Hugging Face's logo

nhat117
/
dica-llama2-13b-chat-hf

PEFT
Model card Files Community
41
nhat117 commited on
Commit
86ae2d2
•
1 Parent(s): b3f197d

Upload folder using huggingface_hub

Browse files
Files changed (15) hide show
  1. README.md +20 -0
  2. adapter_config.json +26 -0
  3. adapter_model.bin +3 -0
  4. adapter_model/README.md +20 -0
  5. adapter_model/adapter_config.json +26 -0
  6. adapter_model/adapter_model.bin +3 -0
  7. added_tokens.json +3 -0
  8. optimizer.pt +3 -0
  9. rng_state.pth +3 -0
  10. scheduler.pt +3 -0
  11. special_tokens_map.json +6 -0
  12. tokenizer.model +3 -0
  13. tokenizer_config.json +35 -0
  14. trainer_state.json +3316 -0
  15. training_args.bin +3 -0
README.md ADDED
@@ -0,0 +1,20 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: peft
3
+ ---
4
+ ## Training procedure
5
+
6
+
7
+ The following `bitsandbytes` quantization config was used during training:
8
+ - load_in_8bit: False
9
+ - load_in_4bit: True
10
+ - llm_int8_threshold: 6.0
11
+ - llm_int8_skip_modules: None
12
+ - llm_int8_enable_fp32_cpu_offload: False
13
+ - llm_int8_has_fp16_weight: False
14
+ - bnb_4bit_quant_type: nf4
15
+ - bnb_4bit_use_double_quant: True
16
+ - bnb_4bit_compute_dtype: float32
17
+ ### Framework versions
18
+
19
+
20
+ - PEFT 0.4.0
adapter_config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "auto_mapping": null,
3
+ "base_model_name_or_path": "meta-llama/Llama-2-13b-chat-hf",
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": 0.0,
12
+ "modules_to_save": null,
13
+ "peft_type": "LORA",
14
+ "r": 64,
15
+ "revision": null,
16
+ "target_modules": [
17
+ "v_proj",
18
+ "gate_proj",
19
+ "up_proj",
20
+ "o_proj",
21
+ "q_proj",
22
+ "k_proj",
23
+ "down_proj"
24
+ ],
25
+ "task_type": "CAUSAL_LM"
26
+ }
adapter_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:da3bd81a894218a7869dced13ea6fb4d67d06e6b7d1a24706659d455ee2c112a
3
+ size 1001592586
adapter_model/README.md ADDED
@@ -0,0 +1,20 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: peft
3
+ ---
4
+ ## Training procedure
5
+
6
+
7
+ The following `bitsandbytes` quantization config was used during training:
8
+ - load_in_8bit: False
9
+ - load_in_4bit: True
10
+ - llm_int8_threshold: 6.0
11
+ - llm_int8_skip_modules: None
12
+ - llm_int8_enable_fp32_cpu_offload: False
13
+ - llm_int8_has_fp16_weight: False
14
+ - bnb_4bit_quant_type: nf4
15
+ - bnb_4bit_use_double_quant: True
16
+ - bnb_4bit_compute_dtype: float32
17
+ ### Framework versions
18
+
19
+
20
+ - PEFT 0.4.0
adapter_model/adapter_config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "auto_mapping": null,
3
+ "base_model_name_or_path": "meta-llama/Llama-2-13b-chat-hf",
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": 0.0,
12
+ "modules_to_save": null,
13
+ "peft_type": "LORA",
14
+ "r": 64,
15
+ "revision": null,
16
+ "target_modules": [
17
+ "v_proj",
18
+ "gate_proj",
19
+ "up_proj",
20
+ "o_proj",
21
+ "q_proj",
22
+ "k_proj",
23
+ "down_proj"
24
+ ],
25
+ "task_type": "CAUSAL_LM"
26
+ }
adapter_model/adapter_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:da3bd81a894218a7869dced13ea6fb4d67d06e6b7d1a24706659d455ee2c112a
3
+ size 1001592586
added_tokens.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ {
2
+ "[PAD]": 32000
3
+ }
optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bdaa8965fead55ac2bbe4d91929a581bb37348c6bbb381385792e0a7a5d16bac
3
+ size 2003127538
rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1ff264f99d31b522cc7e2a4eac9d38606d0c58a34c0adc74d71e0ca8b371dc36
3
+ size 14244
scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e997e0d6d40a2b0f2de1526ef37099f7ade8dbd5e6bc24d505316c3435bbf100
3
+ size 1064
special_tokens_map.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "eos_token": "</s>",
4
+ "pad_token": "[PAD]",
5
+ "unk_token": "<unk>"
6
+ }
tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
3
+ size 499723
tokenizer_config.json ADDED
@@ -0,0 +1,35 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": true,
3
+ "add_eos_token": false,
4
+ "bos_token": {
5
+ "__type": "AddedToken",
6
+ "content": "<s>",
7
+ "lstrip": false,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false
11
+ },
12
+ "clean_up_tokenization_spaces": false,
13
+ "eos_token": {
14
+ "__type": "AddedToken",
15
+ "content": "</s>",
16
+ "lstrip": false,
17
+ "normalized": false,
18
+ "rstrip": false,
19
+ "single_word": false
20
+ },
21
+ "legacy": false,
22
+ "model_max_length": 1000000000000000019884624838656,
23
+ "pad_token": null,
24
+ "padding_side": "right",
25
+ "sp_model_kwargs": {},
26
+ "tokenizer_class": "LlamaTokenizer",
27
+ "unk_token": {
28
+ "__type": "AddedToken",
29
+ "content": "<unk>",
30
+ "lstrip": false,
31
+ "normalized": false,
32
+ "rstrip": false,
33
+ "single_word": false
34
+ }
35
+ }
trainer_state.json ADDED
@@ -0,0 +1,3316 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 0.5298332229514119,
5
+ "global_step": 5500,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.0,
12
+ "learning_rate": 0.0002,
13
+ "loss": 1.1938,
14
+ "step": 10
15
+ },
16
+ {
17
+ "epoch": 0.0,
18
+ "learning_rate": 0.0002,
19
+ "loss": 0.9654,
20
+ "step": 20
21
+ },
22
+ {
23
+ "epoch": 0.0,
24
+ "learning_rate": 0.0002,
25
+ "loss": 0.8956,
26
+ "step": 30
27
+ },
28
+ {
29
+ "epoch": 0.0,
30
+ "learning_rate": 0.0002,
31
+ "loss": 0.8471,
32
+ "step": 40
33
+ },
34
+ {
35
+ "epoch": 0.0,
36
+ "learning_rate": 0.0002,
37
+ "loss": 0.8339,
38
+ "step": 50
39
+ },
40
+ {
41
+ "epoch": 0.01,
42
+ "learning_rate": 0.0002,
43
+ "loss": 0.8938,
44
+ "step": 60
45
+ },
46
+ {
47
+ "epoch": 0.01,
48
+ "learning_rate": 0.0002,
49
+ "loss": 0.8326,
50
+ "step": 70
51
+ },
52
+ {
53
+ "epoch": 0.01,
54
+ "learning_rate": 0.0002,
55
+ "loss": 0.8485,
56
+ "step": 80
57
+ },
58
+ {
59
+ "epoch": 0.01,
60
+ "learning_rate": 0.0002,
61
+ "loss": 0.8344,
62
+ "step": 90
63
+ },
64
+ {
65
+ "epoch": 0.01,
66
+ "learning_rate": 0.0002,
67
+ "loss": 0.7906,
68
+ "step": 100
69
+ },
70
+ {
71
+ "epoch": 0.01,
72
+ "learning_rate": 0.0002,
73
+ "loss": 0.8281,
74
+ "step": 110
75
+ },
76
+ {
77
+ "epoch": 0.01,
78
+ "learning_rate": 0.0002,
79
+ "loss": 0.7754,
80
+ "step": 120
81
+ },
82
+ {
83
+ "epoch": 0.01,
84
+ "learning_rate": 0.0002,
85
+ "loss": 0.756,
86
+ "step": 130
87
+ },
88
+ {
89
+ "epoch": 0.01,
90
+ "learning_rate": 0.0002,
91
+ "loss": 0.7897,
92
+ "step": 140
93
+ },
94
+ {
95
+ "epoch": 0.01,
96
+ "learning_rate": 0.0002,
97
+ "loss": 0.7877,
98
+ "step": 150
99
+ },
100
+ {
101
+ "epoch": 0.02,
102
+ "learning_rate": 0.0002,
103
+ "loss": 0.7702,
104
+ "step": 160
105
+ },
106
+ {
107
+ "epoch": 0.02,
108
+ "learning_rate": 0.0002,
109
+ "loss": 0.7708,
110
+ "step": 170
111
+ },
112
+ {
113
+ "epoch": 0.02,
114
+ "learning_rate": 0.0002,
115
+ "loss": 0.7661,
116
+ "step": 180
117
+ },
118
+ {
119
+ "epoch": 0.02,
120
+ "learning_rate": 0.0002,
121
+ "loss": 0.7778,
122
+ "step": 190
123
+ },
124
+ {
125
+ "epoch": 0.02,
126
+ "learning_rate": 0.0002,
127
+ "loss": 0.7886,
128
+ "step": 200
129
+ },
130
+ {
131
+ "epoch": 0.02,
132
+ "learning_rate": 0.0002,
133
+ "loss": 0.7938,
134
+ "step": 210
135
+ },
136
+ {
137
+ "epoch": 0.02,
138
+ "learning_rate": 0.0002,
139
+ "loss": 0.7891,
140
+ "step": 220
141
+ },
142
+ {
143
+ "epoch": 0.02,
144
+ "learning_rate": 0.0002,
145
+ "loss": 0.7624,
146
+ "step": 230
147
+ },
148
+ {
149
+ "epoch": 0.02,
150
+ "learning_rate": 0.0002,
151
+ "loss": 0.7464,
152
+ "step": 240
153
+ },
154
+ {
155
+ "epoch": 0.02,
156
+ "learning_rate": 0.0002,
157
+ "loss": 0.7609,
158
+ "step": 250
159
+ },
160
+ {
161
+ "epoch": 0.03,
162
+ "learning_rate": 0.0002,
163
+ "loss": 0.7524,
164
+ "step": 260
165
+ },
166
+ {
167
+ "epoch": 0.03,
168
+ "learning_rate": 0.0002,
169
+ "loss": 0.8035,
170
+ "step": 270
171
+ },
172
+ {
173
+ "epoch": 0.03,
174
+ "learning_rate": 0.0002,
175
+ "loss": 0.7652,
176
+ "step": 280
177
+ },
178
+ {
179
+ "epoch": 0.03,
180
+ "learning_rate": 0.0002,
181
+ "loss": 0.729,
182
+ "step": 290
183
+ },
184
+ {
185
+ "epoch": 0.03,
186
+ "learning_rate": 0.0002,
187
+ "loss": 0.74,
188
+ "step": 300
189
+ },
190
+ {
191
+ "epoch": 0.03,
192
+ "learning_rate": 0.0002,
193
+ "loss": 0.7707,
194
+ "step": 310
195
+ },
196
+ {
197
+ "epoch": 0.03,
198
+ "learning_rate": 0.0002,
199
+ "loss": 0.7492,
200
+ "step": 320
201
+ },
202
+ {
203
+ "epoch": 0.03,
204
+ "learning_rate": 0.0002,
205
+ "loss": 0.764,
206
+ "step": 330
207
+ },
208
+ {
209
+ "epoch": 0.03,
210
+ "learning_rate": 0.0002,
211
+ "loss": 0.8004,
212
+ "step": 340
213
+ },
214
+ {
215
+ "epoch": 0.03,
216
+ "learning_rate": 0.0002,
217
+ "loss": 0.7152,
218
+ "step": 350
219
+ },
220
+ {
221
+ "epoch": 0.03,
222
+ "learning_rate": 0.0002,
223
+ "loss": 0.7655,
224
+ "step": 360
225
+ },
226
+ {
227
+ "epoch": 0.04,
228
+ "learning_rate": 0.0002,
229
+ "loss": 0.7631,
230
+ "step": 370
231
+ },
232
+ {
233
+ "epoch": 0.04,
234
+ "learning_rate": 0.0002,
235
+ "loss": 0.7939,
236
+ "step": 380
237
+ },
238
+ {
239
+ "epoch": 0.04,
240
+ "learning_rate": 0.0002,
241
+ "loss": 0.7283,
242
+ "step": 390
243
+ },
244
+ {
245
+ "epoch": 0.04,
246
+ "learning_rate": 0.0002,
247
+ "loss": 0.7662,
248
+ "step": 400
249
+ },
250
+ {
251
+ "epoch": 0.04,
252
+ "learning_rate": 0.0002,
253
+ "loss": 0.763,
254
+ "step": 410
255
+ },
256
+ {
257
+ "epoch": 0.04,
258
+ "learning_rate": 0.0002,
259
+ "loss": 0.755,
260
+ "step": 420
261
+ },
262
+ {
263
+ "epoch": 0.04,
264
+ "learning_rate": 0.0002,
265
+ "loss": 0.7703,
266
+ "step": 430
267
+ },
268
+ {
269
+ "epoch": 0.04,
270
+ "learning_rate": 0.0002,
271
+ "loss": 0.7685,
272
+ "step": 440
273
+ },
274
+ {
275
+ "epoch": 0.04,
276
+ "learning_rate": 0.0002,
277
+ "loss": 0.73,
278
+ "step": 450
279
+ },
280
+ {
281
+ "epoch": 0.04,
282
+ "learning_rate": 0.0002,
283
+ "loss": 0.745,
284
+ "step": 460
285
+ },
286
+ {
287
+ "epoch": 0.05,
288
+ "learning_rate": 0.0002,
289
+ "loss": 0.7501,
290
+ "step": 470
291
+ },
292
+ {
293
+ "epoch": 0.05,
294
+ "learning_rate": 0.0002,
295
+ "loss": 0.7431,
296
+ "step": 480
297
+ },
298
+ {
299
+ "epoch": 0.05,
300
+ "learning_rate": 0.0002,
301
+ "loss": 0.7346,
302
+ "step": 490
303
+ },
304
+ {
305
+ "epoch": 0.05,
306
+ "learning_rate": 0.0002,
307
+ "loss": 0.7409,
308
+ "step": 500
309
+ },
310
+ {
311
+ "epoch": 0.05,
312
+ "learning_rate": 0.0002,
313
+ "loss": 0.7505,
314
+ "step": 510
315
+ },
316
+ {
317
+ "epoch": 0.05,
318
+ "learning_rate": 0.0002,
319
+ "loss": 0.7504,
320
+ "step": 520
321
+ },
322
+ {
323
+ "epoch": 0.05,
324
+ "learning_rate": 0.0002,
325
+ "loss": 0.7602,
326
+ "step": 530
327
+ },
328
+ {
329
+ "epoch": 0.05,
330
+ "learning_rate": 0.0002,
331
+ "loss": 0.7239,
332
+ "step": 540
333
+ },
334
+ {
335
+ "epoch": 0.05,
336
+ "learning_rate": 0.0002,
337
+ "loss": 0.7386,
338
+ "step": 550
339
+ },
340
+ {
341
+ "epoch": 0.05,
342
+ "learning_rate": 0.0002,
343
+ "loss": 0.7223,
344
+ "step": 560
345
+ },
346
+ {
347
+ "epoch": 0.05,
348
+ "learning_rate": 0.0002,
349
+ "loss": 0.7629,
350
+ "step": 570
351
+ },
352
+ {
353
+ "epoch": 0.06,
354
+ "learning_rate": 0.0002,
355
+ "loss": 0.7385,
356
+ "step": 580
357
+ },
358
+ {
359
+ "epoch": 0.06,
360
+ "learning_rate": 0.0002,
361
+ "loss": 0.7367,
362
+ "step": 590
363
+ },
364
+ {
365
+ "epoch": 0.06,
366
+ "learning_rate": 0.0002,
367
+ "loss": 0.7453,
368
+ "step": 600
369
+ },
370
+ {
371
+ "epoch": 0.06,
372
+ "learning_rate": 0.0002,
373
+ "loss": 0.7534,
374
+ "step": 610
375
+ },
376
+ {
377
+ "epoch": 0.06,
378
+ "learning_rate": 0.0002,
379
+ "loss": 0.7314,
380
+ "step": 620
381
+ },
382
+ {
383
+ "epoch": 0.06,
384
+ "learning_rate": 0.0002,
385
+ "loss": 0.75,
386
+ "step": 630
387
+ },
388
+ {
389
+ "epoch": 0.06,
390
+ "learning_rate": 0.0002,
391
+ "loss": 0.729,
392
+ "step": 640
393
+ },
394
+ {
395
+ "epoch": 0.06,
396
+ "learning_rate": 0.0002,
397
+ "loss": 0.7084,
398
+ "step": 650
399
+ },
400
+ {
401
+ "epoch": 0.06,
402
+ "learning_rate": 0.0002,
403
+ "loss": 0.7478,
404
+ "step": 660
405
+ },
406
+ {
407
+ "epoch": 0.06,
408
+ "learning_rate": 0.0002,
409
+ "loss": 0.7316,
410
+ "step": 670
411
+ },
412
+ {
413
+ "epoch": 0.07,
414
+ "learning_rate": 0.0002,
415
+ "loss": 0.74,
416
+ "step": 680
417
+ },
418
+ {
419
+ "epoch": 0.07,
420
+ "learning_rate": 0.0002,
421
+ "loss": 0.7033,
422
+ "step": 690
423
+ },
424
+ {
425
+ "epoch": 0.07,
426
+ "learning_rate": 0.0002,
427
+ "loss": 0.7491,
428
+ "step": 700
429
+ },
430
+ {
431
+ "epoch": 0.07,
432
+ "learning_rate": 0.0002,
433
+ "loss": 0.7395,
434
+ "step": 710
435
+ },
436
+ {
437
+ "epoch": 0.07,
438
+ "learning_rate": 0.0002,
439
+ "loss": 0.7168,
440
+ "step": 720
441
+ },
442
+ {
443
+ "epoch": 0.07,
444
+ "learning_rate": 0.0002,
445
+ "loss": 0.7537,
446
+ "step": 730
447
+ },
448
+ {
449
+ "epoch": 0.07,
450
+ "learning_rate": 0.0002,
451
+ "loss": 0.7021,
452
+ "step": 740
453
+ },
454
+ {
455
+ "epoch": 0.07,
456
+ "learning_rate": 0.0002,
457
+ "loss": 0.743,
458
+ "step": 750
459
+ },
460
+ {
461
+ "epoch": 0.07,
462
+ "learning_rate": 0.0002,
463
+ "loss": 0.7121,
464
+ "step": 760
465
+ },
466
+ {
467
+ "epoch": 0.07,
468
+ "learning_rate": 0.0002,
469
+ "loss": 0.7209,
470
+ "step": 770
471
+ },
472
+ {
473
+ "epoch": 0.08,
474
+ "learning_rate": 0.0002,
475
+ "loss": 0.7105,
476
+ "step": 780
477
+ },
478
+ {
479
+ "epoch": 0.08,
480
+ "learning_rate": 0.0002,
481
+ "loss": 0.7695,
482
+ "step": 790
483
+ },
484
+ {
485
+ "epoch": 0.08,
486
+ "learning_rate": 0.0002,
487
+ "loss": 0.7322,
488
+ "step": 800
489
+ },
490
+ {
491
+ "epoch": 0.08,
492
+ "learning_rate": 0.0002,
493
+ "loss": 0.709,
494
+ "step": 810
495
+ },
496
+ {
497
+ "epoch": 0.08,
498
+ "learning_rate": 0.0002,
499
+ "loss": 0.7186,
500
+ "step": 820
501
+ },
502
+ {
503
+ "epoch": 0.08,
504
+ "learning_rate": 0.0002,
505
+ "loss": 0.7244,
506
+ "step": 830
507
+ },
508
+ {
509
+ "epoch": 0.08,
510
+ "learning_rate": 0.0002,
511
+ "loss": 0.6892,
512
+ "step": 840
513
+ },
514
+ {
515
+ "epoch": 0.08,
516
+ "learning_rate": 0.0002,
517
+ "loss": 0.7405,
518
+ "step": 850
519
+ },
520
+ {
521
+ "epoch": 0.08,
522
+ "learning_rate": 0.0002,
523
+ "loss": 0.7345,
524
+ "step": 860
525
+ },
526
+ {
527
+ "epoch": 0.08,
528
+ "learning_rate": 0.0002,
529
+ "loss": 0.724,
530
+ "step": 870
531
+ },
532
+ {
533
+ "epoch": 0.08,
534
+ "learning_rate": 0.0002,
535
+ "loss": 0.7511,
536
+ "step": 880
537
+ },
538
+ {
539
+ "epoch": 0.09,
540
+ "learning_rate": 0.0002,
541
+ "loss": 0.7498,
542
+ "step": 890
543
+ },
544
+ {
545
+ "epoch": 0.09,
546
+ "learning_rate": 0.0002,
547
+ "loss": 0.726,
548
+ "step": 900
549
+ },
550
+ {
551
+ "epoch": 0.09,
552
+ "learning_rate": 0.0002,
553
+ "loss": 0.7419,
554
+ "step": 910
555
+ },
556
+ {
557
+ "epoch": 0.09,
558
+ "learning_rate": 0.0002,
559
+ "loss": 0.7117,
560
+ "step": 920
561
+ },
562
+ {
563
+ "epoch": 0.09,
564
+ "learning_rate": 0.0002,
565
+ "loss": 0.7193,
566
+ "step": 930
567
+ },
568
+ {
569
+ "epoch": 0.09,
570
+ "learning_rate": 0.0002,
571
+ "loss": 0.6877,
572
+ "step": 940
573
+ },
574
+ {
575
+ "epoch": 0.09,
576
+ "learning_rate": 0.0002,
577
+ "loss": 0.7084,
578
+ "step": 950
579
+ },
580
+ {
581
+ "epoch": 0.09,
582
+ "learning_rate": 0.0002,
583
+ "loss": 0.7087,
584
+ "step": 960
585
+ },
586
+ {
587
+ "epoch": 0.09,
588
+ "learning_rate": 0.0002,
589
+ "loss": 0.7354,
590
+ "step": 970
591
+ },
592
+ {
593
+ "epoch": 0.09,
594
+ "learning_rate": 0.0002,
595
+ "loss": 0.7164,
596
+ "step": 980
597
+ },
598
+ {
599
+ "epoch": 0.1,
600
+ "learning_rate": 0.0002,
601
+ "loss": 0.7076,
602
+ "step": 990
603
+ },
604
+ {
605
+ "epoch": 0.1,
606
+ "learning_rate": 0.0002,
607
+ "loss": 0.7031,
608
+ "step": 1000
609
+ },
610
+ {
611
+ "epoch": 0.1,
612
+ "learning_rate": 0.0002,
613
+ "loss": 0.7186,
614
+ "step": 1010
615
+ },
616
+ {
617
+ "epoch": 0.1,
618
+ "learning_rate": 0.0002,
619
+ "loss": 0.7311,
620
+ "step": 1020
621
+ },
622
+ {
623
+ "epoch": 0.1,
624
+ "learning_rate": 0.0002,
625
+ "loss": 0.7047,
626
+ "step": 1030
627
+ },
628
+ {
629
+ "epoch": 0.1,
630
+ "learning_rate": 0.0002,
631
+ "loss": 0.7054,
632
+ "step": 1040
633
+ },
634
+ {
635
+ "epoch": 0.1,
636
+ "learning_rate": 0.0002,
637
+ "loss": 0.7221,
638
+ "step": 1050
639
+ },
640
+ {
641
+ "epoch": 0.1,
642
+ "learning_rate": 0.0002,
643
+ "loss": 0.7158,
644
+ "step": 1060
645
+ },
646
+ {
647
+ "epoch": 0.1,
648
+ "learning_rate": 0.0002,
649
+ "loss": 0.6933,
650
+ "step": 1070
651
+ },
652
+ {
653
+ "epoch": 0.1,
654
+ "learning_rate": 0.0002,
655
+ "loss": 0.6992,
656
+ "step": 1080
657
+ },
658
+ {
659
+ "epoch": 0.11,
660
+ "learning_rate": 0.0002,
661
+ "loss": 0.6806,
662
+ "step": 1090
663
+ },
664
+ {
665
+ "epoch": 0.11,
666
+ "learning_rate": 0.0002,
667
+ "loss": 0.7053,
668
+ "step": 1100
669
+ },
670
+ {
671
+ "epoch": 0.11,
672
+ "learning_rate": 0.0002,
673
+ "loss": 0.7124,
674
+ "step": 1110
675
+ },
676
+ {
677
+ "epoch": 0.11,
678
+ "learning_rate": 0.0002,
679
+ "loss": 0.7011,
680
+ "step": 1120
681
+ },
682
+ {
683
+ "epoch": 0.11,
684
+ "learning_rate": 0.0002,
685
+ "loss": 0.7036,
686
+ "step": 1130
687
+ },
688
+ {
689
+ "epoch": 0.11,
690
+ "learning_rate": 0.0002,
691
+ "loss": 0.7195,
692
+ "step": 1140
693
+ },
694
+ {
695
+ "epoch": 0.11,
696
+ "learning_rate": 0.0002,
697
+ "loss": 0.706,
698
+ "step": 1150
699
+ },
700
+ {
701
+ "epoch": 0.11,
702
+ "learning_rate": 0.0002,
703
+ "loss": 0.6977,
704
+ "step": 1160
705
+ },
706
+ {
707
+ "epoch": 0.11,
708
+ "learning_rate": 0.0002,
709
+ "loss": 0.709,
710
+ "step": 1170
711
+ },
712
+ {
713
+ "epoch": 0.11,
714
+ "learning_rate": 0.0002,
715
+ "loss": 0.6983,
716
+ "step": 1180
717
+ },
718
+ {
719
+ "epoch": 0.11,
720
+ "learning_rate": 0.0002,
721
+ "loss": 0.7074,
722
+ "step": 1190
723
+ },
724
+ {
725
+ "epoch": 0.12,
726
+ "learning_rate": 0.0002,
727
+ "loss": 0.7214,
728
+ "step": 1200
729
+ },
730
+ {
731
+ "epoch": 0.12,
732
+ "learning_rate": 0.0002,
733
+ "loss": 0.7162,
734
+ "step": 1210
735
+ },
736
+ {
737
+ "epoch": 0.12,
738
+ "learning_rate": 0.0002,
739
+ "loss": 0.7396,
740
+ "step": 1220
741
+ },
742
+ {
743
+ "epoch": 0.12,
744
+ "learning_rate": 0.0002,
745
+ "loss": 0.7097,
746
+ "step": 1230
747
+ },
748
+ {
749
+ "epoch": 0.12,
750
+ "learning_rate": 0.0002,
751
+ "loss": 0.696,
752
+ "step": 1240
753
+ },
754
+ {
755
+ "epoch": 0.12,
756
+ "learning_rate": 0.0002,
757
+ "loss": 0.6924,
758
+ "step": 1250
759
+ },
760
+ {
761
+ "epoch": 0.12,
762
+ "learning_rate": 0.0002,
763
+ "loss": 0.7157,
764
+ "step": 1260
765
+ },
766
+ {
767
+ "epoch": 0.12,
768
+ "learning_rate": 0.0002,
769
+ "loss": 0.6844,
770
+ "step": 1270
771
+ },
772
+ {
773
+ "epoch": 0.12,
774
+ "learning_rate": 0.0002,
775
+ "loss": 0.7076,
776
+ "step": 1280
777
+ },
778
+ {
779
+ "epoch": 0.12,
780
+ "learning_rate": 0.0002,
781
+ "loss": 0.7204,
782
+ "step": 1290
783
+ },
784
+ {
785
+ "epoch": 0.13,
786
+ "learning_rate": 0.0002,
787
+ "loss": 0.7031,
788
+ "step": 1300
789
+ },
790
+ {
791
+ "epoch": 0.13,
792
+ "learning_rate": 0.0002,
793
+ "loss": 0.7215,
794
+ "step": 1310
795
+ },
796
+ {
797
+ "epoch": 0.13,
798
+ "learning_rate": 0.0002,
799
+ "loss": 0.7195,
800
+ "step": 1320
801
+ },
802
+ {
803
+ "epoch": 0.13,
804
+ "learning_rate": 0.0002,
805
+ "loss": 0.7149,
806
+ "step": 1330
807
+ },
808
+ {
809
+ "epoch": 0.13,
810
+ "learning_rate": 0.0002,
811
+ "loss": 0.6728,
812
+ "step": 1340
813
+ },
814
+ {
815
+ "epoch": 0.13,
816
+ "learning_rate": 0.0002,
817
+ "loss": 0.7129,
818
+ "step": 1350
819
+ },
820
+ {
821
+ "epoch": 0.13,
822
+ "learning_rate": 0.0002,
823
+ "loss": 0.6875,
824
+ "step": 1360
825
+ },
826
+ {
827
+ "epoch": 0.13,
828
+ "learning_rate": 0.0002,
829
+ "loss": 0.7116,
830
+ "step": 1370
831
+ },
832
+ {
833
+ "epoch": 0.13,
834
+ "learning_rate": 0.0002,
835
+ "loss": 0.7241,
836
+ "step": 1380
837
+ },
838
+ {
839
+ "epoch": 0.13,
840
+ "learning_rate": 0.0002,
841
+ "loss": 0.6876,
842
+ "step": 1390
843
+ },
844
+ {
845
+ "epoch": 0.13,
846
+ "learning_rate": 0.0002,
847
+ "loss": 0.7109,
848
+ "step": 1400
849
+ },
850
+ {
851
+ "epoch": 0.14,
852
+ "learning_rate": 0.0002,
853
+ "loss": 0.6953,
854
+ "step": 1410
855
+ },
856
+ {
857
+ "epoch": 0.14,
858
+ "learning_rate": 0.0002,
859
+ "loss": 0.7023,
860
+ "step": 1420
861
+ },
862
+ {
863
+ "epoch": 0.14,
864
+ "learning_rate": 0.0002,
865
+ "loss": 0.7145,
866
+ "step": 1430
867
+ },
868
+ {
869
+ "epoch": 0.14,
870
+ "learning_rate": 0.0002,
871
+ "loss": 0.6958,
872
+ "step": 1440
873
+ },
874
+ {
875
+ "epoch": 0.14,
876
+ "learning_rate": 0.0002,
877
+ "loss": 0.6825,
878
+ "step": 1450
879
+ },
880
+ {
881
+ "epoch": 0.14,
882
+ "learning_rate": 0.0002,
883
+ "loss": 0.6811,
884
+ "step": 1460
885
+ },
886
+ {
887
+ "epoch": 0.14,
888
+ "learning_rate": 0.0002,
889
+ "loss": 0.6851,
890
+ "step": 1470
891
+ },
892
+ {
893
+ "epoch": 0.14,
894
+ "learning_rate": 0.0002,
895
+ "loss": 0.6866,
896
+ "step": 1480
897
+ },
898
+ {
899
+ "epoch": 0.14,
900
+ "learning_rate": 0.0002,
901
+ "loss": 0.7145,
902
+ "step": 1490
903
+ },
904
+ {
905
+ "epoch": 0.14,
906
+ "learning_rate": 0.0002,
907
+ "loss": 0.672,
908
+ "step": 1500
909
+ },
910
+ {
911
+ "epoch": 0.15,
912
+ "learning_rate": 0.0002,
913
+ "loss": 0.7374,
914
+ "step": 1510
915
+ },
916
+ {
917
+ "epoch": 0.15,
918
+ "learning_rate": 0.0002,
919
+ "loss": 0.6459,
920
+ "step": 1520
921
+ },
922
+ {
923
+ "epoch": 0.15,
924
+ "learning_rate": 0.0002,
925
+ "loss": 0.6859,
926
+ "step": 1530
927
+ },
928
+ {
929
+ "epoch": 0.15,
930
+ "learning_rate": 0.0002,
931
+ "loss": 0.7014,
932
+ "step": 1540
933
+ },
934
+ {
935
+ "epoch": 0.15,
936
+ "learning_rate": 0.0002,
937
+ "loss": 0.6974,
938
+ "step": 1550
939
+ },
940
+ {
941
+ "epoch": 0.15,
942
+ "learning_rate": 0.0002,
943
+ "loss": 0.6764,
944
+ "step": 1560
945
+ },
946
+ {
947
+ "epoch": 0.15,
948
+ "learning_rate": 0.0002,
949
+ "loss": 0.6902,
950
+ "step": 1570
951
+ },
952
+ {
953
+ "epoch": 0.15,
954
+ "learning_rate": 0.0002,
955
+ "loss": 0.7007,
956
+ "step": 1580
957
+ },
958
+ {
959
+ "epoch": 0.15,
960
+ "learning_rate": 0.0002,
961
+ "loss": 0.6808,
962
+ "step": 1590
963
+ },
964
+ {
965
+ "epoch": 0.15,
966
+ "learning_rate": 0.0002,
967
+ "loss": 0.7018,
968
+ "step": 1600
969
+ },
970
+ {
971
+ "epoch": 0.16,
972
+ "learning_rate": 0.0002,
973
+ "loss": 0.6991,
974
+ "step": 1610
975
+ },
976
+ {
977
+ "epoch": 0.16,
978
+ "learning_rate": 0.0002,
979
+ "loss": 0.687,
980
+ "step": 1620
981
+ },
982
+ {
983
+ "epoch": 0.16,
984
+ "learning_rate": 0.0002,
985
+ "loss": 0.6815,
986
+ "step": 1630
987
+ },
988
+ {
989
+ "epoch": 0.16,
990
+ "learning_rate": 0.0002,
991
+ "loss": 0.6754,
992
+ "step": 1640
993
+ },
994
+ {
995
+ "epoch": 0.16,
996
+ "learning_rate": 0.0002,
997
+ "loss": 0.7208,
998
+ "step": 1650
999
+ },
1000
+ {
1001
+ "epoch": 0.16,
1002
+ "learning_rate": 0.0002,
1003
+ "loss": 0.6822,
1004
+ "step": 1660
1005
+ },
1006
+ {
1007
+ "epoch": 0.16,
1008
+ "learning_rate": 0.0002,
1009
+ "loss": 0.6964,
1010
+ "step": 1670
1011
+ },
1012
+ {
1013
+ "epoch": 0.16,
1014
+ "learning_rate": 0.0002,
1015
+ "loss": 0.6762,
1016
+ "step": 1680
1017
+ },
1018
+ {
1019
+ "epoch": 0.16,
1020
+ "learning_rate": 0.0002,
1021
+ "loss": 0.6847,
1022
+ "step": 1690
1023
+ },
1024
+ {
1025
+ "epoch": 0.16,
1026
+ "learning_rate": 0.0002,
1027
+ "loss": 0.698,
1028
+ "step": 1700
1029
+ },
1030
+ {
1031
+ "epoch": 0.16,
1032
+ "learning_rate": 0.0002,
1033
+ "loss": 0.6905,
1034
+ "step": 1710
1035
+ },
1036
+ {
1037
+ "epoch": 0.17,
1038
+ "learning_rate": 0.0002,
1039
+ "loss": 0.6765,
1040
+ "step": 1720
1041
+ },
1042
+ {
1043
+ "epoch": 0.17,
1044
+ "learning_rate": 0.0002,
1045
+ "loss": 0.662,
1046
+ "step": 1730
1047
+ },
1048
+ {
1049
+ "epoch": 0.17,
1050
+ "learning_rate": 0.0002,
1051
+ "loss": 0.6654,
1052
+ "step": 1740
1053
+ },
1054
+ {
1055
+ "epoch": 0.17,
1056
+ "learning_rate": 0.0002,
1057
+ "loss": 0.7016,
1058
+ "step": 1750
1059
+ },
1060
+ {
1061
+ "epoch": 0.17,
1062
+ "learning_rate": 0.0002,
1063
+ "loss": 0.6975,
1064
+ "step": 1760
1065
+ },
1066
+ {
1067
+ "epoch": 0.17,
1068
+ "learning_rate": 0.0002,
1069
+ "loss": 0.6733,
1070
+ "step": 1770
1071
+ },
1072
+ {
1073
+ "epoch": 0.17,
1074
+ "learning_rate": 0.0002,
1075
+ "loss": 0.698,
1076
+ "step": 1780
1077
+ },
1078
+ {
1079
+ "epoch": 0.17,
1080
+ "learning_rate": 0.0002,
1081
+ "loss": 0.6764,
1082
+ "step": 1790
1083
+ },
1084
+ {
1085
+ "epoch": 0.17,
1086
+ "learning_rate": 0.0002,
1087
+ "loss": 0.727,
1088
+ "step": 1800
1089
+ },
1090
+ {
1091
+ "epoch": 0.17,
1092
+ "learning_rate": 0.0002,
1093
+ "loss": 0.7089,
1094
+ "step": 1810
1095
+ },
1096
+ {
1097
+ "epoch": 0.18,
1098
+ "learning_rate": 0.0002,
1099
+ "loss": 0.6938,
1100
+ "step": 1820
1101
+ },
1102
+ {
1103
+ "epoch": 0.18,
1104
+ "learning_rate": 0.0002,
1105
+ "loss": 0.6931,
1106
+ "step": 1830
1107
+ },
1108
+ {
1109
+ "epoch": 0.18,
1110
+ "learning_rate": 0.0002,
1111
+ "loss": 0.7009,
1112
+ "step": 1840
1113
+ },
1114
+ {
1115
+ "epoch": 0.18,
1116
+ "learning_rate": 0.0002,
1117
+ "loss": 0.723,
1118
+ "step": 1850
1119
+ },
1120
+ {
1121
+ "epoch": 0.18,
1122
+ "learning_rate": 0.0002,
1123
+ "loss": 0.6701,
1124
+ "step": 1860
1125
+ },
1126
+ {
1127
+ "epoch": 0.18,
1128
+ "learning_rate": 0.0002,
1129
+ "loss": 0.6604,
1130
+ "step": 1870
1131
+ },
1132
+ {
1133
+ "epoch": 0.18,
1134
+ "learning_rate": 0.0002,
1135
+ "loss": 0.6837,
1136
+ "step": 1880
1137
+ },
1138
+ {
1139
+ "epoch": 0.18,
1140
+ "learning_rate": 0.0002,
1141
+ "loss": 0.6793,
1142
+ "step": 1890
1143
+ },
1144
+ {
1145
+ "epoch": 0.18,
1146
+ "learning_rate": 0.0002,
1147
+ "loss": 0.6826,
1148
+ "step": 1900
1149
+ },
1150
+ {
1151
+ "epoch": 0.18,
1152
+ "learning_rate": 0.0002,
1153
+ "loss": 0.6722,
1154
+ "step": 1910
1155
+ },
1156
+ {
1157
+ "epoch": 0.18,
1158
+ "learning_rate": 0.0002,
1159
+ "loss": 0.6859,
1160
+ "step": 1920
1161
+ },
1162
+ {
1163
+ "epoch": 0.19,
1164
+ "learning_rate": 0.0002,
1165
+ "loss": 0.6744,
1166
+ "step": 1930
1167
+ },
1168
+ {
1169
+ "epoch": 0.19,
1170
+ "learning_rate": 0.0002,
1171
+ "loss": 0.7037,
1172
+ "step": 1940
1173
+ },
1174
+ {
1175
+ "epoch": 0.19,
1176
+ "learning_rate": 0.0002,
1177
+ "loss": 0.6826,
1178
+ "step": 1950
1179
+ },
1180
+ {
1181
+ "epoch": 0.19,
1182
+ "learning_rate": 0.0002,
1183
+ "loss": 0.6663,
1184
+ "step": 1960
1185
+ },
1186
+ {
1187
+ "epoch": 0.19,
1188
+ "learning_rate": 0.0002,
1189
+ "loss": 0.6986,
1190
+ "step": 1970
1191
+ },
1192
+ {
1193
+ "epoch": 0.19,
1194
+ "learning_rate": 0.0002,
1195
+ "loss": 0.6701,
1196
+ "step": 1980
1197
+ },
1198
+ {
1199
+ "epoch": 0.19,
1200
+ "learning_rate": 0.0002,
1201
+ "loss": 0.6932,
1202
+ "step": 1990
1203
+ },
1204
+ {
1205
+ "epoch": 0.19,
1206
+ "learning_rate": 0.0002,
1207
+ "loss": 0.6902,
1208
+ "step": 2000
1209
+ },
1210
+ {
1211
+ "epoch": 0.19,
1212
+ "learning_rate": 0.0002,
1213
+ "loss": 0.6797,
1214
+ "step": 2010
1215
+ },
1216
+ {
1217
+ "epoch": 0.19,
1218
+ "learning_rate": 0.0002,
1219
+ "loss": 0.6729,
1220
+ "step": 2020
1221
+ },
1222
+ {
1223
+ "epoch": 0.2,
1224
+ "learning_rate": 0.0002,
1225
+ "loss": 0.6662,
1226
+ "step": 2030
1227
+ },
1228
+ {
1229
+ "epoch": 0.2,
1230
+ "learning_rate": 0.0002,
1231
+ "loss": 0.6977,
1232
+ "step": 2040
1233
+ },
1234
+ {
1235
+ "epoch": 0.2,
1236
+ "learning_rate": 0.0002,
1237
+ "loss": 0.6798,
1238
+ "step": 2050
1239
+ },
1240
+ {
1241
+ "epoch": 0.2,
1242
+ "learning_rate": 0.0002,
1243
+ "loss": 0.6549,
1244
+ "step": 2060
1245
+ },
1246
+ {
1247
+ "epoch": 0.2,
1248
+ "learning_rate": 0.0002,
1249
+ "loss": 0.7058,
1250
+ "step": 2070
1251
+ },
1252
+ {
1253
+ "epoch": 0.2,
1254
+ "learning_rate": 0.0002,
1255
+ "loss": 0.6936,
1256
+ "step": 2080
1257
+ },
1258
+ {
1259
+ "epoch": 0.2,
1260
+ "learning_rate": 0.0002,
1261
+ "loss": 0.6549,
1262
+ "step": 2090
1263
+ },
1264
+ {
1265
+ "epoch": 0.2,
1266
+ "learning_rate": 0.0002,
1267
+ "loss": 0.6969,
1268
+ "step": 2100
1269
+ },
1270
+ {
1271
+ "epoch": 0.2,
1272
+ "learning_rate": 0.0002,
1273
+ "loss": 0.6804,
1274
+ "step": 2110
1275
+ },
1276
+ {
1277
+ "epoch": 0.2,
1278
+ "learning_rate": 0.0002,
1279
+ "loss": 0.6737,
1280
+ "step": 2120
1281
+ },
1282
+ {
1283
+ "epoch": 0.21,
1284
+ "learning_rate": 0.0002,
1285
+ "loss": 0.6796,
1286
+ "step": 2130
1287
+ },
1288
+ {
1289
+ "epoch": 0.21,
1290
+ "learning_rate": 0.0002,
1291
+ "loss": 0.6777,
1292
+ "step": 2140
1293
+ },
1294
+ {
1295
+ "epoch": 0.21,
1296
+ "learning_rate": 0.0002,
1297
+ "loss": 0.6447,
1298
+ "step": 2150
1299
+ },
1300
+ {
1301
+ "epoch": 0.21,
1302
+ "learning_rate": 0.0002,
1303
+ "loss": 0.689,
1304
+ "step": 2160
1305
+ },
1306
+ {
1307
+ "epoch": 0.21,
1308
+ "learning_rate": 0.0002,
1309
+ "loss": 0.6698,
1310
+ "step": 2170
1311
+ },
1312
+ {
1313
+ "epoch": 0.21,
1314
+ "learning_rate": 0.0002,
1315
+ "loss": 0.6826,
1316
+ "step": 2180
1317
+ },
1318
+ {
1319
+ "epoch": 0.21,
1320
+ "learning_rate": 0.0002,
1321
+ "loss": 0.6801,
1322
+ "step": 2190
1323
+ },
1324
+ {
1325
+ "epoch": 0.21,
1326
+ "learning_rate": 0.0002,
1327
+ "loss": 0.6362,
1328
+ "step": 2200
1329
+ },
1330
+ {
1331
+ "epoch": 0.21,
1332
+ "learning_rate": 0.0002,
1333
+ "loss": 0.6868,
1334
+ "step": 2210
1335
+ },
1336
+ {
1337
+ "epoch": 0.21,
1338
+ "learning_rate": 0.0002,
1339
+ "loss": 0.6581,
1340
+ "step": 2220
1341
+ },
1342
+ {
1343
+ "epoch": 0.21,
1344
+ "learning_rate": 0.0002,
1345
+ "loss": 0.6759,
1346
+ "step": 2230
1347
+ },
1348
+ {
1349
+ "epoch": 0.22,
1350
+ "learning_rate": 0.0002,
1351
+ "loss": 0.6813,
1352
+ "step": 2240
1353
+ },
1354
+ {
1355
+ "epoch": 0.22,
1356
+ "learning_rate": 0.0002,
1357
+ "loss": 0.7001,
1358
+ "step": 2250
1359
+ },
1360
+ {
1361
+ "epoch": 0.22,
1362
+ "learning_rate": 0.0002,
1363
+ "loss": 0.6569,
1364
+ "step": 2260
1365
+ },
1366
+ {
1367
+ "epoch": 0.22,
1368
+ "learning_rate": 0.0002,
1369
+ "loss": 0.6986,
1370
+ "step": 2270
1371
+ },
1372
+ {
1373
+ "epoch": 0.22,
1374
+ "learning_rate": 0.0002,
1375
+ "loss": 0.6887,
1376
+ "step": 2280
1377
+ },
1378
+ {
1379
+ "epoch": 0.22,
1380
+ "learning_rate": 0.0002,
1381
+ "loss": 0.6711,
1382
+ "step": 2290
1383
+ },
1384
+ {
1385
+ "epoch": 0.22,
1386
+ "learning_rate": 0.0002,
1387
+ "loss": 0.666,
1388
+ "step": 2300
1389
+ },
1390
+ {
1391
+ "epoch": 0.22,
1392
+ "learning_rate": 0.0002,
1393
+ "loss": 0.6572,
1394
+ "step": 2310
1395
+ },
1396
+ {
1397
+ "epoch": 0.22,
1398
+ "learning_rate": 0.0002,
1399
+ "loss": 0.6839,
1400
+ "step": 2320
1401
+ },
1402
+ {
1403
+ "epoch": 0.22,
1404
+ "learning_rate": 0.0002,
1405
+ "loss": 0.6466,
1406
+ "step": 2330
1407
+ },
1408
+ {
1409
+ "epoch": 0.23,
1410
+ "learning_rate": 0.0002,
1411
+ "loss": 0.6916,
1412
+ "step": 2340
1413
+ },
1414
+ {
1415
+ "epoch": 0.23,
1416
+ "learning_rate": 0.0002,
1417
+ "loss": 0.6941,
1418
+ "step": 2350
1419
+ },
1420
+ {
1421
+ "epoch": 0.23,
1422
+ "learning_rate": 0.0002,
1423
+ "loss": 0.6797,
1424
+ "step": 2360
1425
+ },
1426
+ {
1427
+ "epoch": 0.23,
1428
+ "learning_rate": 0.0002,
1429
+ "loss": 0.6653,
1430
+ "step": 2370
1431
+ },
1432
+ {
1433
+ "epoch": 0.23,
1434
+ "learning_rate": 0.0002,
1435
+ "loss": 0.6759,
1436
+ "step": 2380
1437
+ },
1438
+ {
1439
+ "epoch": 0.23,
1440
+ "learning_rate": 0.0002,
1441
+ "loss": 0.6639,
1442
+ "step": 2390
1443
+ },
1444
+ {
1445
+ "epoch": 0.23,
1446
+ "learning_rate": 0.0002,
1447
+ "loss": 0.6566,
1448
+ "step": 2400
1449
+ },
1450
+ {
1451
+ "epoch": 0.23,
1452
+ "learning_rate": 0.0002,
1453
+ "loss": 0.6564,
1454
+ "step": 2410
1455
+ },
1456
+ {
1457
+ "epoch": 0.23,
1458
+ "learning_rate": 0.0002,
1459
+ "loss": 0.6667,
1460
+ "step": 2420
1461
+ },
1462
+ {
1463
+ "epoch": 0.23,
1464
+ "learning_rate": 0.0002,
1465
+ "loss": 0.7028,
1466
+ "step": 2430
1467
+ },
1468
+ {
1469
+ "epoch": 0.24,
1470
+ "learning_rate": 0.0002,
1471
+ "loss": 0.6602,
1472
+ "step": 2440
1473
+ },
1474
+ {
1475
+ "epoch": 0.24,
1476
+ "learning_rate": 0.0002,
1477
+ "loss": 0.6432,
1478
+ "step": 2450
1479
+ },
1480
+ {
1481
+ "epoch": 0.24,
1482
+ "learning_rate": 0.0002,
1483
+ "loss": 0.659,
1484
+ "step": 2460
1485
+ },
1486
+ {
1487
+ "epoch": 0.24,
1488
+ "learning_rate": 0.0002,
1489
+ "loss": 0.6631,
1490
+ "step": 2470
1491
+ },
1492
+ {
1493
+ "epoch": 0.24,
1494
+ "learning_rate": 0.0002,
1495
+ "loss": 0.6816,
1496
+ "step": 2480
1497
+ },
1498
+ {
1499
+ "epoch": 0.24,
1500
+ "learning_rate": 0.0002,
1501
+ "loss": 0.668,
1502
+ "step": 2490
1503
+ },
1504
+ {
1505
+ "epoch": 0.24,
1506
+ "learning_rate": 0.0002,
1507
+ "loss": 0.674,
1508
+ "step": 2500
1509
+ },
1510
+ {
1511
+ "epoch": 0.24,
1512
+ "learning_rate": 0.0002,
1513
+ "loss": 0.6973,
1514
+ "step": 2510
1515
+ },
1516
+ {
1517
+ "epoch": 0.24,
1518
+ "learning_rate": 0.0002,
1519
+ "loss": 0.6602,
1520
+ "step": 2520
1521
+ },
1522
+ {
1523
+ "epoch": 0.24,
1524
+ "learning_rate": 0.0002,
1525
+ "loss": 0.6489,
1526
+ "step": 2530
1527
+ },
1528
+ {
1529
+ "epoch": 0.24,
1530
+ "learning_rate": 0.0002,
1531
+ "loss": 0.6589,
1532
+ "step": 2540
1533
+ },
1534
+ {
1535
+ "epoch": 0.25,
1536
+ "learning_rate": 0.0002,
1537
+ "loss": 0.6595,
1538
+ "step": 2550
1539
+ },
1540
+ {
1541
+ "epoch": 0.25,
1542
+ "learning_rate": 0.0002,
1543
+ "loss": 0.647,
1544
+ "step": 2560
1545
+ },
1546
+ {
1547
+ "epoch": 0.25,
1548
+ "learning_rate": 0.0002,
1549
+ "loss": 0.671,
1550
+ "step": 2570
1551
+ },
1552
+ {
1553
+ "epoch": 0.25,
1554
+ "learning_rate": 0.0002,
1555
+ "loss": 0.6623,
1556
+ "step": 2580
1557
+ },
1558
+ {
1559
+ "epoch": 0.25,
1560
+ "learning_rate": 0.0002,
1561
+ "loss": 0.6985,
1562
+ "step": 2590
1563
+ },
1564
+ {
1565
+ "epoch": 0.25,
1566
+ "learning_rate": 0.0002,
1567
+ "loss": 0.684,
1568
+ "step": 2600
1569
+ },
1570
+ {
1571
+ "epoch": 0.25,
1572
+ "learning_rate": 0.0002,
1573
+ "loss": 0.6947,
1574
+ "step": 2610
1575
+ },
1576
+ {
1577
+ "epoch": 0.25,
1578
+ "learning_rate": 0.0002,
1579
+ "loss": 0.672,
1580
+ "step": 2620
1581
+ },
1582
+ {
1583
+ "epoch": 0.25,
1584
+ "learning_rate": 0.0002,
1585
+ "loss": 0.7097,
1586
+ "step": 2630
1587
+ },
1588
+ {
1589
+ "epoch": 0.25,
1590
+ "learning_rate": 0.0002,
1591
+ "loss": 0.6685,
1592
+ "step": 2640
1593
+ },
1594
+ {
1595
+ "epoch": 0.26,
1596
+ "learning_rate": 0.0002,
1597
+ "loss": 0.6572,
1598
+ "step": 2650
1599
+ },
1600
+ {
1601
+ "epoch": 0.26,
1602
+ "learning_rate": 0.0002,
1603
+ "loss": 0.6503,
1604
+ "step": 2660
1605
+ },
1606
+ {
1607
+ "epoch": 0.26,
1608
+ "learning_rate": 0.0002,
1609
+ "loss": 0.6454,
1610
+ "step": 2670
1611
+ },
1612
+ {
1613
+ "epoch": 0.26,
1614
+ "learning_rate": 0.0002,
1615
+ "loss": 0.6792,
1616
+ "step": 2680
1617
+ },
1618
+ {
1619
+ "epoch": 0.26,
1620
+ "learning_rate": 0.0002,
1621
+ "loss": 0.6816,
1622
+ "step": 2690
1623
+ },
1624
+ {
1625
+ "epoch": 0.26,
1626
+ "learning_rate": 0.0002,
1627
+ "loss": 0.6713,
1628
+ "step": 2700
1629
+ },
1630
+ {
1631
+ "epoch": 0.26,
1632
+ "learning_rate": 0.0002,
1633
+ "loss": 0.652,
1634
+ "step": 2710
1635
+ },
1636
+ {
1637
+ "epoch": 0.26,
1638
+ "learning_rate": 0.0002,
1639
+ "loss": 0.6558,
1640
+ "step": 2720
1641
+ },
1642
+ {
1643
+ "epoch": 0.26,
1644
+ "learning_rate": 0.0002,
1645
+ "loss": 0.6696,
1646
+ "step": 2730
1647
+ },
1648
+ {
1649
+ "epoch": 0.26,
1650
+ "learning_rate": 0.0002,
1651
+ "loss": 0.6561,
1652
+ "step": 2740
1653
+ },
1654
+ {
1655
+ "epoch": 0.26,
1656
+ "learning_rate": 0.0002,
1657
+ "loss": 0.6527,
1658
+ "step": 2750
1659
+ },
1660
+ {
1661
+ "epoch": 0.27,
1662
+ "learning_rate": 0.0002,
1663
+ "loss": 0.6705,
1664
+ "step": 2760
1665
+ },
1666
+ {
1667
+ "epoch": 0.27,
1668
+ "learning_rate": 0.0002,
1669
+ "loss": 0.6734,
1670
+ "step": 2770
1671
+ },
1672
+ {
1673
+ "epoch": 0.27,
1674
+ "learning_rate": 0.0002,
1675
+ "loss": 0.644,
1676
+ "step": 2780
1677
+ },
1678
+ {
1679
+ "epoch": 0.27,
1680
+ "learning_rate": 0.0002,
1681
+ "loss": 0.6615,
1682
+ "step": 2790
1683
+ },
1684
+ {
1685
+ "epoch": 0.27,
1686
+ "learning_rate": 0.0002,
1687
+ "loss": 0.683,
1688
+ "step": 2800
1689
+ },
1690
+ {
1691
+ "epoch": 0.27,
1692
+ "learning_rate": 0.0002,
1693
+ "loss": 0.658,
1694
+ "step": 2810
1695
+ },
1696
+ {
1697
+ "epoch": 0.27,
1698
+ "learning_rate": 0.0002,
1699
+ "loss": 0.651,
1700
+ "step": 2820
1701
+ },
1702
+ {
1703
+ "epoch": 0.27,
1704
+ "learning_rate": 0.0002,
1705
+ "loss": 0.6695,
1706
+ "step": 2830
1707
+ },
1708
+ {
1709
+ "epoch": 0.27,
1710
+ "learning_rate": 0.0002,
1711
+ "loss": 0.6792,
1712
+ "step": 2840
1713
+ },
1714
+ {
1715
+ "epoch": 0.27,
1716
+ "learning_rate": 0.0002,
1717
+ "loss": 0.6694,
1718
+ "step": 2850
1719
+ },
1720
+ {
1721
+ "epoch": 0.28,
1722
+ "learning_rate": 0.0002,
1723
+ "loss": 0.6626,
1724
+ "step": 2860
1725
+ },
1726
+ {
1727
+ "epoch": 0.28,
1728
+ "learning_rate": 0.0002,
1729
+ "loss": 0.6794,
1730
+ "step": 2870
1731
+ },
1732
+ {
1733
+ "epoch": 0.28,
1734
+ "learning_rate": 0.0002,
1735
+ "loss": 0.6948,
1736
+ "step": 2880
1737
+ },
1738
+ {
1739
+ "epoch": 0.28,
1740
+ "learning_rate": 0.0002,
1741
+ "loss": 0.6449,
1742
+ "step": 2890
1743
+ },
1744
+ {
1745
+ "epoch": 0.28,
1746
+ "learning_rate": 0.0002,
1747
+ "loss": 0.6466,
1748
+ "step": 2900
1749
+ },
1750
+ {
1751
+ "epoch": 0.28,
1752
+ "learning_rate": 0.0002,
1753
+ "loss": 0.6588,
1754
+ "step": 2910
1755
+ },
1756
+ {
1757
+ "epoch": 0.28,
1758
+ "learning_rate": 0.0002,
1759
+ "loss": 0.6452,
1760
+ "step": 2920
1761
+ },
1762
+ {
1763
+ "epoch": 0.28,
1764
+ "learning_rate": 0.0002,
1765
+ "loss": 0.6818,
1766
+ "step": 2930
1767
+ },
1768
+ {
1769
+ "epoch": 0.28,
1770
+ "learning_rate": 0.0002,
1771
+ "loss": 0.6749,
1772
+ "step": 2940
1773
+ },
1774
+ {
1775
+ "epoch": 0.28,
1776
+ "learning_rate": 0.0002,
1777
+ "loss": 0.6458,
1778
+ "step": 2950
1779
+ },
1780
+ {
1781
+ "epoch": 0.29,
1782
+ "learning_rate": 0.0002,
1783
+ "loss": 0.6842,
1784
+ "step": 2960
1785
+ },
1786
+ {
1787
+ "epoch": 0.29,
1788
+ "learning_rate": 0.0002,
1789
+ "loss": 0.652,
1790
+ "step": 2970
1791
+ },
1792
+ {
1793
+ "epoch": 0.29,
1794
+ "learning_rate": 0.0002,
1795
+ "loss": 0.6633,
1796
+ "step": 2980
1797
+ },
1798
+ {
1799
+ "epoch": 0.29,
1800
+ "learning_rate": 0.0002,
1801
+ "loss": 0.6717,
1802
+ "step": 2990
1803
+ },
1804
+ {
1805
+ "epoch": 0.29,
1806
+ "learning_rate": 0.0002,
1807
+ "loss": 0.6573,
1808
+ "step": 3000
1809
+ },
1810
+ {
1811
+ "epoch": 0.29,
1812
+ "learning_rate": 0.0002,
1813
+ "loss": 0.6254,
1814
+ "step": 3010
1815
+ },
1816
+ {
1817
+ "epoch": 0.29,
1818
+ "learning_rate": 0.0002,
1819
+ "loss": 0.6701,
1820
+ "step": 3020
1821
+ },
1822
+ {
1823
+ "epoch": 0.29,
1824
+ "learning_rate": 0.0002,
1825
+ "loss": 0.6781,
1826
+ "step": 3030
1827
+ },
1828
+ {
1829
+ "epoch": 0.29,
1830
+ "learning_rate": 0.0002,
1831
+ "loss": 0.6534,
1832
+ "step": 3040
1833
+ },
1834
+ {
1835
+ "epoch": 0.29,
1836
+ "learning_rate": 0.0002,
1837
+ "loss": 0.6925,
1838
+ "step": 3050
1839
+ },
1840
+ {
1841
+ "epoch": 0.29,
1842
+ "learning_rate": 0.0002,
1843
+ "loss": 0.6727,
1844
+ "step": 3060
1845
+ },
1846
+ {
1847
+ "epoch": 0.3,
1848
+ "learning_rate": 0.0002,
1849
+ "loss": 0.6548,
1850
+ "step": 3070
1851
+ },
1852
+ {
1853
+ "epoch": 0.3,
1854
+ "learning_rate": 0.0002,
1855
+ "loss": 0.6666,
1856
+ "step": 3080
1857
+ },
1858
+ {
1859
+ "epoch": 0.3,
1860
+ "learning_rate": 0.0002,
1861
+ "loss": 0.6757,
1862
+ "step": 3090
1863
+ },
1864
+ {
1865
+ "epoch": 0.3,
1866
+ "learning_rate": 0.0002,
1867
+ "loss": 0.662,
1868
+ "step": 3100
1869
+ },
1870
+ {
1871
+ "epoch": 0.3,
1872
+ "learning_rate": 0.0002,
1873
+ "loss": 0.6293,
1874
+ "step": 3110
1875
+ },
1876
+ {
1877
+ "epoch": 0.3,
1878
+ "learning_rate": 0.0002,
1879
+ "loss": 0.6786,
1880
+ "step": 3120
1881
+ },
1882
+ {
1883
+ "epoch": 0.3,
1884
+ "learning_rate": 0.0002,
1885
+ "loss": 0.6802,
1886
+ "step": 3130
1887
+ },
1888
+ {
1889
+ "epoch": 0.3,
1890
+ "learning_rate": 0.0002,
1891
+ "loss": 0.6698,
1892
+ "step": 3140
1893
+ },
1894
+ {
1895
+ "epoch": 0.3,
1896
+ "learning_rate": 0.0002,
1897
+ "loss": 0.678,
1898
+ "step": 3150
1899
+ },
1900
+ {
1901
+ "epoch": 0.3,
1902
+ "learning_rate": 0.0002,
1903
+ "loss": 0.646,
1904
+ "step": 3160
1905
+ },
1906
+ {
1907
+ "epoch": 0.31,
1908
+ "learning_rate": 0.0002,
1909
+ "loss": 0.6518,
1910
+ "step": 3170
1911
+ },
1912
+ {
1913
+ "epoch": 0.31,
1914
+ "learning_rate": 0.0002,
1915
+ "loss": 0.6808,
1916
+ "step": 3180
1917
+ },
1918
+ {
1919
+ "epoch": 0.31,
1920
+ "learning_rate": 0.0002,
1921
+ "loss": 0.6446,
1922
+ "step": 3190
1923
+ },
1924
+ {
1925
+ "epoch": 0.31,
1926
+ "learning_rate": 0.0002,
1927
+ "loss": 0.6747,
1928
+ "step": 3200
1929
+ },
1930
+ {
1931
+ "epoch": 0.31,
1932
+ "learning_rate": 0.0002,
1933
+ "loss": 0.6614,
1934
+ "step": 3210
1935
+ },
1936
+ {
1937
+ "epoch": 0.31,
1938
+ "learning_rate": 0.0002,
1939
+ "loss": 0.6425,
1940
+ "step": 3220
1941
+ },
1942
+ {
1943
+ "epoch": 0.31,
1944
+ "learning_rate": 0.0002,
1945
+ "loss": 0.6641,
1946
+ "step": 3230
1947
+ },
1948
+ {
1949
+ "epoch": 0.31,
1950
+ "learning_rate": 0.0002,
1951
+ "loss": 0.6637,
1952
+ "step": 3240
1953
+ },
1954
+ {
1955
+ "epoch": 0.31,
1956
+ "learning_rate": 0.0002,
1957
+ "loss": 0.6187,
1958
+ "step": 3250
1959
+ },
1960
+ {
1961
+ "epoch": 0.31,
1962
+ "learning_rate": 0.0002,
1963
+ "loss": 0.6592,
1964
+ "step": 3260
1965
+ },
1966
+ {
1967
+ "epoch": 0.32,
1968
+ "learning_rate": 0.0002,
1969
+ "loss": 0.624,
1970
+ "step": 3270
1971
+ },
1972
+ {
1973
+ "epoch": 0.32,
1974
+ "learning_rate": 0.0002,
1975
+ "loss": 0.6945,
1976
+ "step": 3280
1977
+ },
1978
+ {
1979
+ "epoch": 0.32,
1980
+ "learning_rate": 0.0002,
1981
+ "loss": 0.6901,
1982
+ "step": 3290
1983
+ },
1984
+ {
1985
+ "epoch": 0.32,
1986
+ "learning_rate": 0.0002,
1987
+ "loss": 0.6502,
1988
+ "step": 3300
1989
+ },
1990
+ {
1991
+ "epoch": 0.32,
1992
+ "learning_rate": 0.0002,
1993
+ "loss": 0.685,
1994
+ "step": 3310
1995
+ },
1996
+ {
1997
+ "epoch": 0.32,
1998
+ "learning_rate": 0.0002,
1999
+ "loss": 0.6845,
2000
+ "step": 3320
2001
+ },
2002
+ {
2003
+ "epoch": 0.32,
2004
+ "learning_rate": 0.0002,
2005
+ "loss": 0.6993,
2006
+ "step": 3330
2007
+ },
2008
+ {
2009
+ "epoch": 0.32,
2010
+ "learning_rate": 0.0002,
2011
+ "loss": 0.6537,
2012
+ "step": 3340
2013
+ },
2014
+ {
2015
+ "epoch": 0.32,
2016
+ "learning_rate": 0.0002,
2017
+ "loss": 0.665,
2018
+ "step": 3350
2019
+ },
2020
+ {
2021
+ "epoch": 0.32,
2022
+ "learning_rate": 0.0002,
2023
+ "loss": 0.6327,
2024
+ "step": 3360
2025
+ },
2026
+ {
2027
+ "epoch": 0.32,
2028
+ "learning_rate": 0.0002,
2029
+ "loss": 0.6417,
2030
+ "step": 3370
2031
+ },
2032
+ {
2033
+ "epoch": 0.33,
2034
+ "learning_rate": 0.0002,
2035
+ "loss": 0.6702,
2036
+ "step": 3380
2037
+ },
2038
+ {
2039
+ "epoch": 0.33,
2040
+ "learning_rate": 0.0002,
2041
+ "loss": 0.6551,
2042
+ "step": 3390
2043
+ },
2044
+ {
2045
+ "epoch": 0.33,
2046
+ "learning_rate": 0.0002,
2047
+ "loss": 0.6807,
2048
+ "step": 3400
2049
+ },
2050
+ {
2051
+ "epoch": 0.33,
2052
+ "learning_rate": 0.0002,
2053
+ "loss": 0.6569,
2054
+ "step": 3410
2055
+ },
2056
+ {
2057
+ "epoch": 0.33,
2058
+ "learning_rate": 0.0002,
2059
+ "loss": 0.6455,
2060
+ "step": 3420
2061
+ },
2062
+ {
2063
+ "epoch": 0.33,
2064
+ "learning_rate": 0.0002,
2065
+ "loss": 0.6242,
2066
+ "step": 3430
2067
+ },
2068
+ {
2069
+ "epoch": 0.33,
2070
+ "learning_rate": 0.0002,
2071
+ "loss": 0.6415,
2072
+ "step": 3440
2073
+ },
2074
+ {
2075
+ "epoch": 0.33,
2076
+ "learning_rate": 0.0002,
2077
+ "loss": 0.6395,
2078
+ "step": 3450
2079
+ },
2080
+ {
2081
+ "epoch": 0.33,
2082
+ "learning_rate": 0.0002,
2083
+ "loss": 0.6561,
2084
+ "step": 3460
2085
+ },
2086
+ {
2087
+ "epoch": 0.33,
2088
+ "learning_rate": 0.0002,
2089
+ "loss": 0.6292,
2090
+ "step": 3470
2091
+ },
2092
+ {
2093
+ "epoch": 0.34,
2094
+ "learning_rate": 0.0002,
2095
+ "loss": 0.6347,
2096
+ "step": 3480
2097
+ },
2098
+ {
2099
+ "epoch": 0.34,
2100
+ "learning_rate": 0.0002,
2101
+ "loss": 0.6301,
2102
+ "step": 3490
2103
+ },
2104
+ {
2105
+ "epoch": 0.34,
2106
+ "learning_rate": 0.0002,
2107
+ "loss": 0.6155,
2108
+ "step": 3500
2109
+ },
2110
+ {
2111
+ "epoch": 0.34,
2112
+ "learning_rate": 0.0002,
2113
+ "loss": 0.6512,
2114
+ "step": 3510
2115
+ },
2116
+ {
2117
+ "epoch": 0.34,
2118
+ "learning_rate": 0.0002,
2119
+ "loss": 0.6486,
2120
+ "step": 3520
2121
+ },
2122
+ {
2123
+ "epoch": 0.34,
2124
+ "learning_rate": 0.0002,
2125
+ "loss": 0.6402,
2126
+ "step": 3530
2127
+ },
2128
+ {
2129
+ "epoch": 0.34,
2130
+ "learning_rate": 0.0002,
2131
+ "loss": 0.6545,
2132
+ "step": 3540
2133
+ },
2134
+ {
2135
+ "epoch": 0.34,
2136
+ "learning_rate": 0.0002,
2137
+ "loss": 0.6461,
2138
+ "step": 3550
2139
+ },
2140
+ {
2141
+ "epoch": 0.34,
2142
+ "learning_rate": 0.0002,
2143
+ "loss": 0.6368,
2144
+ "step": 3560
2145
+ },
2146
+ {
2147
+ "epoch": 0.34,
2148
+ "learning_rate": 0.0002,
2149
+ "loss": 0.6517,
2150
+ "step": 3570
2151
+ },
2152
+ {
2153
+ "epoch": 0.34,
2154
+ "learning_rate": 0.0002,
2155
+ "loss": 0.6372,
2156
+ "step": 3580
2157
+ },
2158
+ {
2159
+ "epoch": 0.35,
2160
+ "learning_rate": 0.0002,
2161
+ "loss": 0.6551,
2162
+ "step": 3590
2163
+ },
2164
+ {
2165
+ "epoch": 0.35,
2166
+ "learning_rate": 0.0002,
2167
+ "loss": 0.6424,
2168
+ "step": 3600
2169
+ },
2170
+ {
2171
+ "epoch": 0.35,
2172
+ "learning_rate": 0.0002,
2173
+ "loss": 0.627,
2174
+ "step": 3610
2175
+ },
2176
+ {
2177
+ "epoch": 0.35,
2178
+ "learning_rate": 0.0002,
2179
+ "loss": 0.6739,
2180
+ "step": 3620
2181
+ },
2182
+ {
2183
+ "epoch": 0.35,
2184
+ "learning_rate": 0.0002,
2185
+ "loss": 0.6344,
2186
+ "step": 3630
2187
+ },
2188
+ {
2189
+ "epoch": 0.35,
2190
+ "learning_rate": 0.0002,
2191
+ "loss": 0.6328,
2192
+ "step": 3640
2193
+ },
2194
+ {
2195
+ "epoch": 0.35,
2196
+ "learning_rate": 0.0002,
2197
+ "loss": 0.6435,
2198
+ "step": 3650
2199
+ },
2200
+ {
2201
+ "epoch": 0.35,
2202
+ "learning_rate": 0.0002,
2203
+ "loss": 0.6822,
2204
+ "step": 3660
2205
+ },
2206
+ {
2207
+ "epoch": 0.35,
2208
+ "learning_rate": 0.0002,
2209
+ "loss": 0.6217,
2210
+ "step": 3670
2211
+ },
2212
+ {
2213
+ "epoch": 0.35,
2214
+ "learning_rate": 0.0002,
2215
+ "loss": 0.618,
2216
+ "step": 3680
2217
+ },
2218
+ {
2219
+ "epoch": 0.36,
2220
+ "learning_rate": 0.0002,
2221
+ "loss": 0.6594,
2222
+ "step": 3690
2223
+ },
2224
+ {
2225
+ "epoch": 0.36,
2226
+ "learning_rate": 0.0002,
2227
+ "loss": 0.6587,
2228
+ "step": 3700
2229
+ },
2230
+ {
2231
+ "epoch": 0.36,
2232
+ "learning_rate": 0.0002,
2233
+ "loss": 0.6671,
2234
+ "step": 3710
2235
+ },
2236
+ {
2237
+ "epoch": 0.36,
2238
+ "learning_rate": 0.0002,
2239
+ "loss": 0.6364,
2240
+ "step": 3720
2241
+ },
2242
+ {
2243
+ "epoch": 0.36,
2244
+ "learning_rate": 0.0002,
2245
+ "loss": 0.6266,
2246
+ "step": 3730
2247
+ },
2248
+ {
2249
+ "epoch": 0.36,
2250
+ "learning_rate": 0.0002,
2251
+ "loss": 0.6544,
2252
+ "step": 3740
2253
+ },
2254
+ {
2255
+ "epoch": 0.36,
2256
+ "learning_rate": 0.0002,
2257
+ "loss": 0.6445,
2258
+ "step": 3750
2259
+ },
2260
+ {
2261
+ "epoch": 0.36,
2262
+ "learning_rate": 0.0002,
2263
+ "loss": 0.6166,
2264
+ "step": 3760
2265
+ },
2266
+ {
2267
+ "epoch": 0.36,
2268
+ "learning_rate": 0.0002,
2269
+ "loss": 0.6462,
2270
+ "step": 3770
2271
+ },
2272
+ {
2273
+ "epoch": 0.36,
2274
+ "learning_rate": 0.0002,
2275
+ "loss": 0.6353,
2276
+ "step": 3780
2277
+ },
2278
+ {
2279
+ "epoch": 0.37,
2280
+ "learning_rate": 0.0002,
2281
+ "loss": 0.6381,
2282
+ "step": 3790
2283
+ },
2284
+ {
2285
+ "epoch": 0.37,
2286
+ "learning_rate": 0.0002,
2287
+ "loss": 0.6474,
2288
+ "step": 3800
2289
+ },
2290
+ {
2291
+ "epoch": 0.37,
2292
+ "learning_rate": 0.0002,
2293
+ "loss": 0.641,
2294
+ "step": 3810
2295
+ },
2296
+ {
2297
+ "epoch": 0.37,
2298
+ "learning_rate": 0.0002,
2299
+ "loss": 0.6506,
2300
+ "step": 3820
2301
+ },
2302
+ {
2303
+ "epoch": 0.37,
2304
+ "learning_rate": 0.0002,
2305
+ "loss": 0.6301,
2306
+ "step": 3830
2307
+ },
2308
+ {
2309
+ "epoch": 0.37,
2310
+ "learning_rate": 0.0002,
2311
+ "loss": 0.6408,
2312
+ "step": 3840
2313
+ },
2314
+ {
2315
+ "epoch": 0.37,
2316
+ "learning_rate": 0.0002,
2317
+ "loss": 0.65,
2318
+ "step": 3850
2319
+ },
2320
+ {
2321
+ "epoch": 0.37,
2322
+ "learning_rate": 0.0002,
2323
+ "loss": 0.6383,
2324
+ "step": 3860
2325
+ },
2326
+ {
2327
+ "epoch": 0.37,
2328
+ "learning_rate": 0.0002,
2329
+ "loss": 0.6274,
2330
+ "step": 3870
2331
+ },
2332
+ {
2333
+ "epoch": 0.37,
2334
+ "learning_rate": 0.0002,
2335
+ "loss": 0.6417,
2336
+ "step": 3880
2337
+ },
2338
+ {
2339
+ "epoch": 0.37,
2340
+ "learning_rate": 0.0002,
2341
+ "loss": 0.6476,
2342
+ "step": 3890
2343
+ },
2344
+ {
2345
+ "epoch": 0.38,
2346
+ "learning_rate": 0.0002,
2347
+ "loss": 0.6334,
2348
+ "step": 3900
2349
+ },
2350
+ {
2351
+ "epoch": 0.38,
2352
+ "learning_rate": 0.0002,
2353
+ "loss": 0.6893,
2354
+ "step": 3910
2355
+ },
2356
+ {
2357
+ "epoch": 0.38,
2358
+ "learning_rate": 0.0002,
2359
+ "loss": 0.6548,
2360
+ "step": 3920
2361
+ },
2362
+ {
2363
+ "epoch": 0.38,
2364
+ "learning_rate": 0.0002,
2365
+ "loss": 0.6401,
2366
+ "step": 3930
2367
+ },
2368
+ {
2369
+ "epoch": 0.38,
2370
+ "learning_rate": 0.0002,
2371
+ "loss": 0.5962,
2372
+ "step": 3940
2373
+ },
2374
+ {
2375
+ "epoch": 0.38,
2376
+ "learning_rate": 0.0002,
2377
+ "loss": 0.6673,
2378
+ "step": 3950
2379
+ },
2380
+ {
2381
+ "epoch": 0.38,
2382
+ "learning_rate": 0.0002,
2383
+ "loss": 0.6749,
2384
+ "step": 3960
2385
+ },
2386
+ {
2387
+ "epoch": 0.38,
2388
+ "learning_rate": 0.0002,
2389
+ "loss": 0.6276,
2390
+ "step": 3970
2391
+ },
2392
+ {
2393
+ "epoch": 0.38,
2394
+ "learning_rate": 0.0002,
2395
+ "loss": 0.6652,
2396
+ "step": 3980
2397
+ },
2398
+ {
2399
+ "epoch": 0.38,
2400
+ "learning_rate": 0.0002,
2401
+ "loss": 0.6337,
2402
+ "step": 3990
2403
+ },
2404
+ {
2405
+ "epoch": 0.39,
2406
+ "learning_rate": 0.0002,
2407
+ "loss": 0.6439,
2408
+ "step": 4000
2409
+ },
2410
+ {
2411
+ "epoch": 0.39,
2412
+ "learning_rate": 0.0002,
2413
+ "loss": 0.6533,
2414
+ "step": 4010
2415
+ },
2416
+ {
2417
+ "epoch": 0.39,
2418
+ "learning_rate": 0.0002,
2419
+ "loss": 0.6455,
2420
+ "step": 4020
2421
+ },
2422
+ {
2423
+ "epoch": 0.39,
2424
+ "learning_rate": 0.0002,
2425
+ "loss": 0.6745,
2426
+ "step": 4030
2427
+ },
2428
+ {
2429
+ "epoch": 0.39,
2430
+ "learning_rate": 0.0002,
2431
+ "loss": 0.6481,
2432
+ "step": 4040
2433
+ },
2434
+ {
2435
+ "epoch": 0.39,
2436
+ "learning_rate": 0.0002,
2437
+ "loss": 0.6416,
2438
+ "step": 4050
2439
+ },
2440
+ {
2441
+ "epoch": 0.39,
2442
+ "learning_rate": 0.0002,
2443
+ "loss": 0.6227,
2444
+ "step": 4060
2445
+ },
2446
+ {
2447
+ "epoch": 0.39,
2448
+ "learning_rate": 0.0002,
2449
+ "loss": 0.6307,
2450
+ "step": 4070
2451
+ },
2452
+ {
2453
+ "epoch": 0.39,
2454
+ "learning_rate": 0.0002,
2455
+ "loss": 0.6634,
2456
+ "step": 4080
2457
+ },
2458
+ {
2459
+ "epoch": 0.39,
2460
+ "learning_rate": 0.0002,
2461
+ "loss": 0.6468,
2462
+ "step": 4090
2463
+ },
2464
+ {
2465
+ "epoch": 0.39,
2466
+ "learning_rate": 0.0002,
2467
+ "loss": 0.6532,
2468
+ "step": 4100
2469
+ },
2470
+ {
2471
+ "epoch": 0.4,
2472
+ "learning_rate": 0.0002,
2473
+ "loss": 0.6223,
2474
+ "step": 4110
2475
+ },
2476
+ {
2477
+ "epoch": 0.4,
2478
+ "learning_rate": 0.0002,
2479
+ "loss": 0.6036,
2480
+ "step": 4120
2481
+ },
2482
+ {
2483
+ "epoch": 0.4,
2484
+ "learning_rate": 0.0002,
2485
+ "loss": 0.6236,
2486
+ "step": 4130
2487
+ },
2488
+ {
2489
+ "epoch": 0.4,
2490
+ "learning_rate": 0.0002,
2491
+ "loss": 0.6307,
2492
+ "step": 4140
2493
+ },
2494
+ {
2495
+ "epoch": 0.4,
2496
+ "learning_rate": 0.0002,
2497
+ "loss": 0.6681,
2498
+ "step": 4150
2499
+ },
2500
+ {
2501
+ "epoch": 0.4,
2502
+ "learning_rate": 0.0002,
2503
+ "loss": 0.6631,
2504
+ "step": 4160
2505
+ },
2506
+ {
2507
+ "epoch": 0.4,
2508
+ "learning_rate": 0.0002,
2509
+ "loss": 0.6394,
2510
+ "step": 4170
2511
+ },
2512
+ {
2513
+ "epoch": 0.4,
2514
+ "learning_rate": 0.0002,
2515
+ "loss": 0.6262,
2516
+ "step": 4180
2517
+ },
2518
+ {
2519
+ "epoch": 0.4,
2520
+ "learning_rate": 0.0002,
2521
+ "loss": 0.6236,
2522
+ "step": 4190
2523
+ },
2524
+ {
2525
+ "epoch": 0.4,
2526
+ "learning_rate": 0.0002,
2527
+ "loss": 0.6047,
2528
+ "step": 4200
2529
+ },
2530
+ {
2531
+ "epoch": 0.41,
2532
+ "learning_rate": 0.0002,
2533
+ "loss": 0.6444,
2534
+ "step": 4210
2535
+ },
2536
+ {
2537
+ "epoch": 0.41,
2538
+ "learning_rate": 0.0002,
2539
+ "loss": 0.6514,
2540
+ "step": 4220
2541
+ },
2542
+ {
2543
+ "epoch": 0.41,
2544
+ "learning_rate": 0.0002,
2545
+ "loss": 0.6529,
2546
+ "step": 4230
2547
+ },
2548
+ {
2549
+ "epoch": 0.41,
2550
+ "learning_rate": 0.0002,
2551
+ "loss": 0.6567,
2552
+ "step": 4240
2553
+ },
2554
+ {
2555
+ "epoch": 0.41,
2556
+ "learning_rate": 0.0002,
2557
+ "loss": 0.6536,
2558
+ "step": 4250
2559
+ },
2560
+ {
2561
+ "epoch": 0.41,
2562
+ "learning_rate": 0.0002,
2563
+ "loss": 0.6426,
2564
+ "step": 4260
2565
+ },
2566
+ {
2567
+ "epoch": 0.41,
2568
+ "learning_rate": 0.0002,
2569
+ "loss": 0.6396,
2570
+ "step": 4270
2571
+ },
2572
+ {
2573
+ "epoch": 0.41,
2574
+ "learning_rate": 0.0002,
2575
+ "loss": 0.6398,
2576
+ "step": 4280
2577
+ },
2578
+ {
2579
+ "epoch": 0.41,
2580
+ "learning_rate": 0.0002,
2581
+ "loss": 0.6604,
2582
+ "step": 4290
2583
+ },
2584
+ {
2585
+ "epoch": 0.41,
2586
+ "learning_rate": 0.0002,
2587
+ "loss": 0.6288,
2588
+ "step": 4300
2589
+ },
2590
+ {
2591
+ "epoch": 0.42,
2592
+ "learning_rate": 0.0002,
2593
+ "loss": 0.6277,
2594
+ "step": 4310
2595
+ },
2596
+ {
2597
+ "epoch": 0.42,
2598
+ "learning_rate": 0.0002,
2599
+ "loss": 0.6434,
2600
+ "step": 4320
2601
+ },
2602
+ {
2603
+ "epoch": 0.42,
2604
+ "learning_rate": 0.0002,
2605
+ "loss": 0.6302,
2606
+ "step": 4330
2607
+ },
2608
+ {
2609
+ "epoch": 0.42,
2610
+ "learning_rate": 0.0002,
2611
+ "loss": 0.6415,
2612
+ "step": 4340
2613
+ },
2614
+ {
2615
+ "epoch": 0.42,
2616
+ "learning_rate": 0.0002,
2617
+ "loss": 0.6332,
2618
+ "step": 4350
2619
+ },
2620
+ {
2621
+ "epoch": 0.42,
2622
+ "learning_rate": 0.0002,
2623
+ "loss": 0.6346,
2624
+ "step": 4360
2625
+ },
2626
+ {
2627
+ "epoch": 0.42,
2628
+ "learning_rate": 0.0002,
2629
+ "loss": 0.6435,
2630
+ "step": 4370
2631
+ },
2632
+ {
2633
+ "epoch": 0.42,
2634
+ "learning_rate": 0.0002,
2635
+ "loss": 0.6205,
2636
+ "step": 4380
2637
+ },
2638
+ {
2639
+ "epoch": 0.42,
2640
+ "learning_rate": 0.0002,
2641
+ "loss": 0.6364,
2642
+ "step": 4390
2643
+ },
2644
+ {
2645
+ "epoch": 0.42,
2646
+ "learning_rate": 0.0002,
2647
+ "loss": 0.6464,
2648
+ "step": 4400
2649
+ },
2650
+ {
2651
+ "epoch": 0.42,
2652
+ "learning_rate": 0.0002,
2653
+ "loss": 0.6111,
2654
+ "step": 4410
2655
+ },
2656
+ {
2657
+ "epoch": 0.43,
2658
+ "learning_rate": 0.0002,
2659
+ "loss": 0.6425,
2660
+ "step": 4420
2661
+ },
2662
+ {
2663
+ "epoch": 0.43,
2664
+ "learning_rate": 0.0002,
2665
+ "loss": 0.6523,
2666
+ "step": 4430
2667
+ },
2668
+ {
2669
+ "epoch": 0.43,
2670
+ "learning_rate": 0.0002,
2671
+ "loss": 0.6396,
2672
+ "step": 4440
2673
+ },
2674
+ {
2675
+ "epoch": 0.43,
2676
+ "learning_rate": 0.0002,
2677
+ "loss": 0.6391,
2678
+ "step": 4450
2679
+ },
2680
+ {
2681
+ "epoch": 0.43,
2682
+ "learning_rate": 0.0002,
2683
+ "loss": 0.6433,
2684
+ "step": 4460
2685
+ },
2686
+ {
2687
+ "epoch": 0.43,
2688
+ "learning_rate": 0.0002,
2689
+ "loss": 0.6625,
2690
+ "step": 4470
2691
+ },
2692
+ {
2693
+ "epoch": 0.43,
2694
+ "learning_rate": 0.0002,
2695
+ "loss": 0.6274,
2696
+ "step": 4480
2697
+ },
2698
+ {
2699
+ "epoch": 0.43,
2700
+ "learning_rate": 0.0002,
2701
+ "loss": 0.6719,
2702
+ "step": 4490
2703
+ },
2704
+ {
2705
+ "epoch": 0.43,
2706
+ "learning_rate": 0.0002,
2707
+ "loss": 0.6659,
2708
+ "step": 4500
2709
+ },
2710
+ {
2711
+ "epoch": 0.43,
2712
+ "learning_rate": 0.0002,
2713
+ "loss": 0.6454,
2714
+ "step": 4510
2715
+ },
2716
+ {
2717
+ "epoch": 0.44,
2718
+ "learning_rate": 0.0002,
2719
+ "loss": 0.6417,
2720
+ "step": 4520
2721
+ },
2722
+ {
2723
+ "epoch": 0.44,
2724
+ "learning_rate": 0.0002,
2725
+ "loss": 0.6159,
2726
+ "step": 4530
2727
+ },
2728
+ {
2729
+ "epoch": 0.44,
2730
+ "learning_rate": 0.0002,
2731
+ "loss": 0.6186,
2732
+ "step": 4540
2733
+ },
2734
+ {
2735
+ "epoch": 0.44,
2736
+ "learning_rate": 0.0002,
2737
+ "loss": 0.6328,
2738
+ "step": 4550
2739
+ },
2740
+ {
2741
+ "epoch": 0.44,
2742
+ "learning_rate": 0.0002,
2743
+ "loss": 0.6494,
2744
+ "step": 4560
2745
+ },
2746
+ {
2747
+ "epoch": 0.44,
2748
+ "learning_rate": 0.0002,
2749
+ "loss": 0.6237,
2750
+ "step": 4570
2751
+ },
2752
+ {
2753
+ "epoch": 0.44,
2754
+ "learning_rate": 0.0002,
2755
+ "loss": 0.6427,
2756
+ "step": 4580
2757
+ },
2758
+ {
2759
+ "epoch": 0.44,
2760
+ "learning_rate": 0.0002,
2761
+ "loss": 0.642,
2762
+ "step": 4590
2763
+ },
2764
+ {
2765
+ "epoch": 0.44,
2766
+ "learning_rate": 0.0002,
2767
+ "loss": 0.6275,
2768
+ "step": 4600
2769
+ },
2770
+ {
2771
+ "epoch": 0.44,
2772
+ "learning_rate": 0.0002,
2773
+ "loss": 0.637,
2774
+ "step": 4610
2775
+ },
2776
+ {
2777
+ "epoch": 0.45,
2778
+ "learning_rate": 0.0002,
2779
+ "loss": 0.6416,
2780
+ "step": 4620
2781
+ },
2782
+ {
2783
+ "epoch": 0.45,
2784
+ "learning_rate": 0.0002,
2785
+ "loss": 0.6332,
2786
+ "step": 4630
2787
+ },
2788
+ {
2789
+ "epoch": 0.45,
2790
+ "learning_rate": 0.0002,
2791
+ "loss": 0.6515,
2792
+ "step": 4640
2793
+ },
2794
+ {
2795
+ "epoch": 0.45,
2796
+ "learning_rate": 0.0002,
2797
+ "loss": 0.6551,
2798
+ "step": 4650
2799
+ },
2800
+ {
2801
+ "epoch": 0.45,
2802
+ "learning_rate": 0.0002,
2803
+ "loss": 0.6643,
2804
+ "step": 4660
2805
+ },
2806
+ {
2807
+ "epoch": 0.45,
2808
+ "learning_rate": 0.0002,
2809
+ "loss": 0.6317,
2810
+ "step": 4670
2811
+ },
2812
+ {
2813
+ "epoch": 0.45,
2814
+ "learning_rate": 0.0002,
2815
+ "loss": 0.6392,
2816
+ "step": 4680
2817
+ },
2818
+ {
2819
+ "epoch": 0.45,
2820
+ "learning_rate": 0.0002,
2821
+ "loss": 0.6447,
2822
+ "step": 4690
2823
+ },
2824
+ {
2825
+ "epoch": 0.45,
2826
+ "learning_rate": 0.0002,
2827
+ "loss": 0.6181,
2828
+ "step": 4700
2829
+ },
2830
+ {
2831
+ "epoch": 0.45,
2832
+ "learning_rate": 0.0002,
2833
+ "loss": 0.6369,
2834
+ "step": 4710
2835
+ },
2836
+ {
2837
+ "epoch": 0.45,
2838
+ "learning_rate": 0.0002,
2839
+ "loss": 0.6565,
2840
+ "step": 4720
2841
+ },
2842
+ {
2843
+ "epoch": 0.46,
2844
+ "learning_rate": 0.0002,
2845
+ "loss": 0.6349,
2846
+ "step": 4730
2847
+ },
2848
+ {
2849
+ "epoch": 0.46,
2850
+ "learning_rate": 0.0002,
2851
+ "loss": 0.6385,
2852
+ "step": 4740
2853
+ },
2854
+ {
2855
+ "epoch": 0.46,
2856
+ "learning_rate": 0.0002,
2857
+ "loss": 0.5835,
2858
+ "step": 4750
2859
+ },
2860
+ {
2861
+ "epoch": 0.46,
2862
+ "learning_rate": 0.0002,
2863
+ "loss": 0.6195,
2864
+ "step": 4760
2865
+ },
2866
+ {
2867
+ "epoch": 0.46,
2868
+ "learning_rate": 0.0002,
2869
+ "loss": 0.6286,
2870
+ "step": 4770
2871
+ },
2872
+ {
2873
+ "epoch": 0.46,
2874
+ "learning_rate": 0.0002,
2875
+ "loss": 0.6153,
2876
+ "step": 4780
2877
+ },
2878
+ {
2879
+ "epoch": 0.46,
2880
+ "learning_rate": 0.0002,
2881
+ "loss": 0.6517,
2882
+ "step": 4790
2883
+ },
2884
+ {
2885
+ "epoch": 0.46,
2886
+ "learning_rate": 0.0002,
2887
+ "loss": 0.6004,
2888
+ "step": 4800
2889
+ },
2890
+ {
2891
+ "epoch": 0.46,
2892
+ "learning_rate": 0.0002,
2893
+ "loss": 0.6299,
2894
+ "step": 4810
2895
+ },
2896
+ {
2897
+ "epoch": 0.46,
2898
+ "learning_rate": 0.0002,
2899
+ "loss": 0.6369,
2900
+ "step": 4820
2901
+ },
2902
+ {
2903
+ "epoch": 0.47,
2904
+ "learning_rate": 0.0002,
2905
+ "loss": 0.6403,
2906
+ "step": 4830
2907
+ },
2908
+ {
2909
+ "epoch": 0.47,
2910
+ "learning_rate": 0.0002,
2911
+ "loss": 0.6443,
2912
+ "step": 4840
2913
+ },
2914
+ {
2915
+ "epoch": 0.47,
2916
+ "learning_rate": 0.0002,
2917
+ "loss": 0.6139,
2918
+ "step": 4850
2919
+ },
2920
+ {
2921
+ "epoch": 0.47,
2922
+ "learning_rate": 0.0002,
2923
+ "loss": 0.6228,
2924
+ "step": 4860
2925
+ },
2926
+ {
2927
+ "epoch": 0.47,
2928
+ "learning_rate": 0.0002,
2929
+ "loss": 0.6212,
2930
+ "step": 4870
2931
+ },
2932
+ {
2933
+ "epoch": 0.47,
2934
+ "learning_rate": 0.0002,
2935
+ "loss": 0.6039,
2936
+ "step": 4880
2937
+ },
2938
+ {
2939
+ "epoch": 0.47,
2940
+ "learning_rate": 0.0002,
2941
+ "loss": 0.6269,
2942
+ "step": 4890
2943
+ },
2944
+ {
2945
+ "epoch": 0.47,
2946
+ "learning_rate": 0.0002,
2947
+ "loss": 0.6302,
2948
+ "step": 4900
2949
+ },
2950
+ {
2951
+ "epoch": 0.47,
2952
+ "learning_rate": 0.0002,
2953
+ "loss": 0.6392,
2954
+ "step": 4910
2955
+ },
2956
+ {
2957
+ "epoch": 0.47,
2958
+ "learning_rate": 0.0002,
2959
+ "loss": 0.6261,
2960
+ "step": 4920
2961
+ },
2962
+ {
2963
+ "epoch": 0.47,
2964
+ "learning_rate": 0.0002,
2965
+ "loss": 0.6201,
2966
+ "step": 4930
2967
+ },
2968
+ {
2969
+ "epoch": 0.48,
2970
+ "learning_rate": 0.0002,
2971
+ "loss": 0.6393,
2972
+ "step": 4940
2973
+ },
2974
+ {
2975
+ "epoch": 0.48,
2976
+ "learning_rate": 0.0002,
2977
+ "loss": 0.6174,
2978
+ "step": 4950
2979
+ },
2980
+ {
2981
+ "epoch": 0.48,
2982
+ "learning_rate": 0.0002,
2983
+ "loss": 0.6158,
2984
+ "step": 4960
2985
+ },
2986
+ {
2987
+ "epoch": 0.48,
2988
+ "learning_rate": 0.0002,
2989
+ "loss": 0.6317,
2990
+ "step": 4970
2991
+ },
2992
+ {
2993
+ "epoch": 0.48,
2994
+ "learning_rate": 0.0002,
2995
+ "loss": 0.6233,
2996
+ "step": 4980
2997
+ },
2998
+ {
2999
+ "epoch": 0.48,
3000
+ "learning_rate": 0.0002,
3001
+ "loss": 0.6733,
3002
+ "step": 4990
3003
+ },
3004
+ {
3005
+ "epoch": 0.48,
3006
+ "learning_rate": 0.0002,
3007
+ "loss": 0.6473,
3008
+ "step": 5000
3009
+ },
3010
+ {
3011
+ "epoch": 0.48,
3012
+ "learning_rate": 0.0002,
3013
+ "loss": 0.6512,
3014
+ "step": 5010
3015
+ },
3016
+ {
3017
+ "epoch": 0.48,
3018
+ "learning_rate": 0.0002,
3019
+ "loss": 0.6293,
3020
+ "step": 5020
3021
+ },
3022
+ {
3023
+ "epoch": 0.48,
3024
+ "learning_rate": 0.0002,
3025
+ "loss": 0.6291,
3026
+ "step": 5030
3027
+ },
3028
+ {
3029
+ "epoch": 0.49,
3030
+ "learning_rate": 0.0002,
3031
+ "loss": 0.6375,
3032
+ "step": 5040
3033
+ },
3034
+ {
3035
+ "epoch": 0.49,
3036
+ "learning_rate": 0.0002,
3037
+ "loss": 0.6087,
3038
+ "step": 5050
3039
+ },
3040
+ {
3041
+ "epoch": 0.49,
3042
+ "learning_rate": 0.0002,
3043
+ "loss": 0.6297,
3044
+ "step": 5060
3045
+ },
3046
+ {
3047
+ "epoch": 0.49,
3048
+ "learning_rate": 0.0002,
3049
+ "loss": 0.6124,
3050
+ "step": 5070
3051
+ },
3052
+ {
3053
+ "epoch": 0.49,
3054
+ "learning_rate": 0.0002,
3055
+ "loss": 0.6128,
3056
+ "step": 5080
3057
+ },
3058
+ {
3059
+ "epoch": 0.49,
3060
+ "learning_rate": 0.0002,
3061
+ "loss": 0.644,
3062
+ "step": 5090
3063
+ },
3064
+ {
3065
+ "epoch": 0.49,
3066
+ "learning_rate": 0.0002,
3067
+ "loss": 0.613,
3068
+ "step": 5100
3069
+ },
3070
+ {
3071
+ "epoch": 0.49,
3072
+ "learning_rate": 0.0002,
3073
+ "loss": 0.601,
3074
+ "step": 5110
3075
+ },
3076
+ {
3077
+ "epoch": 0.49,
3078
+ "learning_rate": 0.0002,
3079
+ "loss": 0.6441,
3080
+ "step": 5120
3081
+ },
3082
+ {
3083
+ "epoch": 0.49,
3084
+ "learning_rate": 0.0002,
3085
+ "loss": 0.6187,
3086
+ "step": 5130
3087
+ },
3088
+ {
3089
+ "epoch": 0.5,
3090
+ "learning_rate": 0.0002,
3091
+ "loss": 0.6355,
3092
+ "step": 5140
3093
+ },
3094
+ {
3095
+ "epoch": 0.5,
3096
+ "learning_rate": 0.0002,
3097
+ "loss": 0.6671,
3098
+ "step": 5150
3099
+ },
3100
+ {
3101
+ "epoch": 0.5,
3102
+ "learning_rate": 0.0002,
3103
+ "loss": 0.6412,
3104
+ "step": 5160
3105
+ },
3106
+ {
3107
+ "epoch": 0.5,
3108
+ "learning_rate": 0.0002,
3109
+ "loss": 0.6154,
3110
+ "step": 5170
3111
+ },
3112
+ {
3113
+ "epoch": 0.5,
3114
+ "learning_rate": 0.0002,
3115
+ "loss": 0.6195,
3116
+ "step": 5180
3117
+ },
3118
+ {
3119
+ "epoch": 0.5,
3120
+ "learning_rate": 0.0002,
3121
+ "loss": 0.6174,
3122
+ "step": 5190
3123
+ },
3124
+ {
3125
+ "epoch": 0.5,
3126
+ "learning_rate": 0.0002,
3127
+ "loss": 0.6402,
3128
+ "step": 5200
3129
+ },
3130
+ {
3131
+ "epoch": 0.5,
3132
+ "learning_rate": 0.0002,
3133
+ "loss": 0.6188,
3134
+ "step": 5210
3135
+ },
3136
+ {
3137
+ "epoch": 0.5,
3138
+ "learning_rate": 0.0002,
3139
+ "loss": 0.6385,
3140
+ "step": 5220
3141
+ },
3142
+ {
3143
+ "epoch": 0.5,
3144
+ "learning_rate": 0.0002,
3145
+ "loss": 0.6402,
3146
+ "step": 5230
3147
+ },
3148
+ {
3149
+ "epoch": 0.5,
3150
+ "learning_rate": 0.0002,
3151
+ "loss": 0.6057,
3152
+ "step": 5240
3153
+ },
3154
+ {
3155
+ "epoch": 0.51,
3156
+ "learning_rate": 0.0002,
3157
+ "loss": 0.6428,
3158
+ "step": 5250
3159
+ },
3160
+ {
3161
+ "epoch": 0.51,
3162
+ "learning_rate": 0.0002,
3163
+ "loss": 0.6161,
3164
+ "step": 5260
3165
+ },
3166
+ {
3167
+ "epoch": 0.51,
3168
+ "learning_rate": 0.0002,
3169
+ "loss": 0.6552,
3170
+ "step": 5270
3171
+ },
3172
+ {
3173
+ "epoch": 0.51,
3174
+ "learning_rate": 0.0002,
3175
+ "loss": 0.6134,
3176
+ "step": 5280
3177
+ },
3178
+ {
3179
+ "epoch": 0.51,
3180
+ "learning_rate": 0.0002,
3181
+ "loss": 0.6543,
3182
+ "step": 5290
3183
+ },
3184
+ {
3185
+ "epoch": 0.51,
3186
+ "learning_rate": 0.0002,
3187
+ "loss": 0.6499,
3188
+ "step": 5300
3189
+ },
3190
+ {
3191
+ "epoch": 0.51,
3192
+ "learning_rate": 0.0002,
3193
+ "loss": 0.6544,
3194
+ "step": 5310
3195
+ },
3196
+ {
3197
+ "epoch": 0.51,
3198
+ "learning_rate": 0.0002,
3199
+ "loss": 0.6269,
3200
+ "step": 5320
3201
+ },
3202
+ {
3203
+ "epoch": 0.51,
3204
+ "learning_rate": 0.0002,
3205
+ "loss": 0.614,
3206
+ "step": 5330
3207
+ },
3208
+ {
3209
+ "epoch": 0.51,
3210
+ "learning_rate": 0.0002,
3211
+ "loss": 0.6233,
3212
+ "step": 5340
3213
+ },
3214
+ {
3215
+ "epoch": 0.52,
3216
+ "learning_rate": 0.0002,
3217
+ "loss": 0.612,
3218
+ "step": 5350
3219
+ },
3220
+ {
3221
+ "epoch": 0.52,
3222
+ "learning_rate": 0.0002,
3223
+ "loss": 0.6396,
3224
+ "step": 5360
3225
+ },
3226
+ {
3227
+ "epoch": 0.52,
3228
+ "learning_rate": 0.0002,
3229
+ "loss": 0.5731,
3230
+ "step": 5370
3231
+ },
3232
+ {
3233
+ "epoch": 0.52,
3234
+ "learning_rate": 0.0002,
3235
+ "loss": 0.6178,
3236
+ "step": 5380
3237
+ },
3238
+ {
3239
+ "epoch": 0.52,
3240
+ "learning_rate": 0.0002,
3241
+ "loss": 0.6245,
3242
+ "step": 5390
3243
+ },
3244
+ {
3245
+ "epoch": 0.52,
3246
+ "learning_rate": 0.0002,
3247
+ "loss": 0.6131,
3248
+ "step": 5400
3249
+ },
3250
+ {
3251
+ "epoch": 0.52,
3252
+ "learning_rate": 0.0002,
3253
+ "loss": 0.642,
3254
+ "step": 5410
3255
+ },
3256
+ {
3257
+ "epoch": 0.52,
3258
+ "learning_rate": 0.0002,
3259
+ "loss": 0.6237,
3260
+ "step": 5420
3261
+ },
3262
+ {
3263
+ "epoch": 0.52,
3264
+ "learning_rate": 0.0002,
3265
+ "loss": 0.5984,
3266
+ "step": 5430
3267
+ },
3268
+ {
3269
+ "epoch": 0.52,
3270
+ "learning_rate": 0.0002,
3271
+ "loss": 0.6262,
3272
+ "step": 5440
3273
+ },
3274
+ {
3275
+ "epoch": 0.53,
3276
+ "learning_rate": 0.0002,
3277
+ "loss": 0.6288,
3278
+ "step": 5450
3279
+ },
3280
+ {
3281
+ "epoch": 0.53,
3282
+ "learning_rate": 0.0002,
3283
+ "loss": 0.6178,
3284
+ "step": 5460
3285
+ },
3286
+ {
3287
+ "epoch": 0.53,
3288
+ "learning_rate": 0.0002,
3289
+ "loss": 0.6479,
3290
+ "step": 5470
3291
+ },
3292
+ {
3293
+ "epoch": 0.53,
3294
+ "learning_rate": 0.0002,
3295
+ "loss": 0.5907,
3296
+ "step": 5480
3297
+ },
3298
+ {
3299
+ "epoch": 0.53,
3300
+ "learning_rate": 0.0002,
3301
+ "loss": 0.6099,
3302
+ "step": 5490
3303
+ },
3304
+ {
3305
+ "epoch": 0.53,
3306
+ "learning_rate": 0.0002,
3307
+ "loss": 0.6218,
3308
+ "step": 5500
3309
+ }
3310
+ ],
3311
+ "max_steps": 62280,
3312
+ "num_train_epochs": 6,
3313
+ "total_flos": 7.776659814923059e+17,
3314
+ "trial_name": null,
3315
+ "trial_params": null
3316
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:24971545237474f7009b32dd5a9c8e66ddf9e9e869713821a98bfc59f651d63c
3
+ size 6200