Hugging Face's logo

ycchen
/
final-lora-ep5

PEFT
Model card Files Community
ycchen commited on
Commit
67637df
•
1 Parent(s): 9925152

Upload folder using huggingface_hub

Browse files
Files changed (9) hide show
  1. README.md +27 -0
  2. adapter_config.json +79 -0
  3. adapter_model.bin +3 -0
  4. qwen.tiktoken +0 -0
  5. scheduler.pt +3 -0
  6. special_tokens_map.json +3 -0
  7. tokenizer_config.json +11 -0
  8. trainer_state.json +4279 -0
  9. training_args.bin +3 -0
README.md ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: peft
3
+ ---
4
+ ## Training procedure
5
+
6
+
7
+ The following `bitsandbytes` quantization config was used during training:
8
+ - quant_method: gptq
9
+ - bits: 8
10
+ - tokenizer: None
11
+ - dataset: None
12
+ - group_size: 128
13
+ - damp_percent: 0.01
14
+ - desc_act: False
15
+ - sym: True
16
+ - true_sequential: True
17
+ - use_cuda_fp16: False
18
+ - model_seqlen: None
19
+ - block_name_to_quantize: None
20
+ - module_name_preceding_first_block: None
21
+ - batch_size: 1
22
+ - pad_token_id: None
23
+ - disable_exllama: False
24
+ ### Framework versions
25
+
26
+
27
+ - PEFT 0.5.0
adapter_config.json ADDED
@@ -0,0 +1,79 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "auto_mapping": null,
3
+ "base_model_name_or_path": "ycchen/yc-test1",
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": 0.05,
12
+ "modules_to_save": null,
13
+ "peft_type": "LORA",
14
+ "r": 32,
15
+ "revision": null,
16
+ "target_modules": [
17
+ "28.attn.c_attn",
18
+ "28.attn.c_proj",
19
+ "28.mlp.c_proj",
20
+ "28.mlp.w1",
21
+ "28.mlp.w2",
22
+ "29.attn.c_attn",
23
+ "29.attn.c_proj",
24
+ "29.mlp.c_proj",
25
+ "29.mlp.w1",
26
+ "29.mlp.w2",
27
+ "30.attn.c_attn",
28
+ "30.attn.c_proj",
29
+ "30.mlp.c_proj",
30
+ "30.mlp.w1",
31
+ "30.mlp.w2",
32
+ "31.attn.c_attn",
33
+ "31.attn.c_proj",
34
+ "31.mlp.c_proj",
35
+ "31.mlp.w1",
36
+ "31.mlp.w2",
37
+ "32.attn.c_attn",
38
+ "32.attn.c_proj",
39
+ "32.mlp.c_proj",
40
+ "32.mlp.w1",
41
+ "32.mlp.w2",
42
+ "33.attn.c_attn",
43
+ "33.attn.c_proj",
44
+ "33.mlp.c_proj",
45
+ "33.mlp.w1",
46
+ "33.mlp.w2",
47
+ "34.attn.c_attn",
48
+ "34.attn.c_proj",
49
+ "34.mlp.c_proj",
50
+ "34.mlp.w1",
51
+ "34.mlp.w2",
52
+ "35.attn.c_attn",
53
+ "35.attn.c_proj",
54
+ "35.mlp.c_proj",
55
+ "35.mlp.w1",
56
+ "35.mlp.w2",
57
+ "36.attn.c_attn",
58
+ "36.attn.c_proj",
59
+ "36.mlp.c_proj",
60
+ "36.mlp.w1",
61
+ "36.mlp.w2",
62
+ "37.attn.c_attn",
63
+ "37.attn.c_proj",
64
+ "37.mlp.c_proj",
65
+ "37.mlp.w1",
66
+ "37.mlp.w2",
67
+ "38.attn.c_attn",
68
+ "38.attn.c_proj",
69
+ "38.mlp.c_proj",
70
+ "38.mlp.w1",
71
+ "38.mlp.w2",
72
+ "39.attn.c_attn",
73
+ "39.attn.c_proj",
74
+ "39.mlp.c_proj",
75
+ "39.mlp.w1",
76
+ "39.mlp.w2"
77
+ ],
78
+ "task_type": "CAUSAL_LM"
79
+ }
adapter_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8758712b1e134ce440dc7448e50918531e4ca6acf28221371ff236c0292d3dc1
3
+ size 133932461
qwen.tiktoken ADDED
The diff for this file is too large to render. See raw diff
 
scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:35babed89faaebedd8e7a5ae31267b775791ff6d62ae8bd0f003c0f775d15457
3
+ size 563
special_tokens_map.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ {
2
+ "pad_token": "<|extra_21|>"
3
+ }
tokenizer_config.json ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "auto_map": {
3
+ "AutoTokenizer": [
4
+ "Qwen/Qwen-14B--tokenization_qwen.QWenTokenizer",
5
+ null
6
+ ]
7
+ },
8
+ "clean_up_tokenization_spaces": true,
9
+ "model_max_length": 8192,
10
+ "tokenizer_class": "QWenTokenizer"
11
+ }
trainer_state.json ADDED
@@ -0,0 +1,4279 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 5.0,
5
+ "eval_steps": 500,
6
+ "global_step": 710,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.01,
13
+ "learning_rate": 2.938151902453357e-07,
14
+ "loss": 1.5844,
15
+ "step": 1
16
+ },
17
+ {
18
+ "epoch": 0.01,
19
+ "learning_rate": 5.876303804906714e-07,
20
+ "loss": 1.1717,
21
+ "step": 2
22
+ },
23
+ {
24
+ "epoch": 0.02,
25
+ "learning_rate": 8.814455707360072e-07,
26
+ "loss": 1.1709,
27
+ "step": 3
28
+ },
29
+ {
30
+ "epoch": 0.03,
31
+ "learning_rate": 1.1752607609813429e-06,
32
+ "loss": 1.5208,
33
+ "step": 4
34
+ },
35
+ {
36
+ "epoch": 0.04,
37
+ "learning_rate": 1.4690759512266787e-06,
38
+ "loss": 1.2788,
39
+ "step": 5
40
+ },
41
+ {
42
+ "epoch": 0.04,
43
+ "learning_rate": 1.7628911414720144e-06,
44
+ "loss": 1.39,
45
+ "step": 6
46
+ },
47
+ {
48
+ "epoch": 0.05,
49
+ "learning_rate": 2.05670633171735e-06,
50
+ "loss": 1.4615,
51
+ "step": 7
52
+ },
53
+ {
54
+ "epoch": 0.06,
55
+ "learning_rate": 2.3505215219626858e-06,
56
+ "loss": 1.2082,
57
+ "step": 8
58
+ },
59
+ {
60
+ "epoch": 0.06,
61
+ "learning_rate": 2.6443367122080216e-06,
62
+ "loss": 1.3854,
63
+ "step": 9
64
+ },
65
+ {
66
+ "epoch": 0.07,
67
+ "learning_rate": 2.9381519024533573e-06,
68
+ "loss": 1.1711,
69
+ "step": 10
70
+ },
71
+ {
72
+ "epoch": 0.08,
73
+ "learning_rate": 3.2319670926986927e-06,
74
+ "loss": 1.3714,
75
+ "step": 11
76
+ },
77
+ {
78
+ "epoch": 0.08,
79
+ "learning_rate": 3.525782282944029e-06,
80
+ "loss": 1.6743,
81
+ "step": 12
82
+ },
83
+ {
84
+ "epoch": 0.09,
85
+ "learning_rate": 3.819597473189365e-06,
86
+ "loss": 1.2937,
87
+ "step": 13
88
+ },
89
+ {
90
+ "epoch": 0.1,
91
+ "learning_rate": 4.1134126634347e-06,
92
+ "loss": 1.2628,
93
+ "step": 14
94
+ },
95
+ {
96
+ "epoch": 0.11,
97
+ "learning_rate": 4.407227853680035e-06,
98
+ "loss": 1.3822,
99
+ "step": 15
100
+ },
101
+ {
102
+ "epoch": 0.11,
103
+ "learning_rate": 4.7010430439253716e-06,
104
+ "loss": 1.3788,
105
+ "step": 16
106
+ },
107
+ {
108
+ "epoch": 0.12,
109
+ "learning_rate": 4.994858234170707e-06,
110
+ "loss": 1.5392,
111
+ "step": 17
112
+ },
113
+ {
114
+ "epoch": 0.13,
115
+ "learning_rate": 5.288673424416043e-06,
116
+ "loss": 1.3841,
117
+ "step": 18
118
+ },
119
+ {
120
+ "epoch": 0.13,
121
+ "learning_rate": 5.582488614661379e-06,
122
+ "loss": 1.3883,
123
+ "step": 19
124
+ },
125
+ {
126
+ "epoch": 0.14,
127
+ "learning_rate": 5.876303804906715e-06,
128
+ "loss": 1.444,
129
+ "step": 20
130
+ },
131
+ {
132
+ "epoch": 0.15,
133
+ "learning_rate": 6.17011899515205e-06,
134
+ "loss": 1.0966,
135
+ "step": 21
136
+ },
137
+ {
138
+ "epoch": 0.15,
139
+ "learning_rate": 6.463934185397385e-06,
140
+ "loss": 1.4783,
141
+ "step": 22
142
+ },
143
+ {
144
+ "epoch": 0.16,
145
+ "learning_rate": 6.757749375642722e-06,
146
+ "loss": 1.3966,
147
+ "step": 23
148
+ },
149
+ {
150
+ "epoch": 0.17,
151
+ "learning_rate": 7.051564565888058e-06,
152
+ "loss": 1.602,
153
+ "step": 24
154
+ },
155
+ {
156
+ "epoch": 0.18,
157
+ "learning_rate": 7.345379756133393e-06,
158
+ "loss": 1.4453,
159
+ "step": 25
160
+ },
161
+ {
162
+ "epoch": 0.18,
163
+ "learning_rate": 7.63919494637873e-06,
164
+ "loss": 1.2995,
165
+ "step": 26
166
+ },
167
+ {
168
+ "epoch": 0.19,
169
+ "learning_rate": 7.933010136624066e-06,
170
+ "loss": 1.318,
171
+ "step": 27
172
+ },
173
+ {
174
+ "epoch": 0.2,
175
+ "learning_rate": 8.2268253268694e-06,
176
+ "loss": 1.1796,
177
+ "step": 28
178
+ },
179
+ {
180
+ "epoch": 0.2,
181
+ "learning_rate": 8.520640517114736e-06,
182
+ "loss": 1.31,
183
+ "step": 29
184
+ },
185
+ {
186
+ "epoch": 0.21,
187
+ "learning_rate": 8.81445570736007e-06,
188
+ "loss": 1.4279,
189
+ "step": 30
190
+ },
191
+ {
192
+ "epoch": 0.22,
193
+ "learning_rate": 9.108270897605407e-06,
194
+ "loss": 1.5326,
195
+ "step": 31
196
+ },
197
+ {
198
+ "epoch": 0.23,
199
+ "learning_rate": 9.402086087850743e-06,
200
+ "loss": 1.2816,
201
+ "step": 32
202
+ },
203
+ {
204
+ "epoch": 0.23,
205
+ "learning_rate": 9.695901278096078e-06,
206
+ "loss": 1.4097,
207
+ "step": 33
208
+ },
209
+ {
210
+ "epoch": 0.24,
211
+ "learning_rate": 9.989716468341414e-06,
212
+ "loss": 1.3857,
213
+ "step": 34
214
+ },
215
+ {
216
+ "epoch": 0.25,
217
+ "learning_rate": 1.028353165858675e-05,
218
+ "loss": 1.5109,
219
+ "step": 35
220
+ },
221
+ {
222
+ "epoch": 0.25,
223
+ "learning_rate": 1.0577346848832086e-05,
224
+ "loss": 1.5855,
225
+ "step": 36
226
+ },
227
+ {
228
+ "epoch": 0.26,
229
+ "learning_rate": 1.0871162039077422e-05,
230
+ "loss": 1.4847,
231
+ "step": 37
232
+ },
233
+ {
234
+ "epoch": 0.27,
235
+ "learning_rate": 1.1164977229322759e-05,
236
+ "loss": 1.3497,
237
+ "step": 38
238
+ },
239
+ {
240
+ "epoch": 0.27,
241
+ "learning_rate": 1.1458792419568093e-05,
242
+ "loss": 1.3788,
243
+ "step": 39
244
+ },
245
+ {
246
+ "epoch": 0.28,
247
+ "learning_rate": 1.175260760981343e-05,
248
+ "loss": 1.4044,
249
+ "step": 40
250
+ },
251
+ {
252
+ "epoch": 0.29,
253
+ "learning_rate": 1.2046422800058766e-05,
254
+ "loss": 1.3204,
255
+ "step": 41
256
+ },
257
+ {
258
+ "epoch": 0.3,
259
+ "learning_rate": 1.23402379903041e-05,
260
+ "loss": 1.4074,
261
+ "step": 42
262
+ },
263
+ {
264
+ "epoch": 0.3,
265
+ "learning_rate": 1.2634053180549435e-05,
266
+ "loss": 1.3244,
267
+ "step": 43
268
+ },
269
+ {
270
+ "epoch": 0.31,
271
+ "learning_rate": 1.292786837079477e-05,
272
+ "loss": 1.4781,
273
+ "step": 44
274
+ },
275
+ {
276
+ "epoch": 0.32,
277
+ "learning_rate": 1.3221683561040107e-05,
278
+ "loss": 1.166,
279
+ "step": 45
280
+ },
281
+ {
282
+ "epoch": 0.32,
283
+ "learning_rate": 1.3515498751285443e-05,
284
+ "loss": 1.5956,
285
+ "step": 46
286
+ },
287
+ {
288
+ "epoch": 0.33,
289
+ "learning_rate": 1.380931394153078e-05,
290
+ "loss": 1.3736,
291
+ "step": 47
292
+ },
293
+ {
294
+ "epoch": 0.34,
295
+ "learning_rate": 1.4103129131776116e-05,
296
+ "loss": 1.5388,
297
+ "step": 48
298
+ },
299
+ {
300
+ "epoch": 0.35,
301
+ "learning_rate": 1.439694432202145e-05,
302
+ "loss": 1.4948,
303
+ "step": 49
304
+ },
305
+ {
306
+ "epoch": 0.35,
307
+ "learning_rate": 1.4690759512266786e-05,
308
+ "loss": 1.3604,
309
+ "step": 50
310
+ },
311
+ {
312
+ "epoch": 0.36,
313
+ "learning_rate": 1.4984574702512122e-05,
314
+ "loss": 1.5697,
315
+ "step": 51
316
+ },
317
+ {
318
+ "epoch": 0.37,
319
+ "learning_rate": 1.527838989275746e-05,
320
+ "loss": 1.3266,
321
+ "step": 52
322
+ },
323
+ {
324
+ "epoch": 0.37,
325
+ "learning_rate": 1.5572205083002795e-05,
326
+ "loss": 1.2487,
327
+ "step": 53
328
+ },
329
+ {
330
+ "epoch": 0.38,
331
+ "learning_rate": 1.586602027324813e-05,
332
+ "loss": 1.3963,
333
+ "step": 54
334
+ },
335
+ {
336
+ "epoch": 0.39,
337
+ "learning_rate": 1.6159835463493464e-05,
338
+ "loss": 1.3293,
339
+ "step": 55
340
+ },
341
+ {
342
+ "epoch": 0.39,
343
+ "learning_rate": 1.64536506537388e-05,
344
+ "loss": 1.5669,
345
+ "step": 56
346
+ },
347
+ {
348
+ "epoch": 0.4,
349
+ "learning_rate": 1.6747465843984136e-05,
350
+ "loss": 1.3252,
351
+ "step": 57
352
+ },
353
+ {
354
+ "epoch": 0.41,
355
+ "learning_rate": 1.7041281034229472e-05,
356
+ "loss": 1.3639,
357
+ "step": 58
358
+ },
359
+ {
360
+ "epoch": 0.42,
361
+ "learning_rate": 1.733509622447481e-05,
362
+ "loss": 1.3889,
363
+ "step": 59
364
+ },
365
+ {
366
+ "epoch": 0.42,
367
+ "learning_rate": 1.762891141472014e-05,
368
+ "loss": 1.3824,
369
+ "step": 60
370
+ },
371
+ {
372
+ "epoch": 0.43,
373
+ "learning_rate": 1.7922726604965478e-05,
374
+ "loss": 1.4756,
375
+ "step": 61
376
+ },
377
+ {
378
+ "epoch": 0.44,
379
+ "learning_rate": 1.8216541795210814e-05,
380
+ "loss": 1.2453,
381
+ "step": 62
382
+ },
383
+ {
384
+ "epoch": 0.44,
385
+ "learning_rate": 1.851035698545615e-05,
386
+ "loss": 1.454,
387
+ "step": 63
388
+ },
389
+ {
390
+ "epoch": 0.45,
391
+ "learning_rate": 1.8804172175701486e-05,
392
+ "loss": 1.26,
393
+ "step": 64
394
+ },
395
+ {
396
+ "epoch": 0.46,
397
+ "learning_rate": 1.9097987365946822e-05,
398
+ "loss": 1.2652,
399
+ "step": 65
400
+ },
401
+ {
402
+ "epoch": 0.46,
403
+ "learning_rate": 1.9391802556192155e-05,
404
+ "loss": 1.424,
405
+ "step": 66
406
+ },
407
+ {
408
+ "epoch": 0.47,
409
+ "learning_rate": 1.968561774643749e-05,
410
+ "loss": 1.4516,
411
+ "step": 67
412
+ },
413
+ {
414
+ "epoch": 0.48,
415
+ "learning_rate": 1.9979432936682828e-05,
416
+ "loss": 1.4799,
417
+ "step": 68
418
+ },
419
+ {
420
+ "epoch": 0.49,
421
+ "learning_rate": 2.0273248126928164e-05,
422
+ "loss": 1.3161,
423
+ "step": 69
424
+ },
425
+ {
426
+ "epoch": 0.49,
427
+ "learning_rate": 2.05670633171735e-05,
428
+ "loss": 1.4528,
429
+ "step": 70
430
+ },
431
+ {
432
+ "epoch": 0.5,
433
+ "learning_rate": 2.0860878507418836e-05,
434
+ "loss": 1.54,
435
+ "step": 71
436
+ },
437
+ {
438
+ "epoch": 0.51,
439
+ "learning_rate": 2.1154693697664172e-05,
440
+ "loss": 1.3282,
441
+ "step": 72
442
+ },
443
+ {
444
+ "epoch": 0.51,
445
+ "learning_rate": 2.144850888790951e-05,
446
+ "loss": 1.2715,
447
+ "step": 73
448
+ },
449
+ {
450
+ "epoch": 0.52,
451
+ "learning_rate": 2.1742324078154845e-05,
452
+ "loss": 1.616,
453
+ "step": 74
454
+ },
455
+ {
456
+ "epoch": 0.53,
457
+ "learning_rate": 2.203613926840018e-05,
458
+ "loss": 1.451,
459
+ "step": 75
460
+ },
461
+ {
462
+ "epoch": 0.54,
463
+ "learning_rate": 2.2329954458645517e-05,
464
+ "loss": 1.3391,
465
+ "step": 76
466
+ },
467
+ {
468
+ "epoch": 0.54,
469
+ "learning_rate": 2.2623769648890853e-05,
470
+ "loss": 1.3733,
471
+ "step": 77
472
+ },
473
+ {
474
+ "epoch": 0.55,
475
+ "learning_rate": 2.2917584839136186e-05,
476
+ "loss": 1.2611,
477
+ "step": 78
478
+ },
479
+ {
480
+ "epoch": 0.56,
481
+ "learning_rate": 2.3211400029381523e-05,
482
+ "loss": 1.2367,
483
+ "step": 79
484
+ },
485
+ {
486
+ "epoch": 0.56,
487
+ "learning_rate": 2.350521521962686e-05,
488
+ "loss": 1.3784,
489
+ "step": 80
490
+ },
491
+ {
492
+ "epoch": 0.57,
493
+ "learning_rate": 2.3799030409872195e-05,
494
+ "loss": 1.2371,
495
+ "step": 81
496
+ },
497
+ {
498
+ "epoch": 0.58,
499
+ "learning_rate": 2.409284560011753e-05,
500
+ "loss": 1.326,
501
+ "step": 82
502
+ },
503
+ {
504
+ "epoch": 0.58,
505
+ "learning_rate": 2.4386660790362864e-05,
506
+ "loss": 1.5421,
507
+ "step": 83
508
+ },
509
+ {
510
+ "epoch": 0.59,
511
+ "learning_rate": 2.46804759806082e-05,
512
+ "loss": 1.5432,
513
+ "step": 84
514
+ },
515
+ {
516
+ "epoch": 0.6,
517
+ "learning_rate": 2.4974291170853536e-05,
518
+ "loss": 1.4241,
519
+ "step": 85
520
+ },
521
+ {
522
+ "epoch": 0.61,
523
+ "learning_rate": 2.526810636109887e-05,
524
+ "loss": 1.4539,
525
+ "step": 86
526
+ },
527
+ {
528
+ "epoch": 0.61,
529
+ "learning_rate": 2.556192155134421e-05,
530
+ "loss": 1.3261,
531
+ "step": 87
532
+ },
533
+ {
534
+ "epoch": 0.62,
535
+ "learning_rate": 2.585573674158954e-05,
536
+ "loss": 1.3152,
537
+ "step": 88
538
+ },
539
+ {
540
+ "epoch": 0.63,
541
+ "learning_rate": 2.614955193183488e-05,
542
+ "loss": 1.4292,
543
+ "step": 89
544
+ },
545
+ {
546
+ "epoch": 0.63,
547
+ "learning_rate": 2.6443367122080214e-05,
548
+ "loss": 1.1705,
549
+ "step": 90
550
+ },
551
+ {
552
+ "epoch": 0.64,
553
+ "learning_rate": 2.6737182312325554e-05,
554
+ "loss": 1.1443,
555
+ "step": 91
556
+ },
557
+ {
558
+ "epoch": 0.65,
559
+ "learning_rate": 2.7030997502570886e-05,
560
+ "loss": 1.4355,
561
+ "step": 92
562
+ },
563
+ {
564
+ "epoch": 0.65,
565
+ "learning_rate": 2.732481269281622e-05,
566
+ "loss": 1.3192,
567
+ "step": 93
568
+ },
569
+ {
570
+ "epoch": 0.66,
571
+ "learning_rate": 2.761862788306156e-05,
572
+ "loss": 1.4356,
573
+ "step": 94
574
+ },
575
+ {
576
+ "epoch": 0.67,
577
+ "learning_rate": 2.791244307330689e-05,
578
+ "loss": 1.1797,
579
+ "step": 95
580
+ },
581
+ {
582
+ "epoch": 0.68,
583
+ "learning_rate": 2.820625826355223e-05,
584
+ "loss": 1.1635,
585
+ "step": 96
586
+ },
587
+ {
588
+ "epoch": 0.68,
589
+ "learning_rate": 2.8500073453797564e-05,
590
+ "loss": 1.2257,
591
+ "step": 97
592
+ },
593
+ {
594
+ "epoch": 0.69,
595
+ "learning_rate": 2.87938886440429e-05,
596
+ "loss": 1.322,
597
+ "step": 98
598
+ },
599
+ {
600
+ "epoch": 0.7,
601
+ "learning_rate": 2.9087703834288236e-05,
602
+ "loss": 1.3281,
603
+ "step": 99
604
+ },
605
+ {
606
+ "epoch": 0.7,
607
+ "learning_rate": 2.9381519024533573e-05,
608
+ "loss": 1.0323,
609
+ "step": 100
610
+ },
611
+ {
612
+ "epoch": 0.71,
613
+ "learning_rate": 2.967533421477891e-05,
614
+ "loss": 1.1299,
615
+ "step": 101
616
+ },
617
+ {
618
+ "epoch": 0.72,
619
+ "learning_rate": 2.9969149405024245e-05,
620
+ "loss": 1.2833,
621
+ "step": 102
622
+ },
623
+ {
624
+ "epoch": 0.73,
625
+ "learning_rate": 3.0262964595269578e-05,
626
+ "loss": 1.3501,
627
+ "step": 103
628
+ },
629
+ {
630
+ "epoch": 0.73,
631
+ "learning_rate": 3.055677978551492e-05,
632
+ "loss": 1.1508,
633
+ "step": 104
634
+ },
635
+ {
636
+ "epoch": 0.74,
637
+ "learning_rate": 3.085059497576025e-05,
638
+ "loss": 1.418,
639
+ "step": 105
640
+ },
641
+ {
642
+ "epoch": 0.75,
643
+ "learning_rate": 3.114441016600559e-05,
644
+ "loss": 1.2717,
645
+ "step": 106
646
+ },
647
+ {
648
+ "epoch": 0.75,
649
+ "learning_rate": 3.143822535625092e-05,
650
+ "loss": 1.3274,
651
+ "step": 107
652
+ },
653
+ {
654
+ "epoch": 0.76,
655
+ "learning_rate": 3.173204054649626e-05,
656
+ "loss": 1.3633,
657
+ "step": 108
658
+ },
659
+ {
660
+ "epoch": 0.77,
661
+ "learning_rate": 3.202585573674159e-05,
662
+ "loss": 1.2618,
663
+ "step": 109
664
+ },
665
+ {
666
+ "epoch": 0.77,
667
+ "learning_rate": 3.231967092698693e-05,
668
+ "loss": 1.4994,
669
+ "step": 110
670
+ },
671
+ {
672
+ "epoch": 0.78,
673
+ "learning_rate": 3.2613486117232264e-05,
674
+ "loss": 1.4325,
675
+ "step": 111
676
+ },
677
+ {
678
+ "epoch": 0.79,
679
+ "learning_rate": 3.29073013074776e-05,
680
+ "loss": 1.3578,
681
+ "step": 112
682
+ },
683
+ {
684
+ "epoch": 0.8,
685
+ "learning_rate": 3.3201116497722936e-05,
686
+ "loss": 1.2844,
687
+ "step": 113
688
+ },
689
+ {
690
+ "epoch": 0.8,
691
+ "learning_rate": 3.349493168796827e-05,
692
+ "loss": 1.1555,
693
+ "step": 114
694
+ },
695
+ {
696
+ "epoch": 0.81,
697
+ "learning_rate": 3.378874687821361e-05,
698
+ "loss": 1.3242,
699
+ "step": 115
700
+ },
701
+ {
702
+ "epoch": 0.82,
703
+ "learning_rate": 3.4082562068458945e-05,
704
+ "loss": 1.1545,
705
+ "step": 116
706
+ },
707
+ {
708
+ "epoch": 0.82,
709
+ "learning_rate": 3.437637725870428e-05,
710
+ "loss": 1.1827,
711
+ "step": 117
712
+ },
713
+ {
714
+ "epoch": 0.83,
715
+ "learning_rate": 3.467019244894962e-05,
716
+ "loss": 1.5443,
717
+ "step": 118
718
+ },
719
+ {
720
+ "epoch": 0.84,
721
+ "learning_rate": 3.4964007639194954e-05,
722
+ "loss": 1.1105,
723
+ "step": 119
724
+ },
725
+ {
726
+ "epoch": 0.85,
727
+ "learning_rate": 3.525782282944028e-05,
728
+ "loss": 1.3643,
729
+ "step": 120
730
+ },
731
+ {
732
+ "epoch": 0.85,
733
+ "learning_rate": 3.5551638019685626e-05,
734
+ "loss": 1.335,
735
+ "step": 121
736
+ },
737
+ {
738
+ "epoch": 0.86,
739
+ "learning_rate": 3.5845453209930955e-05,
740
+ "loss": 1.2855,
741
+ "step": 122
742
+ },
743
+ {
744
+ "epoch": 0.87,
745
+ "learning_rate": 3.61392684001763e-05,
746
+ "loss": 1.5569,
747
+ "step": 123
748
+ },
749
+ {
750
+ "epoch": 0.87,
751
+ "learning_rate": 3.643308359042163e-05,
752
+ "loss": 1.3648,
753
+ "step": 124
754
+ },
755
+ {
756
+ "epoch": 0.88,
757
+ "learning_rate": 3.672689878066697e-05,
758
+ "loss": 1.1981,
759
+ "step": 125
760
+ },
761
+ {
762
+ "epoch": 0.89,
763
+ "learning_rate": 3.70207139709123e-05,
764
+ "loss": 1.281,
765
+ "step": 126
766
+ },
767
+ {
768
+ "epoch": 0.89,
769
+ "learning_rate": 3.7314529161157636e-05,
770
+ "loss": 1.2883,
771
+ "step": 127
772
+ },
773
+ {
774
+ "epoch": 0.9,
775
+ "learning_rate": 3.760834435140297e-05,
776
+ "loss": 1.4385,
777
+ "step": 128
778
+ },
779
+ {
780
+ "epoch": 0.91,
781
+ "learning_rate": 3.790215954164831e-05,
782
+ "loss": 1.4013,
783
+ "step": 129
784
+ },
785
+ {
786
+ "epoch": 0.92,
787
+ "learning_rate": 3.8195974731893645e-05,
788
+ "loss": 1.3303,
789
+ "step": 130
790
+ },
791
+ {
792
+ "epoch": 0.92,
793
+ "learning_rate": 3.848978992213898e-05,
794
+ "loss": 1.3222,
795
+ "step": 131
796
+ },
797
+ {
798
+ "epoch": 0.93,
799
+ "learning_rate": 3.878360511238431e-05,
800
+ "loss": 1.3686,
801
+ "step": 132
802
+ },
803
+ {
804
+ "epoch": 0.94,
805
+ "learning_rate": 3.9077420302629654e-05,
806
+ "loss": 1.3684,
807
+ "step": 133
808
+ },
809
+ {
810
+ "epoch": 0.94,
811
+ "learning_rate": 3.937123549287498e-05,
812
+ "loss": 1.3496,
813
+ "step": 134
814
+ },
815
+ {
816
+ "epoch": 0.95,
817
+ "learning_rate": 3.9665050683120326e-05,
818
+ "loss": 1.2483,
819
+ "step": 135
820
+ },
821
+ {
822
+ "epoch": 0.96,
823
+ "learning_rate": 3.9958865873365655e-05,
824
+ "loss": 1.3186,
825
+ "step": 136
826
+ },
827
+ {
828
+ "epoch": 0.96,
829
+ "learning_rate": 4.025268106361099e-05,
830
+ "loss": 1.3585,
831
+ "step": 137
832
+ },
833
+ {
834
+ "epoch": 0.97,
835
+ "learning_rate": 4.054649625385633e-05,
836
+ "loss": 1.1332,
837
+ "step": 138
838
+ },
839
+ {
840
+ "epoch": 0.98,
841
+ "learning_rate": 4.0840311444101664e-05,
842
+ "loss": 1.3868,
843
+ "step": 139
844
+ },
845
+ {
846
+ "epoch": 0.99,
847
+ "learning_rate": 4.1134126634347e-05,
848
+ "loss": 1.5011,
849
+ "step": 140
850
+ },
851
+ {
852
+ "epoch": 0.99,
853
+ "learning_rate": 4.1427941824592336e-05,
854
+ "loss": 1.2125,
855
+ "step": 141
856
+ },
857
+ {
858
+ "epoch": 1.0,
859
+ "learning_rate": 4.172175701483767e-05,
860
+ "loss": 1.2079,
861
+ "step": 142
862
+ },
863
+ {
864
+ "epoch": 1.01,
865
+ "learning_rate": 4.201557220508301e-05,
866
+ "loss": 1.5137,
867
+ "step": 143
868
+ },
869
+ {
870
+ "epoch": 1.01,
871
+ "learning_rate": 4.2309387395328345e-05,
872
+ "loss": 1.0962,
873
+ "step": 144
874
+ },
875
+ {
876
+ "epoch": 1.02,
877
+ "learning_rate": 4.260320258557368e-05,
878
+ "loss": 1.0864,
879
+ "step": 145
880
+ },
881
+ {
882
+ "epoch": 1.03,
883
+ "learning_rate": 4.289701777581902e-05,
884
+ "loss": 1.433,
885
+ "step": 146
886
+ },
887
+ {
888
+ "epoch": 1.04,
889
+ "learning_rate": 4.3190832966064354e-05,
890
+ "loss": 1.1769,
891
+ "step": 147
892
+ },
893
+ {
894
+ "epoch": 1.04,
895
+ "learning_rate": 4.348464815630969e-05,
896
+ "loss": 1.3221,
897
+ "step": 148
898
+ },
899
+ {
900
+ "epoch": 1.05,
901
+ "learning_rate": 4.377846334655502e-05,
902
+ "loss": 1.3977,
903
+ "step": 149
904
+ },
905
+ {
906
+ "epoch": 1.06,
907
+ "learning_rate": 4.407227853680036e-05,
908
+ "loss": 1.1165,
909
+ "step": 150
910
+ },
911
+ {
912
+ "epoch": 1.06,
913
+ "learning_rate": 4.436609372704569e-05,
914
+ "loss": 1.2971,
915
+ "step": 151
916
+ },
917
+ {
918
+ "epoch": 1.07,
919
+ "learning_rate": 4.4659908917291035e-05,
920
+ "loss": 1.0846,
921
+ "step": 152
922
+ },
923
+ {
924
+ "epoch": 1.08,
925
+ "learning_rate": 4.4953724107536364e-05,
926
+ "loss": 1.2692,
927
+ "step": 153
928
+ },
929
+ {
930
+ "epoch": 1.08,
931
+ "learning_rate": 4.524753929778171e-05,
932
+ "loss": 1.6022,
933
+ "step": 154
934
+ },
935
+ {
936
+ "epoch": 1.09,
937
+ "learning_rate": 4.5541354488027036e-05,
938
+ "loss": 1.2185,
939
+ "step": 155
940
+ },
941
+ {
942
+ "epoch": 1.1,
943
+ "learning_rate": 4.583516967827237e-05,
944
+ "loss": 1.1814,
945
+ "step": 156
946
+ },
947
+ {
948
+ "epoch": 1.11,
949
+ "learning_rate": 4.612898486851771e-05,
950
+ "loss": 1.2761,
951
+ "step": 157
952
+ },
953
+ {
954
+ "epoch": 1.11,
955
+ "learning_rate": 4.6422800058763045e-05,
956
+ "loss": 1.3127,
957
+ "step": 158
958
+ },
959
+ {
960
+ "epoch": 1.12,
961
+ "learning_rate": 4.671661524900838e-05,
962
+ "loss": 1.4407,
963
+ "step": 159
964
+ },
965
+ {
966
+ "epoch": 1.13,
967
+ "learning_rate": 4.701043043925372e-05,
968
+ "loss": 1.3185,
969
+ "step": 160
970
+ },
971
+ {
972
+ "epoch": 1.13,
973
+ "learning_rate": 4.730424562949905e-05,
974
+ "loss": 1.2999,
975
+ "step": 161
976
+ },
977
+ {
978
+ "epoch": 1.14,
979
+ "learning_rate": 4.759806081974439e-05,
980
+ "loss": 1.364,
981
+ "step": 162
982
+ },
983
+ {
984
+ "epoch": 1.15,
985
+ "learning_rate": 4.789187600998972e-05,
986
+ "loss": 1.0288,
987
+ "step": 163
988
+ },
989
+ {
990
+ "epoch": 1.15,
991
+ "learning_rate": 4.818569120023506e-05,
992
+ "loss": 1.4062,
993
+ "step": 164
994
+ },
995
+ {
996
+ "epoch": 1.16,
997
+ "learning_rate": 4.847950639048039e-05,
998
+ "loss": 1.3316,
999
+ "step": 165
1000
+ },
1001
+ {
1002
+ "epoch": 1.17,
1003
+ "learning_rate": 4.877332158072573e-05,
1004
+ "loss": 1.5267,
1005
+ "step": 166
1006
+ },
1007
+ {
1008
+ "epoch": 1.18,
1009
+ "learning_rate": 4.9067136770971064e-05,
1010
+ "loss": 1.3439,
1011
+ "step": 167
1012
+ },
1013
+ {
1014
+ "epoch": 1.18,
1015
+ "learning_rate": 4.93609519612164e-05,
1016
+ "loss": 1.2067,
1017
+ "step": 168
1018
+ },
1019
+ {
1020
+ "epoch": 1.19,
1021
+ "learning_rate": 4.9654767151461736e-05,
1022
+ "loss": 1.2325,
1023
+ "step": 169
1024
+ },
1025
+ {
1026
+ "epoch": 1.2,
1027
+ "learning_rate": 4.994858234170707e-05,
1028
+ "loss": 1.082,
1029
+ "step": 170
1030
+ },
1031
+ {
1032
+ "epoch": 1.2,
1033
+ "learning_rate": 5.024239753195241e-05,
1034
+ "loss": 1.2253,
1035
+ "step": 171
1036
+ },
1037
+ {
1038
+ "epoch": 1.21,
1039
+ "learning_rate": 5.053621272219774e-05,
1040
+ "loss": 1.3254,
1041
+ "step": 172
1042
+ },
1043
+ {
1044
+ "epoch": 1.22,
1045
+ "learning_rate": 5.083002791244308e-05,
1046
+ "loss": 1.4549,
1047
+ "step": 173
1048
+ },
1049
+ {
1050
+ "epoch": 1.23,
1051
+ "learning_rate": 5.112384310268842e-05,
1052
+ "loss": 1.1586,
1053
+ "step": 174
1054
+ },
1055
+ {
1056
+ "epoch": 1.23,
1057
+ "learning_rate": 5.1417658292933754e-05,
1058
+ "loss": 1.3143,
1059
+ "step": 175
1060
+ },
1061
+ {
1062
+ "epoch": 1.24,
1063
+ "learning_rate": 5.171147348317908e-05,
1064
+ "loss": 1.2859,
1065
+ "step": 176
1066
+ },
1067
+ {
1068
+ "epoch": 1.25,
1069
+ "learning_rate": 5.2005288673424426e-05,
1070
+ "loss": 1.4026,
1071
+ "step": 177
1072
+ },
1073
+ {
1074
+ "epoch": 1.25,
1075
+ "learning_rate": 5.229910386366976e-05,
1076
+ "loss": 1.5041,
1077
+ "step": 178
1078
+ },
1079
+ {
1080
+ "epoch": 1.26,
1081
+ "learning_rate": 5.259291905391509e-05,
1082
+ "loss": 1.3742,
1083
+ "step": 179
1084
+ },
1085
+ {
1086
+ "epoch": 1.27,
1087
+ "learning_rate": 5.288673424416043e-05,
1088
+ "loss": 1.2539,
1089
+ "step": 180
1090
+ },
1091
+ {
1092
+ "epoch": 1.27,
1093
+ "learning_rate": 5.318054943440577e-05,
1094
+ "loss": 1.3051,
1095
+ "step": 181
1096
+ },
1097
+ {
1098
+ "epoch": 1.28,
1099
+ "learning_rate": 5.347436462465111e-05,
1100
+ "loss": 1.291,
1101
+ "step": 182
1102
+ },
1103
+ {
1104
+ "epoch": 1.29,
1105
+ "learning_rate": 5.3768179814896436e-05,
1106
+ "loss": 1.2192,
1107
+ "step": 183
1108
+ },
1109
+ {
1110
+ "epoch": 1.3,
1111
+ "learning_rate": 5.406199500514177e-05,
1112
+ "loss": 1.312,
1113
+ "step": 184
1114
+ },
1115
+ {
1116
+ "epoch": 1.3,
1117
+ "learning_rate": 5.4355810195387116e-05,
1118
+ "loss": 1.2242,
1119
+ "step": 185
1120
+ },
1121
+ {
1122
+ "epoch": 1.31,
1123
+ "learning_rate": 5.464962538563244e-05,
1124
+ "loss": 1.4149,
1125
+ "step": 186
1126
+ },
1127
+ {
1128
+ "epoch": 1.32,
1129
+ "learning_rate": 5.494344057587778e-05,
1130
+ "loss": 1.0762,
1131
+ "step": 187
1132
+ },
1133
+ {
1134
+ "epoch": 1.32,
1135
+ "learning_rate": 5.523725576612312e-05,
1136
+ "loss": 1.5351,
1137
+ "step": 188
1138
+ },
1139
+ {
1140
+ "epoch": 1.33,
1141
+ "learning_rate": 5.553107095636845e-05,
1142
+ "loss": 1.2344,
1143
+ "step": 189
1144
+ },
1145
+ {
1146
+ "epoch": 1.34,
1147
+ "learning_rate": 5.582488614661378e-05,
1148
+ "loss": 1.4752,
1149
+ "step": 190
1150
+ },
1151
+ {
1152
+ "epoch": 1.35,
1153
+ "learning_rate": 5.6118701336859126e-05,
1154
+ "loss": 1.4017,
1155
+ "step": 191
1156
+ },
1157
+ {
1158
+ "epoch": 1.35,
1159
+ "learning_rate": 5.641251652710446e-05,
1160
+ "loss": 1.2617,
1161
+ "step": 192
1162
+ },
1163
+ {
1164
+ "epoch": 1.36,
1165
+ "learning_rate": 5.670633171734979e-05,
1166
+ "loss": 1.4561,
1167
+ "step": 193
1168
+ },
1169
+ {
1170
+ "epoch": 1.37,
1171
+ "learning_rate": 5.700014690759513e-05,
1172
+ "loss": 1.2679,
1173
+ "step": 194
1174
+ },
1175
+ {
1176
+ "epoch": 1.37,
1177
+ "learning_rate": 5.729396209784047e-05,
1178
+ "loss": 1.171,
1179
+ "step": 195
1180
+ },
1181
+ {
1182
+ "epoch": 1.38,
1183
+ "learning_rate": 5.75877772880858e-05,
1184
+ "loss": 1.277,
1185
+ "step": 196
1186
+ },
1187
+ {
1188
+ "epoch": 1.39,
1189
+ "learning_rate": 5.7881592478331136e-05,
1190
+ "loss": 1.1648,
1191
+ "step": 197
1192
+ },
1193
+ {
1194
+ "epoch": 1.39,
1195
+ "learning_rate": 5.817540766857647e-05,
1196
+ "loss": 1.4565,
1197
+ "step": 198
1198
+ },
1199
+ {
1200
+ "epoch": 1.4,
1201
+ "learning_rate": 5.8469222858821816e-05,
1202
+ "loss": 1.246,
1203
+ "step": 199
1204
+ },
1205
+ {
1206
+ "epoch": 1.41,
1207
+ "learning_rate": 5.8763038049067145e-05,
1208
+ "loss": 1.2742,
1209
+ "step": 200
1210
+ },
1211
+ {
1212
+ "epoch": 1.42,
1213
+ "learning_rate": 5.905685323931248e-05,
1214
+ "loss": 1.2846,
1215
+ "step": 201
1216
+ },
1217
+ {
1218
+ "epoch": 1.42,
1219
+ "learning_rate": 5.935066842955782e-05,
1220
+ "loss": 1.3021,
1221
+ "step": 202
1222
+ },
1223
+ {
1224
+ "epoch": 1.43,
1225
+ "learning_rate": 5.964448361980315e-05,
1226
+ "loss": 1.4153,
1227
+ "step": 203
1228
+ },
1229
+ {
1230
+ "epoch": 1.44,
1231
+ "learning_rate": 5.993829881004849e-05,
1232
+ "loss": 1.1717,
1233
+ "step": 204
1234
+ },
1235
+ {
1236
+ "epoch": 1.44,
1237
+ "learning_rate": 6.0232114000293826e-05,
1238
+ "loss": 1.3415,
1239
+ "step": 205
1240
+ },
1241
+ {
1242
+ "epoch": 1.45,
1243
+ "learning_rate": 6.0525929190539155e-05,
1244
+ "loss": 1.19,
1245
+ "step": 206
1246
+ },
1247
+ {
1248
+ "epoch": 1.46,
1249
+ "learning_rate": 6.081974438078449e-05,
1250
+ "loss": 1.1805,
1251
+ "step": 207
1252
+ },
1253
+ {
1254
+ "epoch": 1.46,
1255
+ "learning_rate": 6.111355957102983e-05,
1256
+ "loss": 1.3612,
1257
+ "step": 208
1258
+ },
1259
+ {
1260
+ "epoch": 1.47,
1261
+ "learning_rate": 6.140737476127517e-05,
1262
+ "loss": 1.3753,
1263
+ "step": 209
1264
+ },
1265
+ {
1266
+ "epoch": 1.48,
1267
+ "learning_rate": 6.17011899515205e-05,
1268
+ "loss": 1.3976,
1269
+ "step": 210
1270
+ },
1271
+ {
1272
+ "epoch": 1.49,
1273
+ "learning_rate": 6.199500514176584e-05,
1274
+ "loss": 1.2598,
1275
+ "step": 211
1276
+ },
1277
+ {
1278
+ "epoch": 1.49,
1279
+ "learning_rate": 6.228882033201118e-05,
1280
+ "loss": 1.3799,
1281
+ "step": 212
1282
+ },
1283
+ {
1284
+ "epoch": 1.5,
1285
+ "learning_rate": 6.25826355222565e-05,
1286
+ "loss": 1.4741,
1287
+ "step": 213
1288
+ },
1289
+ {
1290
+ "epoch": 1.51,
1291
+ "learning_rate": 6.287645071250184e-05,
1292
+ "loss": 1.2601,
1293
+ "step": 214
1294
+ },
1295
+ {
1296
+ "epoch": 1.51,
1297
+ "learning_rate": 6.317026590274719e-05,
1298
+ "loss": 1.1886,
1299
+ "step": 215
1300
+ },
1301
+ {
1302
+ "epoch": 1.52,
1303
+ "learning_rate": 6.346408109299252e-05,
1304
+ "loss": 1.5535,
1305
+ "step": 216
1306
+ },
1307
+ {
1308
+ "epoch": 1.53,
1309
+ "learning_rate": 6.375789628323785e-05,
1310
+ "loss": 1.3573,
1311
+ "step": 217
1312
+ },
1313
+ {
1314
+ "epoch": 1.54,
1315
+ "learning_rate": 6.405171147348318e-05,
1316
+ "loss": 1.2504,
1317
+ "step": 218
1318
+ },
1319
+ {
1320
+ "epoch": 1.54,
1321
+ "learning_rate": 6.434552666372853e-05,
1322
+ "loss": 1.3212,
1323
+ "step": 219
1324
+ },
1325
+ {
1326
+ "epoch": 1.55,
1327
+ "learning_rate": 6.463934185397386e-05,
1328
+ "loss": 1.177,
1329
+ "step": 220
1330
+ },
1331
+ {
1332
+ "epoch": 1.56,
1333
+ "learning_rate": 6.493315704421919e-05,
1334
+ "loss": 1.1466,
1335
+ "step": 221
1336
+ },
1337
+ {
1338
+ "epoch": 1.56,
1339
+ "learning_rate": 6.522697223446453e-05,
1340
+ "loss": 1.3115,
1341
+ "step": 222
1342
+ },
1343
+ {
1344
+ "epoch": 1.57,
1345
+ "learning_rate": 6.552078742470986e-05,
1346
+ "loss": 1.163,
1347
+ "step": 223
1348
+ },
1349
+ {
1350
+ "epoch": 1.58,
1351
+ "learning_rate": 6.58146026149552e-05,
1352
+ "loss": 1.2517,
1353
+ "step": 224
1354
+ },
1355
+ {
1356
+ "epoch": 1.58,
1357
+ "learning_rate": 6.610841780520054e-05,
1358
+ "loss": 1.4645,
1359
+ "step": 225
1360
+ },
1361
+ {
1362
+ "epoch": 1.59,
1363
+ "learning_rate": 6.640223299544587e-05,
1364
+ "loss": 1.5033,
1365
+ "step": 226
1366
+ },
1367
+ {
1368
+ "epoch": 1.6,
1369
+ "learning_rate": 6.669604818569121e-05,
1370
+ "loss": 1.367,
1371
+ "step": 227
1372
+ },
1373
+ {
1374
+ "epoch": 1.61,
1375
+ "learning_rate": 6.698986337593655e-05,
1376
+ "loss": 1.3936,
1377
+ "step": 228
1378
+ },
1379
+ {
1380
+ "epoch": 1.61,
1381
+ "learning_rate": 6.728367856618188e-05,
1382
+ "loss": 1.2716,
1383
+ "step": 229
1384
+ },
1385
+ {
1386
+ "epoch": 1.62,
1387
+ "learning_rate": 6.757749375642722e-05,
1388
+ "loss": 1.2563,
1389
+ "step": 230
1390
+ },
1391
+ {
1392
+ "epoch": 1.63,
1393
+ "learning_rate": 6.787130894667255e-05,
1394
+ "loss": 1.3766,
1395
+ "step": 231
1396
+ },
1397
+ {
1398
+ "epoch": 1.63,
1399
+ "learning_rate": 6.816512413691789e-05,
1400
+ "loss": 1.1206,
1401
+ "step": 232
1402
+ },
1403
+ {
1404
+ "epoch": 1.64,
1405
+ "learning_rate": 6.845893932716323e-05,
1406
+ "loss": 1.0943,
1407
+ "step": 233
1408
+ },
1409
+ {
1410
+ "epoch": 1.65,
1411
+ "learning_rate": 6.875275451740856e-05,
1412
+ "loss": 1.3942,
1413
+ "step": 234
1414
+ },
1415
+ {
1416
+ "epoch": 1.65,
1417
+ "learning_rate": 6.90465697076539e-05,
1418
+ "loss": 1.2673,
1419
+ "step": 235
1420
+ },
1421
+ {
1422
+ "epoch": 1.66,
1423
+ "learning_rate": 6.934038489789923e-05,
1424
+ "loss": 1.3732,
1425
+ "step": 236
1426
+ },
1427
+ {
1428
+ "epoch": 1.67,
1429
+ "learning_rate": 6.963420008814456e-05,
1430
+ "loss": 1.1018,
1431
+ "step": 237
1432
+ },
1433
+ {
1434
+ "epoch": 1.68,
1435
+ "learning_rate": 6.992801527838991e-05,
1436
+ "loss": 1.1131,
1437
+ "step": 238
1438
+ },
1439
+ {
1440
+ "epoch": 1.68,
1441
+ "learning_rate": 7.022183046863524e-05,
1442
+ "loss": 1.178,
1443
+ "step": 239
1444
+ },
1445
+ {
1446
+ "epoch": 1.69,
1447
+ "learning_rate": 7.051564565888057e-05,
1448
+ "loss": 1.2849,
1449
+ "step": 240
1450
+ },
1451
+ {
1452
+ "epoch": 1.7,
1453
+ "learning_rate": 7.08094608491259e-05,
1454
+ "loss": 1.2847,
1455
+ "step": 241
1456
+ },
1457
+ {
1458
+ "epoch": 1.7,
1459
+ "learning_rate": 7.110327603937125e-05,
1460
+ "loss": 0.9812,
1461
+ "step": 242
1462
+ },
1463
+ {
1464
+ "epoch": 1.71,
1465
+ "learning_rate": 7.139709122961659e-05,
1466
+ "loss": 1.0997,
1467
+ "step": 243
1468
+ },
1469
+ {
1470
+ "epoch": 1.72,
1471
+ "learning_rate": 7.169090641986191e-05,
1472
+ "loss": 1.2575,
1473
+ "step": 244
1474
+ },
1475
+ {
1476
+ "epoch": 1.73,
1477
+ "learning_rate": 7.198472161010725e-05,
1478
+ "loss": 1.3092,
1479
+ "step": 245
1480
+ },
1481
+ {
1482
+ "epoch": 1.73,
1483
+ "learning_rate": 7.22785368003526e-05,
1484
+ "loss": 1.1123,
1485
+ "step": 246
1486
+ },
1487
+ {
1488
+ "epoch": 1.74,
1489
+ "learning_rate": 7.257235199059792e-05,
1490
+ "loss": 1.3787,
1491
+ "step": 247
1492
+ },
1493
+ {
1494
+ "epoch": 1.75,
1495
+ "learning_rate": 7.286616718084326e-05,
1496
+ "loss": 1.2164,
1497
+ "step": 248
1498
+ },
1499
+ {
1500
+ "epoch": 1.75,
1501
+ "learning_rate": 7.315998237108859e-05,
1502
+ "loss": 1.2863,
1503
+ "step": 249
1504
+ },
1505
+ {
1506
+ "epoch": 1.76,
1507
+ "learning_rate": 7.345379756133394e-05,
1508
+ "loss": 1.3277,
1509
+ "step": 250
1510
+ },
1511
+ {
1512
+ "epoch": 1.77,
1513
+ "learning_rate": 7.374761275157926e-05,
1514
+ "loss": 1.2248,
1515
+ "step": 251
1516
+ },
1517
+ {
1518
+ "epoch": 1.77,
1519
+ "learning_rate": 7.40414279418246e-05,
1520
+ "loss": 1.4587,
1521
+ "step": 252
1522
+ },
1523
+ {
1524
+ "epoch": 1.78,
1525
+ "learning_rate": 7.433524313206994e-05,
1526
+ "loss": 1.3788,
1527
+ "step": 253
1528
+ },
1529
+ {
1530
+ "epoch": 1.79,
1531
+ "learning_rate": 7.462905832231527e-05,
1532
+ "loss": 1.3173,
1533
+ "step": 254
1534
+ },
1535
+ {
1536
+ "epoch": 1.8,
1537
+ "learning_rate": 7.492287351256061e-05,
1538
+ "loss": 1.2555,
1539
+ "step": 255
1540
+ },
1541
+ {
1542
+ "epoch": 1.8,
1543
+ "learning_rate": 7.521668870280595e-05,
1544
+ "loss": 1.1032,
1545
+ "step": 256
1546
+ },
1547
+ {
1548
+ "epoch": 1.81,
1549
+ "learning_rate": 7.551050389305128e-05,
1550
+ "loss": 1.271,
1551
+ "step": 257
1552
+ },
1553
+ {
1554
+ "epoch": 1.82,
1555
+ "learning_rate": 7.580431908329662e-05,
1556
+ "loss": 1.1156,
1557
+ "step": 258
1558
+ },
1559
+ {
1560
+ "epoch": 1.82,
1561
+ "learning_rate": 7.609813427354195e-05,
1562
+ "loss": 1.1455,
1563
+ "step": 259
1564
+ },
1565
+ {
1566
+ "epoch": 1.83,
1567
+ "learning_rate": 7.639194946378729e-05,
1568
+ "loss": 1.4856,
1569
+ "step": 260
1570
+ },
1571
+ {
1572
+ "epoch": 1.84,
1573
+ "learning_rate": 7.668576465403263e-05,
1574
+ "loss": 1.0758,
1575
+ "step": 261
1576
+ },
1577
+ {
1578
+ "epoch": 1.85,
1579
+ "learning_rate": 7.697957984427796e-05,
1580
+ "loss": 1.3166,
1581
+ "step": 262
1582
+ },
1583
+ {
1584
+ "epoch": 1.85,
1585
+ "learning_rate": 7.72733950345233e-05,
1586
+ "loss": 1.2848,
1587
+ "step": 263
1588
+ },
1589
+ {
1590
+ "epoch": 1.86,
1591
+ "learning_rate": 7.756721022476862e-05,
1592
+ "loss": 1.2372,
1593
+ "step": 264
1594
+ },
1595
+ {
1596
+ "epoch": 1.87,
1597
+ "learning_rate": 7.786102541501397e-05,
1598
+ "loss": 1.4906,
1599
+ "step": 265
1600
+ },
1601
+ {
1602
+ "epoch": 1.87,
1603
+ "learning_rate": 7.815484060525931e-05,
1604
+ "loss": 1.3265,
1605
+ "step": 266
1606
+ },
1607
+ {
1608
+ "epoch": 1.88,
1609
+ "learning_rate": 7.844865579550464e-05,
1610
+ "loss": 1.1439,
1611
+ "step": 267
1612
+ },
1613
+ {
1614
+ "epoch": 1.89,
1615
+ "learning_rate": 7.874247098574997e-05,
1616
+ "loss": 1.2426,
1617
+ "step": 268
1618
+ },
1619
+ {
1620
+ "epoch": 1.89,
1621
+ "learning_rate": 7.903628617599532e-05,
1622
+ "loss": 1.2351,
1623
+ "step": 269
1624
+ },
1625
+ {
1626
+ "epoch": 1.9,
1627
+ "learning_rate": 7.933010136624065e-05,
1628
+ "loss": 1.3957,
1629
+ "step": 270
1630
+ },
1631
+ {
1632
+ "epoch": 1.91,
1633
+ "learning_rate": 7.962391655648597e-05,
1634
+ "loss": 1.3555,
1635
+ "step": 271
1636
+ },
1637
+ {
1638
+ "epoch": 1.92,
1639
+ "learning_rate": 7.991773174673131e-05,
1640
+ "loss": 1.2983,
1641
+ "step": 272
1642
+ },
1643
+ {
1644
+ "epoch": 1.92,
1645
+ "learning_rate": 8.021154693697666e-05,
1646
+ "loss": 1.2816,
1647
+ "step": 273
1648
+ },
1649
+ {
1650
+ "epoch": 1.93,
1651
+ "learning_rate": 8.050536212722198e-05,
1652
+ "loss": 1.3042,
1653
+ "step": 274
1654
+ },
1655
+ {
1656
+ "epoch": 1.94,
1657
+ "learning_rate": 8.079917731746732e-05,
1658
+ "loss": 1.3328,
1659
+ "step": 275
1660
+ },
1661
+ {
1662
+ "epoch": 1.94,
1663
+ "learning_rate": 8.109299250771266e-05,
1664
+ "loss": 1.2964,
1665
+ "step": 276
1666
+ },
1667
+ {
1668
+ "epoch": 1.95,
1669
+ "learning_rate": 8.1386807697958e-05,
1670
+ "loss": 1.1836,
1671
+ "step": 277
1672
+ },
1673
+ {
1674
+ "epoch": 1.96,
1675
+ "learning_rate": 8.168062288820333e-05,
1676
+ "loss": 1.2704,
1677
+ "step": 278
1678
+ },
1679
+ {
1680
+ "epoch": 1.96,
1681
+ "learning_rate": 8.197443807844866e-05,
1682
+ "loss": 1.3228,
1683
+ "step": 279
1684
+ },
1685
+ {
1686
+ "epoch": 1.97,
1687
+ "learning_rate": 8.2268253268694e-05,
1688
+ "loss": 1.0871,
1689
+ "step": 280
1690
+ },
1691
+ {
1692
+ "epoch": 1.98,
1693
+ "learning_rate": 8.256206845893934e-05,
1694
+ "loss": 1.3571,
1695
+ "step": 281
1696
+ },
1697
+ {
1698
+ "epoch": 1.99,
1699
+ "learning_rate": 8.285588364918467e-05,
1700
+ "loss": 1.4434,
1701
+ "step": 282
1702
+ },
1703
+ {
1704
+ "epoch": 1.99,
1705
+ "learning_rate": 8.314969883943001e-05,
1706
+ "loss": 1.1603,
1707
+ "step": 283
1708
+ },
1709
+ {
1710
+ "epoch": 2.0,
1711
+ "learning_rate": 8.344351402967535e-05,
1712
+ "loss": 1.1378,
1713
+ "step": 284
1714
+ },
1715
+ {
1716
+ "epoch": 2.01,
1717
+ "learning_rate": 8.373732921992068e-05,
1718
+ "loss": 1.4653,
1719
+ "step": 285
1720
+ },
1721
+ {
1722
+ "epoch": 2.01,
1723
+ "learning_rate": 8.403114441016602e-05,
1724
+ "loss": 1.0566,
1725
+ "step": 286
1726
+ },
1727
+ {
1728
+ "epoch": 2.02,
1729
+ "learning_rate": 8.432495960041135e-05,
1730
+ "loss": 1.0505,
1731
+ "step": 287
1732
+ },
1733
+ {
1734
+ "epoch": 2.03,
1735
+ "learning_rate": 8.461877479065669e-05,
1736
+ "loss": 1.3879,
1737
+ "step": 288
1738
+ },
1739
+ {
1740
+ "epoch": 2.04,
1741
+ "learning_rate": 8.491258998090203e-05,
1742
+ "loss": 1.1117,
1743
+ "step": 289
1744
+ },
1745
+ {
1746
+ "epoch": 2.04,
1747
+ "learning_rate": 8.520640517114736e-05,
1748
+ "loss": 1.2697,
1749
+ "step": 290
1750
+ },
1751
+ {
1752
+ "epoch": 2.05,
1753
+ "learning_rate": 8.55002203613927e-05,
1754
+ "loss": 1.3577,
1755
+ "step": 291
1756
+ },
1757
+ {
1758
+ "epoch": 2.06,
1759
+ "learning_rate": 8.579403555163803e-05,
1760
+ "loss": 1.0753,
1761
+ "step": 292
1762
+ },
1763
+ {
1764
+ "epoch": 2.06,
1765
+ "learning_rate": 8.608785074188337e-05,
1766
+ "loss": 1.2443,
1767
+ "step": 293
1768
+ },
1769
+ {
1770
+ "epoch": 2.07,
1771
+ "learning_rate": 8.638166593212871e-05,
1772
+ "loss": 1.047,
1773
+ "step": 294
1774
+ },
1775
+ {
1776
+ "epoch": 2.08,
1777
+ "learning_rate": 8.667548112237403e-05,
1778
+ "loss": 1.2097,
1779
+ "step": 295
1780
+ },
1781
+ {
1782
+ "epoch": 2.08,
1783
+ "learning_rate": 8.696929631261938e-05,
1784
+ "loss": 1.5352,
1785
+ "step": 296
1786
+ },
1787
+ {
1788
+ "epoch": 2.09,
1789
+ "learning_rate": 8.726311150286472e-05,
1790
+ "loss": 1.1819,
1791
+ "step": 297
1792
+ },
1793
+ {
1794
+ "epoch": 2.1,
1795
+ "learning_rate": 8.755692669311004e-05,
1796
+ "loss": 1.1436,
1797
+ "step": 298
1798
+ },
1799
+ {
1800
+ "epoch": 2.11,
1801
+ "learning_rate": 8.785074188335537e-05,
1802
+ "loss": 1.1968,
1803
+ "step": 299
1804
+ },
1805
+ {
1806
+ "epoch": 2.11,
1807
+ "learning_rate": 8.814455707360072e-05,
1808
+ "loss": 1.2693,
1809
+ "step": 300
1810
+ },
1811
+ {
1812
+ "epoch": 2.12,
1813
+ "learning_rate": 8.843837226384606e-05,
1814
+ "loss": 1.3791,
1815
+ "step": 301
1816
+ },
1817
+ {
1818
+ "epoch": 2.13,
1819
+ "learning_rate": 8.873218745409138e-05,
1820
+ "loss": 1.2777,
1821
+ "step": 302
1822
+ },
1823
+ {
1824
+ "epoch": 2.13,
1825
+ "learning_rate": 8.902600264433672e-05,
1826
+ "loss": 1.248,
1827
+ "step": 303
1828
+ },
1829
+ {
1830
+ "epoch": 2.14,
1831
+ "learning_rate": 8.931981783458207e-05,
1832
+ "loss": 1.2936,
1833
+ "step": 304
1834
+ },
1835
+ {
1836
+ "epoch": 2.15,
1837
+ "learning_rate": 8.961363302482739e-05,
1838
+ "loss": 0.9981,
1839
+ "step": 305
1840
+ },
1841
+ {
1842
+ "epoch": 2.15,
1843
+ "learning_rate": 8.990744821507273e-05,
1844
+ "loss": 1.3413,
1845
+ "step": 306
1846
+ },
1847
+ {
1848
+ "epoch": 2.16,
1849
+ "learning_rate": 9.020126340531806e-05,
1850
+ "loss": 1.2833,
1851
+ "step": 307
1852
+ },
1853
+ {
1854
+ "epoch": 2.17,
1855
+ "learning_rate": 9.049507859556341e-05,
1856
+ "loss": 1.4853,
1857
+ "step": 308
1858
+ },
1859
+ {
1860
+ "epoch": 2.18,
1861
+ "learning_rate": 9.078889378580874e-05,
1862
+ "loss": 1.2807,
1863
+ "step": 309
1864
+ },
1865
+ {
1866
+ "epoch": 2.18,
1867
+ "learning_rate": 9.108270897605407e-05,
1868
+ "loss": 1.1716,
1869
+ "step": 310
1870
+ },
1871
+ {
1872
+ "epoch": 2.19,
1873
+ "learning_rate": 9.137652416629941e-05,
1874
+ "loss": 1.1964,
1875
+ "step": 311
1876
+ },
1877
+ {
1878
+ "epoch": 2.2,
1879
+ "learning_rate": 9.167033935654475e-05,
1880
+ "loss": 1.0411,
1881
+ "step": 312
1882
+ },
1883
+ {
1884
+ "epoch": 2.2,
1885
+ "learning_rate": 9.196415454679008e-05,
1886
+ "loss": 1.1827,
1887
+ "step": 313
1888
+ },
1889
+ {
1890
+ "epoch": 2.21,
1891
+ "learning_rate": 9.225796973703542e-05,
1892
+ "loss": 1.2726,
1893
+ "step": 314
1894
+ },
1895
+ {
1896
+ "epoch": 2.22,
1897
+ "learning_rate": 9.255178492728075e-05,
1898
+ "loss": 1.4049,
1899
+ "step": 315
1900
+ },
1901
+ {
1902
+ "epoch": 2.23,
1903
+ "learning_rate": 9.284560011752609e-05,
1904
+ "loss": 1.09,
1905
+ "step": 316
1906
+ },
1907
+ {
1908
+ "epoch": 2.23,
1909
+ "learning_rate": 9.313941530777143e-05,
1910
+ "loss": 1.2616,
1911
+ "step": 317
1912
+ },
1913
+ {
1914
+ "epoch": 2.24,
1915
+ "learning_rate": 9.343323049801676e-05,
1916
+ "loss": 1.2126,
1917
+ "step": 318
1918
+ },
1919
+ {
1920
+ "epoch": 2.25,
1921
+ "learning_rate": 9.37270456882621e-05,
1922
+ "loss": 1.3458,
1923
+ "step": 319
1924
+ },
1925
+ {
1926
+ "epoch": 2.25,
1927
+ "learning_rate": 9.402086087850743e-05,
1928
+ "loss": 1.4486,
1929
+ "step": 320
1930
+ },
1931
+ {
1932
+ "epoch": 2.26,
1933
+ "learning_rate": 9.431467606875277e-05,
1934
+ "loss": 1.2901,
1935
+ "step": 321
1936
+ },
1937
+ {
1938
+ "epoch": 2.27,
1939
+ "learning_rate": 9.46084912589981e-05,
1940
+ "loss": 1.2061,
1941
+ "step": 322
1942
+ },
1943
+ {
1944
+ "epoch": 2.27,
1945
+ "learning_rate": 9.490230644924344e-05,
1946
+ "loss": 1.2587,
1947
+ "step": 323
1948
+ },
1949
+ {
1950
+ "epoch": 2.28,
1951
+ "learning_rate": 9.519612163948878e-05,
1952
+ "loss": 1.2236,
1953
+ "step": 324
1954
+ },
1955
+ {
1956
+ "epoch": 2.29,
1957
+ "learning_rate": 9.548993682973412e-05,
1958
+ "loss": 1.1454,
1959
+ "step": 325
1960
+ },
1961
+ {
1962
+ "epoch": 2.3,
1963
+ "learning_rate": 9.578375201997944e-05,
1964
+ "loss": 1.2289,
1965
+ "step": 326
1966
+ },
1967
+ {
1968
+ "epoch": 2.3,
1969
+ "learning_rate": 9.607756721022479e-05,
1970
+ "loss": 1.1934,
1971
+ "step": 327
1972
+ },
1973
+ {
1974
+ "epoch": 2.31,
1975
+ "learning_rate": 9.637138240047012e-05,
1976
+ "loss": 1.3799,
1977
+ "step": 328
1978
+ },
1979
+ {
1980
+ "epoch": 2.32,
1981
+ "learning_rate": 9.666519759071545e-05,
1982
+ "loss": 1.0258,
1983
+ "step": 329
1984
+ },
1985
+ {
1986
+ "epoch": 2.32,
1987
+ "learning_rate": 9.695901278096078e-05,
1988
+ "loss": 1.4939,
1989
+ "step": 330
1990
+ },
1991
+ {
1992
+ "epoch": 2.33,
1993
+ "learning_rate": 9.725282797120613e-05,
1994
+ "loss": 1.186,
1995
+ "step": 331
1996
+ },
1997
+ {
1998
+ "epoch": 2.34,
1999
+ "learning_rate": 9.754664316145146e-05,
2000
+ "loss": 1.4199,
2001
+ "step": 332
2002
+ },
2003
+ {
2004
+ "epoch": 2.35,
2005
+ "learning_rate": 9.784045835169679e-05,
2006
+ "loss": 1.3008,
2007
+ "step": 333
2008
+ },
2009
+ {
2010
+ "epoch": 2.35,
2011
+ "learning_rate": 9.813427354194213e-05,
2012
+ "loss": 1.1806,
2013
+ "step": 334
2014
+ },
2015
+ {
2016
+ "epoch": 2.36,
2017
+ "learning_rate": 9.842808873218748e-05,
2018
+ "loss": 1.3952,
2019
+ "step": 335
2020
+ },
2021
+ {
2022
+ "epoch": 2.37,
2023
+ "learning_rate": 9.87219039224328e-05,
2024
+ "loss": 1.2403,
2025
+ "step": 336
2026
+ },
2027
+ {
2028
+ "epoch": 2.37,
2029
+ "learning_rate": 9.901571911267814e-05,
2030
+ "loss": 1.1104,
2031
+ "step": 337
2032
+ },
2033
+ {
2034
+ "epoch": 2.38,
2035
+ "learning_rate": 9.930953430292347e-05,
2036
+ "loss": 1.2103,
2037
+ "step": 338
2038
+ },
2039
+ {
2040
+ "epoch": 2.39,
2041
+ "learning_rate": 9.960334949316881e-05,
2042
+ "loss": 1.0564,
2043
+ "step": 339
2044
+ },
2045
+ {
2046
+ "epoch": 2.39,
2047
+ "learning_rate": 9.989716468341415e-05,
2048
+ "loss": 1.3984,
2049
+ "step": 340
2050
+ },
2051
+ {
2052
+ "epoch": 2.4,
2053
+ "learning_rate": 9.99999963334151e-05,
2054
+ "loss": 1.2017,
2055
+ "step": 341
2056
+ },
2057
+ {
2058
+ "epoch": 2.41,
2059
+ "learning_rate": 9.999997637330952e-05,
2060
+ "loss": 1.2251,
2061
+ "step": 342
2062
+ },
2063
+ {
2064
+ "epoch": 2.42,
2065
+ "learning_rate": 9.99999390565975e-05,
2066
+ "loss": 1.2006,
2067
+ "step": 343
2068
+ },
2069
+ {
2070
+ "epoch": 2.42,
2071
+ "learning_rate": 9.999988438329202e-05,
2072
+ "loss": 1.236,
2073
+ "step": 344
2074
+ },
2075
+ {
2076
+ "epoch": 2.43,
2077
+ "learning_rate": 9.999981235341205e-05,
2078
+ "loss": 1.3684,
2079
+ "step": 345
2080
+ },
2081
+ {
2082
+ "epoch": 2.44,
2083
+ "learning_rate": 9.999972296698259e-05,
2084
+ "loss": 1.1272,
2085
+ "step": 346
2086
+ },
2087
+ {
2088
+ "epoch": 2.44,
2089
+ "learning_rate": 9.999961622403467e-05,
2090
+ "loss": 1.2259,
2091
+ "step": 347
2092
+ },
2093
+ {
2094
+ "epoch": 2.45,
2095
+ "learning_rate": 9.999949212460534e-05,
2096
+ "loss": 1.1502,
2097
+ "step": 348
2098
+ },
2099
+ {
2100
+ "epoch": 2.46,
2101
+ "learning_rate": 9.999935066873766e-05,
2102
+ "loss": 1.1134,
2103
+ "step": 349
2104
+ },
2105
+ {
2106
+ "epoch": 2.46,
2107
+ "learning_rate": 9.999919185648079e-05,
2108
+ "loss": 1.306,
2109
+ "step": 350
2110
+ },
2111
+ {
2112
+ "epoch": 2.47,
2113
+ "learning_rate": 9.99990156878898e-05,
2114
+ "loss": 1.3023,
2115
+ "step": 351
2116
+ },
2117
+ {
2118
+ "epoch": 2.48,
2119
+ "learning_rate": 9.999882216302589e-05,
2120
+ "loss": 1.3383,
2121
+ "step": 352
2122
+ },
2123
+ {
2124
+ "epoch": 2.49,
2125
+ "learning_rate": 9.99986112819562e-05,
2126
+ "loss": 1.2076,
2127
+ "step": 353
2128
+ },
2129
+ {
2130
+ "epoch": 2.49,
2131
+ "learning_rate": 9.999838304475396e-05,
2132
+ "loss": 1.3298,
2133
+ "step": 354
2134
+ },
2135
+ {
2136
+ "epoch": 2.5,
2137
+ "learning_rate": 9.99981374514984e-05,
2138
+ "loss": 1.4117,
2139
+ "step": 355
2140
+ },
2141
+ {
2142
+ "epoch": 2.51,
2143
+ "learning_rate": 9.999787450227475e-05,
2144
+ "loss": 1.2157,
2145
+ "step": 356
2146
+ },
2147
+ {
2148
+ "epoch": 2.51,
2149
+ "learning_rate": 9.999759419717429e-05,
2150
+ "loss": 1.1052,
2151
+ "step": 357
2152
+ },
2153
+ {
2154
+ "epoch": 2.52,
2155
+ "learning_rate": 9.999729653629435e-05,
2156
+ "loss": 1.4931,
2157
+ "step": 358
2158
+ },
2159
+ {
2160
+ "epoch": 2.53,
2161
+ "learning_rate": 9.999698151973824e-05,
2162
+ "loss": 1.2367,
2163
+ "step": 359
2164
+ },
2165
+ {
2166
+ "epoch": 2.54,
2167
+ "learning_rate": 9.999664914761532e-05,
2168
+ "loss": 1.1763,
2169
+ "step": 360
2170
+ },
2171
+ {
2172
+ "epoch": 2.54,
2173
+ "learning_rate": 9.999629942004095e-05,
2174
+ "loss": 1.2708,
2175
+ "step": 361
2176
+ },
2177
+ {
2178
+ "epoch": 2.55,
2179
+ "learning_rate": 9.999593233713654e-05,
2180
+ "loss": 1.114,
2181
+ "step": 362
2182
+ },
2183
+ {
2184
+ "epoch": 2.56,
2185
+ "learning_rate": 9.999554789902954e-05,
2186
+ "loss": 1.066,
2187
+ "step": 363
2188
+ },
2189
+ {
2190
+ "epoch": 2.56,
2191
+ "learning_rate": 9.999514610585336e-05,
2192
+ "loss": 1.223,
2193
+ "step": 364
2194
+ },
2195
+ {
2196
+ "epoch": 2.57,
2197
+ "learning_rate": 9.99947269577475e-05,
2198
+ "loss": 1.1002,
2199
+ "step": 365
2200
+ },
2201
+ {
2202
+ "epoch": 2.58,
2203
+ "learning_rate": 9.999429045485746e-05,
2204
+ "loss": 1.1679,
2205
+ "step": 366
2206
+ },
2207
+ {
2208
+ "epoch": 2.58,
2209
+ "learning_rate": 9.999383659733476e-05,
2210
+ "loss": 1.3512,
2211
+ "step": 367
2212
+ },
2213
+ {
2214
+ "epoch": 2.59,
2215
+ "learning_rate": 9.999336538533694e-05,
2216
+ "loss": 1.445,
2217
+ "step": 368
2218
+ },
2219
+ {
2220
+ "epoch": 2.6,
2221
+ "learning_rate": 9.99928768190276e-05,
2222
+ "loss": 1.2979,
2223
+ "step": 369
2224
+ },
2225
+ {
2226
+ "epoch": 2.61,
2227
+ "learning_rate": 9.99923708985763e-05,
2228
+ "loss": 1.3093,
2229
+ "step": 370
2230
+ },
2231
+ {
2232
+ "epoch": 2.61,
2233
+ "learning_rate": 9.999184762415869e-05,
2234
+ "loss": 1.2169,
2235
+ "step": 371
2236
+ },
2237
+ {
2238
+ "epoch": 2.62,
2239
+ "learning_rate": 9.99913069959564e-05,
2240
+ "loss": 1.1727,
2241
+ "step": 372
2242
+ },
2243
+ {
2244
+ "epoch": 2.63,
2245
+ "learning_rate": 9.99907490141571e-05,
2246
+ "loss": 1.2972,
2247
+ "step": 373
2248
+ },
2249
+ {
2250
+ "epoch": 2.63,
2251
+ "learning_rate": 9.999017367895449e-05,
2252
+ "loss": 1.0836,
2253
+ "step": 374
2254
+ },
2255
+ {
2256
+ "epoch": 2.64,
2257
+ "learning_rate": 9.99895809905483e-05,
2258
+ "loss": 1.0422,
2259
+ "step": 375
2260
+ },
2261
+ {
2262
+ "epoch": 2.65,
2263
+ "learning_rate": 9.998897094914424e-05,
2264
+ "loss": 1.3384,
2265
+ "step": 376
2266
+ },
2267
+ {
2268
+ "epoch": 2.65,
2269
+ "learning_rate": 9.99883435549541e-05,
2270
+ "loss": 1.19,
2271
+ "step": 377
2272
+ },
2273
+ {
2274
+ "epoch": 2.66,
2275
+ "learning_rate": 9.998769880819565e-05,
2276
+ "loss": 1.2808,
2277
+ "step": 378
2278
+ },
2279
+ {
2280
+ "epoch": 2.67,
2281
+ "learning_rate": 9.998703670909273e-05,
2282
+ "loss": 1.0036,
2283
+ "step": 379
2284
+ },
2285
+ {
2286
+ "epoch": 2.68,
2287
+ "learning_rate": 9.998635725787515e-05,
2288
+ "loss": 1.0307,
2289
+ "step": 380
2290
+ },
2291
+ {
2292
+ "epoch": 2.68,
2293
+ "learning_rate": 9.998566045477877e-05,
2294
+ "loss": 1.1171,
2295
+ "step": 381
2296
+ },
2297
+ {
2298
+ "epoch": 2.69,
2299
+ "learning_rate": 9.99849463000455e-05,
2300
+ "loss": 1.2063,
2301
+ "step": 382
2302
+ },
2303
+ {
2304
+ "epoch": 2.7,
2305
+ "learning_rate": 9.998421479392322e-05,
2306
+ "loss": 1.2263,
2307
+ "step": 383
2308
+ },
2309
+ {
2310
+ "epoch": 2.7,
2311
+ "learning_rate": 9.998346593666587e-05,
2312
+ "loss": 0.8941,
2313
+ "step": 384
2314
+ },
2315
+ {
2316
+ "epoch": 2.71,
2317
+ "learning_rate": 9.99826997285334e-05,
2318
+ "loss": 1.0527,
2319
+ "step": 385
2320
+ },
2321
+ {
2322
+ "epoch": 2.72,
2323
+ "learning_rate": 9.998191616979178e-05,
2324
+ "loss": 1.2203,
2325
+ "step": 386
2326
+ },
2327
+ {
2328
+ "epoch": 2.73,
2329
+ "learning_rate": 9.9981115260713e-05,
2330
+ "loss": 1.2373,
2331
+ "step": 387
2332
+ },
2333
+ {
2334
+ "epoch": 2.73,
2335
+ "learning_rate": 9.998029700157512e-05,
2336
+ "loss": 1.0372,
2337
+ "step": 388
2338
+ },
2339
+ {
2340
+ "epoch": 2.74,
2341
+ "learning_rate": 9.997946139266214e-05,
2342
+ "loss": 1.3097,
2343
+ "step": 389
2344
+ },
2345
+ {
2346
+ "epoch": 2.75,
2347
+ "learning_rate": 9.997860843426416e-05,
2348
+ "loss": 1.1405,
2349
+ "step": 390
2350
+ },
2351
+ {
2352
+ "epoch": 2.75,
2353
+ "learning_rate": 9.997773812667726e-05,
2354
+ "loss": 1.2198,
2355
+ "step": 391
2356
+ },
2357
+ {
2358
+ "epoch": 2.76,
2359
+ "learning_rate": 9.997685047020353e-05,
2360
+ "loss": 1.2622,
2361
+ "step": 392
2362
+ },
2363
+ {
2364
+ "epoch": 2.77,
2365
+ "learning_rate": 9.997594546515112e-05,
2366
+ "loss": 1.1772,
2367
+ "step": 393
2368
+ },
2369
+ {
2370
+ "epoch": 2.77,
2371
+ "learning_rate": 9.997502311183421e-05,
2372
+ "loss": 1.3354,
2373
+ "step": 394
2374
+ },
2375
+ {
2376
+ "epoch": 2.78,
2377
+ "learning_rate": 9.997408341057295e-05,
2378
+ "loss": 1.2733,
2379
+ "step": 395
2380
+ },
2381
+ {
2382
+ "epoch": 2.79,
2383
+ "learning_rate": 9.997312636169354e-05,
2384
+ "loss": 1.2603,
2385
+ "step": 396
2386
+ },
2387
+ {
2388
+ "epoch": 2.8,
2389
+ "learning_rate": 9.99721519655282e-05,
2390
+ "loss": 1.182,
2391
+ "step": 397
2392
+ },
2393
+ {
2394
+ "epoch": 2.8,
2395
+ "learning_rate": 9.997116022241519e-05,
2396
+ "loss": 1.001,
2397
+ "step": 398
2398
+ },
2399
+ {
2400
+ "epoch": 2.81,
2401
+ "learning_rate": 9.997015113269878e-05,
2402
+ "loss": 1.1661,
2403
+ "step": 399
2404
+ },
2405
+ {
2406
+ "epoch": 2.82,
2407
+ "learning_rate": 9.996912469672925e-05,
2408
+ "loss": 1.0415,
2409
+ "step": 400
2410
+ },
2411
+ {
2412
+ "epoch": 2.82,
2413
+ "learning_rate": 9.99680809148629e-05,
2414
+ "loss": 1.0874,
2415
+ "step": 401
2416
+ },
2417
+ {
2418
+ "epoch": 2.83,
2419
+ "learning_rate": 9.996701978746207e-05,
2420
+ "loss": 1.3411,
2421
+ "step": 402
2422
+ },
2423
+ {
2424
+ "epoch": 2.84,
2425
+ "learning_rate": 9.996594131489511e-05,
2426
+ "loss": 1.0126,
2427
+ "step": 403
2428
+ },
2429
+ {
2430
+ "epoch": 2.85,
2431
+ "learning_rate": 9.99648454975364e-05,
2432
+ "loss": 1.2103,
2433
+ "step": 404
2434
+ },
2435
+ {
2436
+ "epoch": 2.85,
2437
+ "learning_rate": 9.996373233576631e-05,
2438
+ "loss": 1.1725,
2439
+ "step": 405
2440
+ },
2441
+ {
2442
+ "epoch": 2.86,
2443
+ "learning_rate": 9.996260182997126e-05,
2444
+ "loss": 1.1493,
2445
+ "step": 406
2446
+ },
2447
+ {
2448
+ "epoch": 2.87,
2449
+ "learning_rate": 9.996145398054372e-05,
2450
+ "loss": 1.3638,
2451
+ "step": 407
2452
+ },
2453
+ {
2454
+ "epoch": 2.87,
2455
+ "learning_rate": 9.99602887878821e-05,
2456
+ "loss": 1.2561,
2457
+ "step": 408
2458
+ },
2459
+ {
2460
+ "epoch": 2.88,
2461
+ "learning_rate": 9.995910625239092e-05,
2462
+ "loss": 1.0412,
2463
+ "step": 409
2464
+ },
2465
+ {
2466
+ "epoch": 2.89,
2467
+ "learning_rate": 9.995790637448064e-05,
2468
+ "loss": 1.1855,
2469
+ "step": 410
2470
+ },
2471
+ {
2472
+ "epoch": 2.89,
2473
+ "learning_rate": 9.995668915456778e-05,
2474
+ "loss": 1.1156,
2475
+ "step": 411
2476
+ },
2477
+ {
2478
+ "epoch": 2.9,
2479
+ "learning_rate": 9.99554545930749e-05,
2480
+ "loss": 1.317,
2481
+ "step": 412
2482
+ },
2483
+ {
2484
+ "epoch": 2.91,
2485
+ "learning_rate": 9.995420269043053e-05,
2486
+ "loss": 1.3018,
2487
+ "step": 413
2488
+ },
2489
+ {
2490
+ "epoch": 2.92,
2491
+ "learning_rate": 9.995293344706927e-05,
2492
+ "loss": 1.2392,
2493
+ "step": 414
2494
+ },
2495
+ {
2496
+ "epoch": 2.92,
2497
+ "learning_rate": 9.99516468634317e-05,
2498
+ "loss": 1.2128,
2499
+ "step": 415
2500
+ },
2501
+ {
2502
+ "epoch": 2.93,
2503
+ "learning_rate": 9.995034293996442e-05,
2504
+ "loss": 1.173,
2505
+ "step": 416
2506
+ },
2507
+ {
2508
+ "epoch": 2.94,
2509
+ "learning_rate": 9.99490216771201e-05,
2510
+ "loss": 1.2461,
2511
+ "step": 417
2512
+ },
2513
+ {
2514
+ "epoch": 2.94,
2515
+ "learning_rate": 9.994768307535736e-05,
2516
+ "loss": 1.1709,
2517
+ "step": 418
2518
+ },
2519
+ {
2520
+ "epoch": 2.95,
2521
+ "learning_rate": 9.99463271351409e-05,
2522
+ "loss": 1.0691,
2523
+ "step": 419
2524
+ },
2525
+ {
2526
+ "epoch": 2.96,
2527
+ "learning_rate": 9.99449538569414e-05,
2528
+ "loss": 1.1749,
2529
+ "step": 420
2530
+ },
2531
+ {
2532
+ "epoch": 2.96,
2533
+ "learning_rate": 9.994356324123555e-05,
2534
+ "loss": 1.2748,
2535
+ "step": 421
2536
+ },
2537
+ {
2538
+ "epoch": 2.97,
2539
+ "learning_rate": 9.994215528850611e-05,
2540
+ "loss": 0.9984,
2541
+ "step": 422
2542
+ },
2543
+ {
2544
+ "epoch": 2.98,
2545
+ "learning_rate": 9.994072999924178e-05,
2546
+ "loss": 1.288,
2547
+ "step": 423
2548
+ },
2549
+ {
2550
+ "epoch": 2.99,
2551
+ "learning_rate": 9.993928737393738e-05,
2552
+ "loss": 1.3438,
2553
+ "step": 424
2554
+ },
2555
+ {
2556
+ "epoch": 2.99,
2557
+ "learning_rate": 9.993782741309366e-05,
2558
+ "loss": 1.0476,
2559
+ "step": 425
2560
+ },
2561
+ {
2562
+ "epoch": 3.0,
2563
+ "learning_rate": 9.993635011721743e-05,
2564
+ "loss": 1.0269,
2565
+ "step": 426
2566
+ },
2567
+ {
2568
+ "epoch": 3.01,
2569
+ "learning_rate": 9.993485548682149e-05,
2570
+ "loss": 1.3591,
2571
+ "step": 427
2572
+ },
2573
+ {
2574
+ "epoch": 3.01,
2575
+ "learning_rate": 9.993334352242469e-05,
2576
+ "loss": 0.9882,
2577
+ "step": 428
2578
+ },
2579
+ {
2580
+ "epoch": 3.02,
2581
+ "learning_rate": 9.993181422455189e-05,
2582
+ "loss": 0.9844,
2583
+ "step": 429
2584
+ },
2585
+ {
2586
+ "epoch": 3.03,
2587
+ "learning_rate": 9.993026759373394e-05,
2588
+ "loss": 1.2867,
2589
+ "step": 430
2590
+ },
2591
+ {
2592
+ "epoch": 3.04,
2593
+ "learning_rate": 9.992870363050772e-05,
2594
+ "loss": 0.9716,
2595
+ "step": 431
2596
+ },
2597
+ {
2598
+ "epoch": 3.04,
2599
+ "learning_rate": 9.992712233541616e-05,
2600
+ "loss": 1.1783,
2601
+ "step": 432
2602
+ },
2603
+ {
2604
+ "epoch": 3.05,
2605
+ "learning_rate": 9.992552370900816e-05,
2606
+ "loss": 1.3025,
2607
+ "step": 433
2608
+ },
2609
+ {
2610
+ "epoch": 3.06,
2611
+ "learning_rate": 9.992390775183865e-05,
2612
+ "loss": 0.9909,
2613
+ "step": 434
2614
+ },
2615
+ {
2616
+ "epoch": 3.06,
2617
+ "learning_rate": 9.99222744644686e-05,
2618
+ "loss": 1.1289,
2619
+ "step": 435
2620
+ },
2621
+ {
2622
+ "epoch": 3.07,
2623
+ "learning_rate": 9.992062384746496e-05,
2624
+ "loss": 0.9763,
2625
+ "step": 436
2626
+ },
2627
+ {
2628
+ "epoch": 3.08,
2629
+ "learning_rate": 9.991895590140072e-05,
2630
+ "loss": 1.0838,
2631
+ "step": 437
2632
+ },
2633
+ {
2634
+ "epoch": 3.08,
2635
+ "learning_rate": 9.991727062685488e-05,
2636
+ "loss": 1.3713,
2637
+ "step": 438
2638
+ },
2639
+ {
2640
+ "epoch": 3.09,
2641
+ "learning_rate": 9.991556802441244e-05,
2642
+ "loss": 1.1156,
2643
+ "step": 439
2644
+ },
2645
+ {
2646
+ "epoch": 3.1,
2647
+ "learning_rate": 9.991384809466445e-05,
2648
+ "loss": 1.0658,
2649
+ "step": 440
2650
+ },
2651
+ {
2652
+ "epoch": 3.11,
2653
+ "learning_rate": 9.991211083820792e-05,
2654
+ "loss": 1.0693,
2655
+ "step": 441
2656
+ },
2657
+ {
2658
+ "epoch": 3.11,
2659
+ "learning_rate": 9.991035625564593e-05,
2660
+ "loss": 1.1844,
2661
+ "step": 442
2662
+ },
2663
+ {
2664
+ "epoch": 3.12,
2665
+ "learning_rate": 9.990858434758756e-05,
2666
+ "loss": 1.2371,
2667
+ "step": 443
2668
+ },
2669
+ {
2670
+ "epoch": 3.13,
2671
+ "learning_rate": 9.990679511464788e-05,
2672
+ "loss": 1.2049,
2673
+ "step": 444
2674
+ },
2675
+ {
2676
+ "epoch": 3.13,
2677
+ "learning_rate": 9.990498855744802e-05,
2678
+ "loss": 1.1487,
2679
+ "step": 445
2680
+ },
2681
+ {
2682
+ "epoch": 3.14,
2683
+ "learning_rate": 9.990316467661506e-05,
2684
+ "loss": 1.1855,
2685
+ "step": 446
2686
+ },
2687
+ {
2688
+ "epoch": 3.15,
2689
+ "learning_rate": 9.990132347278213e-05,
2690
+ "loss": 0.9415,
2691
+ "step": 447
2692
+ },
2693
+ {
2694
+ "epoch": 3.15,
2695
+ "learning_rate": 9.989946494658839e-05,
2696
+ "loss": 1.2081,
2697
+ "step": 448
2698
+ },
2699
+ {
2700
+ "epoch": 3.16,
2701
+ "learning_rate": 9.989758909867899e-05,
2702
+ "loss": 1.1945,
2703
+ "step": 449
2704
+ },
2705
+ {
2706
+ "epoch": 3.17,
2707
+ "learning_rate": 9.989569592970509e-05,
2708
+ "loss": 1.386,
2709
+ "step": 450
2710
+ },
2711
+ {
2712
+ "epoch": 3.18,
2713
+ "learning_rate": 9.989378544032388e-05,
2714
+ "loss": 1.1616,
2715
+ "step": 451
2716
+ },
2717
+ {
2718
+ "epoch": 3.18,
2719
+ "learning_rate": 9.989185763119853e-05,
2720
+ "loss": 1.0811,
2721
+ "step": 452
2722
+ },
2723
+ {
2724
+ "epoch": 3.19,
2725
+ "learning_rate": 9.988991250299826e-05,
2726
+ "loss": 1.1135,
2727
+ "step": 453
2728
+ },
2729
+ {
2730
+ "epoch": 3.2,
2731
+ "learning_rate": 9.98879500563983e-05,
2732
+ "loss": 0.9678,
2733
+ "step": 454
2734
+ },
2735
+ {
2736
+ "epoch": 3.2,
2737
+ "learning_rate": 9.988597029207986e-05,
2738
+ "loss": 1.1091,
2739
+ "step": 455
2740
+ },
2741
+ {
2742
+ "epoch": 3.21,
2743
+ "learning_rate": 9.98839732107302e-05,
2744
+ "loss": 1.171,
2745
+ "step": 456
2746
+ },
2747
+ {
2748
+ "epoch": 3.22,
2749
+ "learning_rate": 9.988195881304254e-05,
2750
+ "loss": 1.2926,
2751
+ "step": 457
2752
+ },
2753
+ {
2754
+ "epoch": 3.23,
2755
+ "learning_rate": 9.987992709971617e-05,
2756
+ "loss": 0.9554,
2757
+ "step": 458
2758
+ },
2759
+ {
2760
+ "epoch": 3.23,
2761
+ "learning_rate": 9.987787807145634e-05,
2762
+ "loss": 1.1523,
2763
+ "step": 459
2764
+ },
2765
+ {
2766
+ "epoch": 3.24,
2767
+ "learning_rate": 9.987581172897436e-05,
2768
+ "loss": 1.0537,
2769
+ "step": 460
2770
+ },
2771
+ {
2772
+ "epoch": 3.25,
2773
+ "learning_rate": 9.987372807298752e-05,
2774
+ "loss": 1.2286,
2775
+ "step": 461
2776
+ },
2777
+ {
2778
+ "epoch": 3.25,
2779
+ "learning_rate": 9.987162710421909e-05,
2780
+ "loss": 1.3111,
2781
+ "step": 462
2782
+ },
2783
+ {
2784
+ "epoch": 3.26,
2785
+ "learning_rate": 9.986950882339843e-05,
2786
+ "loss": 1.095,
2787
+ "step": 463
2788
+ },
2789
+ {
2790
+ "epoch": 3.27,
2791
+ "learning_rate": 9.986737323126084e-05,
2792
+ "loss": 1.1133,
2793
+ "step": 464
2794
+ },
2795
+ {
2796
+ "epoch": 3.27,
2797
+ "learning_rate": 9.986522032854766e-05,
2798
+ "loss": 1.1405,
2799
+ "step": 465
2800
+ },
2801
+ {
2802
+ "epoch": 3.28,
2803
+ "learning_rate": 9.986305011600621e-05,
2804
+ "loss": 1.0994,
2805
+ "step": 466
2806
+ },
2807
+ {
2808
+ "epoch": 3.29,
2809
+ "learning_rate": 9.986086259438987e-05,
2810
+ "loss": 1.0086,
2811
+ "step": 467
2812
+ },
2813
+ {
2814
+ "epoch": 3.3,
2815
+ "learning_rate": 9.985865776445799e-05,
2816
+ "loss": 1.0992,
2817
+ "step": 468
2818
+ },
2819
+ {
2820
+ "epoch": 3.3,
2821
+ "learning_rate": 9.985643562697595e-05,
2822
+ "loss": 1.1195,
2823
+ "step": 469
2824
+ },
2825
+ {
2826
+ "epoch": 3.31,
2827
+ "learning_rate": 9.98541961827151e-05,
2828
+ "loss": 1.2934,
2829
+ "step": 470
2830
+ },
2831
+ {
2832
+ "epoch": 3.32,
2833
+ "learning_rate": 9.985193943245284e-05,
2834
+ "loss": 0.932,
2835
+ "step": 471
2836
+ },
2837
+ {
2838
+ "epoch": 3.32,
2839
+ "learning_rate": 9.984966537697254e-05,
2840
+ "loss": 1.4145,
2841
+ "step": 472
2842
+ },
2843
+ {
2844
+ "epoch": 3.33,
2845
+ "learning_rate": 9.984737401706364e-05,
2846
+ "loss": 1.0835,
2847
+ "step": 473
2848
+ },
2849
+ {
2850
+ "epoch": 3.34,
2851
+ "learning_rate": 9.98450653535215e-05,
2852
+ "loss": 1.2985,
2853
+ "step": 474
2854
+ },
2855
+ {
2856
+ "epoch": 3.35,
2857
+ "learning_rate": 9.984273938714756e-05,
2858
+ "loss": 1.1208,
2859
+ "step": 475
2860
+ },
2861
+ {
2862
+ "epoch": 3.35,
2863
+ "learning_rate": 9.984039611874922e-05,
2864
+ "loss": 1.0568,
2865
+ "step": 476
2866
+ },
2867
+ {
2868
+ "epoch": 3.36,
2869
+ "learning_rate": 9.983803554913991e-05,
2870
+ "loss": 1.2661,
2871
+ "step": 477
2872
+ },
2873
+ {
2874
+ "epoch": 3.37,
2875
+ "learning_rate": 9.983565767913908e-05,
2876
+ "loss": 1.1787,
2877
+ "step": 478
2878
+ },
2879
+ {
2880
+ "epoch": 3.37,
2881
+ "learning_rate": 9.983326250957215e-05,
2882
+ "loss": 1.0023,
2883
+ "step": 479
2884
+ },
2885
+ {
2886
+ "epoch": 3.38,
2887
+ "learning_rate": 9.983085004127055e-05,
2888
+ "loss": 1.073,
2889
+ "step": 480
2890
+ },
2891
+ {
2892
+ "epoch": 3.39,
2893
+ "learning_rate": 9.982842027507174e-05,
2894
+ "loss": 0.8713,
2895
+ "step": 481
2896
+ },
2897
+ {
2898
+ "epoch": 3.39,
2899
+ "learning_rate": 9.982597321181916e-05,
2900
+ "loss": 1.2682,
2901
+ "step": 482
2902
+ },
2903
+ {
2904
+ "epoch": 3.4,
2905
+ "learning_rate": 9.982350885236227e-05,
2906
+ "loss": 1.1112,
2907
+ "step": 483
2908
+ },
2909
+ {
2910
+ "epoch": 3.41,
2911
+ "learning_rate": 9.982102719755652e-05,
2912
+ "loss": 1.1175,
2913
+ "step": 484
2914
+ },
2915
+ {
2916
+ "epoch": 3.42,
2917
+ "learning_rate": 9.981852824826341e-05,
2918
+ "loss": 1.0672,
2919
+ "step": 485
2920
+ },
2921
+ {
2922
+ "epoch": 3.42,
2923
+ "learning_rate": 9.981601200535036e-05,
2924
+ "loss": 1.121,
2925
+ "step": 486
2926
+ },
2927
+ {
2928
+ "epoch": 3.43,
2929
+ "learning_rate": 9.981347846969086e-05,
2930
+ "loss": 1.2847,
2931
+ "step": 487
2932
+ },
2933
+ {
2934
+ "epoch": 3.44,
2935
+ "learning_rate": 9.981092764216437e-05,
2936
+ "loss": 1.0444,
2937
+ "step": 488
2938
+ },
2939
+ {
2940
+ "epoch": 3.44,
2941
+ "learning_rate": 9.980835952365639e-05,
2942
+ "loss": 1.0468,
2943
+ "step": 489
2944
+ },
2945
+ {
2946
+ "epoch": 3.45,
2947
+ "learning_rate": 9.980577411505836e-05,
2948
+ "loss": 1.0451,
2949
+ "step": 490
2950
+ },
2951
+ {
2952
+ "epoch": 3.46,
2953
+ "learning_rate": 9.98031714172678e-05,
2954
+ "loss": 1.0078,
2955
+ "step": 491
2956
+ },
2957
+ {
2958
+ "epoch": 3.46,
2959
+ "learning_rate": 9.980055143118815e-05,
2960
+ "loss": 1.189,
2961
+ "step": 492
2962
+ },
2963
+ {
2964
+ "epoch": 3.47,
2965
+ "learning_rate": 9.979791415772892e-05,
2966
+ "loss": 1.1625,
2967
+ "step": 493
2968
+ },
2969
+ {
2970
+ "epoch": 3.48,
2971
+ "learning_rate": 9.979525959780558e-05,
2972
+ "loss": 1.2144,
2973
+ "step": 494
2974
+ },
2975
+ {
2976
+ "epoch": 3.49,
2977
+ "learning_rate": 9.979258775233962e-05,
2978
+ "loss": 1.0964,
2979
+ "step": 495
2980
+ },
2981
+ {
2982
+ "epoch": 3.49,
2983
+ "learning_rate": 9.978989862225853e-05,
2984
+ "loss": 1.2053,
2985
+ "step": 496
2986
+ },
2987
+ {
2988
+ "epoch": 3.5,
2989
+ "learning_rate": 9.978719220849577e-05,
2990
+ "loss": 1.2922,
2991
+ "step": 497
2992
+ },
2993
+ {
2994
+ "epoch": 3.51,
2995
+ "learning_rate": 9.978446851199085e-05,
2996
+ "loss": 1.1391,
2997
+ "step": 498
2998
+ },
2999
+ {
3000
+ "epoch": 3.51,
3001
+ "learning_rate": 9.978172753368924e-05,
3002
+ "loss": 0.9737,
3003
+ "step": 499
3004
+ },
3005
+ {
3006
+ "epoch": 3.52,
3007
+ "learning_rate": 9.977896927454242e-05,
3008
+ "loss": 1.3689,
3009
+ "step": 500
3010
+ },
3011
+ {
3012
+ "epoch": 3.53,
3013
+ "learning_rate": 9.977619373550789e-05,
3014
+ "loss": 1.0774,
3015
+ "step": 501
3016
+ },
3017
+ {
3018
+ "epoch": 3.54,
3019
+ "learning_rate": 9.977340091754909e-05,
3020
+ "loss": 1.0158,
3021
+ "step": 502
3022
+ },
3023
+ {
3024
+ "epoch": 3.54,
3025
+ "learning_rate": 9.977059082163554e-05,
3026
+ "loss": 1.1729,
3027
+ "step": 503
3028
+ },
3029
+ {
3030
+ "epoch": 3.55,
3031
+ "learning_rate": 9.976776344874268e-05,
3032
+ "loss": 1.0058,
3033
+ "step": 504
3034
+ },
3035
+ {
3036
+ "epoch": 3.56,
3037
+ "learning_rate": 9.9764918799852e-05,
3038
+ "loss": 0.9511,
3039
+ "step": 505
3040
+ },
3041
+ {
3042
+ "epoch": 3.56,
3043
+ "learning_rate": 9.976205687595098e-05,
3044
+ "loss": 1.0977,
3045
+ "step": 506
3046
+ },
3047
+ {
3048
+ "epoch": 3.57,
3049
+ "learning_rate": 9.975917767803307e-05,
3050
+ "loss": 0.9736,
3051
+ "step": 507
3052
+ },
3053
+ {
3054
+ "epoch": 3.58,
3055
+ "learning_rate": 9.975628120709774e-05,
3056
+ "loss": 1.0244,
3057
+ "step": 508
3058
+ },
3059
+ {
3060
+ "epoch": 3.58,
3061
+ "learning_rate": 9.975336746415044e-05,
3062
+ "loss": 1.1551,
3063
+ "step": 509
3064
+ },
3065
+ {
3066
+ "epoch": 3.59,
3067
+ "learning_rate": 9.975043645020264e-05,
3068
+ "loss": 1.3528,
3069
+ "step": 510
3070
+ },
3071
+ {
3072
+ "epoch": 3.6,
3073
+ "learning_rate": 9.974748816627176e-05,
3074
+ "loss": 1.1664,
3075
+ "step": 511
3076
+ },
3077
+ {
3078
+ "epoch": 3.61,
3079
+ "learning_rate": 9.974452261338128e-05,
3080
+ "loss": 1.1727,
3081
+ "step": 512
3082
+ },
3083
+ {
3084
+ "epoch": 3.61,
3085
+ "learning_rate": 9.974153979256062e-05,
3086
+ "loss": 1.1234,
3087
+ "step": 513
3088
+ },
3089
+ {
3090
+ "epoch": 3.62,
3091
+ "learning_rate": 9.973853970484521e-05,
3092
+ "loss": 1.0327,
3093
+ "step": 514
3094
+ },
3095
+ {
3096
+ "epoch": 3.63,
3097
+ "learning_rate": 9.973552235127648e-05,
3098
+ "loss": 1.1534,
3099
+ "step": 515
3100
+ },
3101
+ {
3102
+ "epoch": 3.63,
3103
+ "learning_rate": 9.973248773290186e-05,
3104
+ "loss": 1.0124,
3105
+ "step": 516
3106
+ },
3107
+ {
3108
+ "epoch": 3.64,
3109
+ "learning_rate": 9.972943585077478e-05,
3110
+ "loss": 0.9568,
3111
+ "step": 517
3112
+ },
3113
+ {
3114
+ "epoch": 3.65,
3115
+ "learning_rate": 9.972636670595461e-05,
3116
+ "loss": 1.2397,
3117
+ "step": 518
3118
+ },
3119
+ {
3120
+ "epoch": 3.65,
3121
+ "learning_rate": 9.972328029950676e-05,
3122
+ "loss": 1.036,
3123
+ "step": 519
3124
+ },
3125
+ {
3126
+ "epoch": 3.66,
3127
+ "learning_rate": 9.972017663250264e-05,
3128
+ "loss": 1.1054,
3129
+ "step": 520
3130
+ },
3131
+ {
3132
+ "epoch": 3.67,
3133
+ "learning_rate": 9.971705570601958e-05,
3134
+ "loss": 0.8758,
3135
+ "step": 521
3136
+ },
3137
+ {
3138
+ "epoch": 3.68,
3139
+ "learning_rate": 9.971391752114104e-05,
3140
+ "loss": 0.9002,
3141
+ "step": 522
3142
+ },
3143
+ {
3144
+ "epoch": 3.68,
3145
+ "learning_rate": 9.971076207895632e-05,
3146
+ "loss": 1.0288,
3147
+ "step": 523
3148
+ },
3149
+ {
3150
+ "epoch": 3.69,
3151
+ "learning_rate": 9.97075893805608e-05,
3152
+ "loss": 1.078,
3153
+ "step": 524
3154
+ },
3155
+ {
3156
+ "epoch": 3.7,
3157
+ "learning_rate": 9.970439942705582e-05,
3158
+ "loss": 1.1148,
3159
+ "step": 525
3160
+ },
3161
+ {
3162
+ "epoch": 3.7,
3163
+ "learning_rate": 9.970119221954871e-05,
3164
+ "loss": 0.7688,
3165
+ "step": 526
3166
+ },
3167
+ {
3168
+ "epoch": 3.71,
3169
+ "learning_rate": 9.969796775915281e-05,
3170
+ "loss": 0.9705,
3171
+ "step": 527
3172
+ },
3173
+ {
3174
+ "epoch": 3.72,
3175
+ "learning_rate": 9.969472604698742e-05,
3176
+ "loss": 1.1566,
3177
+ "step": 528
3178
+ },
3179
+ {
3180
+ "epoch": 3.73,
3181
+ "learning_rate": 9.969146708417786e-05,
3182
+ "loss": 1.1484,
3183
+ "step": 529
3184
+ },
3185
+ {
3186
+ "epoch": 3.73,
3187
+ "learning_rate": 9.968819087185539e-05,
3188
+ "loss": 0.9396,
3189
+ "step": 530
3190
+ },
3191
+ {
3192
+ "epoch": 3.74,
3193
+ "learning_rate": 9.968489741115734e-05,
3194
+ "loss": 1.1973,
3195
+ "step": 531
3196
+ },
3197
+ {
3198
+ "epoch": 3.75,
3199
+ "learning_rate": 9.968158670322692e-05,
3200
+ "loss": 1.0191,
3201
+ "step": 532
3202
+ },
3203
+ {
3204
+ "epoch": 3.75,
3205
+ "learning_rate": 9.967825874921341e-05,
3206
+ "loss": 1.1284,
3207
+ "step": 533
3208
+ },
3209
+ {
3210
+ "epoch": 3.76,
3211
+ "learning_rate": 9.967491355027207e-05,
3212
+ "loss": 1.1671,
3213
+ "step": 534
3214
+ },
3215
+ {
3216
+ "epoch": 3.77,
3217
+ "learning_rate": 9.96715511075641e-05,
3218
+ "loss": 1.0915,
3219
+ "step": 535
3220
+ },
3221
+ {
3222
+ "epoch": 3.77,
3223
+ "learning_rate": 9.96681714222567e-05,
3224
+ "loss": 1.1303,
3225
+ "step": 536
3226
+ },
3227
+ {
3228
+ "epoch": 3.78,
3229
+ "learning_rate": 9.966477449552312e-05,
3230
+ "loss": 1.1181,
3231
+ "step": 537
3232
+ },
3233
+ {
3234
+ "epoch": 3.79,
3235
+ "learning_rate": 9.96613603285425e-05,
3236
+ "loss": 1.163,
3237
+ "step": 538
3238
+ },
3239
+ {
3240
+ "epoch": 3.8,
3241
+ "learning_rate": 9.96579289225e-05,
3242
+ "loss": 1.0522,
3243
+ "step": 539
3244
+ },
3245
+ {
3246
+ "epoch": 3.8,
3247
+ "learning_rate": 9.96544802785868e-05,
3248
+ "loss": 0.8393,
3249
+ "step": 540
3250
+ },
3251
+ {
3252
+ "epoch": 3.81,
3253
+ "learning_rate": 9.965101439800002e-05,
3254
+ "loss": 1.028,
3255
+ "step": 541
3256
+ },
3257
+ {
3258
+ "epoch": 3.82,
3259
+ "learning_rate": 9.96475312819428e-05,
3260
+ "loss": 0.9271,
3261
+ "step": 542
3262
+ },
3263
+ {
3264
+ "epoch": 3.82,
3265
+ "learning_rate": 9.964403093162422e-05,
3266
+ "loss": 0.9996,
3267
+ "step": 543
3268
+ },
3269
+ {
3270
+ "epoch": 3.83,
3271
+ "learning_rate": 9.964051334825937e-05,
3272
+ "loss": 1.1694,
3273
+ "step": 544
3274
+ },
3275
+ {
3276
+ "epoch": 3.84,
3277
+ "learning_rate": 9.96369785330693e-05,
3278
+ "loss": 0.9082,
3279
+ "step": 545
3280
+ },
3281
+ {
3282
+ "epoch": 3.85,
3283
+ "learning_rate": 9.96334264872811e-05,
3284
+ "loss": 1.0761,
3285
+ "step": 546
3286
+ },
3287
+ {
3288
+ "epoch": 3.85,
3289
+ "learning_rate": 9.962985721212777e-05,
3290
+ "loss": 0.9997,
3291
+ "step": 547
3292
+ },
3293
+ {
3294
+ "epoch": 3.86,
3295
+ "learning_rate": 9.962627070884832e-05,
3296
+ "loss": 0.9916,
3297
+ "step": 548
3298
+ },
3299
+ {
3300
+ "epoch": 3.87,
3301
+ "learning_rate": 9.962266697868776e-05,
3302
+ "loss": 1.2202,
3303
+ "step": 549
3304
+ },
3305
+ {
3306
+ "epoch": 3.87,
3307
+ "learning_rate": 9.961904602289705e-05,
3308
+ "loss": 1.1519,
3309
+ "step": 550
3310
+ },
3311
+ {
3312
+ "epoch": 3.88,
3313
+ "learning_rate": 9.961540784273311e-05,
3314
+ "loss": 0.9058,
3315
+ "step": 551
3316
+ },
3317
+ {
3318
+ "epoch": 3.89,
3319
+ "learning_rate": 9.961175243945893e-05,
3320
+ "loss": 1.1024,
3321
+ "step": 552
3322
+ },
3323
+ {
3324
+ "epoch": 3.89,
3325
+ "learning_rate": 9.960807981434338e-05,
3326
+ "loss": 0.9425,
3327
+ "step": 553
3328
+ },
3329
+ {
3330
+ "epoch": 3.9,
3331
+ "learning_rate": 9.960438996866134e-05,
3332
+ "loss": 1.1816,
3333
+ "step": 554
3334
+ },
3335
+ {
3336
+ "epoch": 3.91,
3337
+ "learning_rate": 9.960068290369371e-05,
3338
+ "loss": 1.2259,
3339
+ "step": 555
3340
+ },
3341
+ {
3342
+ "epoch": 3.92,
3343
+ "learning_rate": 9.959695862072731e-05,
3344
+ "loss": 1.1489,
3345
+ "step": 556
3346
+ },
3347
+ {
3348
+ "epoch": 3.92,
3349
+ "learning_rate": 9.959321712105496e-05,
3350
+ "loss": 1.1185,
3351
+ "step": 557
3352
+ },
3353
+ {
3354
+ "epoch": 3.93,
3355
+ "learning_rate": 9.958945840597545e-05,
3356
+ "loss": 1.0252,
3357
+ "step": 558
3358
+ },
3359
+ {
3360
+ "epoch": 3.94,
3361
+ "learning_rate": 9.958568247679355e-05,
3362
+ "loss": 1.1161,
3363
+ "step": 559
3364
+ },
3365
+ {
3366
+ "epoch": 3.94,
3367
+ "learning_rate": 9.958188933482003e-05,
3368
+ "loss": 1.0131,
3369
+ "step": 560
3370
+ },
3371
+ {
3372
+ "epoch": 3.95,
3373
+ "learning_rate": 9.957807898137157e-05,
3374
+ "loss": 0.9445,
3375
+ "step": 561
3376
+ },
3377
+ {
3378
+ "epoch": 3.96,
3379
+ "learning_rate": 9.957425141777093e-05,
3380
+ "loss": 1.039,
3381
+ "step": 562
3382
+ },
3383
+ {
3384
+ "epoch": 3.96,
3385
+ "learning_rate": 9.957040664534672e-05,
3386
+ "loss": 1.19,
3387
+ "step": 563
3388
+ },
3389
+ {
3390
+ "epoch": 3.97,
3391
+ "learning_rate": 9.95665446654336e-05,
3392
+ "loss": 0.8734,
3393
+ "step": 564
3394
+ },
3395
+ {
3396
+ "epoch": 3.98,
3397
+ "learning_rate": 9.956266547937222e-05,
3398
+ "loss": 1.1848,
3399
+ "step": 565
3400
+ },
3401
+ {
3402
+ "epoch": 3.99,
3403
+ "learning_rate": 9.955876908850913e-05,
3404
+ "loss": 1.1969,
3405
+ "step": 566
3406
+ },
3407
+ {
3408
+ "epoch": 3.99,
3409
+ "learning_rate": 9.955485549419692e-05,
3410
+ "loss": 0.903,
3411
+ "step": 567
3412
+ },
3413
+ {
3414
+ "epoch": 4.0,
3415
+ "learning_rate": 9.95509246977941e-05,
3416
+ "loss": 0.8832,
3417
+ "step": 568
3418
+ },
3419
+ {
3420
+ "epoch": 4.01,
3421
+ "learning_rate": 9.954697670066522e-05,
3422
+ "loss": 1.194,
3423
+ "step": 569
3424
+ },
3425
+ {
3426
+ "epoch": 4.01,
3427
+ "learning_rate": 9.954301150418071e-05,
3428
+ "loss": 0.9009,
3429
+ "step": 570
3430
+ },
3431
+ {
3432
+ "epoch": 4.02,
3433
+ "learning_rate": 9.953902910971703e-05,
3434
+ "loss": 0.8919,
3435
+ "step": 571
3436
+ },
3437
+ {
3438
+ "epoch": 4.03,
3439
+ "learning_rate": 9.95350295186566e-05,
3440
+ "loss": 1.1544,
3441
+ "step": 572
3442
+ },
3443
+ {
3444
+ "epoch": 4.04,
3445
+ "learning_rate": 9.953101273238782e-05,
3446
+ "loss": 0.8454,
3447
+ "step": 573
3448
+ },
3449
+ {
3450
+ "epoch": 4.04,
3451
+ "learning_rate": 9.952697875230506e-05,
3452
+ "loss": 1.0843,
3453
+ "step": 574
3454
+ },
3455
+ {
3456
+ "epoch": 4.05,
3457
+ "learning_rate": 9.95229275798086e-05,
3458
+ "loss": 1.2351,
3459
+ "step": 575
3460
+ },
3461
+ {
3462
+ "epoch": 4.06,
3463
+ "learning_rate": 9.951885921630475e-05,
3464
+ "loss": 0.8698,
3465
+ "step": 576
3466
+ },
3467
+ {
3468
+ "epoch": 4.06,
3469
+ "learning_rate": 9.951477366320579e-05,
3470
+ "loss": 0.9866,
3471
+ "step": 577
3472
+ },
3473
+ {
3474
+ "epoch": 4.07,
3475
+ "learning_rate": 9.951067092192993e-05,
3476
+ "loss": 0.8646,
3477
+ "step": 578
3478
+ },
3479
+ {
3480
+ "epoch": 4.08,
3481
+ "learning_rate": 9.950655099390136e-05,
3482
+ "loss": 0.9214,
3483
+ "step": 579
3484
+ },
3485
+ {
3486
+ "epoch": 4.08,
3487
+ "learning_rate": 9.950241388055024e-05,
3488
+ "loss": 1.1716,
3489
+ "step": 580
3490
+ },
3491
+ {
3492
+ "epoch": 4.09,
3493
+ "learning_rate": 9.949825958331272e-05,
3494
+ "loss": 1.0097,
3495
+ "step": 581
3496
+ },
3497
+ {
3498
+ "epoch": 4.1,
3499
+ "learning_rate": 9.949408810363088e-05,
3500
+ "loss": 0.9652,
3501
+ "step": 582
3502
+ },
3503
+ {
3504
+ "epoch": 4.11,
3505
+ "learning_rate": 9.948989944295274e-05,
3506
+ "loss": 0.9114,
3507
+ "step": 583
3508
+ },
3509
+ {
3510
+ "epoch": 4.11,
3511
+ "learning_rate": 9.948569360273236e-05,
3512
+ "loss": 1.0803,
3513
+ "step": 584
3514
+ },
3515
+ {
3516
+ "epoch": 4.12,
3517
+ "learning_rate": 9.948147058442973e-05,
3518
+ "loss": 1.1052,
3519
+ "step": 585
3520
+ },
3521
+ {
3522
+ "epoch": 4.13,
3523
+ "learning_rate": 9.947723038951076e-05,
3524
+ "loss": 1.0975,
3525
+ "step": 586
3526
+ },
3527
+ {
3528
+ "epoch": 4.13,
3529
+ "learning_rate": 9.947297301944738e-05,
3530
+ "loss": 1.0353,
3531
+ "step": 587
3532
+ },
3533
+ {
3534
+ "epoch": 4.14,
3535
+ "learning_rate": 9.946869847571745e-05,
3536
+ "loss": 1.0835,
3537
+ "step": 588
3538
+ },
3539
+ {
3540
+ "epoch": 4.15,
3541
+ "learning_rate": 9.946440675980482e-05,
3542
+ "loss": 0.8705,
3543
+ "step": 589
3544
+ },
3545
+ {
3546
+ "epoch": 4.15,
3547
+ "learning_rate": 9.946009787319925e-05,
3548
+ "loss": 1.118,
3549
+ "step": 590
3550
+ },
3551
+ {
3552
+ "epoch": 4.16,
3553
+ "learning_rate": 9.945577181739654e-05,
3554
+ "loss": 1.096,
3555
+ "step": 591
3556
+ },
3557
+ {
3558
+ "epoch": 4.17,
3559
+ "learning_rate": 9.945142859389835e-05,
3560
+ "loss": 1.2616,
3561
+ "step": 592
3562
+ },
3563
+ {
3564
+ "epoch": 4.18,
3565
+ "learning_rate": 9.944706820421242e-05,
3566
+ "loss": 1.03,
3567
+ "step": 593
3568
+ },
3569
+ {
3570
+ "epoch": 4.18,
3571
+ "learning_rate": 9.944269064985232e-05,
3572
+ "loss": 0.9491,
3573
+ "step": 594
3574
+ },
3575
+ {
3576
+ "epoch": 4.19,
3577
+ "learning_rate": 9.943829593233767e-05,
3578
+ "loss": 1.0014,
3579
+ "step": 595
3580
+ },
3581
+ {
3582
+ "epoch": 4.2,
3583
+ "learning_rate": 9.943388405319401e-05,
3584
+ "loss": 0.8753,
3585
+ "step": 596
3586
+ },
3587
+ {
3588
+ "epoch": 4.2,
3589
+ "learning_rate": 9.942945501395284e-05,
3590
+ "loss": 1.0214,
3591
+ "step": 597
3592
+ },
3593
+ {
3594
+ "epoch": 4.21,
3595
+ "learning_rate": 9.942500881615164e-05,
3596
+ "loss": 1.0443,
3597
+ "step": 598
3598
+ },
3599
+ {
3600
+ "epoch": 4.22,
3601
+ "learning_rate": 9.942054546133384e-05,
3602
+ "loss": 1.1553,
3603
+ "step": 599
3604
+ },
3605
+ {
3606
+ "epoch": 4.23,
3607
+ "learning_rate": 9.941606495104877e-05,
3608
+ "loss": 0.8192,
3609
+ "step": 600
3610
+ },
3611
+ {
3612
+ "epoch": 4.23,
3613
+ "learning_rate": 9.94115672868518e-05,
3614
+ "loss": 1.014,
3615
+ "step": 601
3616
+ },
3617
+ {
3618
+ "epoch": 4.24,
3619
+ "learning_rate": 9.940705247030418e-05,
3620
+ "loss": 0.8881,
3621
+ "step": 602
3622
+ },
3623
+ {
3624
+ "epoch": 4.25,
3625
+ "learning_rate": 9.940252050297318e-05,
3626
+ "loss": 1.1146,
3627
+ "step": 603
3628
+ },
3629
+ {
3630
+ "epoch": 4.25,
3631
+ "learning_rate": 9.939797138643198e-05,
3632
+ "loss": 1.1543,
3633
+ "step": 604
3634
+ },
3635
+ {
3636
+ "epoch": 4.26,
3637
+ "learning_rate": 9.939340512225972e-05,
3638
+ "loss": 0.8623,
3639
+ "step": 605
3640
+ },
3641
+ {
3642
+ "epoch": 4.27,
3643
+ "learning_rate": 9.938882171204151e-05,
3644
+ "loss": 0.9957,
3645
+ "step": 606
3646
+ },
3647
+ {
3648
+ "epoch": 4.27,
3649
+ "learning_rate": 9.938422115736839e-05,
3650
+ "loss": 1.0105,
3651
+ "step": 607
3652
+ },
3653
+ {
3654
+ "epoch": 4.28,
3655
+ "learning_rate": 9.937960345983736e-05,
3656
+ "loss": 0.93,
3657
+ "step": 608
3658
+ },
3659
+ {
3660
+ "epoch": 4.29,
3661
+ "learning_rate": 9.937496862105138e-05,
3662
+ "loss": 0.8828,
3663
+ "step": 609
3664
+ },
3665
+ {
3666
+ "epoch": 4.3,
3667
+ "learning_rate": 9.937031664261933e-05,
3668
+ "loss": 1.0038,
3669
+ "step": 610
3670
+ },
3671
+ {
3672
+ "epoch": 4.3,
3673
+ "learning_rate": 9.93656475261561e-05,
3674
+ "loss": 1.0304,
3675
+ "step": 611
3676
+ },
3677
+ {
3678
+ "epoch": 4.31,
3679
+ "learning_rate": 9.936096127328247e-05,
3680
+ "loss": 1.1919,
3681
+ "step": 612
3682
+ },
3683
+ {
3684
+ "epoch": 4.32,
3685
+ "learning_rate": 9.93562578856252e-05,
3686
+ "loss": 0.8084,
3687
+ "step": 613
3688
+ },
3689
+ {
3690
+ "epoch": 4.32,
3691
+ "learning_rate": 9.935153736481696e-05,
3692
+ "loss": 1.3357,
3693
+ "step": 614
3694
+ },
3695
+ {
3696
+ "epoch": 4.33,
3697
+ "learning_rate": 9.934679971249641e-05,
3698
+ "loss": 0.9653,
3699
+ "step": 615
3700
+ },
3701
+ {
3702
+ "epoch": 4.34,
3703
+ "learning_rate": 9.934204493030816e-05,
3704
+ "loss": 1.1869,
3705
+ "step": 616
3706
+ },
3707
+ {
3708
+ "epoch": 4.35,
3709
+ "learning_rate": 9.933727301990272e-05,
3710
+ "loss": 0.9646,
3711
+ "step": 617
3712
+ },
3713
+ {
3714
+ "epoch": 4.35,
3715
+ "learning_rate": 9.93324839829366e-05,
3716
+ "loss": 0.9412,
3717
+ "step": 618
3718
+ },
3719
+ {
3720
+ "epoch": 4.36,
3721
+ "learning_rate": 9.932767782107221e-05,
3722
+ "loss": 1.0985,
3723
+ "step": 619
3724
+ },
3725
+ {
3726
+ "epoch": 4.37,
3727
+ "learning_rate": 9.932285453597794e-05,
3728
+ "loss": 1.1193,
3729
+ "step": 620
3730
+ },
3731
+ {
3732
+ "epoch": 4.37,
3733
+ "learning_rate": 9.931801412932808e-05,
3734
+ "loss": 0.8925,
3735
+ "step": 621
3736
+ },
3737
+ {
3738
+ "epoch": 4.38,
3739
+ "learning_rate": 9.931315660280294e-05,
3740
+ "loss": 0.9306,
3741
+ "step": 622
3742
+ },
3743
+ {
3744
+ "epoch": 4.39,
3745
+ "learning_rate": 9.930828195808868e-05,
3746
+ "loss": 0.7386,
3747
+ "step": 623
3748
+ },
3749
+ {
3750
+ "epoch": 4.39,
3751
+ "learning_rate": 9.930339019687746e-05,
3752
+ "loss": 1.1414,
3753
+ "step": 624
3754
+ },
3755
+ {
3756
+ "epoch": 4.4,
3757
+ "learning_rate": 9.929848132086737e-05,
3758
+ "loss": 1.0167,
3759
+ "step": 625
3760
+ },
3761
+ {
3762
+ "epoch": 4.41,
3763
+ "learning_rate": 9.929355533176244e-05,
3764
+ "loss": 1.0053,
3765
+ "step": 626
3766
+ },
3767
+ {
3768
+ "epoch": 4.42,
3769
+ "learning_rate": 9.928861223127264e-05,
3770
+ "loss": 0.9431,
3771
+ "step": 627
3772
+ },
3773
+ {
3774
+ "epoch": 4.42,
3775
+ "learning_rate": 9.928365202111389e-05,
3776
+ "loss": 1.0089,
3777
+ "step": 628
3778
+ },
3779
+ {
3780
+ "epoch": 4.43,
3781
+ "learning_rate": 9.927867470300803e-05,
3782
+ "loss": 1.1917,
3783
+ "step": 629
3784
+ },
3785
+ {
3786
+ "epoch": 4.44,
3787
+ "learning_rate": 9.927368027868283e-05,
3788
+ "loss": 0.9679,
3789
+ "step": 630
3790
+ },
3791
+ {
3792
+ "epoch": 4.44,
3793
+ "learning_rate": 9.926866874987203e-05,
3794
+ "loss": 0.8943,
3795
+ "step": 631
3796
+ },
3797
+ {
3798
+ "epoch": 4.45,
3799
+ "learning_rate": 9.92636401183153e-05,
3800
+ "loss": 0.9199,
3801
+ "step": 632
3802
+ },
3803
+ {
3804
+ "epoch": 4.46,
3805
+ "learning_rate": 9.925859438575824e-05,
3806
+ "loss": 0.8748,
3807
+ "step": 633
3808
+ },
3809
+ {
3810
+ "epoch": 4.46,
3811
+ "learning_rate": 9.925353155395238e-05,
3812
+ "loss": 1.0573,
3813
+ "step": 634
3814
+ },
3815
+ {
3816
+ "epoch": 4.47,
3817
+ "learning_rate": 9.924845162465522e-05,
3818
+ "loss": 1.0292,
3819
+ "step": 635
3820
+ },
3821
+ {
3822
+ "epoch": 4.48,
3823
+ "learning_rate": 9.924335459963011e-05,
3824
+ "loss": 1.0957,
3825
+ "step": 636
3826
+ },
3827
+ {
3828
+ "epoch": 4.49,
3829
+ "learning_rate": 9.923824048064644e-05,
3830
+ "loss": 0.9933,
3831
+ "step": 637
3832
+ },
3833
+ {
3834
+ "epoch": 4.49,
3835
+ "learning_rate": 9.923310926947947e-05,
3836
+ "loss": 1.0641,
3837
+ "step": 638
3838
+ },
3839
+ {
3840
+ "epoch": 4.5,
3841
+ "learning_rate": 9.922796096791041e-05,
3842
+ "loss": 1.1815,
3843
+ "step": 639
3844
+ },
3845
+ {
3846
+ "epoch": 4.51,
3847
+ "learning_rate": 9.922279557772639e-05,
3848
+ "loss": 1.09,
3849
+ "step": 640
3850
+ },
3851
+ {
3852
+ "epoch": 4.51,
3853
+ "learning_rate": 9.921761310072051e-05,
3854
+ "loss": 0.8597,
3855
+ "step": 641
3856
+ },
3857
+ {
3858
+ "epoch": 4.52,
3859
+ "learning_rate": 9.921241353869175e-05,
3860
+ "loss": 1.2314,
3861
+ "step": 642
3862
+ },
3863
+ {
3864
+ "epoch": 4.53,
3865
+ "learning_rate": 9.920719689344506e-05,
3866
+ "loss": 0.9624,
3867
+ "step": 643
3868
+ },
3869
+ {
3870
+ "epoch": 4.54,
3871
+ "learning_rate": 9.920196316679129e-05,
3872
+ "loss": 0.8595,
3873
+ "step": 644
3874
+ },
3875
+ {
3876
+ "epoch": 4.54,
3877
+ "learning_rate": 9.919671236054726e-05,
3878
+ "loss": 1.0533,
3879
+ "step": 645
3880
+ },
3881
+ {
3882
+ "epoch": 4.55,
3883
+ "learning_rate": 9.919144447653569e-05,
3884
+ "loss": 0.9057,
3885
+ "step": 646
3886
+ },
3887
+ {
3888
+ "epoch": 4.56,
3889
+ "learning_rate": 9.918615951658521e-05,
3890
+ "loss": 0.828,
3891
+ "step": 647
3892
+ },
3893
+ {
3894
+ "epoch": 4.56,
3895
+ "learning_rate": 9.918085748253044e-05,
3896
+ "loss": 0.9885,
3897
+ "step": 648
3898
+ },
3899
+ {
3900
+ "epoch": 4.57,
3901
+ "learning_rate": 9.917553837621183e-05,
3902
+ "loss": 0.8416,
3903
+ "step": 649
3904
+ },
3905
+ {
3906
+ "epoch": 4.58,
3907
+ "learning_rate": 9.917020219947586e-05,
3908
+ "loss": 0.8906,
3909
+ "step": 650
3910
+ },
3911
+ {
3912
+ "epoch": 4.58,
3913
+ "learning_rate": 9.916484895417488e-05,
3914
+ "loss": 0.9593,
3915
+ "step": 651
3916
+ },
3917
+ {
3918
+ "epoch": 4.59,
3919
+ "learning_rate": 9.915947864216716e-05,
3920
+ "loss": 1.2514,
3921
+ "step": 652
3922
+ },
3923
+ {
3924
+ "epoch": 4.6,
3925
+ "learning_rate": 9.915409126531691e-05,
3926
+ "loss": 1.082,
3927
+ "step": 653
3928
+ },
3929
+ {
3930
+ "epoch": 4.61,
3931
+ "learning_rate": 9.91486868254943e-05,
3932
+ "loss": 1.0438,
3933
+ "step": 654
3934
+ },
3935
+ {
3936
+ "epoch": 4.61,
3937
+ "learning_rate": 9.914326532457534e-05,
3938
+ "loss": 1.0188,
3939
+ "step": 655
3940
+ },
3941
+ {
3942
+ "epoch": 4.62,
3943
+ "learning_rate": 9.9137826764442e-05,
3944
+ "loss": 0.8931,
3945
+ "step": 656
3946
+ },
3947
+ {
3948
+ "epoch": 4.63,
3949
+ "learning_rate": 9.913237114698223e-05,
3950
+ "loss": 0.9998,
3951
+ "step": 657
3952
+ },
3953
+ {
3954
+ "epoch": 4.63,
3955
+ "learning_rate": 9.912689847408983e-05,
3956
+ "loss": 0.9381,
3957
+ "step": 658
3958
+ },
3959
+ {
3960
+ "epoch": 4.64,
3961
+ "learning_rate": 9.912140874766453e-05,
3962
+ "loss": 0.8725,
3963
+ "step": 659
3964
+ },
3965
+ {
3966
+ "epoch": 4.65,
3967
+ "learning_rate": 9.9115901969612e-05,
3968
+ "loss": 1.1549,
3969
+ "step": 660
3970
+ },
3971
+ {
3972
+ "epoch": 4.65,
3973
+ "learning_rate": 9.911037814184382e-05,
3974
+ "loss": 0.8704,
3975
+ "step": 661
3976
+ },
3977
+ {
3978
+ "epoch": 4.66,
3979
+ "learning_rate": 9.910483726627749e-05,
3980
+ "loss": 0.9152,
3981
+ "step": 662
3982
+ },
3983
+ {
3984
+ "epoch": 4.67,
3985
+ "learning_rate": 9.90992793448364e-05,
3986
+ "loss": 0.7531,
3987
+ "step": 663
3988
+ },
3989
+ {
3990
+ "epoch": 4.68,
3991
+ "learning_rate": 9.909370437944993e-05,
3992
+ "loss": 0.7568,
3993
+ "step": 664
3994
+ },
3995
+ {
3996
+ "epoch": 4.68,
3997
+ "learning_rate": 9.908811237205331e-05,
3998
+ "loss": 0.9515,
3999
+ "step": 665
4000
+ },
4001
+ {
4002
+ "epoch": 4.69,
4003
+ "learning_rate": 9.90825033245877e-05,
4004
+ "loss": 0.9485,
4005
+ "step": 666
4006
+ },
4007
+ {
4008
+ "epoch": 4.7,
4009
+ "learning_rate": 9.907687723900019e-05,
4010
+ "loss": 1.0008,
4011
+ "step": 667
4012
+ },
4013
+ {
4014
+ "epoch": 4.7,
4015
+ "learning_rate": 9.907123411724377e-05,
4016
+ "loss": 0.6774,
4017
+ "step": 668
4018
+ },
4019
+ {
4020
+ "epoch": 4.71,
4021
+ "learning_rate": 9.906557396127734e-05,
4022
+ "loss": 0.8914,
4023
+ "step": 669
4024
+ },
4025
+ {
4026
+ "epoch": 4.72,
4027
+ "learning_rate": 9.905989677306575e-05,
4028
+ "loss": 1.0804,
4029
+ "step": 670
4030
+ },
4031
+ {
4032
+ "epoch": 4.73,
4033
+ "learning_rate": 9.905420255457971e-05,
4034
+ "loss": 1.0676,
4035
+ "step": 671
4036
+ },
4037
+ {
4038
+ "epoch": 4.73,
4039
+ "learning_rate": 9.904849130779588e-05,
4040
+ "loss": 0.8497,
4041
+ "step": 672
4042
+ },
4043
+ {
4044
+ "epoch": 4.74,
4045
+ "learning_rate": 9.904276303469683e-05,
4046
+ "loss": 1.0927,
4047
+ "step": 673
4048
+ },
4049
+ {
4050
+ "epoch": 4.75,
4051
+ "learning_rate": 9.903701773727101e-05,
4052
+ "loss": 0.9029,
4053
+ "step": 674
4054
+ },
4055
+ {
4056
+ "epoch": 4.75,
4057
+ "learning_rate": 9.903125541751279e-05,
4058
+ "loss": 1.05,
4059
+ "step": 675
4060
+ },
4061
+ {
4062
+ "epoch": 4.76,
4063
+ "learning_rate": 9.902547607742247e-05,
4064
+ "loss": 1.084,
4065
+ "step": 676
4066
+ },
4067
+ {
4068
+ "epoch": 4.77,
4069
+ "learning_rate": 9.901967971900627e-05,
4070
+ "loss": 1.0038,
4071
+ "step": 677
4072
+ },
4073
+ {
4074
+ "epoch": 4.77,
4075
+ "learning_rate": 9.901386634427625e-05,
4076
+ "loss": 0.9117,
4077
+ "step": 678
4078
+ },
4079
+ {
4080
+ "epoch": 4.78,
4081
+ "learning_rate": 9.900803595525044e-05,
4082
+ "loss": 0.9534,
4083
+ "step": 679
4084
+ },
4085
+ {
4086
+ "epoch": 4.79,
4087
+ "learning_rate": 9.900218855395276e-05,
4088
+ "loss": 1.0565,
4089
+ "step": 680
4090
+ },
4091
+ {
4092
+ "epoch": 4.8,
4093
+ "learning_rate": 9.899632414241302e-05,
4094
+ "loss": 0.9274,
4095
+ "step": 681
4096
+ },
4097
+ {
4098
+ "epoch": 4.8,
4099
+ "learning_rate": 9.899044272266696e-05,
4100
+ "loss": 0.7033,
4101
+ "step": 682
4102
+ },
4103
+ {
4104
+ "epoch": 4.81,
4105
+ "learning_rate": 9.89845442967562e-05,
4106
+ "loss": 0.8854,
4107
+ "step": 683
4108
+ },
4109
+ {
4110
+ "epoch": 4.82,
4111
+ "learning_rate": 9.897862886672829e-05,
4112
+ "loss": 0.8273,
4113
+ "step": 684
4114
+ },
4115
+ {
4116
+ "epoch": 4.82,
4117
+ "learning_rate": 9.897269643463663e-05,
4118
+ "loss": 0.8897,
4119
+ "step": 685
4120
+ },
4121
+ {
4122
+ "epoch": 4.83,
4123
+ "learning_rate": 9.896674700254061e-05,
4124
+ "loss": 1.037,
4125
+ "step": 686
4126
+ },
4127
+ {
4128
+ "epoch": 4.84,
4129
+ "learning_rate": 9.896078057250542e-05,
4130
+ "loss": 0.8056,
4131
+ "step": 687
4132
+ },
4133
+ {
4134
+ "epoch": 4.85,
4135
+ "learning_rate": 9.895479714660226e-05,
4136
+ "loss": 0.9452,
4137
+ "step": 688
4138
+ },
4139
+ {
4140
+ "epoch": 4.85,
4141
+ "learning_rate": 9.894879672690811e-05,
4142
+ "loss": 0.8404,
4143
+ "step": 689
4144
+ },
4145
+ {
4146
+ "epoch": 4.86,
4147
+ "learning_rate": 9.894277931550593e-05,
4148
+ "loss": 0.8433,
4149
+ "step": 690
4150
+ },
4151
+ {
4152
+ "epoch": 4.87,
4153
+ "learning_rate": 9.893674491448456e-05,
4154
+ "loss": 1.1066,
4155
+ "step": 691
4156
+ },
4157
+ {
4158
+ "epoch": 4.87,
4159
+ "learning_rate": 9.893069352593876e-05,
4160
+ "loss": 1.0541,
4161
+ "step": 692
4162
+ },
4163
+ {
4164
+ "epoch": 4.88,
4165
+ "learning_rate": 9.892462515196914e-05,
4166
+ "loss": 0.7835,
4167
+ "step": 693
4168
+ },
4169
+ {
4170
+ "epoch": 4.89,
4171
+ "learning_rate": 9.89185397946822e-05,
4172
+ "loss": 1.0211,
4173
+ "step": 694
4174
+ },
4175
+ {
4176
+ "epoch": 4.89,
4177
+ "learning_rate": 9.89124374561904e-05,
4178
+ "loss": 0.7885,
4179
+ "step": 695
4180
+ },
4181
+ {
4182
+ "epoch": 4.9,
4183
+ "learning_rate": 9.890631813861207e-05,
4184
+ "loss": 1.0477,
4185
+ "step": 696
4186
+ },
4187
+ {
4188
+ "epoch": 4.91,
4189
+ "learning_rate": 9.89001818440714e-05,
4190
+ "loss": 1.1342,
4191
+ "step": 697
4192
+ },
4193
+ {
4194
+ "epoch": 4.92,
4195
+ "learning_rate": 9.88940285746985e-05,
4196
+ "loss": 1.0765,
4197
+ "step": 698
4198
+ },
4199
+ {
4200
+ "epoch": 4.92,
4201
+ "learning_rate": 9.888785833262935e-05,
4202
+ "loss": 1.0354,
4203
+ "step": 699
4204
+ },
4205
+ {
4206
+ "epoch": 4.93,
4207
+ "learning_rate": 9.888167112000589e-05,
4208
+ "loss": 0.8909,
4209
+ "step": 700
4210
+ },
4211
+ {
4212
+ "epoch": 4.94,
4213
+ "learning_rate": 9.887546693897585e-05,
4214
+ "loss": 0.9852,
4215
+ "step": 701
4216
+ },
4217
+ {
4218
+ "epoch": 4.94,
4219
+ "learning_rate": 9.886924579169293e-05,
4220
+ "loss": 0.8719,
4221
+ "step": 702
4222
+ },
4223
+ {
4224
+ "epoch": 4.95,
4225
+ "learning_rate": 9.886300768031666e-05,
4226
+ "loss": 0.8452,
4227
+ "step": 703
4228
+ },
4229
+ {
4230
+ "epoch": 4.96,
4231
+ "learning_rate": 9.885675260701254e-05,
4232
+ "loss": 0.9104,
4233
+ "step": 704
4234
+ },
4235
+ {
4236
+ "epoch": 4.96,
4237
+ "learning_rate": 9.885048057395187e-05,
4238
+ "loss": 1.1023,
4239
+ "step": 705
4240
+ },
4241
+ {
4242
+ "epoch": 4.97,
4243
+ "learning_rate": 9.884419158331188e-05,
4244
+ "loss": 0.7713,
4245
+ "step": 706
4246
+ },
4247
+ {
4248
+ "epoch": 4.98,
4249
+ "learning_rate": 9.883788563727569e-05,
4250
+ "loss": 1.088,
4251
+ "step": 707
4252
+ },
4253
+ {
4254
+ "epoch": 4.99,
4255
+ "learning_rate": 9.883156273803231e-05,
4256
+ "loss": 0.9891,
4257
+ "step": 708
4258
+ },
4259
+ {
4260
+ "epoch": 4.99,
4261
+ "learning_rate": 9.88252228877766e-05,
4262
+ "loss": 0.7652,
4263
+ "step": 709
4264
+ },
4265
+ {
4266
+ "epoch": 5.0,
4267
+ "learning_rate": 9.881886608870932e-05,
4268
+ "loss": 0.749,
4269
+ "step": 710
4270
+ }
4271
+ ],
4272
+ "logging_steps": 1,
4273
+ "max_steps": 710,
4274
+ "num_train_epochs": 5,
4275
+ "save_steps": 500,
4276
+ "total_flos": 5.578854905151488e+16,
4277
+ "trial_name": null,
4278
+ "trial_params": null
4279
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5209c0fcb6bd3374435acd71f4e80ba80e72720014c29d39325dcdc84f40a785
3
+ size 4027