
3439
+ "learning_rate": 0.0004,
3440
+ "loss": 0.531,
3441
+ "step": 5720
3442
+ },
3443
+ {
3444
+ "epoch": 0.79,
3445
+ "learning_rate": 0.0004,
3446
+ "loss": 0.5186,
3447
+ "step": 5730
3448
+ },
3449
+ {
3450
+ "epoch": 0.79,
3451
+ "learning_rate": 0.0004,
3452
+ "loss": 0.5124,
3453
+ "step": 5740
3454
+ },
3455
+ {
3456
+ "epoch": 0.79,
3457
+ "learning_rate": 0.0004,
3458
+ "loss": 0.5053,
3459
+ "step": 5750
3460
+ },
3461
+ {
3462
+ "epoch": 0.79,
3463
+ "learning_rate": 0.0004,
3464
+ "loss": 0.5489,
3465
+ "step": 5760
3466
+ },
3467
+ {
3468
+ "epoch": 0.79,
3469
+ "learning_rate": 0.0004,
3470
+ "loss": 0.5227,
3471
+ "step": 5770
3472
+ },
3473
+ {
3474
+ "epoch": 0.8,
3475
+ "learning_rate": 0.0004,
3476
+ "loss": 0.533,
3477
+ "step": 5780
3478
+ },
3479
+ {
3480
+ "epoch": 0.8,
3481
+ "learning_rate": 0.0004,
3482
+ "loss": 0.5366,
3483
+ "step": 5790
3484
+ },
3485
+ {
3486
+ "epoch": 0.8,
3487
+ "learning_rate": 0.0004,
3488
+ "loss": 0.5202,
3489
+ "step": 5800
3490
+ },
3491
+ {
3492
+ "epoch": 0.8,
3493
+ "learning_rate": 0.0004,
3494
+ "loss": 0.5287,
3495
+ "step": 5810
3496
+ },
3497
+ {
3498
+ "epoch": 0.8,
3499
+ "learning_rate": 0.0004,
3500
+ "loss": 0.5121,
3501
+ "step": 5820
3502
+ },
3503
+ {
3504
+ "epoch": 0.8,
3505
+ "learning_rate": 0.0004,
3506
+ "loss": 0.5129,
3507
+ "step": 5830
3508
+ },
3509
+ {
3510
+ "epoch": 0.8,
3511
+ "learning_rate": 0.0004,
3512
+ "loss": 0.5341,
3513
+ "step": 5840
3514
+ },
3515
+ {
3516
+ "epoch": 0.8,
3517
+ "learning_rate": 0.0004,
3518
+ "loss": 0.5213,
3519
+ "step": 5850
3520
+ },
3521
+ {
3522
+ "epoch": 0.81,
3523
+ "learning_rate": 0.0004,
3524
+ "loss": 0.5382,
3525
+ "step": 5860
3526
+ },
3527
+ {
3528
+ "epoch": 0.81,
3529
+ "learning_rate": 0.0004,
3530
+ "loss": 0.5273,
3531
+ "step": 5870
3532
+ },
3533
+ {
3534
+ "epoch": 0.81,
3535
+ "learning_rate": 0.0004,
3536
+ "loss": 0.5123,
3537
+ "step": 5880
3538
+ },
3539
+ {
3540
+ "epoch": 0.81,
3541
+ "learning_rate": 0.0004,
3542
+ "loss": 0.516,
3543
+ "step": 5890
3544
+ },
3545
+ {
3546
+ "epoch": 0.81,
3547
+ "learning_rate": 0.0004,
3548
+ "loss": 0.5114,
3549
+ "step": 5900
3550
+ },
3551
+ {
3552
+ "epoch": 0.81,
3553
+ "learning_rate": 0.0004,
3554
+ "loss": 0.5114,
3555
+ "step": 5910
3556
+ },
3557
+ {
3558
+ "epoch": 0.81,
3559
+ "learning_rate": 0.0004,
3560
+ "loss": 0.5234,
3561
+ "step": 5920
3562
+ },
3563
+ {
3564
+ "epoch": 0.82,
3565
+ "learning_rate": 0.0004,
3566
+ "loss": 0.5261,
3567
+ "step": 5930
3568
+ },
3569
+ {
3570
+ "epoch": 0.82,
3571
+ "learning_rate": 0.0004,
3572
+ "loss": 0.4916,
3573
+ "step": 5940
3574
+ },
3575
+ {
3576
+ "epoch": 0.82,
3577
+ "learning_rate": 0.0004,
3578
+ "loss": 0.5155,
3579
+ "step": 5950
3580
+ },
3581
+ {
3582
+ "epoch": 0.82,
3583
+ "learning_rate": 0.0004,
3584
+ "loss": 0.534,
3585
+ "step": 5960
3586
+ },
3587
+ {
3588
+ "epoch": 0.82,
3589
+ "learning_rate": 0.0004,
3590
+ "loss": 0.525,
3591
+ "step": 5970
3592
+ },
3593
+ {
3594
+ "epoch": 0.82,
3595
+ "learning_rate": 0.0004,
3596
+ "loss": 0.5319,
3597
+ "step": 5980
3598
+ },
3599
+ {
3600
+ "epoch": 0.82,
3601
+ "learning_rate": 0.0004,
3602
+ "loss": 0.5112,
3603
+ "step": 5990
3604
+ },
3605
+ {
3606
+ "epoch": 0.83,
3607
+ "learning_rate": 0.0004,
3608
+ "loss": 0.5161,
3609
+ "step": 6000
3610
+ },
3611
+ {
3612
+ "epoch": 0.83,
3613
+ "learning_rate": 0.0004,
3614
+ "loss": 0.5046,
3615
+ "step": 6010
3616
+ },
3617
+ {
3618
+ "epoch": 0.83,
3619
+ "learning_rate": 0.0004,
3620
+ "loss": 0.5339,
3621
+ "step": 6020
3622
+ },
3623
+ {
3624
+ "epoch": 0.83,
3625
+ "learning_rate": 0.0004,
3626
+ "loss": 0.5197,
3627
+ "step": 6030
3628
+ },
3629
+ {
3630
+ "epoch": 0.83,
3631
+ "learning_rate": 0.0004,
3632
+ "loss": 0.5181,
3633
+ "step": 6040
3634
+ },
3635
+ {
3636
+ "epoch": 0.83,
3637
+ "learning_rate": 0.0004,
3638
+ "loss": 0.517,
3639
+ "step": 6050
3640
+ },
3641
+ {
3642
+ "epoch": 0.83,
3643
+ "learning_rate": 0.0004,
3644
+ "loss": 0.5166,
3645
+ "step": 6060
3646
+ },
3647
+ {
3648
+ "epoch": 0.84,
3649
+ "learning_rate": 0.0004,
3650
+ "loss": 0.5259,
3651
+ "step": 6070
3652
+ },
3653
+ {
3654
+ "epoch": 0.84,
3655
+ "learning_rate": 0.0004,
3656
+ "loss": 0.516,
3657
+ "step": 6080
3658
+ },
3659
+ {
3660
+ "epoch": 0.84,
3661
+ "learning_rate": 0.0004,
3662
+ "loss": 0.5026,
3663
+ "step": 6090
3664
+ },
3665
+ {
3666
+ "epoch": 0.84,
3667
+ "learning_rate": 0.0004,
3668
+ "loss": 0.5209,
3669
+ "step": 6100
3670
+ },
3671
+ {
3672
+ "epoch": 0.84,
3673
+ "learning_rate": 0.0004,
3674
+ "loss": 0.5083,
3675
+ "step": 6110
3676
+ },
3677
+ {
3678
+ "epoch": 0.84,
3679
+ "learning_rate": 0.0004,
3680
+ "loss": 0.543,
3681
+ "step": 6120
3682
+ },
3683
+ {
3684
+ "epoch": 0.84,
3685
+ "learning_rate": 0.0004,
3686
+ "loss": 0.5313,
3687
+ "step": 6130
3688
+ },
3689
+ {
3690
+ "epoch": 0.84,
3691
+ "learning_rate": 0.0004,
3692
+ "loss": 0.539,
3693
+ "step": 6140
3694
+ },
3695
+ {
3696
+ "epoch": 0.85,
3697
+ "learning_rate": 0.0004,
3698
+ "loss": 0.5165,
3699
+ "step": 6150
3700
+ },
3701
+ {
3702
+ "epoch": 0.85,
3703
+ "learning_rate": 0.0004,
3704
+ "loss": 0.5185,
3705
+ "step": 6160
3706
+ },
3707
+ {
3708
+ "epoch": 0.85,
3709
+ "learning_rate": 0.0004,
3710
+ "loss": 0.5286,
3711
+ "step": 6170
3712
+ },
3713
+ {
3714
+ "epoch": 0.85,
3715
+ "learning_rate": 0.0004,
3716
+ "loss": 0.5229,
3717
+ "step": 6180
3718
+ },
3719
+ {
3720
+ "epoch": 0.85,
3721
+ "learning_rate": 0.0004,
3722
+ "loss": 0.5128,
3723
+ "step": 6190
3724
+ },
3725
+ {
3726
+ "epoch": 0.85,
3727
+ "learning_rate": 0.0004,
3728
+ "loss": 0.5178,
3729
+ "step": 6200
3730
+ },
3731
+ {
3732
+ "epoch": 0.85,
3733
+ "learning_rate": 0.0004,
3734
+ "loss": 0.4972,
3735
+ "step": 6210
3736
+ },
3737
+ {
3738
+ "epoch": 0.86,
3739
+ "learning_rate": 0.0004,
3740
+ "loss": 0.5277,
3741
+ "step": 6220
3742
+ },
3743
+ {
3744
+ "epoch": 0.86,
3745
+ "learning_rate": 0.0004,
3746
+ "loss": 0.4977,
3747
+ "step": 6230
3748
+ },
3749
+ {
3750
+ "epoch": 0.86,
3751
+ "learning_rate": 0.0004,
3752
+ "loss": 0.5226,
3753
+ "step": 6240
3754
+ },
3755
+ {
3756
+ "epoch": 0.86,
3757
+ "learning_rate": 0.0004,
3758
+ "loss": 0.5125,
3759
+ "step": 6250
3760
+ },
3761
+ {
3762
+ "epoch": 0.86,
3763
+ "learning_rate": 0.0004,
3764
+ "loss": 0.5209,
3765
+ "step": 6260
3766
+ },
3767
+ {
3768
+ "epoch": 0.86,
3769
+ "learning_rate": 0.0004,
3770
+ "loss": 0.4912,
3771
+ "step": 6270
3772
+ },
3773
+ {
3774
+ "epoch": 0.86,
3775
+ "learning_rate": 0.0004,
3776
+ "loss": 0.4965,
3777
+ "step": 6280
3778
+ },
3779
+ {
3780
+ "epoch": 0.87,
3781
+ "learning_rate": 0.0004,
3782
+ "loss": 0.511,
3783
+ "step": 6290
3784
+ },
3785
+ {
3786
+ "epoch": 0.87,
3787
+ "learning_rate": 0.0004,
3788
+ "loss": 0.506,
3789
+ "step": 6300
3790
+ },
3791
+ {
3792
+ "epoch": 0.87,
3793
+ "learning_rate": 0.0004,
3794
+ "loss": 0.529,
3795
+ "step": 6310
3796
+ },
3797
+ {
3798
+ "epoch": 0.87,
3799
+ "learning_rate": 0.0004,
3800
+ "loss": 0.5106,
3801
+ "step": 6320
3802
+ },
3803
+ {
3804
+ "epoch": 0.87,
3805
+ "learning_rate": 0.0004,
3806
+ "loss": 0.5028,
3807
+ "step": 6330
3808
+ },
3809
+ {
3810
+ "epoch": 0.87,
3811
+ "learning_rate": 0.0004,
3812
+ "loss": 0.5159,
3813
+ "step": 6340
3814
+ },
3815
+ {
3816
+ "epoch": 0.87,
3817
+ "learning_rate": 0.0004,
3818
+ "loss": 0.5232,
3819
+ "step": 6350
3820
+ },
3821
+ {
3822
+ "epoch": 0.88,
3823
+ "learning_rate": 0.0004,
3824
+ "loss": 0.52,
3825
+ "step": 6360
3826
+ },
3827
+ {
3828
+ "epoch": 0.88,
3829
+ "learning_rate": 0.0004,
3830
+ "loss": 0.5273,
3831
+ "step": 6370
3832
+ },
3833
+ {
3834
+ "epoch": 0.88,
3835
+ "learning_rate": 0.0004,
3836
+ "loss": 0.5095,
3837
+ "step": 6380
3838
+ },
3839
+ {
3840
+ "epoch": 0.88,
3841
+ "learning_rate": 0.0004,
3842
+ "loss": 0.5151,
3843
+ "step": 6390
3844
+ },
3845
+ {
3846
+ "epoch": 0.88,
3847
+ "learning_rate": 0.0004,
3848
+ "loss": 0.5207,
3849
+ "step": 6400
3850
+ },
3851
+ {
3852
+ "epoch": 0.88,
3853
+ "learning_rate": 0.0004,
3854
+ "loss": 0.5113,
3855
+ "step": 6410
3856
+ },
3857
+ {
3858
+ "epoch": 0.88,
3859
+ "learning_rate": 0.0004,
3860
+ "loss": 0.5193,
3861
+ "step": 6420
3862
+ },
3863
+ {
3864
+ "epoch": 0.88,
3865
+ "learning_rate": 0.0004,
3866
+ "loss": 0.5304,
3867
+ "step": 6430
3868
+ },
3869
+ {
3870
+ "epoch": 0.89,
3871
+ "learning_rate": 0.0004,
3872
+ "loss": 0.5271,
3873
+ "step": 6440
3874
+ },
3875
+ {
3876
+ "epoch": 0.89,
3877
+ "learning_rate": 0.0004,
3878
+ "loss": 0.5155,
3879
+ "step": 6450
3880
+ },
3881
+ {
3882
+ "epoch": 0.89,
3883
+ "learning_rate": 0.0004,
3884
+ "loss": 0.5228,
3885
+ "step": 6460
3886
+ },
3887
+ {
3888
+ "epoch": 0.89,
3889
+ "learning_rate": 0.0004,
3890
+ "loss": 0.5087,
3891
+ "step": 6470
3892
+ },
3893
+ {
3894
+ "epoch": 0.89,
3895
+ "learning_rate": 0.0004,
3896
+ "loss": 0.5199,
3897
+ "step": 6480
3898
+ },
3899
+ {
3900
+ "epoch": 0.89,
3901
+ "learning_rate": 0.0004,
3902
+ "loss": 0.5059,
3903
+ "step": 6490
3904
+ },
3905
+ {
3906
+ "epoch": 0.89,
3907
+ "learning_rate": 0.0004,
3908
+ "loss": 0.5128,
3909
+ "step": 6500
3910
+ },
3911
+ {
3912
+ "epoch": 0.9,
3913
+ "learning_rate": 0.0004,
3914
+ "loss": 0.536,
3915
+ "step": 6510
3916
+ },
3917
+ {
3918
+ "epoch": 0.9,
3919
+ "learning_rate": 0.0004,
3920
+ "loss": 0.5008,
3921
+ "step": 6520
3922
+ },
3923
+ {
3924
+ "epoch": 0.9,
3925
+ "learning_rate": 0.0004,
3926
+ "loss": 0.5167,
3927
+ "step": 6530
3928
+ },
3929
+ {
3930
+ "epoch": 0.9,
3931
+ "learning_rate": 0.0004,
3932
+ "loss": 0.5151,
3933
+ "step": 6540
3934
+ },
3935
+ {
3936
+ "epoch": 0.9,
3937
+ "learning_rate": 0.0004,
3938
+ "loss": 0.5151,
3939
+ "step": 6550
3940
+ },
3941
+ {
3942
+ "epoch": 0.9,
3943
+ "learning_rate": 0.0004,
3944
+ "loss": 0.507,
3945
+ "step": 6560
3946
+ },
3947
+ {
3948
+ "epoch": 0.9,
3949
+ "learning_rate": 0.0004,
3950
+ "loss": 0.5274,
3951
+ "step": 6570
3952
+ },
3953
+ {
3954
+ "epoch": 0.91,
3955
+ "learning_rate": 0.0004,
3956
+ "loss": 0.5088,
3957
+ "step": 6580
3958
+ },
3959
+ {
3960
+ "epoch": 0.91,
3961
+ "learning_rate": 0.0004,
3962
+ "loss": 0.5268,
3963
+ "step": 6590
3964
+ },
3965
+ {
3966
+ "epoch": 0.91,
3967
+ "learning_rate": 0.0004,
3968
+ "loss": 0.5175,
3969
+ "step": 6600
3970
+ },
3971
+ {
3972
+ "epoch": 0.91,
3973
+ "learning_rate": 0.0004,
3974
+ "loss": 0.5205,
3975
+ "step": 6610
3976
+ },
3977
+ {
3978
+ "epoch": 0.91,
3979
+ "learning_rate": 0.0004,
3980
+ "loss": 0.5267,
3981
+ "step": 6620
3982
+ },
3983
+ {
3984
+ "epoch": 0.91,
3985
+ "learning_rate": 0.0004,
3986
+ "loss": 0.5122,
3987
+ "step": 6630
3988
+ },
3989
+ {
3990
+ "epoch": 0.91,
3991
+ "learning_rate": 0.0004,
3992
+ "loss": 0.5058,
3993
+ "step": 6640
3994
+ },
3995
+ {
3996
+ "epoch": 0.91,
3997
+ "learning_rate": 0.0004,
3998
+ "loss": 0.5131,
3999
+ "step": 6650
4000
+ },
4001
+ {
4002
+ "epoch": 0.92,
4003
+ "learning_rate": 0.0004,
4004
+ "loss": 0.5009,
4005
+ "step": 6660
4006
+ },
4007
+ {
4008
+ "epoch": 0.92,
4009
+ "learning_rate": 0.0004,
4010
+ "loss": 0.5174,
4011
+ "step": 6670
4012
+ },
4013
+ {
4014
+ "epoch": 0.92,
4015
+ "learning_rate": 0.0004,
4016
+ "loss": 0.4971,
4017
+ "step": 6680
4018
+ },
4019
+ {
4020
+ "epoch": 0.92,
4021
+ "learning_rate": 0.0004,
4022
+ "loss": 0.5045,
4023
+ "step": 6690
4024
+ },
4025
+ {
4026
+ "epoch": 0.92,
4027
+ "learning_rate": 0.0004,
4028
+ "loss": 0.5129,
4029
+ "step": 6700
4030
+ },
4031
+ {
4032
+ "epoch": 0.92,
4033
+ "learning_rate": 0.0004,
4034
+ "loss": 0.5183,
4035
+ "step": 6710
4036
+ },
4037
+ {
4038
+ "epoch": 0.92,
4039
+ "learning_rate": 0.0004,
4040
+ "loss": 0.5168,
4041
+ "step": 6720
4042
+ },
4043
+ {
4044
+ "epoch": 0.93,
4045
+ "learning_rate": 0.0004,
4046
+ "loss": 0.5219,
4047
+ "step": 6730
4048
+ },
4049
+ {
4050
+ "epoch": 0.93,
4051
+ "learning_rate": 0.0004,
4052
+ "loss": 0.4939,
4053
+ "step": 6740
4054
+ },
4055
+ {
4056
+ "epoch": 0.93,
4057
+ "learning_rate": 0.0004,
4058
+ "loss": 0.5048,
4059
+ "step": 6750
4060
+ },
4061
+ {
4062
+ "epoch": 0.93,
4063
+ "learning_rate": 0.0004,
4064
+ "loss": 0.5144,
4065
+ "step": 6760
4066
+ },
4067
+ {
4068
+ "epoch": 0.93,
4069
+ "learning_rate": 0.0004,
4070
+ "loss": 0.4923,
4071
+ "step": 6770
4072
+ },
4073
+ {
4074
+ "epoch": 0.93,
4075
+ "learning_rate": 0.0004,
4076
+ "loss": 0.5264,
4077
+ "step": 6780
4078
+ },
4079
+ {
4080
+ "epoch": 0.93,
4081
+ "learning_rate": 0.0004,
4082
+ "loss": 0.5089,
4083
+ "step": 6790
4084
+ },
4085
+ {
4086
+ "epoch": 0.94,
4087
+ "learning_rate": 0.0004,
4088
+ "loss": 0.4996,
4089
+ "step": 6800
4090
+ },
4091
+ {
4092
+ "epoch": 0.94,
4093
+ "learning_rate": 0.0004,
4094
+ "loss": 0.5014,
4095
+ "step": 6810
4096
+ },
4097
+ {
4098
+ "epoch": 0.94,
4099
+ "learning_rate": 0.0004,
4100
+ "loss": 0.5054,
4101
+ "step": 6820
4102
+ },
4103
+ {
4104
+ "epoch": 0.94,
4105
+ "learning_rate": 0.0004,
4106
+ "loss": 0.5125,
4107
+ "step": 6830
4108
+ },
4109
+ {
4110
+ "epoch": 0.94,
4111
+ "learning_rate": 0.0004,
4112
+ "loss": 0.508,
4113
+ "step": 6840
4114
+ },
4115
+ {
4116
+ "epoch": 0.94,
4117
+ "learning_rate": 0.0004,
4118
+ "loss": 0.5164,
4119
+ "step": 6850
4120
+ },
4121
+ {
4122
+ "epoch": 0.94,
4123
+ "learning_rate": 0.0004,
4124
+ "loss": 0.5086,
4125
+ "step": 6860
4126
+ },
4127
+ {
4128
+ "epoch": 0.95,
4129
+ "learning_rate": 0.0004,
4130
+ "loss": 0.5078,
4131
+ "step": 6870
4132
+ },
4133
+ {
4134
+ "epoch": 0.95,
4135
+ "learning_rate": 0.0004,
4136
+ "loss": 0.5094,
4137
+ "step": 6880
4138
+ },
4139
+ {
4140
+ "epoch": 0.95,
4141
+ "learning_rate": 0.0004,
4142
+ "loss": 0.4825,
4143
+ "step": 6890
4144
+ },
4145
+ {
4146
+ "epoch": 0.95,
4147
+ "learning_rate": 0.0004,
4148
+ "loss": 0.5068,
4149
+ "step": 6900
4150
+ },
4151
+ {
4152
+ "epoch": 0.95,
4153
+ "learning_rate": 0.0004,
4154
+ "loss": 0.524,
4155
+ "step": 6910
4156
+ },
4157
+ {
4158
+ "epoch": 0.95,
4159
+ "learning_rate": 0.0004,
4160
+ "loss": 0.5097,
4161
+ "step": 6920
4162
+ },
4163
+ {
4164
+ "epoch": 0.95,
4165
+ "learning_rate": 0.0004,
4166
+ "loss": 0.5109,
4167
+ "step": 6930
4168
+ },
4169
+ {
4170
+ "epoch": 0.95,
4171
+ "learning_rate": 0.0004,
4172
+ "loss": 0.5073,
4173
+ "step": 6940
4174
+ },
4175
+ {
4176
+ "epoch": 0.96,
4177
+ "learning_rate": 0.0004,
4178
+ "loss": 0.5154,
4179
+ "step": 6950
4180
+ },
4181
+ {
4182
+ "epoch": 0.96,
4183
+ "learning_rate": 0.0004,
4184
+ "loss": 0.5133,
4185
+ "step": 6960
4186
+ },
4187
+ {
4188
+ "epoch": 0.96,
4189
+ "learning_rate": 0.0004,
4190
+ "loss": 0.5203,
4191
+ "step": 6970
4192
+ },
4193
+ {
4194
+ "epoch": 0.96,
4195
+ "learning_rate": 0.0004,
4196
+ "loss": 0.5185,
4197
+ "step": 6980
4198
+ },
4199
+ {
4200
+ "epoch": 0.96,
4201
+ "learning_rate": 0.0004,
4202
+ "loss": 0.5115,
4203
+ "step": 6990
4204
+ },
4205
+ {
4206
+ "epoch": 0.96,
4207
+ "learning_rate": 0.0004,
4208
+ "loss": 0.5172,
4209
+ "step": 7000
4210
+ },
4211
+ {
4212
+ "epoch": 0.96,
4213
+ "learning_rate": 0.0004,
4214
+ "loss": 0.5361,
4215
+ "step": 7010
4216
+ },
4217
+ {
4218
+ "epoch": 0.97,
4219
+ "learning_rate": 0.0004,
4220
+ "loss": 0.5243,
4221
+ "step": 7020
4222
+ },
4223
+ {
4224
+ "epoch": 0.97,
4225
+ "learning_rate": 0.0004,
4226
+ "loss": 0.4933,
4227
+ "step": 7030
4228
+ },
4229
+ {
4230
+ "epoch": 0.97,
4231
+ "learning_rate": 0.0004,
4232
+ "loss": 0.5159,
4233
+ "step": 7040
4234
+ },
4235
+ {
4236
+ "epoch": 0.97,
4237
+ "learning_rate": 0.0004,
4238
+ "loss": 0.5087,
4239
+ "step": 7050
4240
+ },
4241
+ {
4242
+ "epoch": 0.97,
4243
+ "learning_rate": 0.0004,
4244
+ "loss": 0.4991,
4245
+ "step": 7060
4246
+ },
4247
+ {
4248
+ "epoch": 0.97,
4249
+ "learning_rate": 0.0004,
4250
+ "loss": 0.5025,
4251
+ "step": 7070
4252
+ },
4253
+ {
4254
+ "epoch": 0.97,
4255
+ "learning_rate": 0.0004,
4256
+ "loss": 0.5142,
4257
+ "step": 7080
4258
+ },
4259
+ {
4260
+ "epoch": 0.98,
4261
+ "learning_rate": 0.0004,
4262
+ "loss": 0.4961,
4263
+ "step": 7090
4264
+ },
4265
+ {
4266
+ "epoch": 0.98,
4267
+ "learning_rate": 0.0004,
4268
+ "loss": 0.5148,
4269
+ "step": 7100
4270
+ },
4271
+ {
4272
+ "epoch": 0.98,
4273
+ "learning_rate": 0.0004,
4274
+ "loss": 0.4982,
4275
+ "step": 7110
4276
+ },
4277
+ {
4278
+ "epoch": 0.98,
4279
+ "learning_rate": 0.0004,
4280
+ "loss": 0.5185,
4281
+ "step": 7120
4282
+ },
4283
+ {
4284
+ "epoch": 0.98,
4285
+ "learning_rate": 0.0004,
4286
+ "loss": 0.5072,
4287
+ "step": 7130
4288
+ },
4289
+ {
4290
+ "epoch": 0.98,
4291
+ "learning_rate": 0.0004,
4292
+ "loss": 0.4863,
4293
+ "step": 7140
4294
+ },
4295
+ {
4296
+ "epoch": 0.98,
4297
+ "learning_rate": 0.0004,
4298
+ "loss": 0.5181,
4299
+ "step": 7150
4300
+ },
4301
+ {
4302
+ "epoch": 0.99,
4303
+ "learning_rate": 0.0004,
4304
+ "loss": 0.5071,
4305
+ "step": 7160
4306
+ },
4307
+ {
4308
+ "epoch": 0.99,
4309
+ "learning_rate": 0.0004,
4310
+ "loss": 0.4807,
4311
+ "step": 7170
4312
+ },
4313
+ {
4314
+ "epoch": 0.99,
4315
+ "learning_rate": 0.0004,
4316
+ "loss": 0.5052,
4317
+ "step": 7180
4318
+ },
4319
+ {
4320
+ "epoch": 0.99,
4321
+ "learning_rate": 0.0004,
4322
+ "loss": 0.5187,
4323
+ "step": 7190
4324
+ },
4325
+ {
4326
+ "epoch": 0.99,
4327
+ "learning_rate": 0.0004,
4328
+ "loss": 0.5243,
4329
+ "step": 7200
4330
+ },
4331
+ {
4332
+ "epoch": 0.99,
4333
+ "learning_rate": 0.0004,
4334
+ "loss": 0.5058,
4335
+ "step": 7210
4336
+ },
4337
+ {
4338
+ "epoch": 0.99,
4339
+ "learning_rate": 0.0004,
4340
+ "loss": 0.4977,
4341
+ "step": 7220
4342
+ },
4343
+ {
4344
+ "epoch": 0.99,
4345
+ "learning_rate": 0.0004,
4346
+ "loss": 0.5155,
4347
+ "step": 7230
4348
+ },
4349
+ {
4350
+ "epoch": 1.0,
4351
+ "learning_rate": 0.0004,
4352
+ "loss": 0.5198,
4353
+ "step": 7240
4354
+ },
4355
+ {
4356
+ "epoch": 1.0,
4357
+ "learning_rate": 0.0004,
4358
+ "loss": 0.5313,
4359
+ "step": 7250
4360
+ },
4361
+ {
4362
+ "epoch": 1.0,
4363
+ "learning_rate": 0.0004,
4364
+ "loss": 0.5083,
4365
+ "step": 7260
4366
+ },
4367
+ {
4368
+ "epoch": 1.0,
4369
+ "learning_rate": 0.0004,
4370
+ "loss": 0.5122,
4371
+ "step": 7270
4372
+ },
4373
+ {
4374
+ "epoch": 1.0,
4375
+ "learning_rate": 0.0004,
4376
+ "loss": 0.5152,
4377
+ "step": 7280
4378
+ },
4379
+ {
4380
+ "epoch": 1.0,
4381
+ "learning_rate": 0.0004,
4382
+ "loss": 0.522,
4383
+ "step": 7290
4384
+ },
4385
+ {
4386
+ "epoch": 1.0,
4387
+ "learning_rate": 0.0004,
4388
+ "loss": 0.5113,
4389
+ "step": 7300
4390
+ },
4391
+ {
4392
+ "epoch": 1.01,
4393
+ "learning_rate": 0.0004,
4394
+ "loss": 0.5033,
4395
+ "step": 7310
4396
+ },
4397
+ {
4398
+ "epoch": 1.01,
4399
+ "learning_rate": 0.0004,
4400
+ "loss": 0.5097,
4401
+ "step": 7320
4402
+ },
4403
+ {
4404
+ "epoch": 1.01,
4405
+ "learning_rate": 0.0004,
4406
+ "loss": 0.5233,
4407
+ "step": 7330
4408
+ },
4409
+ {
4410
+ "epoch": 1.01,
4411
+ "learning_rate": 0.0004,
4412
+ "loss": 0.5039,
4413
+ "step": 7340
4414
+ },
4415
+ {
4416
+ "epoch": 1.01,
4417
+ "learning_rate": 0.0004,
4418
+ "loss": 0.5091,
4419
+ "step": 7350
4420
+ },
4421
+ {
4422
+ "epoch": 1.01,
4423
+ "learning_rate": 0.0004,
4424
+ "loss": 0.5215,
4425
+ "step": 7360
4426
+ },
4427
+ {
4428
+ "epoch": 1.01,
4429
+ "learning_rate": 0.0004,
4430
+ "loss": 0.4979,
4431
+ "step": 7370
4432
+ },
4433
+ {
4434
+ "epoch": 1.02,
4435
+ "learning_rate": 0.0004,
4436
+ "loss": 0.4949,
4437
+ "step": 7380
4438
+ },
4439
+ {
4440
+ "epoch": 1.02,
4441
+ "learning_rate": 0.0004,
4442
+ "loss": 0.5163,
4443
+ "step": 7390
4444
+ },
4445
+ {
4446
+ "epoch": 1.02,
4447
+ "learning_rate": 0.0004,
4448
+ "loss": 0.5033,
4449
+ "step": 7400
4450
+ },
4451
+ {
4452
+ "epoch": 1.02,
4453
+ "learning_rate": 0.0004,
4454
+ "loss": 0.5141,
4455
+ "step": 7410
4456
+ },
4457
+ {
4458
+ "epoch": 1.02,
4459
+ "learning_rate": 0.0004,
4460
+ "loss": 0.4948,
4461
+ "step": 7420
4462
+ },
4463
+ {
4464
+ "epoch": 1.02,
4465
+ "learning_rate": 0.0004,
4466
+ "loss": 0.4724,
4467
+ "step": 7430
4468
+ },
4469
+ {
4470
+ "epoch": 1.02,
4471
+ "learning_rate": 0.0004,
4472
+ "loss": 0.4959,
4473
+ "step": 7440
4474
+ },
4475
+ {
4476
+ "epoch": 1.02,
4477
+ "learning_rate": 0.0004,
4478
+ "loss": 0.4982,
4479
+ "step": 7450
4480
+ },
4481
+ {
4482
+ "epoch": 1.03,
4483
+ "learning_rate": 0.0004,
4484
+ "loss": 0.5074,
4485
+ "step": 7460
4486
+ },
4487
+ {
4488
+ "epoch": 1.03,
4489
+ "learning_rate": 0.0004,
4490
+ "loss": 0.5181,
4491
+ "step": 7470
4492
+ },
4493
+ {
4494
+ "epoch": 1.03,
4495
+ "learning_rate": 0.0004,
4496
+ "loss": 0.479,
4497
+ "step": 7480
4498
+ },
4499
+ {
4500
+ "epoch": 1.03,
4501
+ "learning_rate": 0.0004,
4502
+ "loss": 0.51,
4503
+ "step": 7490
4504
+ },
4505
+ {
4506
+ "epoch": 1.03,
4507
+ "learning_rate": 0.0004,
4508
+ "loss": 0.4975,
4509
+ "step": 7500
4510
+ },
4511
+ {
4512
+ "epoch": 1.03,
4513
+ "learning_rate": 0.0004,
4514
+ "loss": 0.5088,
4515
+ "step": 7510
4516
+ },
4517
+ {
4518
+ "epoch": 1.03,
4519
+ "learning_rate": 0.0004,
4520
+ "loss": 0.5214,
4521
+ "step": 7520
4522
+ },
4523
+ {
4524
+ "epoch": 1.04,
4525
+ "learning_rate": 0.0004,
4526
+ "loss": 0.5131,
4527
+ "step": 7530
4528
+ },
4529
+ {
4530
+ "epoch": 1.04,
4531
+ "learning_rate": 0.0004,
4532
+ "loss": 0.483,
4533
+ "step": 7540
4534
+ },
4535
+ {
4536
+ "epoch": 1.04,
4537
+ "learning_rate": 0.0004,
4538
+ "loss": 0.5104,
4539
+ "step": 7550
4540
+ },
4541
+ {
4542
+ "epoch": 1.04,
4543
+ "learning_rate": 0.0004,
4544
+ "loss": 0.4812,
4545
+ "step": 7560
4546
+ },
4547
+ {
4548
+ "epoch": 1.04,
4549
+ "learning_rate": 0.0004,
4550
+ "loss": 0.5016,
4551
+ "step": 7570
4552
+ },
4553
+ {
4554
+ "epoch": 1.04,
4555
+ "learning_rate": 0.0004,
4556
+ "loss": 0.4848,
4557
+ "step": 7580
4558
+ },
4559
+ {
4560
+ "epoch": 1.04,
4561
+ "learning_rate": 0.0004,
4562
+ "loss": 0.5095,
4563
+ "step": 7590
4564
+ },
4565
+ {
4566
+ "epoch": 1.05,
4567
+ "learning_rate": 0.0004,
4568
+ "loss": 0.5105,
4569
+ "step": 7600
4570
+ },
4571
+ {
4572
+ "epoch": 1.05,
4573
+ "learning_rate": 0.0004,
4574
+ "loss": 0.503,
4575
+ "step": 7610
4576
+ },
4577
+ {
4578
+ "epoch": 1.05,
4579
+ "learning_rate": 0.0004,
4580
+ "loss": 0.4859,
4581
+ "step": 7620
4582
+ },
4583
+ {
4584
+ "epoch": 1.05,
4585
+ "learning_rate": 0.0004,
4586
+ "loss": 0.5161,
4587
+ "step": 7630
4588
+ },
4589
+ {
4590
+ "epoch": 1.05,
4591
+ "learning_rate": 0.0004,
4592
+ "loss": 0.5081,
4593
+ "step": 7640
4594
+ },
4595
+ {
4596
+ "epoch": 1.05,
4597
+ "learning_rate": 0.0004,
4598
+ "loss": 0.5174,
4599
+ "step": 7650
4600
+ },
4601
+ {
4602
+ "epoch": 1.05,
4603
+ "learning_rate": 0.0004,
4604
+ "loss": 0.5133,
4605
+ "step": 7660
4606
+ },
4607
+ {
4608
+ "epoch": 1.06,
4609
+ "learning_rate": 0.0004,
4610
+ "loss": 0.5143,
4611
+ "step": 7670
4612
+ },
4613
+ {
4614
+ "epoch": 1.06,
4615
+ "learning_rate": 0.0004,
4616
+ "loss": 0.4895,
4617
+ "step": 7680
4618
+ },
4619
+ {
4620
+ "epoch": 1.06,
4621
+ "learning_rate": 0.0004,
4622
+ "loss": 0.5093,
4623
+ "step": 7690
4624
+ },
4625
+ {
4626
+ "epoch": 1.06,
4627
+ "learning_rate": 0.0004,
4628
+ "loss": 0.5011,
4629
+ "step": 7700
4630
+ },
4631
+ {
4632
+ "epoch": 1.06,
4633
+ "learning_rate": 0.0004,
4634
+ "loss": 0.4987,
4635
+ "step": 7710
4636
+ },
4637
+ {
4638
+ "epoch": 1.06,
4639
+ "learning_rate": 0.0004,
4640
+ "loss": 0.5093,
4641
+ "step": 7720
4642
+ },
4643
+ {
4644
+ "epoch": 1.06,
4645
+ "learning_rate": 0.0004,
4646
+ "loss": 0.4935,
4647
+ "step": 7730
4648
+ },
4649
+ {
4650
+ "epoch": 1.06,
4651
+ "learning_rate": 0.0004,
4652
+ "loss": 0.5175,
4653
+ "step": 7740
4654
+ },
4655
+ {
4656
+ "epoch": 1.07,
4657
+ "learning_rate": 0.0004,
4658
+ "loss": 0.5026,
4659
+ "step": 7750
4660
+ },
4661
+ {
4662
+ "epoch": 1.07,
4663
+ "learning_rate": 0.0004,
4664
+ "loss": 0.4963,
4665
+ "step": 7760
4666
+ },
4667
+ {
4668
+ "epoch": 1.07,
4669
+ "learning_rate": 0.0004,
4670
+ "loss": 0.4784,
4671
+ "step": 7770
4672
+ },
4673
+ {
4674
+ "epoch": 1.07,
4675
+ "learning_rate": 0.0004,
4676
+ "loss": 0.488,
4677
+ "step": 7780
4678
+ },
4679
+ {
4680
+ "epoch": 1.07,
4681
+ "learning_rate": 0.0004,
4682
+ "loss": 0.4973,
4683
+ "step": 7790
4684
+ },
4685
+ {
4686
+ "epoch": 1.07,
4687
+ "learning_rate": 0.0004,
4688
+ "loss": 0.4876,
4689
+ "step": 7800
4690
+ },
4691
+ {
4692
+ "epoch": 1.07,
4693
+ "learning_rate": 0.0004,
4694
+ "loss": 0.5124,
4695
+ "step": 7810
4696
+ },
4697
+ {
4698
+ "epoch": 1.08,
4699
+ "learning_rate": 0.0004,
4700
+ "loss": 0.5014,
4701
+ "step": 7820
4702
+ },
4703
+ {
4704
+ "epoch": 1.08,
4705
+ "learning_rate": 0.0004,
4706
+ "loss": 0.4924,
4707
+ "step": 7830
4708
+ },
4709
+ {
4710
+ "epoch": 1.08,
4711
+ "learning_rate": 0.0004,
4712
+ "loss": 0.5041,
4713
+ "step": 7840
4714
+ },
4715
+ {
4716
+ "epoch": 1.08,
4717
+ "learning_rate": 0.0004,
4718
+ "loss": 0.5101,
4719
+ "step": 7850
4720
+ },
4721
+ {
4722
+ "epoch": 1.08,
4723
+ "learning_rate": 0.0004,
4724
+ "loss": 0.5236,
4725
+ "step": 7860
4726
+ },
4727
+ {
4728
+ "epoch": 1.08,
4729
+ "learning_rate": 0.0004,
4730
+ "loss": 0.5086,
4731
+ "step": 7870
4732
+ },
4733
+ {
4734
+ "epoch": 1.08,
4735
+ "learning_rate": 0.0004,
4736
+ "loss": 0.5032,
4737
+ "step": 7880
4738
+ },
4739
+ {
4740
+ "epoch": 1.09,
4741
+ "learning_rate": 0.0004,
4742
+ "loss": 0.5093,
4743
+ "step": 7890
4744
+ },
4745
+ {
4746
+ "epoch": 1.09,
4747
+ "learning_rate": 0.0004,
4748
+ "loss": 0.5155,
4749
+ "step": 7900
4750
+ },
4751
+ {
4752
+ "epoch": 1.09,
4753
+ "learning_rate": 0.0004,
4754
+ "loss": 0.5072,
4755
+ "step": 7910
4756
+ },
4757
+ {
4758
+ "epoch": 1.09,
4759
+ "learning_rate": 0.0004,
4760
+ "loss": 0.5199,
4761
+ "step": 7920
4762
+ },
4763
+ {
4764
+ "epoch": 1.09,
4765
+ "learning_rate": 0.0004,
4766
+ "loss": 0.4829,
4767
+ "step": 7930
4768
+ },
4769
+ {
4770
+ "epoch": 1.09,
4771
+ "learning_rate": 0.0004,
4772
+ "loss": 0.5146,
4773
+ "step": 7940
4774
+ },
4775
+ {
4776
+ "epoch": 1.09,
4777
+ "learning_rate": 0.0004,
4778
+ "loss": 0.504,
4779
+ "step": 7950
4780
+ },
4781
+ {
4782
+ "epoch": 1.1,
4783
+ "learning_rate": 0.0004,
4784
+ "loss": 0.5176,
4785
+ "step": 7960
4786
+ },
4787
+ {
4788
+ "epoch": 1.1,
4789
+ "learning_rate": 0.0004,
4790
+ "loss": 0.4947,
4791
+ "step": 7970
4792
+ },
4793
+ {
4794
+ "epoch": 1.1,
4795
+ "learning_rate": 0.0004,
4796
+ "loss": 0.5089,
4797
+ "step": 7980
4798
+ },
4799
+ {
4800
+ "epoch": 1.1,
4801
+ "learning_rate": 0.0004,
4802
+ "loss": 0.5092,
4803
+ "step": 7990
4804
+ },
4805
+ {
4806
+ "epoch": 1.1,
4807
+ "learning_rate": 0.0004,
4808
+ "loss": 0.4978,
4809
+ "step": 8000
4810
+ }
4811
+ ],
4812
+ "logging_steps": 10,
4813
+ "max_steps": 8000,
4814
+ "num_train_epochs": 2,
4815
+ "save_steps": 500,
4816
+ "total_flos": 2.79870537990144e+18,
4817
+ "trial_name": null,
4818
+ "trial_params": null
4819
+ }
checkpoint-8000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:58e089f4f6c22d5aa543c563a53411c24ff19725869b66d35dca24302d3df43d
3
+ size 4347