Hugging Face's logo

Tugay
/
clickbait_spoiling_classification

Transformers
PyTorch
t5
text2text-generation
text-generation-inference
Model card Files
xet
Community
Tugay Bilgiş commited on
Commit
e98a7e9
·
1 Parent(s): 2d0c609

initial

Browse files
Files changed (12) hide show
  1. config.json +60 -0
  2. generation_config.json +7 -0
  3. optimizer.pt +3 -0
  4. pytorch_model.bin +3 -0
  5. rng_state.pth +3 -0
  6. scheduler.pt +3 -0
  7. special_tokens_map.json +107 -0
  8. spiece.model +3 -0
  9. tokenizer.json +0 -0
  10. tokenizer_config.json +112 -0
  11. trainer_state.json +3652 -0
  12. training_args.bin +3 -0
config.json ADDED
@@ -0,0 +1,60 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "t5-large",
3
+ "architectures": [
4
+ "T5ForConditionalGeneration"
5
+ ],
6
+ "d_ff": 4096,
7
+ "d_kv": 64,
8
+ "d_model": 1024,
9
+ "decoder_start_token_id": 0,
10
+ "dense_act_fn": "relu",
11
+ "dropout_rate": 0.1,
12
+ "eos_token_id": 1,
13
+ "feed_forward_proj": "relu",
14
+ "initializer_factor": 1.0,
15
+ "is_encoder_decoder": true,
16
+ "is_gated_act": false,
17
+ "layer_norm_epsilon": 1e-06,
18
+ "model_type": "t5",
19
+ "n_positions": 512,
20
+ "num_decoder_layers": 24,
21
+ "num_heads": 16,
22
+ "num_layers": 24,
23
+ "output_past": true,
24
+ "pad_token_id": 0,
25
+ "relative_attention_max_distance": 128,
26
+ "relative_attention_num_buckets": 32,
27
+ "task_specific_params": {
28
+ "summarization": {
29
+ "early_stopping": true,
30
+ "length_penalty": 2.0,
31
+ "max_length": 200,
32
+ "min_length": 30,
33
+ "no_repeat_ngram_size": 3,
34
+ "num_beams": 4,
35
+ "prefix": "summarize: "
36
+ },
37
+ "translation_en_to_de": {
38
+ "early_stopping": true,
39
+ "max_length": 300,
40
+ "num_beams": 4,
41
+ "prefix": "translate English to German: "
42
+ },
43
+ "translation_en_to_fr": {
44
+ "early_stopping": true,
45
+ "max_length": 300,
46
+ "num_beams": 4,
47
+ "prefix": "translate English to French: "
48
+ },
49
+ "translation_en_to_ro": {
50
+ "early_stopping": true,
51
+ "max_length": 300,
52
+ "num_beams": 4,
53
+ "prefix": "translate English to Romanian: "
54
+ }
55
+ },
56
+ "torch_dtype": "float32",
57
+ "transformers_version": "4.26.0",
58
+ "use_cache": true,
59
+ "vocab_size": 32128
60
+ }
generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "decoder_start_token_id": 0,
4
+ "eos_token_id": 1,
5
+ "pad_token_id": 0,
6
+ "transformers_version": "4.26.0"
7
+ }
optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f5ec61e8182d5189ec497168a245a93db2f4de994c22fb19000aca1f3784e167
3
+ size 5901650259
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6cff018f0286f0ddbe81814b806ba1433b342602bbb6275f3ec79c17fc368dfe
3
+ size 2950844807
rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0f6fcfeebe4f6472c1bba98ec40c6aa8ab24ef4787abf637e4e420769fe794d9
3
+ size 14503
scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:34e4b4ee1825584d1e578f9117585ad7337b770ff25496950404aa042d6a14cb
3
+ size 623
special_tokens_map.json ADDED
@@ -0,0 +1,107 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<extra_id_0>",
4
+ "<extra_id_1>",
5
+ "<extra_id_2>",
6
+ "<extra_id_3>",
7
+ "<extra_id_4>",
8
+ "<extra_id_5>",
9
+ "<extra_id_6>",
10
+ "<extra_id_7>",
11
+ "<extra_id_8>",
12
+ "<extra_id_9>",
13
+ "<extra_id_10>",
14
+ "<extra_id_11>",
15
+ "<extra_id_12>",
16
+ "<extra_id_13>",
17
+ "<extra_id_14>",
18
+ "<extra_id_15>",
19
+ "<extra_id_16>",
20
+ "<extra_id_17>",
21
+ "<extra_id_18>",
22
+ "<extra_id_19>",
23
+ "<extra_id_20>",
24
+ "<extra_id_21>",
25
+ "<extra_id_22>",
26
+ "<extra_id_23>",
27
+ "<extra_id_24>",
28
+ "<extra_id_25>",
29
+ "<extra_id_26>",
30
+ "<extra_id_27>",
31
+ "<extra_id_28>",
32
+ "<extra_id_29>",
33
+ "<extra_id_30>",
34
+ "<extra_id_31>",
35
+ "<extra_id_32>",
36
+ "<extra_id_33>",
37
+ "<extra_id_34>",
38
+ "<extra_id_35>",
39
+ "<extra_id_36>",
40
+ "<extra_id_37>",
41
+ "<extra_id_38>",
42
+ "<extra_id_39>",
43
+ "<extra_id_40>",
44
+ "<extra_id_41>",
45
+ "<extra_id_42>",
46
+ "<extra_id_43>",
47
+ "<extra_id_44>",
48
+ "<extra_id_45>",
49
+ "<extra_id_46>",
50
+ "<extra_id_47>",
51
+ "<extra_id_48>",
52
+ "<extra_id_49>",
53
+ "<extra_id_50>",
54
+ "<extra_id_51>",
55
+ "<extra_id_52>",
56
+ "<extra_id_53>",
57
+ "<extra_id_54>",
58
+ "<extra_id_55>",
59
+ "<extra_id_56>",
60
+ "<extra_id_57>",
61
+ "<extra_id_58>",
62
+ "<extra_id_59>",
63
+ "<extra_id_60>",
64
+ "<extra_id_61>",
65
+ "<extra_id_62>",
66
+ "<extra_id_63>",
67
+ "<extra_id_64>",
68
+ "<extra_id_65>",
69
+ "<extra_id_66>",
70
+ "<extra_id_67>",
71
+ "<extra_id_68>",
72
+ "<extra_id_69>",
73
+ "<extra_id_70>",
74
+ "<extra_id_71>",
75
+ "<extra_id_72>",
76
+ "<extra_id_73>",
77
+ "<extra_id_74>",
78
+ "<extra_id_75>",
79
+ "<extra_id_76>",
80
+ "<extra_id_77>",
81
+ "<extra_id_78>",
82
+ "<extra_id_79>",
83
+ "<extra_id_80>",
84
+ "<extra_id_81>",
85
+ "<extra_id_82>",
86
+ "<extra_id_83>",
87
+ "<extra_id_84>",
88
+ "<extra_id_85>",
89
+ "<extra_id_86>",
90
+ "<extra_id_87>",
91
+ "<extra_id_88>",
92
+ "<extra_id_89>",
93
+ "<extra_id_90>",
94
+ "<extra_id_91>",
95
+ "<extra_id_92>",
96
+ "<extra_id_93>",
97
+ "<extra_id_94>",
98
+ "<extra_id_95>",
99
+ "<extra_id_96>",
100
+ "<extra_id_97>",
101
+ "<extra_id_98>",
102
+ "<extra_id_99>"
103
+ ],
104
+ "eos_token": "</s>",
105
+ "pad_token": "<pad>",
106
+ "unk_token": "<unk>"
107
+ }
spiece.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d60acb128cf7b7f2536e8f38a5b18a05535c9e14c7a355904270e15b0945ea86
3
+ size 791656
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,112 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<extra_id_0>",
4
+ "<extra_id_1>",
5
+ "<extra_id_2>",
6
+ "<extra_id_3>",
7
+ "<extra_id_4>",
8
+ "<extra_id_5>",
9
+ "<extra_id_6>",
10
+ "<extra_id_7>",
11
+ "<extra_id_8>",
12
+ "<extra_id_9>",
13
+ "<extra_id_10>",
14
+ "<extra_id_11>",
15
+ "<extra_id_12>",
16
+ "<extra_id_13>",
17
+ "<extra_id_14>",
18
+ "<extra_id_15>",
19
+ "<extra_id_16>",
20
+ "<extra_id_17>",
21
+ "<extra_id_18>",
22
+ "<extra_id_19>",
23
+ "<extra_id_20>",
24
+ "<extra_id_21>",
25
+ "<extra_id_22>",
26
+ "<extra_id_23>",
27
+ "<extra_id_24>",
28
+ "<extra_id_25>",
29
+ "<extra_id_26>",
30
+ "<extra_id_27>",
31
+ "<extra_id_28>",
32
+ "<extra_id_29>",
33
+ "<extra_id_30>",
34
+ "<extra_id_31>",
35
+ "<extra_id_32>",
36
+ "<extra_id_33>",
37
+ "<extra_id_34>",
38
+ "<extra_id_35>",
39
+ "<extra_id_36>",
40
+ "<extra_id_37>",
41
+ "<extra_id_38>",
42
+ "<extra_id_39>",
43
+ "<extra_id_40>",
44
+ "<extra_id_41>",
45
+ "<extra_id_42>",
46
+ "<extra_id_43>",
47
+ "<extra_id_44>",
48
+ "<extra_id_45>",
49
+ "<extra_id_46>",
50
+ "<extra_id_47>",
51
+ "<extra_id_48>",
52
+ "<extra_id_49>",
53
+ "<extra_id_50>",
54
+ "<extra_id_51>",
55
+ "<extra_id_52>",
56
+ "<extra_id_53>",
57
+ "<extra_id_54>",
58
+ "<extra_id_55>",
59
+ "<extra_id_56>",
60
+ "<extra_id_57>",
61
+ "<extra_id_58>",
62
+ "<extra_id_59>",
63
+ "<extra_id_60>",
64
+ "<extra_id_61>",
65
+ "<extra_id_62>",
66
+ "<extra_id_63>",
67
+ "<extra_id_64>",
68
+ "<extra_id_65>",
69
+ "<extra_id_66>",
70
+ "<extra_id_67>",
71
+ "<extra_id_68>",
72
+ "<extra_id_69>",
73
+ "<extra_id_70>",
74
+ "<extra_id_71>",
75
+ "<extra_id_72>",
76
+ "<extra_id_73>",
77
+ "<extra_id_74>",
78
+ "<extra_id_75>",
79
+ "<extra_id_76>",
80
+ "<extra_id_77>",
81
+ "<extra_id_78>",
82
+ "<extra_id_79>",
83
+ "<extra_id_80>",
84
+ "<extra_id_81>",
85
+ "<extra_id_82>",
86
+ "<extra_id_83>",
87
+ "<extra_id_84>",
88
+ "<extra_id_85>",
89
+ "<extra_id_86>",
90
+ "<extra_id_87>",
91
+ "<extra_id_88>",
92
+ "<extra_id_89>",
93
+ "<extra_id_90>",
94
+ "<extra_id_91>",
95
+ "<extra_id_92>",
96
+ "<extra_id_93>",
97
+ "<extra_id_94>",
98
+ "<extra_id_95>",
99
+ "<extra_id_96>",
100
+ "<extra_id_97>",
101
+ "<extra_id_98>",
102
+ "<extra_id_99>"
103
+ ],
104
+ "eos_token": "</s>",
105
+ "extra_ids": 100,
106
+ "model_max_length": 512,
107
+ "name_or_path": "t5-large",
108
+ "pad_token": "<pad>",
109
+ "special_tokens_map_file": null,
110
+ "tokenizer_class": "T5Tokenizer",
111
+ "unk_token": "<unk>"
112
+ }
trainer_state.json ADDED
@@ -0,0 +1,3652 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 3.0,
5
+ "global_step": 1200,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.01,
12
+ "learning_rate": 4.9975e-05,
13
+ "loss": 14.5662,
14
+ "step": 2
15
+ },
16
+ {
17
+ "epoch": 0.01,
18
+ "learning_rate": 4.995e-05,
19
+ "loss": 7.1236,
20
+ "step": 4
21
+ },
22
+ {
23
+ "epoch": 0.01,
24
+ "learning_rate": 4.992500000000001e-05,
25
+ "loss": 5.3202,
26
+ "step": 6
27
+ },
28
+ {
29
+ "epoch": 0.02,
30
+ "learning_rate": 4.99e-05,
31
+ "loss": 3.9452,
32
+ "step": 8
33
+ },
34
+ {
35
+ "epoch": 0.03,
36
+ "learning_rate": 4.9875000000000006e-05,
37
+ "loss": 3.4958,
38
+ "step": 10
39
+ },
40
+ {
41
+ "epoch": 0.03,
42
+ "learning_rate": 4.9850000000000006e-05,
43
+ "loss": 2.8169,
44
+ "step": 12
45
+ },
46
+ {
47
+ "epoch": 0.04,
48
+ "learning_rate": 4.9825000000000005e-05,
49
+ "loss": 1.9955,
50
+ "step": 14
51
+ },
52
+ {
53
+ "epoch": 0.04,
54
+ "learning_rate": 4.9800000000000004e-05,
55
+ "loss": 1.5664,
56
+ "step": 16
57
+ },
58
+ {
59
+ "epoch": 0.04,
60
+ "learning_rate": 4.9775000000000004e-05,
61
+ "loss": 1.3016,
62
+ "step": 18
63
+ },
64
+ {
65
+ "epoch": 0.05,
66
+ "learning_rate": 4.975e-05,
67
+ "loss": 0.8614,
68
+ "step": 20
69
+ },
70
+ {
71
+ "epoch": 0.06,
72
+ "learning_rate": 4.9725e-05,
73
+ "loss": 1.0415,
74
+ "step": 22
75
+ },
76
+ {
77
+ "epoch": 0.06,
78
+ "learning_rate": 4.97e-05,
79
+ "loss": 0.5895,
80
+ "step": 24
81
+ },
82
+ {
83
+ "epoch": 0.07,
84
+ "learning_rate": 4.967500000000001e-05,
85
+ "loss": 0.4986,
86
+ "step": 26
87
+ },
88
+ {
89
+ "epoch": 0.07,
90
+ "learning_rate": 4.965e-05,
91
+ "loss": 0.8056,
92
+ "step": 28
93
+ },
94
+ {
95
+ "epoch": 0.07,
96
+ "learning_rate": 4.962500000000001e-05,
97
+ "loss": 0.8067,
98
+ "step": 30
99
+ },
100
+ {
101
+ "epoch": 0.08,
102
+ "learning_rate": 4.96e-05,
103
+ "loss": 0.5058,
104
+ "step": 32
105
+ },
106
+ {
107
+ "epoch": 0.09,
108
+ "learning_rate": 4.9575000000000006e-05,
109
+ "loss": 0.6796,
110
+ "step": 34
111
+ },
112
+ {
113
+ "epoch": 0.09,
114
+ "learning_rate": 4.9550000000000005e-05,
115
+ "loss": 0.8734,
116
+ "step": 36
117
+ },
118
+ {
119
+ "epoch": 0.1,
120
+ "learning_rate": 4.9525000000000004e-05,
121
+ "loss": 0.9313,
122
+ "step": 38
123
+ },
124
+ {
125
+ "epoch": 0.1,
126
+ "learning_rate": 4.9500000000000004e-05,
127
+ "loss": 0.5948,
128
+ "step": 40
129
+ },
130
+ {
131
+ "epoch": 0.1,
132
+ "learning_rate": 4.9475e-05,
133
+ "loss": 0.6952,
134
+ "step": 42
135
+ },
136
+ {
137
+ "epoch": 0.11,
138
+ "learning_rate": 4.945e-05,
139
+ "loss": 0.562,
140
+ "step": 44
141
+ },
142
+ {
143
+ "epoch": 0.12,
144
+ "learning_rate": 4.9425e-05,
145
+ "loss": 0.7806,
146
+ "step": 46
147
+ },
148
+ {
149
+ "epoch": 0.12,
150
+ "learning_rate": 4.94e-05,
151
+ "loss": 0.5262,
152
+ "step": 48
153
+ },
154
+ {
155
+ "epoch": 0.12,
156
+ "learning_rate": 4.937500000000001e-05,
157
+ "loss": 0.7553,
158
+ "step": 50
159
+ },
160
+ {
161
+ "epoch": 0.13,
162
+ "learning_rate": 4.935e-05,
163
+ "loss": 0.7249,
164
+ "step": 52
165
+ },
166
+ {
167
+ "epoch": 0.14,
168
+ "learning_rate": 4.9325000000000006e-05,
169
+ "loss": 0.7524,
170
+ "step": 54
171
+ },
172
+ {
173
+ "epoch": 0.14,
174
+ "learning_rate": 4.93e-05,
175
+ "loss": 0.6196,
176
+ "step": 56
177
+ },
178
+ {
179
+ "epoch": 0.14,
180
+ "learning_rate": 4.9275000000000005e-05,
181
+ "loss": 0.7026,
182
+ "step": 58
183
+ },
184
+ {
185
+ "epoch": 0.15,
186
+ "learning_rate": 4.9250000000000004e-05,
187
+ "loss": 0.7122,
188
+ "step": 60
189
+ },
190
+ {
191
+ "epoch": 0.15,
192
+ "learning_rate": 4.9225000000000004e-05,
193
+ "loss": 0.7,
194
+ "step": 62
195
+ },
196
+ {
197
+ "epoch": 0.16,
198
+ "learning_rate": 4.92e-05,
199
+ "loss": 0.6362,
200
+ "step": 64
201
+ },
202
+ {
203
+ "epoch": 0.17,
204
+ "learning_rate": 4.9175e-05,
205
+ "loss": 0.5532,
206
+ "step": 66
207
+ },
208
+ {
209
+ "epoch": 0.17,
210
+ "learning_rate": 4.915e-05,
211
+ "loss": 0.735,
212
+ "step": 68
213
+ },
214
+ {
215
+ "epoch": 0.17,
216
+ "learning_rate": 4.9125e-05,
217
+ "loss": 0.6036,
218
+ "step": 70
219
+ },
220
+ {
221
+ "epoch": 0.18,
222
+ "learning_rate": 4.91e-05,
223
+ "loss": 0.583,
224
+ "step": 72
225
+ },
226
+ {
227
+ "epoch": 0.18,
228
+ "learning_rate": 4.907500000000001e-05,
229
+ "loss": 0.5929,
230
+ "step": 74
231
+ },
232
+ {
233
+ "epoch": 0.19,
234
+ "learning_rate": 4.905e-05,
235
+ "loss": 0.7786,
236
+ "step": 76
237
+ },
238
+ {
239
+ "epoch": 0.2,
240
+ "learning_rate": 4.9025000000000006e-05,
241
+ "loss": 0.6554,
242
+ "step": 78
243
+ },
244
+ {
245
+ "epoch": 0.2,
246
+ "learning_rate": 4.9e-05,
247
+ "loss": 0.6304,
248
+ "step": 80
249
+ },
250
+ {
251
+ "epoch": 0.2,
252
+ "learning_rate": 4.8975000000000005e-05,
253
+ "loss": 0.6701,
254
+ "step": 82
255
+ },
256
+ {
257
+ "epoch": 0.21,
258
+ "learning_rate": 4.8950000000000004e-05,
259
+ "loss": 0.6078,
260
+ "step": 84
261
+ },
262
+ {
263
+ "epoch": 0.21,
264
+ "learning_rate": 4.8925e-05,
265
+ "loss": 0.6914,
266
+ "step": 86
267
+ },
268
+ {
269
+ "epoch": 0.22,
270
+ "learning_rate": 4.89e-05,
271
+ "loss": 0.6281,
272
+ "step": 88
273
+ },
274
+ {
275
+ "epoch": 0.23,
276
+ "learning_rate": 4.8875e-05,
277
+ "loss": 0.6858,
278
+ "step": 90
279
+ },
280
+ {
281
+ "epoch": 0.23,
282
+ "learning_rate": 4.885e-05,
283
+ "loss": 0.5675,
284
+ "step": 92
285
+ },
286
+ {
287
+ "epoch": 0.23,
288
+ "learning_rate": 4.8825e-05,
289
+ "loss": 0.7266,
290
+ "step": 94
291
+ },
292
+ {
293
+ "epoch": 0.24,
294
+ "learning_rate": 4.88e-05,
295
+ "loss": 0.9016,
296
+ "step": 96
297
+ },
298
+ {
299
+ "epoch": 0.24,
300
+ "learning_rate": 4.8775000000000007e-05,
301
+ "loss": 0.5519,
302
+ "step": 98
303
+ },
304
+ {
305
+ "epoch": 0.25,
306
+ "learning_rate": 4.875e-05,
307
+ "loss": 0.5509,
308
+ "step": 100
309
+ },
310
+ {
311
+ "epoch": 0.26,
312
+ "learning_rate": 4.8725000000000005e-05,
313
+ "loss": 1.0457,
314
+ "step": 102
315
+ },
316
+ {
317
+ "epoch": 0.26,
318
+ "learning_rate": 4.87e-05,
319
+ "loss": 0.8419,
320
+ "step": 104
321
+ },
322
+ {
323
+ "epoch": 0.27,
324
+ "learning_rate": 4.8675000000000004e-05,
325
+ "loss": 0.6735,
326
+ "step": 106
327
+ },
328
+ {
329
+ "epoch": 0.27,
330
+ "learning_rate": 4.8650000000000003e-05,
331
+ "loss": 0.5402,
332
+ "step": 108
333
+ },
334
+ {
335
+ "epoch": 0.28,
336
+ "learning_rate": 4.8625e-05,
337
+ "loss": 0.5331,
338
+ "step": 110
339
+ },
340
+ {
341
+ "epoch": 0.28,
342
+ "learning_rate": 4.86e-05,
343
+ "loss": 0.5752,
344
+ "step": 112
345
+ },
346
+ {
347
+ "epoch": 0.28,
348
+ "learning_rate": 4.8575e-05,
349
+ "loss": 0.5226,
350
+ "step": 114
351
+ },
352
+ {
353
+ "epoch": 0.29,
354
+ "learning_rate": 4.855e-05,
355
+ "loss": 0.6734,
356
+ "step": 116
357
+ },
358
+ {
359
+ "epoch": 0.29,
360
+ "learning_rate": 4.8525e-05,
361
+ "loss": 0.4719,
362
+ "step": 118
363
+ },
364
+ {
365
+ "epoch": 0.3,
366
+ "learning_rate": 4.85e-05,
367
+ "loss": 0.5542,
368
+ "step": 120
369
+ },
370
+ {
371
+ "epoch": 0.3,
372
+ "learning_rate": 4.8475000000000006e-05,
373
+ "loss": 0.7252,
374
+ "step": 122
375
+ },
376
+ {
377
+ "epoch": 0.31,
378
+ "learning_rate": 4.845e-05,
379
+ "loss": 0.5549,
380
+ "step": 124
381
+ },
382
+ {
383
+ "epoch": 0.32,
384
+ "learning_rate": 4.8425000000000005e-05,
385
+ "loss": 0.4797,
386
+ "step": 126
387
+ },
388
+ {
389
+ "epoch": 0.32,
390
+ "learning_rate": 4.8400000000000004e-05,
391
+ "loss": 0.6463,
392
+ "step": 128
393
+ },
394
+ {
395
+ "epoch": 0.33,
396
+ "learning_rate": 4.8375000000000004e-05,
397
+ "loss": 0.6462,
398
+ "step": 130
399
+ },
400
+ {
401
+ "epoch": 0.33,
402
+ "learning_rate": 4.835e-05,
403
+ "loss": 0.5555,
404
+ "step": 132
405
+ },
406
+ {
407
+ "epoch": 0.34,
408
+ "learning_rate": 4.8325e-05,
409
+ "loss": 0.7847,
410
+ "step": 134
411
+ },
412
+ {
413
+ "epoch": 0.34,
414
+ "learning_rate": 4.83e-05,
415
+ "loss": 0.6703,
416
+ "step": 136
417
+ },
418
+ {
419
+ "epoch": 0.34,
420
+ "learning_rate": 4.8275e-05,
421
+ "loss": 0.5331,
422
+ "step": 138
423
+ },
424
+ {
425
+ "epoch": 0.35,
426
+ "learning_rate": 4.825e-05,
427
+ "loss": 0.5694,
428
+ "step": 140
429
+ },
430
+ {
431
+ "epoch": 0.35,
432
+ "learning_rate": 4.822500000000001e-05,
433
+ "loss": 0.6958,
434
+ "step": 142
435
+ },
436
+ {
437
+ "epoch": 0.36,
438
+ "learning_rate": 4.82e-05,
439
+ "loss": 0.6098,
440
+ "step": 144
441
+ },
442
+ {
443
+ "epoch": 0.36,
444
+ "learning_rate": 4.8175000000000005e-05,
445
+ "loss": 0.5491,
446
+ "step": 146
447
+ },
448
+ {
449
+ "epoch": 0.37,
450
+ "learning_rate": 4.815e-05,
451
+ "loss": 0.6715,
452
+ "step": 148
453
+ },
454
+ {
455
+ "epoch": 0.38,
456
+ "learning_rate": 4.8125000000000004e-05,
457
+ "loss": 0.3916,
458
+ "step": 150
459
+ },
460
+ {
461
+ "epoch": 0.38,
462
+ "learning_rate": 4.8100000000000004e-05,
463
+ "loss": 0.6793,
464
+ "step": 152
465
+ },
466
+ {
467
+ "epoch": 0.39,
468
+ "learning_rate": 4.8075e-05,
469
+ "loss": 0.5973,
470
+ "step": 154
471
+ },
472
+ {
473
+ "epoch": 0.39,
474
+ "learning_rate": 4.805e-05,
475
+ "loss": 1.2975,
476
+ "step": 156
477
+ },
478
+ {
479
+ "epoch": 0.4,
480
+ "learning_rate": 4.8025e-05,
481
+ "loss": 0.6328,
482
+ "step": 158
483
+ },
484
+ {
485
+ "epoch": 0.4,
486
+ "learning_rate": 4.8e-05,
487
+ "loss": 0.4737,
488
+ "step": 160
489
+ },
490
+ {
491
+ "epoch": 0.41,
492
+ "learning_rate": 4.7975e-05,
493
+ "loss": 0.6352,
494
+ "step": 162
495
+ },
496
+ {
497
+ "epoch": 0.41,
498
+ "learning_rate": 4.795e-05,
499
+ "loss": 0.4901,
500
+ "step": 164
501
+ },
502
+ {
503
+ "epoch": 0.41,
504
+ "learning_rate": 4.7925000000000006e-05,
505
+ "loss": 0.5981,
506
+ "step": 166
507
+ },
508
+ {
509
+ "epoch": 0.42,
510
+ "learning_rate": 4.79e-05,
511
+ "loss": 0.507,
512
+ "step": 168
513
+ },
514
+ {
515
+ "epoch": 0.42,
516
+ "learning_rate": 4.7875000000000005e-05,
517
+ "loss": 0.8278,
518
+ "step": 170
519
+ },
520
+ {
521
+ "epoch": 0.43,
522
+ "learning_rate": 4.785e-05,
523
+ "loss": 0.8347,
524
+ "step": 172
525
+ },
526
+ {
527
+ "epoch": 0.43,
528
+ "learning_rate": 4.7825000000000004e-05,
529
+ "loss": 0.5688,
530
+ "step": 174
531
+ },
532
+ {
533
+ "epoch": 0.44,
534
+ "learning_rate": 4.78e-05,
535
+ "loss": 0.5499,
536
+ "step": 176
537
+ },
538
+ {
539
+ "epoch": 0.45,
540
+ "learning_rate": 4.7775e-05,
541
+ "loss": 0.8471,
542
+ "step": 178
543
+ },
544
+ {
545
+ "epoch": 0.45,
546
+ "learning_rate": 4.775e-05,
547
+ "loss": 0.917,
548
+ "step": 180
549
+ },
550
+ {
551
+ "epoch": 0.46,
552
+ "learning_rate": 4.7725e-05,
553
+ "loss": 0.6619,
554
+ "step": 182
555
+ },
556
+ {
557
+ "epoch": 0.46,
558
+ "learning_rate": 4.77e-05,
559
+ "loss": 0.7828,
560
+ "step": 184
561
+ },
562
+ {
563
+ "epoch": 0.47,
564
+ "learning_rate": 4.7675e-05,
565
+ "loss": 0.5367,
566
+ "step": 186
567
+ },
568
+ {
569
+ "epoch": 0.47,
570
+ "learning_rate": 4.765e-05,
571
+ "loss": 0.4836,
572
+ "step": 188
573
+ },
574
+ {
575
+ "epoch": 0.47,
576
+ "learning_rate": 4.7625000000000006e-05,
577
+ "loss": 0.5622,
578
+ "step": 190
579
+ },
580
+ {
581
+ "epoch": 0.48,
582
+ "learning_rate": 4.76e-05,
583
+ "loss": 0.9229,
584
+ "step": 192
585
+ },
586
+ {
587
+ "epoch": 0.48,
588
+ "learning_rate": 4.7575000000000004e-05,
589
+ "loss": 0.5629,
590
+ "step": 194
591
+ },
592
+ {
593
+ "epoch": 0.49,
594
+ "learning_rate": 4.755e-05,
595
+ "loss": 0.4308,
596
+ "step": 196
597
+ },
598
+ {
599
+ "epoch": 0.49,
600
+ "learning_rate": 4.7525e-05,
601
+ "loss": 0.4076,
602
+ "step": 198
603
+ },
604
+ {
605
+ "epoch": 0.5,
606
+ "learning_rate": 4.75e-05,
607
+ "loss": 0.4774,
608
+ "step": 200
609
+ },
610
+ {
611
+ "epoch": 0.51,
612
+ "learning_rate": 4.7475e-05,
613
+ "loss": 0.5469,
614
+ "step": 202
615
+ },
616
+ {
617
+ "epoch": 0.51,
618
+ "learning_rate": 4.745e-05,
619
+ "loss": 0.5152,
620
+ "step": 204
621
+ },
622
+ {
623
+ "epoch": 0.52,
624
+ "learning_rate": 4.7425e-05,
625
+ "loss": 0.6016,
626
+ "step": 206
627
+ },
628
+ {
629
+ "epoch": 0.52,
630
+ "learning_rate": 4.74e-05,
631
+ "loss": 0.5576,
632
+ "step": 208
633
+ },
634
+ {
635
+ "epoch": 0.53,
636
+ "learning_rate": 4.7375e-05,
637
+ "loss": 0.6042,
638
+ "step": 210
639
+ },
640
+ {
641
+ "epoch": 0.53,
642
+ "learning_rate": 4.735e-05,
643
+ "loss": 0.4603,
644
+ "step": 212
645
+ },
646
+ {
647
+ "epoch": 0.54,
648
+ "learning_rate": 4.7325000000000005e-05,
649
+ "loss": 0.6274,
650
+ "step": 214
651
+ },
652
+ {
653
+ "epoch": 0.54,
654
+ "learning_rate": 4.73e-05,
655
+ "loss": 0.4953,
656
+ "step": 216
657
+ },
658
+ {
659
+ "epoch": 0.55,
660
+ "learning_rate": 4.7275000000000004e-05,
661
+ "loss": 0.5774,
662
+ "step": 218
663
+ },
664
+ {
665
+ "epoch": 0.55,
666
+ "learning_rate": 4.7249999999999997e-05,
667
+ "loss": 0.7247,
668
+ "step": 220
669
+ },
670
+ {
671
+ "epoch": 0.56,
672
+ "learning_rate": 4.7225e-05,
673
+ "loss": 0.541,
674
+ "step": 222
675
+ },
676
+ {
677
+ "epoch": 0.56,
678
+ "learning_rate": 4.72e-05,
679
+ "loss": 0.8272,
680
+ "step": 224
681
+ },
682
+ {
683
+ "epoch": 0.56,
684
+ "learning_rate": 4.7175e-05,
685
+ "loss": 0.7274,
686
+ "step": 226
687
+ },
688
+ {
689
+ "epoch": 0.57,
690
+ "learning_rate": 4.715e-05,
691
+ "loss": 0.5989,
692
+ "step": 228
693
+ },
694
+ {
695
+ "epoch": 0.57,
696
+ "learning_rate": 4.7125e-05,
697
+ "loss": 0.4759,
698
+ "step": 230
699
+ },
700
+ {
701
+ "epoch": 0.58,
702
+ "learning_rate": 4.71e-05,
703
+ "loss": 0.5049,
704
+ "step": 232
705
+ },
706
+ {
707
+ "epoch": 0.58,
708
+ "learning_rate": 4.7075e-05,
709
+ "loss": 0.4231,
710
+ "step": 234
711
+ },
712
+ {
713
+ "epoch": 0.59,
714
+ "learning_rate": 4.705e-05,
715
+ "loss": 0.6353,
716
+ "step": 236
717
+ },
718
+ {
719
+ "epoch": 0.59,
720
+ "learning_rate": 4.7025000000000005e-05,
721
+ "loss": 0.7109,
722
+ "step": 238
723
+ },
724
+ {
725
+ "epoch": 0.6,
726
+ "learning_rate": 4.7e-05,
727
+ "loss": 0.4595,
728
+ "step": 240
729
+ },
730
+ {
731
+ "epoch": 0.6,
732
+ "learning_rate": 4.6975000000000003e-05,
733
+ "loss": 0.4064,
734
+ "step": 242
735
+ },
736
+ {
737
+ "epoch": 0.61,
738
+ "learning_rate": 4.695e-05,
739
+ "loss": 0.5207,
740
+ "step": 244
741
+ },
742
+ {
743
+ "epoch": 0.61,
744
+ "learning_rate": 4.6925e-05,
745
+ "loss": 0.5171,
746
+ "step": 246
747
+ },
748
+ {
749
+ "epoch": 0.62,
750
+ "learning_rate": 4.69e-05,
751
+ "loss": 0.5339,
752
+ "step": 248
753
+ },
754
+ {
755
+ "epoch": 0.62,
756
+ "learning_rate": 4.6875e-05,
757
+ "loss": 0.6237,
758
+ "step": 250
759
+ },
760
+ {
761
+ "epoch": 0.63,
762
+ "learning_rate": 4.685000000000001e-05,
763
+ "loss": 0.459,
764
+ "step": 252
765
+ },
766
+ {
767
+ "epoch": 0.64,
768
+ "learning_rate": 4.6825e-05,
769
+ "loss": 0.4735,
770
+ "step": 254
771
+ },
772
+ {
773
+ "epoch": 0.64,
774
+ "learning_rate": 4.6800000000000006e-05,
775
+ "loss": 0.623,
776
+ "step": 256
777
+ },
778
+ {
779
+ "epoch": 0.65,
780
+ "learning_rate": 4.6775000000000005e-05,
781
+ "loss": 0.5452,
782
+ "step": 258
783
+ },
784
+ {
785
+ "epoch": 0.65,
786
+ "learning_rate": 4.6750000000000005e-05,
787
+ "loss": 0.5576,
788
+ "step": 260
789
+ },
790
+ {
791
+ "epoch": 0.66,
792
+ "learning_rate": 4.6725000000000004e-05,
793
+ "loss": 0.4818,
794
+ "step": 262
795
+ },
796
+ {
797
+ "epoch": 0.66,
798
+ "learning_rate": 4.6700000000000003e-05,
799
+ "loss": 0.5624,
800
+ "step": 264
801
+ },
802
+ {
803
+ "epoch": 0.67,
804
+ "learning_rate": 4.6675e-05,
805
+ "loss": 0.6891,
806
+ "step": 266
807
+ },
808
+ {
809
+ "epoch": 0.67,
810
+ "learning_rate": 4.665e-05,
811
+ "loss": 0.5276,
812
+ "step": 268
813
+ },
814
+ {
815
+ "epoch": 0.68,
816
+ "learning_rate": 4.6625e-05,
817
+ "loss": 0.4512,
818
+ "step": 270
819
+ },
820
+ {
821
+ "epoch": 0.68,
822
+ "learning_rate": 4.660000000000001e-05,
823
+ "loss": 0.5978,
824
+ "step": 272
825
+ },
826
+ {
827
+ "epoch": 0.69,
828
+ "learning_rate": 4.6575e-05,
829
+ "loss": 0.5461,
830
+ "step": 274
831
+ },
832
+ {
833
+ "epoch": 0.69,
834
+ "learning_rate": 4.655000000000001e-05,
835
+ "loss": 0.5528,
836
+ "step": 276
837
+ },
838
+ {
839
+ "epoch": 0.69,
840
+ "learning_rate": 4.6525e-05,
841
+ "loss": 0.4676,
842
+ "step": 278
843
+ },
844
+ {
845
+ "epoch": 0.7,
846
+ "learning_rate": 4.6500000000000005e-05,
847
+ "loss": 0.4727,
848
+ "step": 280
849
+ },
850
+ {
851
+ "epoch": 0.7,
852
+ "learning_rate": 4.6475000000000005e-05,
853
+ "loss": 0.4363,
854
+ "step": 282
855
+ },
856
+ {
857
+ "epoch": 0.71,
858
+ "learning_rate": 4.6450000000000004e-05,
859
+ "loss": 0.5155,
860
+ "step": 284
861
+ },
862
+ {
863
+ "epoch": 0.71,
864
+ "learning_rate": 4.6425000000000004e-05,
865
+ "loss": 0.593,
866
+ "step": 286
867
+ },
868
+ {
869
+ "epoch": 0.72,
870
+ "learning_rate": 4.64e-05,
871
+ "loss": 0.4865,
872
+ "step": 288
873
+ },
874
+ {
875
+ "epoch": 0.72,
876
+ "learning_rate": 4.6375e-05,
877
+ "loss": 0.5427,
878
+ "step": 290
879
+ },
880
+ {
881
+ "epoch": 0.73,
882
+ "learning_rate": 4.635e-05,
883
+ "loss": 0.3649,
884
+ "step": 292
885
+ },
886
+ {
887
+ "epoch": 0.73,
888
+ "learning_rate": 4.6325e-05,
889
+ "loss": 0.3859,
890
+ "step": 294
891
+ },
892
+ {
893
+ "epoch": 0.74,
894
+ "learning_rate": 4.630000000000001e-05,
895
+ "loss": 0.4382,
896
+ "step": 296
897
+ },
898
+ {
899
+ "epoch": 0.74,
900
+ "learning_rate": 4.6275e-05,
901
+ "loss": 0.5052,
902
+ "step": 298
903
+ },
904
+ {
905
+ "epoch": 0.75,
906
+ "learning_rate": 4.6250000000000006e-05,
907
+ "loss": 0.7267,
908
+ "step": 300
909
+ },
910
+ {
911
+ "epoch": 0.76,
912
+ "learning_rate": 4.6225e-05,
913
+ "loss": 0.3374,
914
+ "step": 302
915
+ },
916
+ {
917
+ "epoch": 0.76,
918
+ "learning_rate": 4.6200000000000005e-05,
919
+ "loss": 0.4778,
920
+ "step": 304
921
+ },
922
+ {
923
+ "epoch": 0.77,
924
+ "learning_rate": 4.6175000000000004e-05,
925
+ "loss": 0.4859,
926
+ "step": 306
927
+ },
928
+ {
929
+ "epoch": 0.77,
930
+ "learning_rate": 4.6150000000000004e-05,
931
+ "loss": 0.279,
932
+ "step": 308
933
+ },
934
+ {
935
+ "epoch": 0.78,
936
+ "learning_rate": 4.6125e-05,
937
+ "loss": 1.4163,
938
+ "step": 310
939
+ },
940
+ {
941
+ "epoch": 0.78,
942
+ "learning_rate": 4.61e-05,
943
+ "loss": 0.3552,
944
+ "step": 312
945
+ },
946
+ {
947
+ "epoch": 0.79,
948
+ "learning_rate": 4.6075e-05,
949
+ "loss": 0.4339,
950
+ "step": 314
951
+ },
952
+ {
953
+ "epoch": 0.79,
954
+ "learning_rate": 4.605e-05,
955
+ "loss": 0.5127,
956
+ "step": 316
957
+ },
958
+ {
959
+ "epoch": 0.8,
960
+ "learning_rate": 4.6025e-05,
961
+ "loss": 0.3658,
962
+ "step": 318
963
+ },
964
+ {
965
+ "epoch": 0.8,
966
+ "learning_rate": 4.600000000000001e-05,
967
+ "loss": 0.339,
968
+ "step": 320
969
+ },
970
+ {
971
+ "epoch": 0.81,
972
+ "learning_rate": 4.5975e-05,
973
+ "loss": 0.5611,
974
+ "step": 322
975
+ },
976
+ {
977
+ "epoch": 0.81,
978
+ "learning_rate": 4.5950000000000006e-05,
979
+ "loss": 0.493,
980
+ "step": 324
981
+ },
982
+ {
983
+ "epoch": 0.81,
984
+ "learning_rate": 4.5925e-05,
985
+ "loss": 0.5616,
986
+ "step": 326
987
+ },
988
+ {
989
+ "epoch": 0.82,
990
+ "learning_rate": 4.5900000000000004e-05,
991
+ "loss": 0.8039,
992
+ "step": 328
993
+ },
994
+ {
995
+ "epoch": 0.82,
996
+ "learning_rate": 4.5875000000000004e-05,
997
+ "loss": 0.4011,
998
+ "step": 330
999
+ },
1000
+ {
1001
+ "epoch": 0.83,
1002
+ "learning_rate": 4.585e-05,
1003
+ "loss": 0.5168,
1004
+ "step": 332
1005
+ },
1006
+ {
1007
+ "epoch": 0.83,
1008
+ "learning_rate": 4.5825e-05,
1009
+ "loss": 0.5843,
1010
+ "step": 334
1011
+ },
1012
+ {
1013
+ "epoch": 0.84,
1014
+ "learning_rate": 4.58e-05,
1015
+ "loss": 0.5999,
1016
+ "step": 336
1017
+ },
1018
+ {
1019
+ "epoch": 0.84,
1020
+ "learning_rate": 4.5775e-05,
1021
+ "loss": 0.5357,
1022
+ "step": 338
1023
+ },
1024
+ {
1025
+ "epoch": 0.85,
1026
+ "learning_rate": 4.575e-05,
1027
+ "loss": 0.3802,
1028
+ "step": 340
1029
+ },
1030
+ {
1031
+ "epoch": 0.85,
1032
+ "learning_rate": 4.5725e-05,
1033
+ "loss": 0.5216,
1034
+ "step": 342
1035
+ },
1036
+ {
1037
+ "epoch": 0.86,
1038
+ "learning_rate": 4.5700000000000006e-05,
1039
+ "loss": 0.4038,
1040
+ "step": 344
1041
+ },
1042
+ {
1043
+ "epoch": 0.86,
1044
+ "learning_rate": 4.5675e-05,
1045
+ "loss": 0.3344,
1046
+ "step": 346
1047
+ },
1048
+ {
1049
+ "epoch": 0.87,
1050
+ "learning_rate": 4.5650000000000005e-05,
1051
+ "loss": 0.5792,
1052
+ "step": 348
1053
+ },
1054
+ {
1055
+ "epoch": 0.88,
1056
+ "learning_rate": 4.5625e-05,
1057
+ "loss": 0.453,
1058
+ "step": 350
1059
+ },
1060
+ {
1061
+ "epoch": 0.88,
1062
+ "learning_rate": 4.5600000000000004e-05,
1063
+ "loss": 0.3997,
1064
+ "step": 352
1065
+ },
1066
+ {
1067
+ "epoch": 0.89,
1068
+ "learning_rate": 4.5575e-05,
1069
+ "loss": 0.3398,
1070
+ "step": 354
1071
+ },
1072
+ {
1073
+ "epoch": 0.89,
1074
+ "learning_rate": 4.555e-05,
1075
+ "loss": 0.3204,
1076
+ "step": 356
1077
+ },
1078
+ {
1079
+ "epoch": 0.9,
1080
+ "learning_rate": 4.5525e-05,
1081
+ "loss": 0.291,
1082
+ "step": 358
1083
+ },
1084
+ {
1085
+ "epoch": 0.9,
1086
+ "learning_rate": 4.55e-05,
1087
+ "loss": 0.3974,
1088
+ "step": 360
1089
+ },
1090
+ {
1091
+ "epoch": 0.91,
1092
+ "learning_rate": 4.5475e-05,
1093
+ "loss": 0.3647,
1094
+ "step": 362
1095
+ },
1096
+ {
1097
+ "epoch": 0.91,
1098
+ "learning_rate": 4.545000000000001e-05,
1099
+ "loss": 0.3516,
1100
+ "step": 364
1101
+ },
1102
+ {
1103
+ "epoch": 0.92,
1104
+ "learning_rate": 4.5425e-05,
1105
+ "loss": 0.3865,
1106
+ "step": 366
1107
+ },
1108
+ {
1109
+ "epoch": 0.92,
1110
+ "learning_rate": 4.5400000000000006e-05,
1111
+ "loss": 0.4976,
1112
+ "step": 368
1113
+ },
1114
+ {
1115
+ "epoch": 0.93,
1116
+ "learning_rate": 4.5375e-05,
1117
+ "loss": 0.3554,
1118
+ "step": 370
1119
+ },
1120
+ {
1121
+ "epoch": 0.93,
1122
+ "learning_rate": 4.5350000000000005e-05,
1123
+ "loss": 0.526,
1124
+ "step": 372
1125
+ },
1126
+ {
1127
+ "epoch": 0.94,
1128
+ "learning_rate": 4.5325000000000004e-05,
1129
+ "loss": 0.5028,
1130
+ "step": 374
1131
+ },
1132
+ {
1133
+ "epoch": 0.94,
1134
+ "learning_rate": 4.53e-05,
1135
+ "loss": 0.2651,
1136
+ "step": 376
1137
+ },
1138
+ {
1139
+ "epoch": 0.94,
1140
+ "learning_rate": 4.5275e-05,
1141
+ "loss": 0.2753,
1142
+ "step": 378
1143
+ },
1144
+ {
1145
+ "epoch": 0.95,
1146
+ "learning_rate": 4.525e-05,
1147
+ "loss": 0.3614,
1148
+ "step": 380
1149
+ },
1150
+ {
1151
+ "epoch": 0.95,
1152
+ "learning_rate": 4.5225e-05,
1153
+ "loss": 0.4194,
1154
+ "step": 382
1155
+ },
1156
+ {
1157
+ "epoch": 0.96,
1158
+ "learning_rate": 4.52e-05,
1159
+ "loss": 0.4073,
1160
+ "step": 384
1161
+ },
1162
+ {
1163
+ "epoch": 0.96,
1164
+ "learning_rate": 4.5175e-05,
1165
+ "loss": 0.8908,
1166
+ "step": 386
1167
+ },
1168
+ {
1169
+ "epoch": 0.97,
1170
+ "learning_rate": 4.5150000000000006e-05,
1171
+ "loss": 0.5112,
1172
+ "step": 388
1173
+ },
1174
+ {
1175
+ "epoch": 0.97,
1176
+ "learning_rate": 4.5125e-05,
1177
+ "loss": 0.4367,
1178
+ "step": 390
1179
+ },
1180
+ {
1181
+ "epoch": 0.98,
1182
+ "learning_rate": 4.5100000000000005e-05,
1183
+ "loss": 0.5748,
1184
+ "step": 392
1185
+ },
1186
+ {
1187
+ "epoch": 0.98,
1188
+ "learning_rate": 4.5075e-05,
1189
+ "loss": 0.4903,
1190
+ "step": 394
1191
+ },
1192
+ {
1193
+ "epoch": 0.99,
1194
+ "learning_rate": 4.5050000000000004e-05,
1195
+ "loss": 0.4007,
1196
+ "step": 396
1197
+ },
1198
+ {
1199
+ "epoch": 0.99,
1200
+ "learning_rate": 4.5025000000000003e-05,
1201
+ "loss": 0.3218,
1202
+ "step": 398
1203
+ },
1204
+ {
1205
+ "epoch": 1.0,
1206
+ "learning_rate": 4.5e-05,
1207
+ "loss": 0.2959,
1208
+ "step": 400
1209
+ },
1210
+ {
1211
+ "epoch": 1.0,
1212
+ "eval_accuracy": 0.6915902829595236,
1213
+ "eval_f1": 0.7120379490516452,
1214
+ "eval_loss": 0.3397860825061798,
1215
+ "eval_precision": 0.7520909138849339,
1216
+ "eval_recall": 0.6915902829595236,
1217
+ "eval_runtime": 46.5586,
1218
+ "eval_samples_per_second": 17.183,
1219
+ "eval_steps_per_second": 2.148,
1220
+ "step": 400
1221
+ },
1222
+ {
1223
+ "epoch": 1.0,
1224
+ "learning_rate": 4.4975e-05,
1225
+ "loss": 0.4021,
1226
+ "step": 402
1227
+ },
1228
+ {
1229
+ "epoch": 1.01,
1230
+ "learning_rate": 4.495e-05,
1231
+ "loss": 0.3147,
1232
+ "step": 404
1233
+ },
1234
+ {
1235
+ "epoch": 1.01,
1236
+ "learning_rate": 4.4925e-05,
1237
+ "loss": 0.7101,
1238
+ "step": 406
1239
+ },
1240
+ {
1241
+ "epoch": 1.02,
1242
+ "learning_rate": 4.49e-05,
1243
+ "loss": 0.4026,
1244
+ "step": 408
1245
+ },
1246
+ {
1247
+ "epoch": 1.02,
1248
+ "learning_rate": 4.4875e-05,
1249
+ "loss": 0.3797,
1250
+ "step": 410
1251
+ },
1252
+ {
1253
+ "epoch": 1.03,
1254
+ "learning_rate": 4.4850000000000006e-05,
1255
+ "loss": 0.6129,
1256
+ "step": 412
1257
+ },
1258
+ {
1259
+ "epoch": 1.03,
1260
+ "learning_rate": 4.4825e-05,
1261
+ "loss": 0.4431,
1262
+ "step": 414
1263
+ },
1264
+ {
1265
+ "epoch": 1.04,
1266
+ "learning_rate": 4.4800000000000005e-05,
1267
+ "loss": 0.381,
1268
+ "step": 416
1269
+ },
1270
+ {
1271
+ "epoch": 1.04,
1272
+ "learning_rate": 4.4775e-05,
1273
+ "loss": 0.2991,
1274
+ "step": 418
1275
+ },
1276
+ {
1277
+ "epoch": 1.05,
1278
+ "learning_rate": 4.4750000000000004e-05,
1279
+ "loss": 0.5143,
1280
+ "step": 420
1281
+ },
1282
+ {
1283
+ "epoch": 1.05,
1284
+ "learning_rate": 4.4725e-05,
1285
+ "loss": 0.3687,
1286
+ "step": 422
1287
+ },
1288
+ {
1289
+ "epoch": 1.06,
1290
+ "learning_rate": 4.47e-05,
1291
+ "loss": 0.3229,
1292
+ "step": 424
1293
+ },
1294
+ {
1295
+ "epoch": 1.06,
1296
+ "learning_rate": 4.4675e-05,
1297
+ "loss": 0.3542,
1298
+ "step": 426
1299
+ },
1300
+ {
1301
+ "epoch": 1.07,
1302
+ "learning_rate": 4.465e-05,
1303
+ "loss": 0.3149,
1304
+ "step": 428
1305
+ },
1306
+ {
1307
+ "epoch": 1.07,
1308
+ "learning_rate": 4.4625e-05,
1309
+ "loss": 0.4093,
1310
+ "step": 430
1311
+ },
1312
+ {
1313
+ "epoch": 1.08,
1314
+ "learning_rate": 4.46e-05,
1315
+ "loss": 0.3791,
1316
+ "step": 432
1317
+ },
1318
+ {
1319
+ "epoch": 1.08,
1320
+ "learning_rate": 4.4575e-05,
1321
+ "loss": 0.2603,
1322
+ "step": 434
1323
+ },
1324
+ {
1325
+ "epoch": 1.09,
1326
+ "learning_rate": 4.4550000000000005e-05,
1327
+ "loss": 0.4405,
1328
+ "step": 436
1329
+ },
1330
+ {
1331
+ "epoch": 1.09,
1332
+ "learning_rate": 4.4525e-05,
1333
+ "loss": 0.4683,
1334
+ "step": 438
1335
+ },
1336
+ {
1337
+ "epoch": 1.1,
1338
+ "learning_rate": 4.4500000000000004e-05,
1339
+ "loss": 0.3554,
1340
+ "step": 440
1341
+ },
1342
+ {
1343
+ "epoch": 1.1,
1344
+ "learning_rate": 4.4475e-05,
1345
+ "loss": 0.5901,
1346
+ "step": 442
1347
+ },
1348
+ {
1349
+ "epoch": 1.11,
1350
+ "learning_rate": 4.445e-05,
1351
+ "loss": 0.4715,
1352
+ "step": 444
1353
+ },
1354
+ {
1355
+ "epoch": 1.11,
1356
+ "learning_rate": 4.4425e-05,
1357
+ "loss": 0.2257,
1358
+ "step": 446
1359
+ },
1360
+ {
1361
+ "epoch": 1.12,
1362
+ "learning_rate": 4.44e-05,
1363
+ "loss": 0.3887,
1364
+ "step": 448
1365
+ },
1366
+ {
1367
+ "epoch": 1.12,
1368
+ "learning_rate": 4.4375e-05,
1369
+ "loss": 0.3492,
1370
+ "step": 450
1371
+ },
1372
+ {
1373
+ "epoch": 1.13,
1374
+ "learning_rate": 4.435e-05,
1375
+ "loss": 0.3245,
1376
+ "step": 452
1377
+ },
1378
+ {
1379
+ "epoch": 1.14,
1380
+ "learning_rate": 4.4325e-05,
1381
+ "loss": 0.6609,
1382
+ "step": 454
1383
+ },
1384
+ {
1385
+ "epoch": 1.14,
1386
+ "learning_rate": 4.43e-05,
1387
+ "loss": 0.1563,
1388
+ "step": 456
1389
+ },
1390
+ {
1391
+ "epoch": 1.15,
1392
+ "learning_rate": 4.4275e-05,
1393
+ "loss": 0.3054,
1394
+ "step": 458
1395
+ },
1396
+ {
1397
+ "epoch": 1.15,
1398
+ "learning_rate": 4.4250000000000005e-05,
1399
+ "loss": 0.3336,
1400
+ "step": 460
1401
+ },
1402
+ {
1403
+ "epoch": 1.16,
1404
+ "learning_rate": 4.4225e-05,
1405
+ "loss": 0.2735,
1406
+ "step": 462
1407
+ },
1408
+ {
1409
+ "epoch": 1.16,
1410
+ "learning_rate": 4.4200000000000004e-05,
1411
+ "loss": 0.5098,
1412
+ "step": 464
1413
+ },
1414
+ {
1415
+ "epoch": 1.17,
1416
+ "learning_rate": 4.4174999999999996e-05,
1417
+ "loss": 0.517,
1418
+ "step": 466
1419
+ },
1420
+ {
1421
+ "epoch": 1.17,
1422
+ "learning_rate": 4.415e-05,
1423
+ "loss": 0.4756,
1424
+ "step": 468
1425
+ },
1426
+ {
1427
+ "epoch": 1.18,
1428
+ "learning_rate": 4.4125e-05,
1429
+ "loss": 0.386,
1430
+ "step": 470
1431
+ },
1432
+ {
1433
+ "epoch": 1.18,
1434
+ "learning_rate": 4.41e-05,
1435
+ "loss": 0.6908,
1436
+ "step": 472
1437
+ },
1438
+ {
1439
+ "epoch": 1.19,
1440
+ "learning_rate": 4.4075e-05,
1441
+ "loss": 0.3154,
1442
+ "step": 474
1443
+ },
1444
+ {
1445
+ "epoch": 1.19,
1446
+ "learning_rate": 4.405e-05,
1447
+ "loss": 0.7584,
1448
+ "step": 476
1449
+ },
1450
+ {
1451
+ "epoch": 1.2,
1452
+ "learning_rate": 4.4025e-05,
1453
+ "loss": 0.3773,
1454
+ "step": 478
1455
+ },
1456
+ {
1457
+ "epoch": 1.2,
1458
+ "learning_rate": 4.4000000000000006e-05,
1459
+ "loss": 0.3767,
1460
+ "step": 480
1461
+ },
1462
+ {
1463
+ "epoch": 1.21,
1464
+ "learning_rate": 4.3975e-05,
1465
+ "loss": 0.3647,
1466
+ "step": 482
1467
+ },
1468
+ {
1469
+ "epoch": 1.21,
1470
+ "learning_rate": 4.3950000000000004e-05,
1471
+ "loss": 0.297,
1472
+ "step": 484
1473
+ },
1474
+ {
1475
+ "epoch": 1.22,
1476
+ "learning_rate": 4.3925e-05,
1477
+ "loss": 0.3088,
1478
+ "step": 486
1479
+ },
1480
+ {
1481
+ "epoch": 1.22,
1482
+ "learning_rate": 4.39e-05,
1483
+ "loss": 0.4385,
1484
+ "step": 488
1485
+ },
1486
+ {
1487
+ "epoch": 1.23,
1488
+ "learning_rate": 4.3875e-05,
1489
+ "loss": 0.2397,
1490
+ "step": 490
1491
+ },
1492
+ {
1493
+ "epoch": 1.23,
1494
+ "learning_rate": 4.385e-05,
1495
+ "loss": 0.3477,
1496
+ "step": 492
1497
+ },
1498
+ {
1499
+ "epoch": 1.23,
1500
+ "learning_rate": 4.3825e-05,
1501
+ "loss": 0.3544,
1502
+ "step": 494
1503
+ },
1504
+ {
1505
+ "epoch": 1.24,
1506
+ "learning_rate": 4.38e-05,
1507
+ "loss": 0.3124,
1508
+ "step": 496
1509
+ },
1510
+ {
1511
+ "epoch": 1.25,
1512
+ "learning_rate": 4.3775e-05,
1513
+ "loss": 0.5746,
1514
+ "step": 498
1515
+ },
1516
+ {
1517
+ "epoch": 1.25,
1518
+ "learning_rate": 4.375e-05,
1519
+ "loss": 0.2082,
1520
+ "step": 500
1521
+ },
1522
+ {
1523
+ "epoch": 1.25,
1524
+ "learning_rate": 4.3725000000000006e-05,
1525
+ "loss": 0.421,
1526
+ "step": 502
1527
+ },
1528
+ {
1529
+ "epoch": 1.26,
1530
+ "learning_rate": 4.3700000000000005e-05,
1531
+ "loss": 0.2457,
1532
+ "step": 504
1533
+ },
1534
+ {
1535
+ "epoch": 1.27,
1536
+ "learning_rate": 4.3675000000000005e-05,
1537
+ "loss": 0.5156,
1538
+ "step": 506
1539
+ },
1540
+ {
1541
+ "epoch": 1.27,
1542
+ "learning_rate": 4.3650000000000004e-05,
1543
+ "loss": 0.3815,
1544
+ "step": 508
1545
+ },
1546
+ {
1547
+ "epoch": 1.27,
1548
+ "learning_rate": 4.3625e-05,
1549
+ "loss": 0.4878,
1550
+ "step": 510
1551
+ },
1552
+ {
1553
+ "epoch": 1.28,
1554
+ "learning_rate": 4.36e-05,
1555
+ "loss": 0.3831,
1556
+ "step": 512
1557
+ },
1558
+ {
1559
+ "epoch": 1.28,
1560
+ "learning_rate": 4.3575e-05,
1561
+ "loss": 0.4634,
1562
+ "step": 514
1563
+ },
1564
+ {
1565
+ "epoch": 1.29,
1566
+ "learning_rate": 4.355e-05,
1567
+ "loss": 0.4368,
1568
+ "step": 516
1569
+ },
1570
+ {
1571
+ "epoch": 1.29,
1572
+ "learning_rate": 4.352500000000001e-05,
1573
+ "loss": 0.3983,
1574
+ "step": 518
1575
+ },
1576
+ {
1577
+ "epoch": 1.3,
1578
+ "learning_rate": 4.35e-05,
1579
+ "loss": 0.4583,
1580
+ "step": 520
1581
+ },
1582
+ {
1583
+ "epoch": 1.3,
1584
+ "learning_rate": 4.3475000000000006e-05,
1585
+ "loss": 0.5595,
1586
+ "step": 522
1587
+ },
1588
+ {
1589
+ "epoch": 1.31,
1590
+ "learning_rate": 4.345e-05,
1591
+ "loss": 0.3677,
1592
+ "step": 524
1593
+ },
1594
+ {
1595
+ "epoch": 1.31,
1596
+ "learning_rate": 4.3425000000000005e-05,
1597
+ "loss": 0.4267,
1598
+ "step": 526
1599
+ },
1600
+ {
1601
+ "epoch": 1.32,
1602
+ "learning_rate": 4.3400000000000005e-05,
1603
+ "loss": 0.3406,
1604
+ "step": 528
1605
+ },
1606
+ {
1607
+ "epoch": 1.32,
1608
+ "learning_rate": 4.3375000000000004e-05,
1609
+ "loss": 0.4084,
1610
+ "step": 530
1611
+ },
1612
+ {
1613
+ "epoch": 1.33,
1614
+ "learning_rate": 4.335e-05,
1615
+ "loss": 0.3459,
1616
+ "step": 532
1617
+ },
1618
+ {
1619
+ "epoch": 1.33,
1620
+ "learning_rate": 4.3325e-05,
1621
+ "loss": 0.4401,
1622
+ "step": 534
1623
+ },
1624
+ {
1625
+ "epoch": 1.34,
1626
+ "learning_rate": 4.33e-05,
1627
+ "loss": 0.6531,
1628
+ "step": 536
1629
+ },
1630
+ {
1631
+ "epoch": 1.34,
1632
+ "learning_rate": 4.3275e-05,
1633
+ "loss": 0.2975,
1634
+ "step": 538
1635
+ },
1636
+ {
1637
+ "epoch": 1.35,
1638
+ "learning_rate": 4.325e-05,
1639
+ "loss": 0.2649,
1640
+ "step": 540
1641
+ },
1642
+ {
1643
+ "epoch": 1.35,
1644
+ "learning_rate": 4.322500000000001e-05,
1645
+ "loss": 0.3199,
1646
+ "step": 542
1647
+ },
1648
+ {
1649
+ "epoch": 1.36,
1650
+ "learning_rate": 4.32e-05,
1651
+ "loss": 0.342,
1652
+ "step": 544
1653
+ },
1654
+ {
1655
+ "epoch": 1.36,
1656
+ "learning_rate": 4.3175000000000006e-05,
1657
+ "loss": 0.631,
1658
+ "step": 546
1659
+ },
1660
+ {
1661
+ "epoch": 1.37,
1662
+ "learning_rate": 4.315e-05,
1663
+ "loss": 0.5276,
1664
+ "step": 548
1665
+ },
1666
+ {
1667
+ "epoch": 1.38,
1668
+ "learning_rate": 4.3125000000000005e-05,
1669
+ "loss": 0.2495,
1670
+ "step": 550
1671
+ },
1672
+ {
1673
+ "epoch": 1.38,
1674
+ "learning_rate": 4.3100000000000004e-05,
1675
+ "loss": 0.3855,
1676
+ "step": 552
1677
+ },
1678
+ {
1679
+ "epoch": 1.39,
1680
+ "learning_rate": 4.3075000000000003e-05,
1681
+ "loss": 0.599,
1682
+ "step": 554
1683
+ },
1684
+ {
1685
+ "epoch": 1.39,
1686
+ "learning_rate": 4.305e-05,
1687
+ "loss": 0.4825,
1688
+ "step": 556
1689
+ },
1690
+ {
1691
+ "epoch": 1.4,
1692
+ "learning_rate": 4.3025e-05,
1693
+ "loss": 0.4482,
1694
+ "step": 558
1695
+ },
1696
+ {
1697
+ "epoch": 1.4,
1698
+ "learning_rate": 4.3e-05,
1699
+ "loss": 0.456,
1700
+ "step": 560
1701
+ },
1702
+ {
1703
+ "epoch": 1.41,
1704
+ "learning_rate": 4.2975e-05,
1705
+ "loss": 0.299,
1706
+ "step": 562
1707
+ },
1708
+ {
1709
+ "epoch": 1.41,
1710
+ "learning_rate": 4.295e-05,
1711
+ "loss": 0.3304,
1712
+ "step": 564
1713
+ },
1714
+ {
1715
+ "epoch": 1.42,
1716
+ "learning_rate": 4.2925000000000007e-05,
1717
+ "loss": 0.3654,
1718
+ "step": 566
1719
+ },
1720
+ {
1721
+ "epoch": 1.42,
1722
+ "learning_rate": 4.29e-05,
1723
+ "loss": 0.1515,
1724
+ "step": 568
1725
+ },
1726
+ {
1727
+ "epoch": 1.43,
1728
+ "learning_rate": 4.2875000000000005e-05,
1729
+ "loss": 0.2824,
1730
+ "step": 570
1731
+ },
1732
+ {
1733
+ "epoch": 1.43,
1734
+ "learning_rate": 4.285e-05,
1735
+ "loss": 0.4656,
1736
+ "step": 572
1737
+ },
1738
+ {
1739
+ "epoch": 1.44,
1740
+ "learning_rate": 4.2825000000000004e-05,
1741
+ "loss": 0.5197,
1742
+ "step": 574
1743
+ },
1744
+ {
1745
+ "epoch": 1.44,
1746
+ "learning_rate": 4.2800000000000004e-05,
1747
+ "loss": 0.5618,
1748
+ "step": 576
1749
+ },
1750
+ {
1751
+ "epoch": 1.45,
1752
+ "learning_rate": 4.2775e-05,
1753
+ "loss": 0.4091,
1754
+ "step": 578
1755
+ },
1756
+ {
1757
+ "epoch": 1.45,
1758
+ "learning_rate": 4.275e-05,
1759
+ "loss": 0.6243,
1760
+ "step": 580
1761
+ },
1762
+ {
1763
+ "epoch": 1.46,
1764
+ "learning_rate": 4.2725e-05,
1765
+ "loss": 0.2485,
1766
+ "step": 582
1767
+ },
1768
+ {
1769
+ "epoch": 1.46,
1770
+ "learning_rate": 4.27e-05,
1771
+ "loss": 0.3898,
1772
+ "step": 584
1773
+ },
1774
+ {
1775
+ "epoch": 1.47,
1776
+ "learning_rate": 4.2675e-05,
1777
+ "loss": 0.373,
1778
+ "step": 586
1779
+ },
1780
+ {
1781
+ "epoch": 1.47,
1782
+ "learning_rate": 4.265e-05,
1783
+ "loss": 0.4957,
1784
+ "step": 588
1785
+ },
1786
+ {
1787
+ "epoch": 1.48,
1788
+ "learning_rate": 4.2625000000000006e-05,
1789
+ "loss": 0.3787,
1790
+ "step": 590
1791
+ },
1792
+ {
1793
+ "epoch": 1.48,
1794
+ "learning_rate": 4.26e-05,
1795
+ "loss": 0.3621,
1796
+ "step": 592
1797
+ },
1798
+ {
1799
+ "epoch": 1.48,
1800
+ "learning_rate": 4.2575000000000005e-05,
1801
+ "loss": 0.3356,
1802
+ "step": 594
1803
+ },
1804
+ {
1805
+ "epoch": 1.49,
1806
+ "learning_rate": 4.2550000000000004e-05,
1807
+ "loss": 0.58,
1808
+ "step": 596
1809
+ },
1810
+ {
1811
+ "epoch": 1.5,
1812
+ "learning_rate": 4.2525000000000004e-05,
1813
+ "loss": 0.3739,
1814
+ "step": 598
1815
+ },
1816
+ {
1817
+ "epoch": 1.5,
1818
+ "learning_rate": 4.25e-05,
1819
+ "loss": 0.265,
1820
+ "step": 600
1821
+ },
1822
+ {
1823
+ "epoch": 1.5,
1824
+ "learning_rate": 4.2475e-05,
1825
+ "loss": 0.541,
1826
+ "step": 602
1827
+ },
1828
+ {
1829
+ "epoch": 1.51,
1830
+ "learning_rate": 4.245e-05,
1831
+ "loss": 0.6195,
1832
+ "step": 604
1833
+ },
1834
+ {
1835
+ "epoch": 1.52,
1836
+ "learning_rate": 4.2425e-05,
1837
+ "loss": 0.2676,
1838
+ "step": 606
1839
+ },
1840
+ {
1841
+ "epoch": 1.52,
1842
+ "learning_rate": 4.24e-05,
1843
+ "loss": 0.4074,
1844
+ "step": 608
1845
+ },
1846
+ {
1847
+ "epoch": 1.52,
1848
+ "learning_rate": 4.237500000000001e-05,
1849
+ "loss": 0.4149,
1850
+ "step": 610
1851
+ },
1852
+ {
1853
+ "epoch": 1.53,
1854
+ "learning_rate": 4.235e-05,
1855
+ "loss": 0.3282,
1856
+ "step": 612
1857
+ },
1858
+ {
1859
+ "epoch": 1.54,
1860
+ "learning_rate": 4.2325000000000006e-05,
1861
+ "loss": 0.3182,
1862
+ "step": 614
1863
+ },
1864
+ {
1865
+ "epoch": 1.54,
1866
+ "learning_rate": 4.23e-05,
1867
+ "loss": 0.2904,
1868
+ "step": 616
1869
+ },
1870
+ {
1871
+ "epoch": 1.54,
1872
+ "learning_rate": 4.2275000000000004e-05,
1873
+ "loss": 0.3478,
1874
+ "step": 618
1875
+ },
1876
+ {
1877
+ "epoch": 1.55,
1878
+ "learning_rate": 4.2250000000000004e-05,
1879
+ "loss": 0.3693,
1880
+ "step": 620
1881
+ },
1882
+ {
1883
+ "epoch": 1.56,
1884
+ "learning_rate": 4.2225e-05,
1885
+ "loss": 0.4881,
1886
+ "step": 622
1887
+ },
1888
+ {
1889
+ "epoch": 1.56,
1890
+ "learning_rate": 4.22e-05,
1891
+ "loss": 0.2592,
1892
+ "step": 624
1893
+ },
1894
+ {
1895
+ "epoch": 1.56,
1896
+ "learning_rate": 4.2175e-05,
1897
+ "loss": 0.4696,
1898
+ "step": 626
1899
+ },
1900
+ {
1901
+ "epoch": 1.57,
1902
+ "learning_rate": 4.215e-05,
1903
+ "loss": 0.2685,
1904
+ "step": 628
1905
+ },
1906
+ {
1907
+ "epoch": 1.57,
1908
+ "learning_rate": 4.2125e-05,
1909
+ "loss": 0.3528,
1910
+ "step": 630
1911
+ },
1912
+ {
1913
+ "epoch": 1.58,
1914
+ "learning_rate": 4.21e-05,
1915
+ "loss": 0.3008,
1916
+ "step": 632
1917
+ },
1918
+ {
1919
+ "epoch": 1.58,
1920
+ "learning_rate": 4.2075000000000006e-05,
1921
+ "loss": 0.2616,
1922
+ "step": 634
1923
+ },
1924
+ {
1925
+ "epoch": 1.59,
1926
+ "learning_rate": 4.205e-05,
1927
+ "loss": 0.6541,
1928
+ "step": 636
1929
+ },
1930
+ {
1931
+ "epoch": 1.59,
1932
+ "learning_rate": 4.2025000000000005e-05,
1933
+ "loss": 0.2476,
1934
+ "step": 638
1935
+ },
1936
+ {
1937
+ "epoch": 1.6,
1938
+ "learning_rate": 4.2e-05,
1939
+ "loss": 0.2073,
1940
+ "step": 640
1941
+ },
1942
+ {
1943
+ "epoch": 1.6,
1944
+ "learning_rate": 4.1975000000000004e-05,
1945
+ "loss": 0.4824,
1946
+ "step": 642
1947
+ },
1948
+ {
1949
+ "epoch": 1.61,
1950
+ "learning_rate": 4.195e-05,
1951
+ "loss": 0.338,
1952
+ "step": 644
1953
+ },
1954
+ {
1955
+ "epoch": 1.61,
1956
+ "learning_rate": 4.1925e-05,
1957
+ "loss": 0.3602,
1958
+ "step": 646
1959
+ },
1960
+ {
1961
+ "epoch": 1.62,
1962
+ "learning_rate": 4.19e-05,
1963
+ "loss": 0.4082,
1964
+ "step": 648
1965
+ },
1966
+ {
1967
+ "epoch": 1.62,
1968
+ "learning_rate": 4.1875e-05,
1969
+ "loss": 0.2916,
1970
+ "step": 650
1971
+ },
1972
+ {
1973
+ "epoch": 1.63,
1974
+ "learning_rate": 4.185e-05,
1975
+ "loss": 0.2902,
1976
+ "step": 652
1977
+ },
1978
+ {
1979
+ "epoch": 1.64,
1980
+ "learning_rate": 4.1825e-05,
1981
+ "loss": 0.3261,
1982
+ "step": 654
1983
+ },
1984
+ {
1985
+ "epoch": 1.64,
1986
+ "learning_rate": 4.18e-05,
1987
+ "loss": 0.2599,
1988
+ "step": 656
1989
+ },
1990
+ {
1991
+ "epoch": 1.65,
1992
+ "learning_rate": 4.1775000000000006e-05,
1993
+ "loss": 0.4284,
1994
+ "step": 658
1995
+ },
1996
+ {
1997
+ "epoch": 1.65,
1998
+ "learning_rate": 4.175e-05,
1999
+ "loss": 0.3981,
2000
+ "step": 660
2001
+ },
2002
+ {
2003
+ "epoch": 1.66,
2004
+ "learning_rate": 4.1725000000000005e-05,
2005
+ "loss": 0.3607,
2006
+ "step": 662
2007
+ },
2008
+ {
2009
+ "epoch": 1.66,
2010
+ "learning_rate": 4.17e-05,
2011
+ "loss": 0.3214,
2012
+ "step": 664
2013
+ },
2014
+ {
2015
+ "epoch": 1.67,
2016
+ "learning_rate": 4.1675e-05,
2017
+ "loss": 0.5022,
2018
+ "step": 666
2019
+ },
2020
+ {
2021
+ "epoch": 1.67,
2022
+ "learning_rate": 4.165e-05,
2023
+ "loss": 0.2203,
2024
+ "step": 668
2025
+ },
2026
+ {
2027
+ "epoch": 1.68,
2028
+ "learning_rate": 4.1625e-05,
2029
+ "loss": 0.2736,
2030
+ "step": 670
2031
+ },
2032
+ {
2033
+ "epoch": 1.68,
2034
+ "learning_rate": 4.16e-05,
2035
+ "loss": 0.2597,
2036
+ "step": 672
2037
+ },
2038
+ {
2039
+ "epoch": 1.69,
2040
+ "learning_rate": 4.1575e-05,
2041
+ "loss": 0.3038,
2042
+ "step": 674
2043
+ },
2044
+ {
2045
+ "epoch": 1.69,
2046
+ "learning_rate": 4.155e-05,
2047
+ "loss": 0.3396,
2048
+ "step": 676
2049
+ },
2050
+ {
2051
+ "epoch": 1.69,
2052
+ "learning_rate": 4.1525e-05,
2053
+ "loss": 0.3447,
2054
+ "step": 678
2055
+ },
2056
+ {
2057
+ "epoch": 1.7,
2058
+ "learning_rate": 4.15e-05,
2059
+ "loss": 0.3168,
2060
+ "step": 680
2061
+ },
2062
+ {
2063
+ "epoch": 1.71,
2064
+ "learning_rate": 4.1475000000000005e-05,
2065
+ "loss": 0.3432,
2066
+ "step": 682
2067
+ },
2068
+ {
2069
+ "epoch": 1.71,
2070
+ "learning_rate": 4.145e-05,
2071
+ "loss": 0.3895,
2072
+ "step": 684
2073
+ },
2074
+ {
2075
+ "epoch": 1.71,
2076
+ "learning_rate": 4.1425000000000004e-05,
2077
+ "loss": 0.387,
2078
+ "step": 686
2079
+ },
2080
+ {
2081
+ "epoch": 1.72,
2082
+ "learning_rate": 4.14e-05,
2083
+ "loss": 0.2479,
2084
+ "step": 688
2085
+ },
2086
+ {
2087
+ "epoch": 1.73,
2088
+ "learning_rate": 4.1375e-05,
2089
+ "loss": 0.3109,
2090
+ "step": 690
2091
+ },
2092
+ {
2093
+ "epoch": 1.73,
2094
+ "learning_rate": 4.135e-05,
2095
+ "loss": 0.2703,
2096
+ "step": 692
2097
+ },
2098
+ {
2099
+ "epoch": 1.73,
2100
+ "learning_rate": 4.1325e-05,
2101
+ "loss": 0.3302,
2102
+ "step": 694
2103
+ },
2104
+ {
2105
+ "epoch": 1.74,
2106
+ "learning_rate": 4.13e-05,
2107
+ "loss": 0.1983,
2108
+ "step": 696
2109
+ },
2110
+ {
2111
+ "epoch": 1.75,
2112
+ "learning_rate": 4.1275e-05,
2113
+ "loss": 0.293,
2114
+ "step": 698
2115
+ },
2116
+ {
2117
+ "epoch": 1.75,
2118
+ "learning_rate": 4.125e-05,
2119
+ "loss": 0.5641,
2120
+ "step": 700
2121
+ },
2122
+ {
2123
+ "epoch": 1.75,
2124
+ "learning_rate": 4.1225e-05,
2125
+ "loss": 0.3746,
2126
+ "step": 702
2127
+ },
2128
+ {
2129
+ "epoch": 1.76,
2130
+ "learning_rate": 4.12e-05,
2131
+ "loss": 0.1783,
2132
+ "step": 704
2133
+ },
2134
+ {
2135
+ "epoch": 1.77,
2136
+ "learning_rate": 4.1175000000000005e-05,
2137
+ "loss": 0.2852,
2138
+ "step": 706
2139
+ },
2140
+ {
2141
+ "epoch": 1.77,
2142
+ "learning_rate": 4.115e-05,
2143
+ "loss": 0.3112,
2144
+ "step": 708
2145
+ },
2146
+ {
2147
+ "epoch": 1.77,
2148
+ "learning_rate": 4.1125000000000004e-05,
2149
+ "loss": 0.1936,
2150
+ "step": 710
2151
+ },
2152
+ {
2153
+ "epoch": 1.78,
2154
+ "learning_rate": 4.11e-05,
2155
+ "loss": 0.3486,
2156
+ "step": 712
2157
+ },
2158
+ {
2159
+ "epoch": 1.79,
2160
+ "learning_rate": 4.1075e-05,
2161
+ "loss": 0.2528,
2162
+ "step": 714
2163
+ },
2164
+ {
2165
+ "epoch": 1.79,
2166
+ "learning_rate": 4.105e-05,
2167
+ "loss": 0.3082,
2168
+ "step": 716
2169
+ },
2170
+ {
2171
+ "epoch": 1.79,
2172
+ "learning_rate": 4.1025e-05,
2173
+ "loss": 0.5328,
2174
+ "step": 718
2175
+ },
2176
+ {
2177
+ "epoch": 1.8,
2178
+ "learning_rate": 4.1e-05,
2179
+ "loss": 0.3125,
2180
+ "step": 720
2181
+ },
2182
+ {
2183
+ "epoch": 1.81,
2184
+ "learning_rate": 4.0975e-05,
2185
+ "loss": 0.2036,
2186
+ "step": 722
2187
+ },
2188
+ {
2189
+ "epoch": 1.81,
2190
+ "learning_rate": 4.095e-05,
2191
+ "loss": 0.295,
2192
+ "step": 724
2193
+ },
2194
+ {
2195
+ "epoch": 1.81,
2196
+ "learning_rate": 4.0925000000000005e-05,
2197
+ "loss": 0.3459,
2198
+ "step": 726
2199
+ },
2200
+ {
2201
+ "epoch": 1.82,
2202
+ "learning_rate": 4.09e-05,
2203
+ "loss": 0.3806,
2204
+ "step": 728
2205
+ },
2206
+ {
2207
+ "epoch": 1.82,
2208
+ "learning_rate": 4.0875000000000004e-05,
2209
+ "loss": 0.34,
2210
+ "step": 730
2211
+ },
2212
+ {
2213
+ "epoch": 1.83,
2214
+ "learning_rate": 4.085e-05,
2215
+ "loss": 0.4861,
2216
+ "step": 732
2217
+ },
2218
+ {
2219
+ "epoch": 1.83,
2220
+ "learning_rate": 4.0825e-05,
2221
+ "loss": 0.231,
2222
+ "step": 734
2223
+ },
2224
+ {
2225
+ "epoch": 1.84,
2226
+ "learning_rate": 4.08e-05,
2227
+ "loss": 0.3442,
2228
+ "step": 736
2229
+ },
2230
+ {
2231
+ "epoch": 1.84,
2232
+ "learning_rate": 4.0775e-05,
2233
+ "loss": 0.5056,
2234
+ "step": 738
2235
+ },
2236
+ {
2237
+ "epoch": 1.85,
2238
+ "learning_rate": 4.075e-05,
2239
+ "loss": 0.468,
2240
+ "step": 740
2241
+ },
2242
+ {
2243
+ "epoch": 1.85,
2244
+ "learning_rate": 4.0725e-05,
2245
+ "loss": 0.2309,
2246
+ "step": 742
2247
+ },
2248
+ {
2249
+ "epoch": 1.86,
2250
+ "learning_rate": 4.07e-05,
2251
+ "loss": 0.2692,
2252
+ "step": 744
2253
+ },
2254
+ {
2255
+ "epoch": 1.86,
2256
+ "learning_rate": 4.0675e-05,
2257
+ "loss": 0.3229,
2258
+ "step": 746
2259
+ },
2260
+ {
2261
+ "epoch": 1.87,
2262
+ "learning_rate": 4.065e-05,
2263
+ "loss": 0.4786,
2264
+ "step": 748
2265
+ },
2266
+ {
2267
+ "epoch": 1.88,
2268
+ "learning_rate": 4.0625000000000005e-05,
2269
+ "loss": 0.3494,
2270
+ "step": 750
2271
+ },
2272
+ {
2273
+ "epoch": 1.88,
2274
+ "learning_rate": 4.0600000000000004e-05,
2275
+ "loss": 0.2746,
2276
+ "step": 752
2277
+ },
2278
+ {
2279
+ "epoch": 1.89,
2280
+ "learning_rate": 4.0575000000000004e-05,
2281
+ "loss": 0.4341,
2282
+ "step": 754
2283
+ },
2284
+ {
2285
+ "epoch": 1.89,
2286
+ "learning_rate": 4.055e-05,
2287
+ "loss": 0.2186,
2288
+ "step": 756
2289
+ },
2290
+ {
2291
+ "epoch": 1.9,
2292
+ "learning_rate": 4.0525e-05,
2293
+ "loss": 0.5763,
2294
+ "step": 758
2295
+ },
2296
+ {
2297
+ "epoch": 1.9,
2298
+ "learning_rate": 4.05e-05,
2299
+ "loss": 0.5783,
2300
+ "step": 760
2301
+ },
2302
+ {
2303
+ "epoch": 1.91,
2304
+ "learning_rate": 4.0475e-05,
2305
+ "loss": 0.3916,
2306
+ "step": 762
2307
+ },
2308
+ {
2309
+ "epoch": 1.91,
2310
+ "learning_rate": 4.045000000000001e-05,
2311
+ "loss": 0.3211,
2312
+ "step": 764
2313
+ },
2314
+ {
2315
+ "epoch": 1.92,
2316
+ "learning_rate": 4.0425e-05,
2317
+ "loss": 0.3737,
2318
+ "step": 766
2319
+ },
2320
+ {
2321
+ "epoch": 1.92,
2322
+ "learning_rate": 4.0400000000000006e-05,
2323
+ "loss": 0.3801,
2324
+ "step": 768
2325
+ },
2326
+ {
2327
+ "epoch": 1.93,
2328
+ "learning_rate": 4.0375e-05,
2329
+ "loss": 0.3124,
2330
+ "step": 770
2331
+ },
2332
+ {
2333
+ "epoch": 1.93,
2334
+ "learning_rate": 4.0350000000000005e-05,
2335
+ "loss": 0.4814,
2336
+ "step": 772
2337
+ },
2338
+ {
2339
+ "epoch": 1.94,
2340
+ "learning_rate": 4.0325000000000004e-05,
2341
+ "loss": 0.4904,
2342
+ "step": 774
2343
+ },
2344
+ {
2345
+ "epoch": 1.94,
2346
+ "learning_rate": 4.0300000000000004e-05,
2347
+ "loss": 0.2338,
2348
+ "step": 776
2349
+ },
2350
+ {
2351
+ "epoch": 1.94,
2352
+ "learning_rate": 4.0275e-05,
2353
+ "loss": 0.2274,
2354
+ "step": 778
2355
+ },
2356
+ {
2357
+ "epoch": 1.95,
2358
+ "learning_rate": 4.025e-05,
2359
+ "loss": 0.3754,
2360
+ "step": 780
2361
+ },
2362
+ {
2363
+ "epoch": 1.96,
2364
+ "learning_rate": 4.0225e-05,
2365
+ "loss": 0.5782,
2366
+ "step": 782
2367
+ },
2368
+ {
2369
+ "epoch": 1.96,
2370
+ "learning_rate": 4.02e-05,
2371
+ "loss": 0.2568,
2372
+ "step": 784
2373
+ },
2374
+ {
2375
+ "epoch": 1.96,
2376
+ "learning_rate": 4.0175e-05,
2377
+ "loss": 0.3817,
2378
+ "step": 786
2379
+ },
2380
+ {
2381
+ "epoch": 1.97,
2382
+ "learning_rate": 4.015000000000001e-05,
2383
+ "loss": 0.1244,
2384
+ "step": 788
2385
+ },
2386
+ {
2387
+ "epoch": 1.98,
2388
+ "learning_rate": 4.0125e-05,
2389
+ "loss": 0.3574,
2390
+ "step": 790
2391
+ },
2392
+ {
2393
+ "epoch": 1.98,
2394
+ "learning_rate": 4.0100000000000006e-05,
2395
+ "loss": 0.184,
2396
+ "step": 792
2397
+ },
2398
+ {
2399
+ "epoch": 1.98,
2400
+ "learning_rate": 4.0075e-05,
2401
+ "loss": 0.5445,
2402
+ "step": 794
2403
+ },
2404
+ {
2405
+ "epoch": 1.99,
2406
+ "learning_rate": 4.0050000000000004e-05,
2407
+ "loss": 0.1936,
2408
+ "step": 796
2409
+ },
2410
+ {
2411
+ "epoch": 2.0,
2412
+ "learning_rate": 4.0025000000000004e-05,
2413
+ "loss": 0.4148,
2414
+ "step": 798
2415
+ },
2416
+ {
2417
+ "epoch": 2.0,
2418
+ "learning_rate": 4e-05,
2419
+ "loss": 0.4133,
2420
+ "step": 800
2421
+ },
2422
+ {
2423
+ "epoch": 2.0,
2424
+ "eval_accuracy": 0.7303129273927457,
2425
+ "eval_f1": 0.7240910454808391,
2426
+ "eval_loss": 0.33349618315696716,
2427
+ "eval_precision": 0.7247629591832809,
2428
+ "eval_recall": 0.7303129273927457,
2429
+ "eval_runtime": 46.5217,
2430
+ "eval_samples_per_second": 17.196,
2431
+ "eval_steps_per_second": 2.15,
2432
+ "step": 800
2433
+ },
2434
+ {
2435
+ "epoch": 2.0,
2436
+ "learning_rate": 3.9975e-05,
2437
+ "loss": 0.4279,
2438
+ "step": 802
2439
+ },
2440
+ {
2441
+ "epoch": 2.01,
2442
+ "learning_rate": 3.995e-05,
2443
+ "loss": 0.3595,
2444
+ "step": 804
2445
+ },
2446
+ {
2447
+ "epoch": 2.02,
2448
+ "learning_rate": 3.9925e-05,
2449
+ "loss": 0.2921,
2450
+ "step": 806
2451
+ },
2452
+ {
2453
+ "epoch": 2.02,
2454
+ "learning_rate": 3.99e-05,
2455
+ "loss": 0.167,
2456
+ "step": 808
2457
+ },
2458
+ {
2459
+ "epoch": 2.02,
2460
+ "learning_rate": 3.9875e-05,
2461
+ "loss": 0.2244,
2462
+ "step": 810
2463
+ },
2464
+ {
2465
+ "epoch": 2.03,
2466
+ "learning_rate": 3.9850000000000006e-05,
2467
+ "loss": 0.1951,
2468
+ "step": 812
2469
+ },
2470
+ {
2471
+ "epoch": 2.04,
2472
+ "learning_rate": 3.9825e-05,
2473
+ "loss": 0.2426,
2474
+ "step": 814
2475
+ },
2476
+ {
2477
+ "epoch": 2.04,
2478
+ "learning_rate": 3.9800000000000005e-05,
2479
+ "loss": 0.3611,
2480
+ "step": 816
2481
+ },
2482
+ {
2483
+ "epoch": 2.04,
2484
+ "learning_rate": 3.9775e-05,
2485
+ "loss": 0.3127,
2486
+ "step": 818
2487
+ },
2488
+ {
2489
+ "epoch": 2.05,
2490
+ "learning_rate": 3.9750000000000004e-05,
2491
+ "loss": 0.2799,
2492
+ "step": 820
2493
+ },
2494
+ {
2495
+ "epoch": 2.06,
2496
+ "learning_rate": 3.9725e-05,
2497
+ "loss": 0.4368,
2498
+ "step": 822
2499
+ },
2500
+ {
2501
+ "epoch": 2.06,
2502
+ "learning_rate": 3.97e-05,
2503
+ "loss": 0.3167,
2504
+ "step": 824
2505
+ },
2506
+ {
2507
+ "epoch": 2.06,
2508
+ "learning_rate": 3.9675e-05,
2509
+ "loss": 0.1099,
2510
+ "step": 826
2511
+ },
2512
+ {
2513
+ "epoch": 2.07,
2514
+ "learning_rate": 3.965e-05,
2515
+ "loss": 0.2974,
2516
+ "step": 828
2517
+ },
2518
+ {
2519
+ "epoch": 2.08,
2520
+ "learning_rate": 3.9625e-05,
2521
+ "loss": 0.2618,
2522
+ "step": 830
2523
+ },
2524
+ {
2525
+ "epoch": 2.08,
2526
+ "learning_rate": 3.960000000000001e-05,
2527
+ "loss": 0.1214,
2528
+ "step": 832
2529
+ },
2530
+ {
2531
+ "epoch": 2.08,
2532
+ "learning_rate": 3.9575e-05,
2533
+ "loss": 0.4326,
2534
+ "step": 834
2535
+ },
2536
+ {
2537
+ "epoch": 2.09,
2538
+ "learning_rate": 3.9550000000000006e-05,
2539
+ "loss": 0.1978,
2540
+ "step": 836
2541
+ },
2542
+ {
2543
+ "epoch": 2.1,
2544
+ "learning_rate": 3.9525e-05,
2545
+ "loss": 0.3145,
2546
+ "step": 838
2547
+ },
2548
+ {
2549
+ "epoch": 2.1,
2550
+ "learning_rate": 3.9500000000000005e-05,
2551
+ "loss": 0.3392,
2552
+ "step": 840
2553
+ },
2554
+ {
2555
+ "epoch": 2.1,
2556
+ "learning_rate": 3.9475000000000004e-05,
2557
+ "loss": 0.2255,
2558
+ "step": 842
2559
+ },
2560
+ {
2561
+ "epoch": 2.11,
2562
+ "learning_rate": 3.9450000000000003e-05,
2563
+ "loss": 0.4223,
2564
+ "step": 844
2565
+ },
2566
+ {
2567
+ "epoch": 2.12,
2568
+ "learning_rate": 3.9425e-05,
2569
+ "loss": 0.2917,
2570
+ "step": 846
2571
+ },
2572
+ {
2573
+ "epoch": 2.12,
2574
+ "learning_rate": 3.94e-05,
2575
+ "loss": 0.3284,
2576
+ "step": 848
2577
+ },
2578
+ {
2579
+ "epoch": 2.12,
2580
+ "learning_rate": 3.9375e-05,
2581
+ "loss": 0.2235,
2582
+ "step": 850
2583
+ },
2584
+ {
2585
+ "epoch": 2.13,
2586
+ "learning_rate": 3.935e-05,
2587
+ "loss": 0.2539,
2588
+ "step": 852
2589
+ },
2590
+ {
2591
+ "epoch": 2.13,
2592
+ "learning_rate": 3.9325e-05,
2593
+ "loss": 0.2643,
2594
+ "step": 854
2595
+ },
2596
+ {
2597
+ "epoch": 2.14,
2598
+ "learning_rate": 3.9300000000000007e-05,
2599
+ "loss": 0.1961,
2600
+ "step": 856
2601
+ },
2602
+ {
2603
+ "epoch": 2.15,
2604
+ "learning_rate": 3.9275e-05,
2605
+ "loss": 0.1986,
2606
+ "step": 858
2607
+ },
2608
+ {
2609
+ "epoch": 2.15,
2610
+ "learning_rate": 3.9250000000000005e-05,
2611
+ "loss": 0.0867,
2612
+ "step": 860
2613
+ },
2614
+ {
2615
+ "epoch": 2.15,
2616
+ "learning_rate": 3.9225e-05,
2617
+ "loss": 0.2181,
2618
+ "step": 862
2619
+ },
2620
+ {
2621
+ "epoch": 2.16,
2622
+ "learning_rate": 3.9200000000000004e-05,
2623
+ "loss": 0.0727,
2624
+ "step": 864
2625
+ },
2626
+ {
2627
+ "epoch": 2.17,
2628
+ "learning_rate": 3.9175000000000004e-05,
2629
+ "loss": 0.2379,
2630
+ "step": 866
2631
+ },
2632
+ {
2633
+ "epoch": 2.17,
2634
+ "learning_rate": 3.915e-05,
2635
+ "loss": 0.2092,
2636
+ "step": 868
2637
+ },
2638
+ {
2639
+ "epoch": 2.17,
2640
+ "learning_rate": 3.9125e-05,
2641
+ "loss": 0.1124,
2642
+ "step": 870
2643
+ },
2644
+ {
2645
+ "epoch": 2.18,
2646
+ "learning_rate": 3.91e-05,
2647
+ "loss": 0.2252,
2648
+ "step": 872
2649
+ },
2650
+ {
2651
+ "epoch": 2.19,
2652
+ "learning_rate": 3.9075e-05,
2653
+ "loss": 0.4025,
2654
+ "step": 874
2655
+ },
2656
+ {
2657
+ "epoch": 2.19,
2658
+ "learning_rate": 3.905e-05,
2659
+ "loss": 0.2782,
2660
+ "step": 876
2661
+ },
2662
+ {
2663
+ "epoch": 2.19,
2664
+ "learning_rate": 3.9025e-05,
2665
+ "loss": 0.5048,
2666
+ "step": 878
2667
+ },
2668
+ {
2669
+ "epoch": 2.2,
2670
+ "learning_rate": 3.9000000000000006e-05,
2671
+ "loss": 0.3301,
2672
+ "step": 880
2673
+ },
2674
+ {
2675
+ "epoch": 2.21,
2676
+ "learning_rate": 3.8975e-05,
2677
+ "loss": 0.3589,
2678
+ "step": 882
2679
+ },
2680
+ {
2681
+ "epoch": 2.21,
2682
+ "learning_rate": 3.8950000000000005e-05,
2683
+ "loss": 0.4942,
2684
+ "step": 884
2685
+ },
2686
+ {
2687
+ "epoch": 2.21,
2688
+ "learning_rate": 3.8925e-05,
2689
+ "loss": 0.4335,
2690
+ "step": 886
2691
+ },
2692
+ {
2693
+ "epoch": 2.22,
2694
+ "learning_rate": 3.8900000000000004e-05,
2695
+ "loss": 0.1538,
2696
+ "step": 888
2697
+ },
2698
+ {
2699
+ "epoch": 2.23,
2700
+ "learning_rate": 3.8875e-05,
2701
+ "loss": 0.442,
2702
+ "step": 890
2703
+ },
2704
+ {
2705
+ "epoch": 2.23,
2706
+ "learning_rate": 3.885e-05,
2707
+ "loss": 0.4319,
2708
+ "step": 892
2709
+ },
2710
+ {
2711
+ "epoch": 2.23,
2712
+ "learning_rate": 3.8825e-05,
2713
+ "loss": 0.3939,
2714
+ "step": 894
2715
+ },
2716
+ {
2717
+ "epoch": 2.24,
2718
+ "learning_rate": 3.88e-05,
2719
+ "loss": 0.1888,
2720
+ "step": 896
2721
+ },
2722
+ {
2723
+ "epoch": 2.25,
2724
+ "learning_rate": 3.8775e-05,
2725
+ "loss": 0.6402,
2726
+ "step": 898
2727
+ },
2728
+ {
2729
+ "epoch": 2.25,
2730
+ "learning_rate": 3.875e-05,
2731
+ "loss": 0.1708,
2732
+ "step": 900
2733
+ },
2734
+ {
2735
+ "epoch": 2.25,
2736
+ "learning_rate": 3.8725e-05,
2737
+ "loss": 0.3315,
2738
+ "step": 902
2739
+ },
2740
+ {
2741
+ "epoch": 2.26,
2742
+ "learning_rate": 3.8700000000000006e-05,
2743
+ "loss": 0.3908,
2744
+ "step": 904
2745
+ },
2746
+ {
2747
+ "epoch": 2.27,
2748
+ "learning_rate": 3.8675e-05,
2749
+ "loss": 0.2653,
2750
+ "step": 906
2751
+ },
2752
+ {
2753
+ "epoch": 2.27,
2754
+ "learning_rate": 3.8650000000000004e-05,
2755
+ "loss": 0.2681,
2756
+ "step": 908
2757
+ },
2758
+ {
2759
+ "epoch": 2.27,
2760
+ "learning_rate": 3.8625e-05,
2761
+ "loss": 0.1869,
2762
+ "step": 910
2763
+ },
2764
+ {
2765
+ "epoch": 2.28,
2766
+ "learning_rate": 3.86e-05,
2767
+ "loss": 0.2696,
2768
+ "step": 912
2769
+ },
2770
+ {
2771
+ "epoch": 2.29,
2772
+ "learning_rate": 3.8575e-05,
2773
+ "loss": 0.4004,
2774
+ "step": 914
2775
+ },
2776
+ {
2777
+ "epoch": 2.29,
2778
+ "learning_rate": 3.855e-05,
2779
+ "loss": 0.2019,
2780
+ "step": 916
2781
+ },
2782
+ {
2783
+ "epoch": 2.29,
2784
+ "learning_rate": 3.8525e-05,
2785
+ "loss": 0.1742,
2786
+ "step": 918
2787
+ },
2788
+ {
2789
+ "epoch": 2.3,
2790
+ "learning_rate": 3.85e-05,
2791
+ "loss": 0.3658,
2792
+ "step": 920
2793
+ },
2794
+ {
2795
+ "epoch": 2.31,
2796
+ "learning_rate": 3.8475e-05,
2797
+ "loss": 0.1322,
2798
+ "step": 922
2799
+ },
2800
+ {
2801
+ "epoch": 2.31,
2802
+ "learning_rate": 3.845e-05,
2803
+ "loss": 0.1761,
2804
+ "step": 924
2805
+ },
2806
+ {
2807
+ "epoch": 2.31,
2808
+ "learning_rate": 3.8425e-05,
2809
+ "loss": 0.2245,
2810
+ "step": 926
2811
+ },
2812
+ {
2813
+ "epoch": 2.32,
2814
+ "learning_rate": 3.8400000000000005e-05,
2815
+ "loss": 0.3481,
2816
+ "step": 928
2817
+ },
2818
+ {
2819
+ "epoch": 2.33,
2820
+ "learning_rate": 3.8375e-05,
2821
+ "loss": 0.326,
2822
+ "step": 930
2823
+ },
2824
+ {
2825
+ "epoch": 2.33,
2826
+ "learning_rate": 3.8350000000000004e-05,
2827
+ "loss": 0.2885,
2828
+ "step": 932
2829
+ },
2830
+ {
2831
+ "epoch": 2.33,
2832
+ "learning_rate": 3.8324999999999996e-05,
2833
+ "loss": 0.1434,
2834
+ "step": 934
2835
+ },
2836
+ {
2837
+ "epoch": 2.34,
2838
+ "learning_rate": 3.83e-05,
2839
+ "loss": 0.2765,
2840
+ "step": 936
2841
+ },
2842
+ {
2843
+ "epoch": 2.34,
2844
+ "learning_rate": 3.8275e-05,
2845
+ "loss": 0.3264,
2846
+ "step": 938
2847
+ },
2848
+ {
2849
+ "epoch": 2.35,
2850
+ "learning_rate": 3.825e-05,
2851
+ "loss": 0.4146,
2852
+ "step": 940
2853
+ },
2854
+ {
2855
+ "epoch": 2.35,
2856
+ "learning_rate": 3.8225e-05,
2857
+ "loss": 0.2461,
2858
+ "step": 942
2859
+ },
2860
+ {
2861
+ "epoch": 2.36,
2862
+ "learning_rate": 3.82e-05,
2863
+ "loss": 0.4797,
2864
+ "step": 944
2865
+ },
2866
+ {
2867
+ "epoch": 2.37,
2868
+ "learning_rate": 3.8175e-05,
2869
+ "loss": 0.3502,
2870
+ "step": 946
2871
+ },
2872
+ {
2873
+ "epoch": 2.37,
2874
+ "learning_rate": 3.8150000000000006e-05,
2875
+ "loss": 0.1427,
2876
+ "step": 948
2877
+ },
2878
+ {
2879
+ "epoch": 2.38,
2880
+ "learning_rate": 3.8125e-05,
2881
+ "loss": 0.164,
2882
+ "step": 950
2883
+ },
2884
+ {
2885
+ "epoch": 2.38,
2886
+ "learning_rate": 3.8100000000000005e-05,
2887
+ "loss": 0.1947,
2888
+ "step": 952
2889
+ },
2890
+ {
2891
+ "epoch": 2.38,
2892
+ "learning_rate": 3.8075e-05,
2893
+ "loss": 0.1943,
2894
+ "step": 954
2895
+ },
2896
+ {
2897
+ "epoch": 2.39,
2898
+ "learning_rate": 3.805e-05,
2899
+ "loss": 0.6346,
2900
+ "step": 956
2901
+ },
2902
+ {
2903
+ "epoch": 2.4,
2904
+ "learning_rate": 3.8025e-05,
2905
+ "loss": 0.6651,
2906
+ "step": 958
2907
+ },
2908
+ {
2909
+ "epoch": 2.4,
2910
+ "learning_rate": 3.8e-05,
2911
+ "loss": 0.4448,
2912
+ "step": 960
2913
+ },
2914
+ {
2915
+ "epoch": 2.41,
2916
+ "learning_rate": 3.7975e-05,
2917
+ "loss": 0.4063,
2918
+ "step": 962
2919
+ },
2920
+ {
2921
+ "epoch": 2.41,
2922
+ "learning_rate": 3.795e-05,
2923
+ "loss": 0.3017,
2924
+ "step": 964
2925
+ },
2926
+ {
2927
+ "epoch": 2.42,
2928
+ "learning_rate": 3.7925e-05,
2929
+ "loss": 0.1966,
2930
+ "step": 966
2931
+ },
2932
+ {
2933
+ "epoch": 2.42,
2934
+ "learning_rate": 3.79e-05,
2935
+ "loss": 0.3202,
2936
+ "step": 968
2937
+ },
2938
+ {
2939
+ "epoch": 2.42,
2940
+ "learning_rate": 3.7875e-05,
2941
+ "loss": 0.1277,
2942
+ "step": 970
2943
+ },
2944
+ {
2945
+ "epoch": 2.43,
2946
+ "learning_rate": 3.7850000000000005e-05,
2947
+ "loss": 0.2548,
2948
+ "step": 972
2949
+ },
2950
+ {
2951
+ "epoch": 2.44,
2952
+ "learning_rate": 3.7825e-05,
2953
+ "loss": 0.1849,
2954
+ "step": 974
2955
+ },
2956
+ {
2957
+ "epoch": 2.44,
2958
+ "learning_rate": 3.7800000000000004e-05,
2959
+ "loss": 0.1185,
2960
+ "step": 976
2961
+ },
2962
+ {
2963
+ "epoch": 2.44,
2964
+ "learning_rate": 3.7775e-05,
2965
+ "loss": 0.4444,
2966
+ "step": 978
2967
+ },
2968
+ {
2969
+ "epoch": 2.45,
2970
+ "learning_rate": 3.775e-05,
2971
+ "loss": 0.2993,
2972
+ "step": 980
2973
+ },
2974
+ {
2975
+ "epoch": 2.46,
2976
+ "learning_rate": 3.7725e-05,
2977
+ "loss": 0.1434,
2978
+ "step": 982
2979
+ },
2980
+ {
2981
+ "epoch": 2.46,
2982
+ "learning_rate": 3.77e-05,
2983
+ "loss": 0.1306,
2984
+ "step": 984
2985
+ },
2986
+ {
2987
+ "epoch": 2.46,
2988
+ "learning_rate": 3.7675e-05,
2989
+ "loss": 0.0721,
2990
+ "step": 986
2991
+ },
2992
+ {
2993
+ "epoch": 2.47,
2994
+ "learning_rate": 3.765e-05,
2995
+ "loss": 0.1188,
2996
+ "step": 988
2997
+ },
2998
+ {
2999
+ "epoch": 2.48,
3000
+ "learning_rate": 3.7625e-05,
3001
+ "loss": 0.128,
3002
+ "step": 990
3003
+ },
3004
+ {
3005
+ "epoch": 2.48,
3006
+ "learning_rate": 3.76e-05,
3007
+ "loss": 0.3053,
3008
+ "step": 992
3009
+ },
3010
+ {
3011
+ "epoch": 2.48,
3012
+ "learning_rate": 3.7575e-05,
3013
+ "loss": 0.293,
3014
+ "step": 994
3015
+ },
3016
+ {
3017
+ "epoch": 2.49,
3018
+ "learning_rate": 3.7550000000000005e-05,
3019
+ "loss": 0.2691,
3020
+ "step": 996
3021
+ },
3022
+ {
3023
+ "epoch": 2.5,
3024
+ "learning_rate": 3.7525e-05,
3025
+ "loss": 0.1903,
3026
+ "step": 998
3027
+ },
3028
+ {
3029
+ "epoch": 2.5,
3030
+ "learning_rate": 3.7500000000000003e-05,
3031
+ "loss": 0.1,
3032
+ "step": 1000
3033
+ },
3034
+ {
3035
+ "epoch": 2.5,
3036
+ "learning_rate": 3.7475e-05,
3037
+ "loss": 0.1671,
3038
+ "step": 1002
3039
+ },
3040
+ {
3041
+ "epoch": 2.51,
3042
+ "learning_rate": 3.745e-05,
3043
+ "loss": 0.3463,
3044
+ "step": 1004
3045
+ },
3046
+ {
3047
+ "epoch": 2.52,
3048
+ "learning_rate": 3.7425e-05,
3049
+ "loss": 0.2526,
3050
+ "step": 1006
3051
+ },
3052
+ {
3053
+ "epoch": 2.52,
3054
+ "learning_rate": 3.74e-05,
3055
+ "loss": 0.4419,
3056
+ "step": 1008
3057
+ },
3058
+ {
3059
+ "epoch": 2.52,
3060
+ "learning_rate": 3.737500000000001e-05,
3061
+ "loss": 0.5626,
3062
+ "step": 1010
3063
+ },
3064
+ {
3065
+ "epoch": 2.53,
3066
+ "learning_rate": 3.735e-05,
3067
+ "loss": 0.2871,
3068
+ "step": 1012
3069
+ },
3070
+ {
3071
+ "epoch": 2.54,
3072
+ "learning_rate": 3.7325000000000006e-05,
3073
+ "loss": 0.0462,
3074
+ "step": 1014
3075
+ },
3076
+ {
3077
+ "epoch": 2.54,
3078
+ "learning_rate": 3.73e-05,
3079
+ "loss": 0.2153,
3080
+ "step": 1016
3081
+ },
3082
+ {
3083
+ "epoch": 2.54,
3084
+ "learning_rate": 3.7275000000000005e-05,
3085
+ "loss": 0.1801,
3086
+ "step": 1018
3087
+ },
3088
+ {
3089
+ "epoch": 2.55,
3090
+ "learning_rate": 3.7250000000000004e-05,
3091
+ "loss": 0.2419,
3092
+ "step": 1020
3093
+ },
3094
+ {
3095
+ "epoch": 2.56,
3096
+ "learning_rate": 3.7225000000000004e-05,
3097
+ "loss": 0.4692,
3098
+ "step": 1022
3099
+ },
3100
+ {
3101
+ "epoch": 2.56,
3102
+ "learning_rate": 3.72e-05,
3103
+ "loss": 0.3793,
3104
+ "step": 1024
3105
+ },
3106
+ {
3107
+ "epoch": 2.56,
3108
+ "learning_rate": 3.7175e-05,
3109
+ "loss": 0.3421,
3110
+ "step": 1026
3111
+ },
3112
+ {
3113
+ "epoch": 2.57,
3114
+ "learning_rate": 3.715e-05,
3115
+ "loss": 0.3306,
3116
+ "step": 1028
3117
+ },
3118
+ {
3119
+ "epoch": 2.58,
3120
+ "learning_rate": 3.7125e-05,
3121
+ "loss": 0.1118,
3122
+ "step": 1030
3123
+ },
3124
+ {
3125
+ "epoch": 2.58,
3126
+ "learning_rate": 3.71e-05,
3127
+ "loss": 0.1968,
3128
+ "step": 1032
3129
+ },
3130
+ {
3131
+ "epoch": 2.58,
3132
+ "learning_rate": 3.707500000000001e-05,
3133
+ "loss": 0.2947,
3134
+ "step": 1034
3135
+ },
3136
+ {
3137
+ "epoch": 2.59,
3138
+ "learning_rate": 3.705e-05,
3139
+ "loss": 0.177,
3140
+ "step": 1036
3141
+ },
3142
+ {
3143
+ "epoch": 2.59,
3144
+ "learning_rate": 3.7025000000000005e-05,
3145
+ "loss": 0.2388,
3146
+ "step": 1038
3147
+ },
3148
+ {
3149
+ "epoch": 2.6,
3150
+ "learning_rate": 3.7e-05,
3151
+ "loss": 0.2208,
3152
+ "step": 1040
3153
+ },
3154
+ {
3155
+ "epoch": 2.6,
3156
+ "learning_rate": 3.6975000000000004e-05,
3157
+ "loss": 0.0863,
3158
+ "step": 1042
3159
+ },
3160
+ {
3161
+ "epoch": 2.61,
3162
+ "learning_rate": 3.6950000000000004e-05,
3163
+ "loss": 0.1208,
3164
+ "step": 1044
3165
+ },
3166
+ {
3167
+ "epoch": 2.62,
3168
+ "learning_rate": 3.6925e-05,
3169
+ "loss": 0.161,
3170
+ "step": 1046
3171
+ },
3172
+ {
3173
+ "epoch": 2.62,
3174
+ "learning_rate": 3.69e-05,
3175
+ "loss": 0.2525,
3176
+ "step": 1048
3177
+ },
3178
+ {
3179
+ "epoch": 2.62,
3180
+ "learning_rate": 3.6875e-05,
3181
+ "loss": 0.0939,
3182
+ "step": 1050
3183
+ },
3184
+ {
3185
+ "epoch": 2.63,
3186
+ "learning_rate": 3.685e-05,
3187
+ "loss": 0.1857,
3188
+ "step": 1052
3189
+ },
3190
+ {
3191
+ "epoch": 2.63,
3192
+ "learning_rate": 3.6825e-05,
3193
+ "loss": 0.4203,
3194
+ "step": 1054
3195
+ },
3196
+ {
3197
+ "epoch": 2.64,
3198
+ "learning_rate": 3.68e-05,
3199
+ "loss": 0.1674,
3200
+ "step": 1056
3201
+ },
3202
+ {
3203
+ "epoch": 2.65,
3204
+ "learning_rate": 3.6775000000000006e-05,
3205
+ "loss": 0.2935,
3206
+ "step": 1058
3207
+ },
3208
+ {
3209
+ "epoch": 2.65,
3210
+ "learning_rate": 3.675e-05,
3211
+ "loss": 0.4283,
3212
+ "step": 1060
3213
+ },
3214
+ {
3215
+ "epoch": 2.66,
3216
+ "learning_rate": 3.6725000000000005e-05,
3217
+ "loss": 0.0551,
3218
+ "step": 1062
3219
+ },
3220
+ {
3221
+ "epoch": 2.66,
3222
+ "learning_rate": 3.6700000000000004e-05,
3223
+ "loss": 0.2137,
3224
+ "step": 1064
3225
+ },
3226
+ {
3227
+ "epoch": 2.67,
3228
+ "learning_rate": 3.6675000000000004e-05,
3229
+ "loss": 0.0585,
3230
+ "step": 1066
3231
+ },
3232
+ {
3233
+ "epoch": 2.67,
3234
+ "learning_rate": 3.665e-05,
3235
+ "loss": 0.5005,
3236
+ "step": 1068
3237
+ },
3238
+ {
3239
+ "epoch": 2.67,
3240
+ "learning_rate": 3.6625e-05,
3241
+ "loss": 0.2685,
3242
+ "step": 1070
3243
+ },
3244
+ {
3245
+ "epoch": 2.68,
3246
+ "learning_rate": 3.66e-05,
3247
+ "loss": 0.2799,
3248
+ "step": 1072
3249
+ },
3250
+ {
3251
+ "epoch": 2.69,
3252
+ "learning_rate": 3.6575e-05,
3253
+ "loss": 0.4154,
3254
+ "step": 1074
3255
+ },
3256
+ {
3257
+ "epoch": 2.69,
3258
+ "learning_rate": 3.655e-05,
3259
+ "loss": 0.3866,
3260
+ "step": 1076
3261
+ },
3262
+ {
3263
+ "epoch": 2.69,
3264
+ "learning_rate": 3.652500000000001e-05,
3265
+ "loss": 0.275,
3266
+ "step": 1078
3267
+ },
3268
+ {
3269
+ "epoch": 2.7,
3270
+ "learning_rate": 3.65e-05,
3271
+ "loss": 0.0749,
3272
+ "step": 1080
3273
+ },
3274
+ {
3275
+ "epoch": 2.71,
3276
+ "learning_rate": 3.6475000000000006e-05,
3277
+ "loss": 0.4298,
3278
+ "step": 1082
3279
+ },
3280
+ {
3281
+ "epoch": 2.71,
3282
+ "learning_rate": 3.645e-05,
3283
+ "loss": 0.3015,
3284
+ "step": 1084
3285
+ },
3286
+ {
3287
+ "epoch": 2.71,
3288
+ "learning_rate": 3.6425000000000004e-05,
3289
+ "loss": 0.4693,
3290
+ "step": 1086
3291
+ },
3292
+ {
3293
+ "epoch": 2.72,
3294
+ "learning_rate": 3.6400000000000004e-05,
3295
+ "loss": 0.5553,
3296
+ "step": 1088
3297
+ },
3298
+ {
3299
+ "epoch": 2.73,
3300
+ "learning_rate": 3.6375e-05,
3301
+ "loss": 0.2838,
3302
+ "step": 1090
3303
+ },
3304
+ {
3305
+ "epoch": 2.73,
3306
+ "learning_rate": 3.635e-05,
3307
+ "loss": 0.2485,
3308
+ "step": 1092
3309
+ },
3310
+ {
3311
+ "epoch": 2.73,
3312
+ "learning_rate": 3.6325e-05,
3313
+ "loss": 0.1477,
3314
+ "step": 1094
3315
+ },
3316
+ {
3317
+ "epoch": 2.74,
3318
+ "learning_rate": 3.63e-05,
3319
+ "loss": 0.1499,
3320
+ "step": 1096
3321
+ },
3322
+ {
3323
+ "epoch": 2.75,
3324
+ "learning_rate": 3.6275e-05,
3325
+ "loss": 0.5901,
3326
+ "step": 1098
3327
+ },
3328
+ {
3329
+ "epoch": 2.75,
3330
+ "learning_rate": 3.625e-05,
3331
+ "loss": 0.3647,
3332
+ "step": 1100
3333
+ },
3334
+ {
3335
+ "epoch": 2.75,
3336
+ "learning_rate": 3.6225000000000006e-05,
3337
+ "loss": 0.29,
3338
+ "step": 1102
3339
+ },
3340
+ {
3341
+ "epoch": 2.76,
3342
+ "learning_rate": 3.62e-05,
3343
+ "loss": 0.2172,
3344
+ "step": 1104
3345
+ },
3346
+ {
3347
+ "epoch": 2.77,
3348
+ "learning_rate": 3.6175000000000005e-05,
3349
+ "loss": 0.2973,
3350
+ "step": 1106
3351
+ },
3352
+ {
3353
+ "epoch": 2.77,
3354
+ "learning_rate": 3.615e-05,
3355
+ "loss": 0.3948,
3356
+ "step": 1108
3357
+ },
3358
+ {
3359
+ "epoch": 2.77,
3360
+ "learning_rate": 3.6125000000000004e-05,
3361
+ "loss": 0.134,
3362
+ "step": 1110
3363
+ },
3364
+ {
3365
+ "epoch": 2.78,
3366
+ "learning_rate": 3.61e-05,
3367
+ "loss": 0.3913,
3368
+ "step": 1112
3369
+ },
3370
+ {
3371
+ "epoch": 2.79,
3372
+ "learning_rate": 3.6075e-05,
3373
+ "loss": 0.167,
3374
+ "step": 1114
3375
+ },
3376
+ {
3377
+ "epoch": 2.79,
3378
+ "learning_rate": 3.605e-05,
3379
+ "loss": 0.3009,
3380
+ "step": 1116
3381
+ },
3382
+ {
3383
+ "epoch": 2.79,
3384
+ "learning_rate": 3.6025e-05,
3385
+ "loss": 0.2548,
3386
+ "step": 1118
3387
+ },
3388
+ {
3389
+ "epoch": 2.8,
3390
+ "learning_rate": 3.6e-05,
3391
+ "loss": 0.1343,
3392
+ "step": 1120
3393
+ },
3394
+ {
3395
+ "epoch": 2.81,
3396
+ "learning_rate": 3.5975e-05,
3397
+ "loss": 0.4397,
3398
+ "step": 1122
3399
+ },
3400
+ {
3401
+ "epoch": 2.81,
3402
+ "learning_rate": 3.595e-05,
3403
+ "loss": 0.2612,
3404
+ "step": 1124
3405
+ },
3406
+ {
3407
+ "epoch": 2.81,
3408
+ "learning_rate": 3.5925000000000006e-05,
3409
+ "loss": 0.3434,
3410
+ "step": 1126
3411
+ },
3412
+ {
3413
+ "epoch": 2.82,
3414
+ "learning_rate": 3.59e-05,
3415
+ "loss": 0.7125,
3416
+ "step": 1128
3417
+ },
3418
+ {
3419
+ "epoch": 2.83,
3420
+ "learning_rate": 3.5875000000000005e-05,
3421
+ "loss": 0.4386,
3422
+ "step": 1130
3423
+ },
3424
+ {
3425
+ "epoch": 2.83,
3426
+ "learning_rate": 3.585e-05,
3427
+ "loss": 0.4036,
3428
+ "step": 1132
3429
+ },
3430
+ {
3431
+ "epoch": 2.83,
3432
+ "learning_rate": 3.5825000000000003e-05,
3433
+ "loss": 0.0801,
3434
+ "step": 1134
3435
+ },
3436
+ {
3437
+ "epoch": 2.84,
3438
+ "learning_rate": 3.58e-05,
3439
+ "loss": 0.1766,
3440
+ "step": 1136
3441
+ },
3442
+ {
3443
+ "epoch": 2.84,
3444
+ "learning_rate": 3.5775e-05,
3445
+ "loss": 0.2278,
3446
+ "step": 1138
3447
+ },
3448
+ {
3449
+ "epoch": 2.85,
3450
+ "learning_rate": 3.575e-05,
3451
+ "loss": 0.2785,
3452
+ "step": 1140
3453
+ },
3454
+ {
3455
+ "epoch": 2.85,
3456
+ "learning_rate": 3.5725e-05,
3457
+ "loss": 0.1986,
3458
+ "step": 1142
3459
+ },
3460
+ {
3461
+ "epoch": 2.86,
3462
+ "learning_rate": 3.57e-05,
3463
+ "loss": 0.3308,
3464
+ "step": 1144
3465
+ },
3466
+ {
3467
+ "epoch": 2.87,
3468
+ "learning_rate": 3.5675e-05,
3469
+ "loss": 0.1467,
3470
+ "step": 1146
3471
+ },
3472
+ {
3473
+ "epoch": 2.87,
3474
+ "learning_rate": 3.565e-05,
3475
+ "loss": 0.2535,
3476
+ "step": 1148
3477
+ },
3478
+ {
3479
+ "epoch": 2.88,
3480
+ "learning_rate": 3.5625000000000005e-05,
3481
+ "loss": 0.2237,
3482
+ "step": 1150
3483
+ },
3484
+ {
3485
+ "epoch": 2.88,
3486
+ "learning_rate": 3.56e-05,
3487
+ "loss": 0.1238,
3488
+ "step": 1152
3489
+ },
3490
+ {
3491
+ "epoch": 2.88,
3492
+ "learning_rate": 3.5575000000000004e-05,
3493
+ "loss": 0.241,
3494
+ "step": 1154
3495
+ },
3496
+ {
3497
+ "epoch": 2.89,
3498
+ "learning_rate": 3.555e-05,
3499
+ "loss": 0.2651,
3500
+ "step": 1156
3501
+ },
3502
+ {
3503
+ "epoch": 2.9,
3504
+ "learning_rate": 3.5525e-05,
3505
+ "loss": 0.3775,
3506
+ "step": 1158
3507
+ },
3508
+ {
3509
+ "epoch": 2.9,
3510
+ "learning_rate": 3.55e-05,
3511
+ "loss": 0.3317,
3512
+ "step": 1160
3513
+ },
3514
+ {
3515
+ "epoch": 2.91,
3516
+ "learning_rate": 3.5475e-05,
3517
+ "loss": 0.2678,
3518
+ "step": 1162
3519
+ },
3520
+ {
3521
+ "epoch": 2.91,
3522
+ "learning_rate": 3.545e-05,
3523
+ "loss": 0.1845,
3524
+ "step": 1164
3525
+ },
3526
+ {
3527
+ "epoch": 2.92,
3528
+ "learning_rate": 3.5425e-05,
3529
+ "loss": 0.1676,
3530
+ "step": 1166
3531
+ },
3532
+ {
3533
+ "epoch": 2.92,
3534
+ "learning_rate": 3.54e-05,
3535
+ "loss": 0.205,
3536
+ "step": 1168
3537
+ },
3538
+ {
3539
+ "epoch": 2.92,
3540
+ "learning_rate": 3.5375e-05,
3541
+ "loss": 0.3689,
3542
+ "step": 1170
3543
+ },
3544
+ {
3545
+ "epoch": 2.93,
3546
+ "learning_rate": 3.535e-05,
3547
+ "loss": 0.0595,
3548
+ "step": 1172
3549
+ },
3550
+ {
3551
+ "epoch": 2.94,
3552
+ "learning_rate": 3.5325000000000005e-05,
3553
+ "loss": 0.1113,
3554
+ "step": 1174
3555
+ },
3556
+ {
3557
+ "epoch": 2.94,
3558
+ "learning_rate": 3.53e-05,
3559
+ "loss": 0.5781,
3560
+ "step": 1176
3561
+ },
3562
+ {
3563
+ "epoch": 2.94,
3564
+ "learning_rate": 3.5275000000000004e-05,
3565
+ "loss": 0.1274,
3566
+ "step": 1178
3567
+ },
3568
+ {
3569
+ "epoch": 2.95,
3570
+ "learning_rate": 3.525e-05,
3571
+ "loss": 0.0737,
3572
+ "step": 1180
3573
+ },
3574
+ {
3575
+ "epoch": 2.96,
3576
+ "learning_rate": 3.5225e-05,
3577
+ "loss": 0.169,
3578
+ "step": 1182
3579
+ },
3580
+ {
3581
+ "epoch": 2.96,
3582
+ "learning_rate": 3.52e-05,
3583
+ "loss": 0.1855,
3584
+ "step": 1184
3585
+ },
3586
+ {
3587
+ "epoch": 2.96,
3588
+ "learning_rate": 3.5175e-05,
3589
+ "loss": 0.1207,
3590
+ "step": 1186
3591
+ },
3592
+ {
3593
+ "epoch": 2.97,
3594
+ "learning_rate": 3.515e-05,
3595
+ "loss": 0.2846,
3596
+ "step": 1188
3597
+ },
3598
+ {
3599
+ "epoch": 2.98,
3600
+ "learning_rate": 3.5125e-05,
3601
+ "loss": 0.2958,
3602
+ "step": 1190
3603
+ },
3604
+ {
3605
+ "epoch": 2.98,
3606
+ "learning_rate": 3.51e-05,
3607
+ "loss": 0.3242,
3608
+ "step": 1192
3609
+ },
3610
+ {
3611
+ "epoch": 2.98,
3612
+ "learning_rate": 3.5075000000000006e-05,
3613
+ "loss": 0.2893,
3614
+ "step": 1194
3615
+ },
3616
+ {
3617
+ "epoch": 2.99,
3618
+ "learning_rate": 3.505e-05,
3619
+ "loss": 0.2842,
3620
+ "step": 1196
3621
+ },
3622
+ {
3623
+ "epoch": 3.0,
3624
+ "learning_rate": 3.5025000000000004e-05,
3625
+ "loss": 0.2714,
3626
+ "step": 1198
3627
+ },
3628
+ {
3629
+ "epoch": 3.0,
3630
+ "learning_rate": 3.5e-05,
3631
+ "loss": 0.3058,
3632
+ "step": 1200
3633
+ },
3634
+ {
3635
+ "epoch": 3.0,
3636
+ "eval_accuracy": 0.7556953969672984,
3637
+ "eval_f1": 0.7496529757688725,
3638
+ "eval_loss": 0.4001426696777344,
3639
+ "eval_precision": 0.7470013420094054,
3640
+ "eval_recall": 0.7556953969672984,
3641
+ "eval_runtime": 46.5671,
3642
+ "eval_samples_per_second": 17.18,
3643
+ "eval_steps_per_second": 2.147,
3644
+ "step": 1200
3645
+ }
3646
+ ],
3647
+ "max_steps": 4000,
3648
+ "num_train_epochs": 10,
3649
+ "total_flos": 2.077633511424e+16,
3650
+ "trial_name": null,
3651
+ "trial_params": null
3652
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:40a21c49ed8493fb0415b2efb279756e5bb99044486927475ddb1c08e1328178
3
+ size 3567