Hugging Face's logo

PurCL
/
codeart-26m-ti-O1

Token Classification
Transformers
PyTorch
codeart
Inference Endpoints
Model card Files Community
sheepy928 commited on
Commit
5d1fe6d
·
verified ·
1 Parent(s): 5eb24cb

Upload 11 files

Browse files
Files changed (11) hide show
  1. config.json +133 -0
  2. optimizer.pt +3 -0
  3. pytorch_model.bin +3 -0
  4. rng_state_0.pth +3 -0
  5. rng_state_1.pth +3 -0
  6. scheduler.pt +3 -0
  7. special_tokens_map.json +7 -0
  8. tokenizer.json +0 -0
  9. tokenizer_config.json +5 -0
  10. trainer_state.json +2592 -0
  11. training_args.bin +3 -0
config.json ADDED
@@ -0,0 +1,133 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "../../save/marinda-default-ep0.4-26m/checkpoint-25000",
3
+ "architectures": [
4
+ "MarindaForTokenClassification"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.1,
7
+ "bos_token_id": 1,
8
+ "classifier_dropout": null,
9
+ "eos_token_id": 2,
10
+ "ep_add_linear_projection": true,
11
+ "hidden_act": "gelu",
12
+ "hidden_dropout_prob": 0.1,
13
+ "hidden_size": 768,
14
+ "id2label": {
15
+ "0": "noacc",
16
+ "1": "base(int)",
17
+ "2": "base(char)**",
18
+ "3": "base(char)*",
19
+ "4": "base(_Bool)",
20
+ "5": "base(unsigned int)",
21
+ "6": "base(long unsigned int)",
22
+ "7": "struct*",
23
+ "8": "void*",
24
+ "9": "base(unsigned char)*",
25
+ "10": "enum",
26
+ "11": "base(unsigned char)",
27
+ "12": "base(char)",
28
+ "13": "base(unsigned int)*",
29
+ "14": "base(long int)",
30
+ "15": "base(short int)*",
31
+ "16": "union*",
32
+ "17": "subroutine*",
33
+ "18": "base(long int)*",
34
+ "19": "base(int)*",
35
+ "20": "array*",
36
+ "21": "struct**",
37
+ "22": "base(long unsigned int)*",
38
+ "23": "base(double)*",
39
+ "24": "void**",
40
+ "25": "base(short unsigned int)",
41
+ "26": "struct",
42
+ "27": "base(_Bool)*",
43
+ "28": "base(char)***",
44
+ "29": "base(long unsigned int)**",
45
+ "30": "base(long double)",
46
+ "31": "struct***",
47
+ "32": "enum*",
48
+ "33": "base(unsigned int)**",
49
+ "34": "base(long double)*",
50
+ "35": "base(float)*",
51
+ "36": "base(long long unsigned int)*",
52
+ "37": "base(short unsigned int)*",
53
+ "38": "base(signed char)*",
54
+ "39": "base(long long int)",
55
+ "40": "array",
56
+ "41": "base(double)",
57
+ "42": "base(unsigned char)**",
58
+ "43": "union",
59
+ "44": "base(short int)",
60
+ "45": "base(long long unsigned int)",
61
+ "46": "subroutine**",
62
+ "47": "array**",
63
+ "48": "base(signed char)",
64
+ "49": "base(float)"
65
+ },
66
+ "initializer_range": 0.02,
67
+ "intermediate_size": 3072,
68
+ "label2id": {
69
+ "array": 40,
70
+ "array*": 20,
71
+ "array**": 47,
72
+ "base(_Bool)": 4,
73
+ "base(_Bool)*": 27,
74
+ "base(char)": 12,
75
+ "base(char)*": 3,
76
+ "base(char)**": 2,
77
+ "base(char)***": 28,
78
+ "base(double)": 41,
79
+ "base(double)*": 23,
80
+ "base(float)": 49,
81
+ "base(float)*": 35,
82
+ "base(int)": 1,
83
+ "base(int)*": 19,
84
+ "base(long double)": 30,
85
+ "base(long double)*": 34,
86
+ "base(long int)": 14,
87
+ "base(long int)*": 18,
88
+ "base(long long int)": 39,
89
+ "base(long long unsigned int)": 45,
90
+ "base(long long unsigned int)*": 36,
91
+ "base(long unsigned int)": 6,
92
+ "base(long unsigned int)*": 22,
93
+ "base(long unsigned int)**": 29,
94
+ "base(short int)": 44,
95
+ "base(short int)*": 15,
96
+ "base(short unsigned int)": 25,
97
+ "base(short unsigned int)*": 37,
98
+ "base(signed char)": 48,
99
+ "base(signed char)*": 38,
100
+ "base(unsigned char)": 11,
101
+ "base(unsigned char)*": 9,
102
+ "base(unsigned char)**": 42,
103
+ "base(unsigned int)": 5,
104
+ "base(unsigned int)*": 13,
105
+ "base(unsigned int)**": 33,
106
+ "enum": 10,
107
+ "enum*": 32,
108
+ "noacc": 0,
109
+ "struct": 26,
110
+ "struct*": 7,
111
+ "struct**": 21,
112
+ "struct***": 31,
113
+ "subroutine*": 17,
114
+ "subroutine**": 46,
115
+ "union": 43,
116
+ "union*": 16,
117
+ "void*": 8,
118
+ "void**": 24
119
+ },
120
+ "layer_norm_eps": 1e-12,
121
+ "max_position_embeddings": 600,
122
+ "max_relative_position_embeddings": 8,
123
+ "model_type": "marinda",
124
+ "num_attention_heads": 12,
125
+ "num_hidden_layers": 12,
126
+ "pad_token_id": 3,
127
+ "position_embedding_type": "mixed",
128
+ "torch_dtype": "float32",
129
+ "transformers_version": "4.31.0",
130
+ "type_vocab_size": 2,
131
+ "use_cache": true,
132
+ "vocab_size": 30523
133
+ }
optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:20967c4f20a058b52ca5a1d439e17ef0952cbeaeec26998ae8ac9a61964f8d17
3
+ size 872119157
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:18c294efe2c2767da850243458291e610c9d653a0f5b5dba728e450f50f4643a
3
+ size 436069275
rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0ccd18f81d3dbdca17929956a74cd990d4f3d14cf0ee198f0800dc432d059417
3
+ size 15523
rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6b8453f020c4c5866fd5f69d9e6bda271b3b78697ca5bff7fe6c5799ce53cbef
3
+ size 15523
scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8e577e5c30634363761c0a2d1ea43c8dcb502d6d9864433a28ad256f7b2f33e6
3
+ size 623
special_tokens_map.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "cls_token": "[CLS]",
3
+ "mask_token": "[MASK]",
4
+ "pad_token": "[PAD]",
5
+ "sep_token": "[SEP]",
6
+ "unk_token": "[UNK]"
7
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "clean_up_tokenization_spaces": true,
3
+ "model_max_length": 512,
4
+ "tokenizer_class": "MarindaTokenizer"
5
+ }
trainer_state.json ADDED
@@ -0,0 +1,2592 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.9480086057353725,
3
+ "best_model_checkpoint": "../save/marinda-default-ep0.4-26m-ti-dbgonly-O1-shuffle/checkpoint-800",
4
+ "epoch": 4.3478260869565215,
5
+ "global_step": 800,
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": 9.978260869565218e-05,
13
+ "loss": 3.7941,
14
+ "step": 2
15
+ },
16
+ {
17
+ "epoch": 0.02,
18
+ "learning_rate": 9.956521739130435e-05,
19
+ "loss": 3.3761,
20
+ "step": 4
21
+ },
22
+ {
23
+ "epoch": 0.03,
24
+ "learning_rate": 9.934782608695653e-05,
25
+ "loss": 2.6042,
26
+ "step": 6
27
+ },
28
+ {
29
+ "epoch": 0.04,
30
+ "learning_rate": 9.91304347826087e-05,
31
+ "loss": 1.3506,
32
+ "step": 8
33
+ },
34
+ {
35
+ "epoch": 0.05,
36
+ "learning_rate": 9.891304347826087e-05,
37
+ "loss": 0.6191,
38
+ "step": 10
39
+ },
40
+ {
41
+ "epoch": 0.07,
42
+ "learning_rate": 9.869565217391305e-05,
43
+ "loss": 0.4785,
44
+ "step": 12
45
+ },
46
+ {
47
+ "epoch": 0.08,
48
+ "learning_rate": 9.847826086956522e-05,
49
+ "loss": 0.4405,
50
+ "step": 14
51
+ },
52
+ {
53
+ "epoch": 0.09,
54
+ "learning_rate": 9.82608695652174e-05,
55
+ "loss": 0.3909,
56
+ "step": 16
57
+ },
58
+ {
59
+ "epoch": 0.1,
60
+ "learning_rate": 9.804347826086957e-05,
61
+ "loss": 0.3607,
62
+ "step": 18
63
+ },
64
+ {
65
+ "epoch": 0.11,
66
+ "learning_rate": 9.782608695652174e-05,
67
+ "loss": 0.3224,
68
+ "step": 20
69
+ },
70
+ {
71
+ "epoch": 0.12,
72
+ "learning_rate": 9.760869565217392e-05,
73
+ "loss": 0.3123,
74
+ "step": 22
75
+ },
76
+ {
77
+ "epoch": 0.13,
78
+ "learning_rate": 9.739130434782609e-05,
79
+ "loss": 0.3061,
80
+ "step": 24
81
+ },
82
+ {
83
+ "epoch": 0.14,
84
+ "learning_rate": 9.717391304347827e-05,
85
+ "loss": 0.2739,
86
+ "step": 26
87
+ },
88
+ {
89
+ "epoch": 0.15,
90
+ "learning_rate": 9.695652173913044e-05,
91
+ "loss": 0.266,
92
+ "step": 28
93
+ },
94
+ {
95
+ "epoch": 0.16,
96
+ "learning_rate": 9.673913043478261e-05,
97
+ "loss": 0.246,
98
+ "step": 30
99
+ },
100
+ {
101
+ "epoch": 0.17,
102
+ "learning_rate": 9.652173913043479e-05,
103
+ "loss": 0.2488,
104
+ "step": 32
105
+ },
106
+ {
107
+ "epoch": 0.18,
108
+ "learning_rate": 9.630434782608696e-05,
109
+ "loss": 0.233,
110
+ "step": 34
111
+ },
112
+ {
113
+ "epoch": 0.2,
114
+ "learning_rate": 9.608695652173914e-05,
115
+ "loss": 0.1973,
116
+ "step": 36
117
+ },
118
+ {
119
+ "epoch": 0.21,
120
+ "learning_rate": 9.586956521739131e-05,
121
+ "loss": 0.1989,
122
+ "step": 38
123
+ },
124
+ {
125
+ "epoch": 0.22,
126
+ "learning_rate": 9.565217391304348e-05,
127
+ "loss": 0.2195,
128
+ "step": 40
129
+ },
130
+ {
131
+ "epoch": 0.23,
132
+ "learning_rate": 9.543478260869566e-05,
133
+ "loss": 0.1891,
134
+ "step": 42
135
+ },
136
+ {
137
+ "epoch": 0.24,
138
+ "learning_rate": 9.521739130434783e-05,
139
+ "loss": 0.1964,
140
+ "step": 44
141
+ },
142
+ {
143
+ "epoch": 0.25,
144
+ "learning_rate": 9.5e-05,
145
+ "loss": 0.1911,
146
+ "step": 46
147
+ },
148
+ {
149
+ "epoch": 0.26,
150
+ "learning_rate": 9.478260869565218e-05,
151
+ "loss": 0.1747,
152
+ "step": 48
153
+ },
154
+ {
155
+ "epoch": 0.27,
156
+ "learning_rate": 9.456521739130435e-05,
157
+ "loss": 0.1626,
158
+ "step": 50
159
+ },
160
+ {
161
+ "epoch": 0.27,
162
+ "eval_f1": 0.3964671547979424,
163
+ "eval_loss": 0.1601678729057312,
164
+ "eval_precision": 0.4276477142864008,
165
+ "eval_recall": 0.36953310617845153,
166
+ "eval_runtime": 13.7757,
167
+ "eval_samples_per_second": 134.73,
168
+ "eval_steps_per_second": 0.581,
169
+ "step": 50
170
+ },
171
+ {
172
+ "epoch": 0.28,
173
+ "learning_rate": 9.434782608695653e-05,
174
+ "loss": 0.1717,
175
+ "step": 52
176
+ },
177
+ {
178
+ "epoch": 0.29,
179
+ "learning_rate": 9.41304347826087e-05,
180
+ "loss": 0.1846,
181
+ "step": 54
182
+ },
183
+ {
184
+ "epoch": 0.3,
185
+ "learning_rate": 9.391304347826087e-05,
186
+ "loss": 0.1602,
187
+ "step": 56
188
+ },
189
+ {
190
+ "epoch": 0.32,
191
+ "learning_rate": 9.369565217391305e-05,
192
+ "loss": 0.1815,
193
+ "step": 58
194
+ },
195
+ {
196
+ "epoch": 0.33,
197
+ "learning_rate": 9.347826086956522e-05,
198
+ "loss": 0.1604,
199
+ "step": 60
200
+ },
201
+ {
202
+ "epoch": 0.34,
203
+ "learning_rate": 9.32608695652174e-05,
204
+ "loss": 0.1519,
205
+ "step": 62
206
+ },
207
+ {
208
+ "epoch": 0.35,
209
+ "learning_rate": 9.304347826086957e-05,
210
+ "loss": 0.1486,
211
+ "step": 64
212
+ },
213
+ {
214
+ "epoch": 0.36,
215
+ "learning_rate": 9.282608695652174e-05,
216
+ "loss": 0.1454,
217
+ "step": 66
218
+ },
219
+ {
220
+ "epoch": 0.37,
221
+ "learning_rate": 9.260869565217392e-05,
222
+ "loss": 0.1415,
223
+ "step": 68
224
+ },
225
+ {
226
+ "epoch": 0.38,
227
+ "learning_rate": 9.239130434782609e-05,
228
+ "loss": 0.129,
229
+ "step": 70
230
+ },
231
+ {
232
+ "epoch": 0.39,
233
+ "learning_rate": 9.217391304347827e-05,
234
+ "loss": 0.1269,
235
+ "step": 72
236
+ },
237
+ {
238
+ "epoch": 0.4,
239
+ "learning_rate": 9.195652173913044e-05,
240
+ "loss": 0.1261,
241
+ "step": 74
242
+ },
243
+ {
244
+ "epoch": 0.41,
245
+ "learning_rate": 9.173913043478261e-05,
246
+ "loss": 0.1232,
247
+ "step": 76
248
+ },
249
+ {
250
+ "epoch": 0.42,
251
+ "learning_rate": 9.152173913043479e-05,
252
+ "loss": 0.1203,
253
+ "step": 78
254
+ },
255
+ {
256
+ "epoch": 0.43,
257
+ "learning_rate": 9.130434782608696e-05,
258
+ "loss": 0.1202,
259
+ "step": 80
260
+ },
261
+ {
262
+ "epoch": 0.45,
263
+ "learning_rate": 9.108695652173914e-05,
264
+ "loss": 0.1348,
265
+ "step": 82
266
+ },
267
+ {
268
+ "epoch": 0.46,
269
+ "learning_rate": 9.086956521739131e-05,
270
+ "loss": 0.1151,
271
+ "step": 84
272
+ },
273
+ {
274
+ "epoch": 0.47,
275
+ "learning_rate": 9.065217391304348e-05,
276
+ "loss": 0.1143,
277
+ "step": 86
278
+ },
279
+ {
280
+ "epoch": 0.48,
281
+ "learning_rate": 9.043478260869566e-05,
282
+ "loss": 0.1099,
283
+ "step": 88
284
+ },
285
+ {
286
+ "epoch": 0.49,
287
+ "learning_rate": 9.021739130434783e-05,
288
+ "loss": 0.1174,
289
+ "step": 90
290
+ },
291
+ {
292
+ "epoch": 0.5,
293
+ "learning_rate": 9e-05,
294
+ "loss": 0.1049,
295
+ "step": 92
296
+ },
297
+ {
298
+ "epoch": 0.51,
299
+ "learning_rate": 8.978260869565218e-05,
300
+ "loss": 0.0966,
301
+ "step": 94
302
+ },
303
+ {
304
+ "epoch": 0.52,
305
+ "learning_rate": 8.956521739130435e-05,
306
+ "loss": 0.1063,
307
+ "step": 96
308
+ },
309
+ {
310
+ "epoch": 0.53,
311
+ "learning_rate": 8.934782608695653e-05,
312
+ "loss": 0.1049,
313
+ "step": 98
314
+ },
315
+ {
316
+ "epoch": 0.54,
317
+ "learning_rate": 8.91304347826087e-05,
318
+ "loss": 0.1005,
319
+ "step": 100
320
+ },
321
+ {
322
+ "epoch": 0.54,
323
+ "eval_f1": 0.6856549592604718,
324
+ "eval_loss": 0.08935681730508804,
325
+ "eval_precision": 0.6768018013783773,
326
+ "eval_recall": 0.6947530661519793,
327
+ "eval_runtime": 14.0272,
328
+ "eval_samples_per_second": 132.314,
329
+ "eval_steps_per_second": 0.57,
330
+ "step": 100
331
+ },
332
+ {
333
+ "epoch": 0.55,
334
+ "learning_rate": 8.891304347826088e-05,
335
+ "loss": 0.1007,
336
+ "step": 102
337
+ },
338
+ {
339
+ "epoch": 0.57,
340
+ "learning_rate": 8.869565217391305e-05,
341
+ "loss": 0.0943,
342
+ "step": 104
343
+ },
344
+ {
345
+ "epoch": 0.58,
346
+ "learning_rate": 8.847826086956522e-05,
347
+ "loss": 0.0872,
348
+ "step": 106
349
+ },
350
+ {
351
+ "epoch": 0.59,
352
+ "learning_rate": 8.82608695652174e-05,
353
+ "loss": 0.0936,
354
+ "step": 108
355
+ },
356
+ {
357
+ "epoch": 0.6,
358
+ "learning_rate": 8.804347826086957e-05,
359
+ "loss": 0.0862,
360
+ "step": 110
361
+ },
362
+ {
363
+ "epoch": 0.61,
364
+ "learning_rate": 8.782608695652174e-05,
365
+ "loss": 0.0871,
366
+ "step": 112
367
+ },
368
+ {
369
+ "epoch": 0.62,
370
+ "learning_rate": 8.760869565217392e-05,
371
+ "loss": 0.0922,
372
+ "step": 114
373
+ },
374
+ {
375
+ "epoch": 0.63,
376
+ "learning_rate": 8.739130434782609e-05,
377
+ "loss": 0.0862,
378
+ "step": 116
379
+ },
380
+ {
381
+ "epoch": 0.64,
382
+ "learning_rate": 8.717391304347827e-05,
383
+ "loss": 0.0824,
384
+ "step": 118
385
+ },
386
+ {
387
+ "epoch": 0.65,
388
+ "learning_rate": 8.695652173913044e-05,
389
+ "loss": 0.0713,
390
+ "step": 120
391
+ },
392
+ {
393
+ "epoch": 0.66,
394
+ "learning_rate": 8.673913043478261e-05,
395
+ "loss": 0.0847,
396
+ "step": 122
397
+ },
398
+ {
399
+ "epoch": 0.67,
400
+ "learning_rate": 8.652173913043479e-05,
401
+ "loss": 0.0896,
402
+ "step": 124
403
+ },
404
+ {
405
+ "epoch": 0.68,
406
+ "learning_rate": 8.630434782608696e-05,
407
+ "loss": 0.0806,
408
+ "step": 126
409
+ },
410
+ {
411
+ "epoch": 0.7,
412
+ "learning_rate": 8.608695652173914e-05,
413
+ "loss": 0.0859,
414
+ "step": 128
415
+ },
416
+ {
417
+ "epoch": 0.71,
418
+ "learning_rate": 8.586956521739131e-05,
419
+ "loss": 0.0898,
420
+ "step": 130
421
+ },
422
+ {
423
+ "epoch": 0.72,
424
+ "learning_rate": 8.565217391304348e-05,
425
+ "loss": 0.0699,
426
+ "step": 132
427
+ },
428
+ {
429
+ "epoch": 0.73,
430
+ "learning_rate": 8.543478260869566e-05,
431
+ "loss": 0.0748,
432
+ "step": 134
433
+ },
434
+ {
435
+ "epoch": 0.74,
436
+ "learning_rate": 8.521739130434783e-05,
437
+ "loss": 0.0767,
438
+ "step": 136
439
+ },
440
+ {
441
+ "epoch": 0.75,
442
+ "learning_rate": 8.5e-05,
443
+ "loss": 0.0794,
444
+ "step": 138
445
+ },
446
+ {
447
+ "epoch": 0.76,
448
+ "learning_rate": 8.478260869565218e-05,
449
+ "loss": 0.0775,
450
+ "step": 140
451
+ },
452
+ {
453
+ "epoch": 0.77,
454
+ "learning_rate": 8.456521739130435e-05,
455
+ "loss": 0.0727,
456
+ "step": 142
457
+ },
458
+ {
459
+ "epoch": 0.78,
460
+ "learning_rate": 8.434782608695653e-05,
461
+ "loss": 0.0723,
462
+ "step": 144
463
+ },
464
+ {
465
+ "epoch": 0.79,
466
+ "learning_rate": 8.41304347826087e-05,
467
+ "loss": 0.0765,
468
+ "step": 146
469
+ },
470
+ {
471
+ "epoch": 0.8,
472
+ "learning_rate": 8.391304347826088e-05,
473
+ "loss": 0.0721,
474
+ "step": 148
475
+ },
476
+ {
477
+ "epoch": 0.82,
478
+ "learning_rate": 8.369565217391305e-05,
479
+ "loss": 0.0691,
480
+ "step": 150
481
+ },
482
+ {
483
+ "epoch": 0.82,
484
+ "eval_f1": 0.7960363629529713,
485
+ "eval_loss": 0.06118812412023544,
486
+ "eval_precision": 0.7864619236687798,
487
+ "eval_recall": 0.805857041419397,
488
+ "eval_runtime": 13.4485,
489
+ "eval_samples_per_second": 138.008,
490
+ "eval_steps_per_second": 0.595,
491
+ "step": 150
492
+ },
493
+ {
494
+ "epoch": 0.83,
495
+ "learning_rate": 8.347826086956521e-05,
496
+ "loss": 0.0603,
497
+ "step": 152
498
+ },
499
+ {
500
+ "epoch": 0.84,
501
+ "learning_rate": 8.32608695652174e-05,
502
+ "loss": 0.0689,
503
+ "step": 154
504
+ },
505
+ {
506
+ "epoch": 0.85,
507
+ "learning_rate": 8.304347826086957e-05,
508
+ "loss": 0.0703,
509
+ "step": 156
510
+ },
511
+ {
512
+ "epoch": 0.86,
513
+ "learning_rate": 8.282608695652175e-05,
514
+ "loss": 0.06,
515
+ "step": 158
516
+ },
517
+ {
518
+ "epoch": 0.87,
519
+ "learning_rate": 8.260869565217392e-05,
520
+ "loss": 0.0679,
521
+ "step": 160
522
+ },
523
+ {
524
+ "epoch": 0.88,
525
+ "learning_rate": 8.23913043478261e-05,
526
+ "loss": 0.0649,
527
+ "step": 162
528
+ },
529
+ {
530
+ "epoch": 0.89,
531
+ "learning_rate": 8.217391304347827e-05,
532
+ "loss": 0.0658,
533
+ "step": 164
534
+ },
535
+ {
536
+ "epoch": 0.9,
537
+ "learning_rate": 8.195652173913044e-05,
538
+ "loss": 0.062,
539
+ "step": 166
540
+ },
541
+ {
542
+ "epoch": 0.91,
543
+ "learning_rate": 8.173913043478262e-05,
544
+ "loss": 0.0676,
545
+ "step": 168
546
+ },
547
+ {
548
+ "epoch": 0.92,
549
+ "learning_rate": 8.152173913043478e-05,
550
+ "loss": 0.0577,
551
+ "step": 170
552
+ },
553
+ {
554
+ "epoch": 0.93,
555
+ "learning_rate": 8.130434782608696e-05,
556
+ "loss": 0.0659,
557
+ "step": 172
558
+ },
559
+ {
560
+ "epoch": 0.95,
561
+ "learning_rate": 8.108695652173914e-05,
562
+ "loss": 0.0555,
563
+ "step": 174
564
+ },
565
+ {
566
+ "epoch": 0.96,
567
+ "learning_rate": 8.086956521739131e-05,
568
+ "loss": 0.0581,
569
+ "step": 176
570
+ },
571
+ {
572
+ "epoch": 0.97,
573
+ "learning_rate": 8.065217391304348e-05,
574
+ "loss": 0.0509,
575
+ "step": 178
576
+ },
577
+ {
578
+ "epoch": 0.98,
579
+ "learning_rate": 8.043478260869566e-05,
580
+ "loss": 0.0651,
581
+ "step": 180
582
+ },
583
+ {
584
+ "epoch": 0.99,
585
+ "learning_rate": 8.021739130434783e-05,
586
+ "loss": 0.0618,
587
+ "step": 182
588
+ },
589
+ {
590
+ "epoch": 1.0,
591
+ "learning_rate": 8e-05,
592
+ "loss": 0.0583,
593
+ "step": 184
594
+ },
595
+ {
596
+ "epoch": 1.01,
597
+ "learning_rate": 7.978260869565217e-05,
598
+ "loss": 0.0548,
599
+ "step": 186
600
+ },
601
+ {
602
+ "epoch": 1.02,
603
+ "learning_rate": 7.956521739130434e-05,
604
+ "loss": 0.0541,
605
+ "step": 188
606
+ },
607
+ {
608
+ "epoch": 1.03,
609
+ "learning_rate": 7.934782608695653e-05,
610
+ "loss": 0.0568,
611
+ "step": 190
612
+ },
613
+ {
614
+ "epoch": 1.04,
615
+ "learning_rate": 7.91304347826087e-05,
616
+ "loss": 0.0516,
617
+ "step": 192
618
+ },
619
+ {
620
+ "epoch": 1.05,
621
+ "learning_rate": 7.891304347826088e-05,
622
+ "loss": 0.0547,
623
+ "step": 194
624
+ },
625
+ {
626
+ "epoch": 1.07,
627
+ "learning_rate": 7.869565217391305e-05,
628
+ "loss": 0.0472,
629
+ "step": 196
630
+ },
631
+ {
632
+ "epoch": 1.08,
633
+ "learning_rate": 7.847826086956522e-05,
634
+ "loss": 0.0474,
635
+ "step": 198
636
+ },
637
+ {
638
+ "epoch": 1.09,
639
+ "learning_rate": 7.82608695652174e-05,
640
+ "loss": 0.0529,
641
+ "step": 200
642
+ },
643
+ {
644
+ "epoch": 1.09,
645
+ "eval_f1": 0.855353337621845,
646
+ "eval_loss": 0.04639717563986778,
647
+ "eval_precision": 0.8488931050924142,
648
+ "eval_recall": 0.8619228046612788,
649
+ "eval_runtime": 13.5842,
650
+ "eval_samples_per_second": 136.629,
651
+ "eval_steps_per_second": 0.589,
652
+ "step": 200
653
+ },
654
+ {
655
+ "epoch": 1.1,
656
+ "learning_rate": 7.804347826086957e-05,
657
+ "loss": 0.0595,
658
+ "step": 202
659
+ },
660
+ {
661
+ "epoch": 1.11,
662
+ "learning_rate": 7.782608695652173e-05,
663
+ "loss": 0.0462,
664
+ "step": 204
665
+ },
666
+ {
667
+ "epoch": 1.12,
668
+ "learning_rate": 7.76086956521739e-05,
669
+ "loss": 0.0543,
670
+ "step": 206
671
+ },
672
+ {
673
+ "epoch": 1.13,
674
+ "learning_rate": 7.73913043478261e-05,
675
+ "loss": 0.0541,
676
+ "step": 208
677
+ },
678
+ {
679
+ "epoch": 1.14,
680
+ "learning_rate": 7.717391304347827e-05,
681
+ "loss": 0.0634,
682
+ "step": 210
683
+ },
684
+ {
685
+ "epoch": 1.15,
686
+ "learning_rate": 7.695652173913044e-05,
687
+ "loss": 0.0547,
688
+ "step": 212
689
+ },
690
+ {
691
+ "epoch": 1.16,
692
+ "learning_rate": 7.673913043478262e-05,
693
+ "loss": 0.0568,
694
+ "step": 214
695
+ },
696
+ {
697
+ "epoch": 1.17,
698
+ "learning_rate": 7.652173913043479e-05,
699
+ "loss": 0.0521,
700
+ "step": 216
701
+ },
702
+ {
703
+ "epoch": 1.18,
704
+ "learning_rate": 7.630434782608696e-05,
705
+ "loss": 0.0565,
706
+ "step": 218
707
+ },
708
+ {
709
+ "epoch": 1.2,
710
+ "learning_rate": 7.608695652173914e-05,
711
+ "loss": 0.048,
712
+ "step": 220
713
+ },
714
+ {
715
+ "epoch": 1.21,
716
+ "learning_rate": 7.58695652173913e-05,
717
+ "loss": 0.0467,
718
+ "step": 222
719
+ },
720
+ {
721
+ "epoch": 1.22,
722
+ "learning_rate": 7.565217391304347e-05,
723
+ "loss": 0.0522,
724
+ "step": 224
725
+ },
726
+ {
727
+ "epoch": 1.23,
728
+ "learning_rate": 7.543478260869566e-05,
729
+ "loss": 0.0504,
730
+ "step": 226
731
+ },
732
+ {
733
+ "epoch": 1.24,
734
+ "learning_rate": 7.521739130434783e-05,
735
+ "loss": 0.053,
736
+ "step": 228
737
+ },
738
+ {
739
+ "epoch": 1.25,
740
+ "learning_rate": 7.500000000000001e-05,
741
+ "loss": 0.0519,
742
+ "step": 230
743
+ },
744
+ {
745
+ "epoch": 1.26,
746
+ "learning_rate": 7.478260869565218e-05,
747
+ "loss": 0.0441,
748
+ "step": 232
749
+ },
750
+ {
751
+ "epoch": 1.27,
752
+ "learning_rate": 7.456521739130435e-05,
753
+ "loss": 0.0469,
754
+ "step": 234
755
+ },
756
+ {
757
+ "epoch": 1.28,
758
+ "learning_rate": 7.434782608695653e-05,
759
+ "loss": 0.0409,
760
+ "step": 236
761
+ },
762
+ {
763
+ "epoch": 1.29,
764
+ "learning_rate": 7.41304347826087e-05,
765
+ "loss": 0.0471,
766
+ "step": 238
767
+ },
768
+ {
769
+ "epoch": 1.3,
770
+ "learning_rate": 7.391304347826086e-05,
771
+ "loss": 0.0523,
772
+ "step": 240
773
+ },
774
+ {
775
+ "epoch": 1.32,
776
+ "learning_rate": 7.369565217391304e-05,
777
+ "loss": 0.0455,
778
+ "step": 242
779
+ },
780
+ {
781
+ "epoch": 1.33,
782
+ "learning_rate": 7.347826086956522e-05,
783
+ "loss": 0.0498,
784
+ "step": 244
785
+ },
786
+ {
787
+ "epoch": 1.34,
788
+ "learning_rate": 7.32608695652174e-05,
789
+ "loss": 0.0468,
790
+ "step": 246
791
+ },
792
+ {
793
+ "epoch": 1.35,
794
+ "learning_rate": 7.304347826086957e-05,
795
+ "loss": 0.0441,
796
+ "step": 248
797
+ },
798
+ {
799
+ "epoch": 1.36,
800
+ "learning_rate": 7.282608695652175e-05,
801
+ "loss": 0.0379,
802
+ "step": 250
803
+ },
804
+ {
805
+ "epoch": 1.36,
806
+ "eval_f1": 0.8870407621848339,
807
+ "eval_loss": 0.0366128571331501,
808
+ "eval_precision": 0.8915342193373116,
809
+ "eval_recall": 0.8826022728902433,
810
+ "eval_runtime": 13.3791,
811
+ "eval_samples_per_second": 138.724,
812
+ "eval_steps_per_second": 0.598,
813
+ "step": 250
814
+ },
815
+ {
816
+ "epoch": 1.37,
817
+ "learning_rate": 7.260869565217392e-05,
818
+ "loss": 0.0467,
819
+ "step": 252
820
+ },
821
+ {
822
+ "epoch": 1.38,
823
+ "learning_rate": 7.23913043478261e-05,
824
+ "loss": 0.0469,
825
+ "step": 254
826
+ },
827
+ {
828
+ "epoch": 1.39,
829
+ "learning_rate": 7.217391304347827e-05,
830
+ "loss": 0.0396,
831
+ "step": 256
832
+ },
833
+ {
834
+ "epoch": 1.4,
835
+ "learning_rate": 7.195652173913043e-05,
836
+ "loss": 0.0449,
837
+ "step": 258
838
+ },
839
+ {
840
+ "epoch": 1.41,
841
+ "learning_rate": 7.17391304347826e-05,
842
+ "loss": 0.0476,
843
+ "step": 260
844
+ },
845
+ {
846
+ "epoch": 1.42,
847
+ "learning_rate": 7.152173913043479e-05,
848
+ "loss": 0.0442,
849
+ "step": 262
850
+ },
851
+ {
852
+ "epoch": 1.43,
853
+ "learning_rate": 7.130434782608696e-05,
854
+ "loss": 0.0504,
855
+ "step": 264
856
+ },
857
+ {
858
+ "epoch": 1.45,
859
+ "learning_rate": 7.108695652173914e-05,
860
+ "loss": 0.0464,
861
+ "step": 266
862
+ },
863
+ {
864
+ "epoch": 1.46,
865
+ "learning_rate": 7.086956521739131e-05,
866
+ "loss": 0.0433,
867
+ "step": 268
868
+ },
869
+ {
870
+ "epoch": 1.47,
871
+ "learning_rate": 7.065217391304349e-05,
872
+ "loss": 0.0388,
873
+ "step": 270
874
+ },
875
+ {
876
+ "epoch": 1.48,
877
+ "learning_rate": 7.043478260869566e-05,
878
+ "loss": 0.0496,
879
+ "step": 272
880
+ },
881
+ {
882
+ "epoch": 1.49,
883
+ "learning_rate": 7.021739130434783e-05,
884
+ "loss": 0.0383,
885
+ "step": 274
886
+ },
887
+ {
888
+ "epoch": 1.5,
889
+ "learning_rate": 7e-05,
890
+ "loss": 0.032,
891
+ "step": 276
892
+ },
893
+ {
894
+ "epoch": 1.51,
895
+ "learning_rate": 6.978260869565217e-05,
896
+ "loss": 0.0368,
897
+ "step": 278
898
+ },
899
+ {
900
+ "epoch": 1.52,
901
+ "learning_rate": 6.956521739130436e-05,
902
+ "loss": 0.0427,
903
+ "step": 280
904
+ },
905
+ {
906
+ "epoch": 1.53,
907
+ "learning_rate": 6.934782608695653e-05,
908
+ "loss": 0.0399,
909
+ "step": 282
910
+ },
911
+ {
912
+ "epoch": 1.54,
913
+ "learning_rate": 6.91304347826087e-05,
914
+ "loss": 0.0409,
915
+ "step": 284
916
+ },
917
+ {
918
+ "epoch": 1.55,
919
+ "learning_rate": 6.891304347826088e-05,
920
+ "loss": 0.0428,
921
+ "step": 286
922
+ },
923
+ {
924
+ "epoch": 1.57,
925
+ "learning_rate": 6.869565217391305e-05,
926
+ "loss": 0.034,
927
+ "step": 288
928
+ },
929
+ {
930
+ "epoch": 1.58,
931
+ "learning_rate": 6.847826086956522e-05,
932
+ "loss": 0.0385,
933
+ "step": 290
934
+ },
935
+ {
936
+ "epoch": 1.59,
937
+ "learning_rate": 6.826086956521739e-05,
938
+ "loss": 0.0325,
939
+ "step": 292
940
+ },
941
+ {
942
+ "epoch": 1.6,
943
+ "learning_rate": 6.804347826086956e-05,
944
+ "loss": 0.041,
945
+ "step": 294
946
+ },
947
+ {
948
+ "epoch": 1.61,
949
+ "learning_rate": 6.782608695652173e-05,
950
+ "loss": 0.0352,
951
+ "step": 296
952
+ },
953
+ {
954
+ "epoch": 1.62,
955
+ "learning_rate": 6.760869565217392e-05,
956
+ "loss": 0.0459,
957
+ "step": 298
958
+ },
959
+ {
960
+ "epoch": 1.63,
961
+ "learning_rate": 6.73913043478261e-05,
962
+ "loss": 0.0336,
963
+ "step": 300
964
+ },
965
+ {
966
+ "epoch": 1.63,
967
+ "eval_f1": 0.9004441256217653,
968
+ "eval_loss": 0.03201919421553612,
969
+ "eval_precision": 0.9060996222486022,
970
+ "eval_recall": 0.8948686655353744,
971
+ "eval_runtime": 13.4043,
972
+ "eval_samples_per_second": 138.463,
973
+ "eval_steps_per_second": 0.597,
974
+ "step": 300
975
+ },
976
+ {
977
+ "epoch": 1.64,
978
+ "learning_rate": 6.717391304347827e-05,
979
+ "loss": 0.0463,
980
+ "step": 302
981
+ },
982
+ {
983
+ "epoch": 1.65,
984
+ "learning_rate": 6.695652173913044e-05,
985
+ "loss": 0.0369,
986
+ "step": 304
987
+ },
988
+ {
989
+ "epoch": 1.66,
990
+ "learning_rate": 6.673913043478262e-05,
991
+ "loss": 0.0374,
992
+ "step": 306
993
+ },
994
+ {
995
+ "epoch": 1.67,
996
+ "learning_rate": 6.652173913043479e-05,
997
+ "loss": 0.0407,
998
+ "step": 308
999
+ },
1000
+ {
1001
+ "epoch": 1.68,
1002
+ "learning_rate": 6.630434782608695e-05,
1003
+ "loss": 0.0375,
1004
+ "step": 310
1005
+ },
1006
+ {
1007
+ "epoch": 1.7,
1008
+ "learning_rate": 6.608695652173912e-05,
1009
+ "loss": 0.0326,
1010
+ "step": 312
1011
+ },
1012
+ {
1013
+ "epoch": 1.71,
1014
+ "learning_rate": 6.58695652173913e-05,
1015
+ "loss": 0.0434,
1016
+ "step": 314
1017
+ },
1018
+ {
1019
+ "epoch": 1.72,
1020
+ "learning_rate": 6.565217391304349e-05,
1021
+ "loss": 0.0352,
1022
+ "step": 316
1023
+ },
1024
+ {
1025
+ "epoch": 1.73,
1026
+ "learning_rate": 6.543478260869566e-05,
1027
+ "loss": 0.0374,
1028
+ "step": 318
1029
+ },
1030
+ {
1031
+ "epoch": 1.74,
1032
+ "learning_rate": 6.521739130434783e-05,
1033
+ "loss": 0.031,
1034
+ "step": 320
1035
+ },
1036
+ {
1037
+ "epoch": 1.75,
1038
+ "learning_rate": 6.500000000000001e-05,
1039
+ "loss": 0.0328,
1040
+ "step": 322
1041
+ },
1042
+ {
1043
+ "epoch": 1.76,
1044
+ "learning_rate": 6.478260869565218e-05,
1045
+ "loss": 0.0318,
1046
+ "step": 324
1047
+ },
1048
+ {
1049
+ "epoch": 1.77,
1050
+ "learning_rate": 6.456521739130436e-05,
1051
+ "loss": 0.0367,
1052
+ "step": 326
1053
+ },
1054
+ {
1055
+ "epoch": 1.78,
1056
+ "learning_rate": 6.434782608695652e-05,
1057
+ "loss": 0.0379,
1058
+ "step": 328
1059
+ },
1060
+ {
1061
+ "epoch": 1.79,
1062
+ "learning_rate": 6.413043478260869e-05,
1063
+ "loss": 0.0331,
1064
+ "step": 330
1065
+ },
1066
+ {
1067
+ "epoch": 1.8,
1068
+ "learning_rate": 6.391304347826086e-05,
1069
+ "loss": 0.0296,
1070
+ "step": 332
1071
+ },
1072
+ {
1073
+ "epoch": 1.82,
1074
+ "learning_rate": 6.369565217391305e-05,
1075
+ "loss": 0.0317,
1076
+ "step": 334
1077
+ },
1078
+ {
1079
+ "epoch": 1.83,
1080
+ "learning_rate": 6.347826086956523e-05,
1081
+ "loss": 0.0328,
1082
+ "step": 336
1083
+ },
1084
+ {
1085
+ "epoch": 1.84,
1086
+ "learning_rate": 6.32608695652174e-05,
1087
+ "loss": 0.0349,
1088
+ "step": 338
1089
+ },
1090
+ {
1091
+ "epoch": 1.85,
1092
+ "learning_rate": 6.304347826086957e-05,
1093
+ "loss": 0.0471,
1094
+ "step": 340
1095
+ },
1096
+ {
1097
+ "epoch": 1.86,
1098
+ "learning_rate": 6.282608695652175e-05,
1099
+ "loss": 0.0321,
1100
+ "step": 342
1101
+ },
1102
+ {
1103
+ "epoch": 1.87,
1104
+ "learning_rate": 6.260869565217392e-05,
1105
+ "loss": 0.0323,
1106
+ "step": 344
1107
+ },
1108
+ {
1109
+ "epoch": 1.88,
1110
+ "learning_rate": 6.239130434782608e-05,
1111
+ "loss": 0.0341,
1112
+ "step": 346
1113
+ },
1114
+ {
1115
+ "epoch": 1.89,
1116
+ "learning_rate": 6.217391304347826e-05,
1117
+ "loss": 0.0312,
1118
+ "step": 348
1119
+ },
1120
+ {
1121
+ "epoch": 1.9,
1122
+ "learning_rate": 6.195652173913043e-05,
1123
+ "loss": 0.032,
1124
+ "step": 350
1125
+ },
1126
+ {
1127
+ "epoch": 1.9,
1128
+ "eval_f1": 0.9094130772705891,
1129
+ "eval_loss": 0.028529103845357895,
1130
+ "eval_precision": 0.9100321537556063,
1131
+ "eval_recall": 0.9088048289070677,
1132
+ "eval_runtime": 13.367,
1133
+ "eval_samples_per_second": 138.849,
1134
+ "eval_steps_per_second": 0.598,
1135
+ "step": 350
1136
+ },
1137
+ {
1138
+ "epoch": 1.91,
1139
+ "learning_rate": 6.173913043478262e-05,
1140
+ "loss": 0.0305,
1141
+ "step": 352
1142
+ },
1143
+ {
1144
+ "epoch": 1.92,
1145
+ "learning_rate": 6.152173913043479e-05,
1146
+ "loss": 0.0287,
1147
+ "step": 354
1148
+ },
1149
+ {
1150
+ "epoch": 1.93,
1151
+ "learning_rate": 6.130434782608696e-05,
1152
+ "loss": 0.038,
1153
+ "step": 356
1154
+ },
1155
+ {
1156
+ "epoch": 1.95,
1157
+ "learning_rate": 6.108695652173914e-05,
1158
+ "loss": 0.0336,
1159
+ "step": 358
1160
+ },
1161
+ {
1162
+ "epoch": 1.96,
1163
+ "learning_rate": 6.086956521739131e-05,
1164
+ "loss": 0.0276,
1165
+ "step": 360
1166
+ },
1167
+ {
1168
+ "epoch": 1.97,
1169
+ "learning_rate": 6.0652173913043487e-05,
1170
+ "loss": 0.0377,
1171
+ "step": 362
1172
+ },
1173
+ {
1174
+ "epoch": 1.98,
1175
+ "learning_rate": 6.0434782608695654e-05,
1176
+ "loss": 0.0331,
1177
+ "step": 364
1178
+ },
1179
+ {
1180
+ "epoch": 1.99,
1181
+ "learning_rate": 6.021739130434783e-05,
1182
+ "loss": 0.0357,
1183
+ "step": 366
1184
+ },
1185
+ {
1186
+ "epoch": 2.0,
1187
+ "learning_rate": 6e-05,
1188
+ "loss": 0.032,
1189
+ "step": 368
1190
+ },
1191
+ {
1192
+ "epoch": 2.01,
1193
+ "learning_rate": 5.9782608695652175e-05,
1194
+ "loss": 0.0243,
1195
+ "step": 370
1196
+ },
1197
+ {
1198
+ "epoch": 2.02,
1199
+ "learning_rate": 5.956521739130435e-05,
1200
+ "loss": 0.0259,
1201
+ "step": 372
1202
+ },
1203
+ {
1204
+ "epoch": 2.03,
1205
+ "learning_rate": 5.934782608695652e-05,
1206
+ "loss": 0.0396,
1207
+ "step": 374
1208
+ },
1209
+ {
1210
+ "epoch": 2.04,
1211
+ "learning_rate": 5.9130434782608704e-05,
1212
+ "loss": 0.0308,
1213
+ "step": 376
1214
+ },
1215
+ {
1216
+ "epoch": 2.05,
1217
+ "learning_rate": 5.891304347826088e-05,
1218
+ "loss": 0.0324,
1219
+ "step": 378
1220
+ },
1221
+ {
1222
+ "epoch": 2.07,
1223
+ "learning_rate": 5.869565217391305e-05,
1224
+ "loss": 0.034,
1225
+ "step": 380
1226
+ },
1227
+ {
1228
+ "epoch": 2.08,
1229
+ "learning_rate": 5.847826086956521e-05,
1230
+ "loss": 0.0316,
1231
+ "step": 382
1232
+ },
1233
+ {
1234
+ "epoch": 2.09,
1235
+ "learning_rate": 5.826086956521739e-05,
1236
+ "loss": 0.0307,
1237
+ "step": 384
1238
+ },
1239
+ {
1240
+ "epoch": 2.1,
1241
+ "learning_rate": 5.804347826086957e-05,
1242
+ "loss": 0.028,
1243
+ "step": 386
1244
+ },
1245
+ {
1246
+ "epoch": 2.11,
1247
+ "learning_rate": 5.782608695652174e-05,
1248
+ "loss": 0.0347,
1249
+ "step": 388
1250
+ },
1251
+ {
1252
+ "epoch": 2.12,
1253
+ "learning_rate": 5.7608695652173915e-05,
1254
+ "loss": 0.0344,
1255
+ "step": 390
1256
+ },
1257
+ {
1258
+ "epoch": 2.13,
1259
+ "learning_rate": 5.739130434782609e-05,
1260
+ "loss": 0.0318,
1261
+ "step": 392
1262
+ },
1263
+ {
1264
+ "epoch": 2.14,
1265
+ "learning_rate": 5.717391304347827e-05,
1266
+ "loss": 0.0309,
1267
+ "step": 394
1268
+ },
1269
+ {
1270
+ "epoch": 2.15,
1271
+ "learning_rate": 5.695652173913044e-05,
1272
+ "loss": 0.0254,
1273
+ "step": 396
1274
+ },
1275
+ {
1276
+ "epoch": 2.16,
1277
+ "learning_rate": 5.673913043478262e-05,
1278
+ "loss": 0.0307,
1279
+ "step": 398
1280
+ },
1281
+ {
1282
+ "epoch": 2.17,
1283
+ "learning_rate": 5.652173913043478e-05,
1284
+ "loss": 0.0306,
1285
+ "step": 400
1286
+ },
1287
+ {
1288
+ "epoch": 2.17,
1289
+ "eval_f1": 0.9192225608721214,
1290
+ "eval_loss": 0.026507139205932617,
1291
+ "eval_precision": 0.9182849446143146,
1292
+ "eval_recall": 0.920172114237896,
1293
+ "eval_runtime": 13.382,
1294
+ "eval_samples_per_second": 138.694,
1295
+ "eval_steps_per_second": 0.598,
1296
+ "step": 400
1297
+ },
1298
+ {
1299
+ "epoch": 2.18,
1300
+ "learning_rate": 5.630434782608696e-05,
1301
+ "loss": 0.03,
1302
+ "step": 402
1303
+ },
1304
+ {
1305
+ "epoch": 2.2,
1306
+ "learning_rate": 5.608695652173913e-05,
1307
+ "loss": 0.03,
1308
+ "step": 404
1309
+ },
1310
+ {
1311
+ "epoch": 2.21,
1312
+ "learning_rate": 5.5869565217391306e-05,
1313
+ "loss": 0.0265,
1314
+ "step": 406
1315
+ },
1316
+ {
1317
+ "epoch": 2.22,
1318
+ "learning_rate": 5.565217391304348e-05,
1319
+ "loss": 0.0254,
1320
+ "step": 408
1321
+ },
1322
+ {
1323
+ "epoch": 2.23,
1324
+ "learning_rate": 5.5434782608695654e-05,
1325
+ "loss": 0.031,
1326
+ "step": 410
1327
+ },
1328
+ {
1329
+ "epoch": 2.24,
1330
+ "learning_rate": 5.5217391304347835e-05,
1331
+ "loss": 0.025,
1332
+ "step": 412
1333
+ },
1334
+ {
1335
+ "epoch": 2.25,
1336
+ "learning_rate": 5.500000000000001e-05,
1337
+ "loss": 0.0278,
1338
+ "step": 414
1339
+ },
1340
+ {
1341
+ "epoch": 2.26,
1342
+ "learning_rate": 5.478260869565217e-05,
1343
+ "loss": 0.0311,
1344
+ "step": 416
1345
+ },
1346
+ {
1347
+ "epoch": 2.27,
1348
+ "learning_rate": 5.456521739130434e-05,
1349
+ "loss": 0.0294,
1350
+ "step": 418
1351
+ },
1352
+ {
1353
+ "epoch": 2.28,
1354
+ "learning_rate": 5.4347826086956524e-05,
1355
+ "loss": 0.0324,
1356
+ "step": 420
1357
+ },
1358
+ {
1359
+ "epoch": 2.29,
1360
+ "learning_rate": 5.41304347826087e-05,
1361
+ "loss": 0.025,
1362
+ "step": 422
1363
+ },
1364
+ {
1365
+ "epoch": 2.3,
1366
+ "learning_rate": 5.391304347826087e-05,
1367
+ "loss": 0.0372,
1368
+ "step": 424
1369
+ },
1370
+ {
1371
+ "epoch": 2.32,
1372
+ "learning_rate": 5.3695652173913046e-05,
1373
+ "loss": 0.0259,
1374
+ "step": 426
1375
+ },
1376
+ {
1377
+ "epoch": 2.33,
1378
+ "learning_rate": 5.347826086956522e-05,
1379
+ "loss": 0.0371,
1380
+ "step": 428
1381
+ },
1382
+ {
1383
+ "epoch": 2.34,
1384
+ "learning_rate": 5.32608695652174e-05,
1385
+ "loss": 0.0273,
1386
+ "step": 430
1387
+ },
1388
+ {
1389
+ "epoch": 2.35,
1390
+ "learning_rate": 5.3043478260869574e-05,
1391
+ "loss": 0.036,
1392
+ "step": 432
1393
+ },
1394
+ {
1395
+ "epoch": 2.36,
1396
+ "learning_rate": 5.2826086956521735e-05,
1397
+ "loss": 0.0287,
1398
+ "step": 434
1399
+ },
1400
+ {
1401
+ "epoch": 2.37,
1402
+ "learning_rate": 5.260869565217391e-05,
1403
+ "loss": 0.0269,
1404
+ "step": 436
1405
+ },
1406
+ {
1407
+ "epoch": 2.38,
1408
+ "learning_rate": 5.239130434782609e-05,
1409
+ "loss": 0.0252,
1410
+ "step": 438
1411
+ },
1412
+ {
1413
+ "epoch": 2.39,
1414
+ "learning_rate": 5.217391304347826e-05,
1415
+ "loss": 0.029,
1416
+ "step": 440
1417
+ },
1418
+ {
1419
+ "epoch": 2.4,
1420
+ "learning_rate": 5.195652173913044e-05,
1421
+ "loss": 0.0306,
1422
+ "step": 442
1423
+ },
1424
+ {
1425
+ "epoch": 2.41,
1426
+ "learning_rate": 5.173913043478261e-05,
1427
+ "loss": 0.0277,
1428
+ "step": 444
1429
+ },
1430
+ {
1431
+ "epoch": 2.42,
1432
+ "learning_rate": 5.1521739130434785e-05,
1433
+ "loss": 0.0263,
1434
+ "step": 446
1435
+ },
1436
+ {
1437
+ "epoch": 2.43,
1438
+ "learning_rate": 5.1304347826086966e-05,
1439
+ "loss": 0.0271,
1440
+ "step": 448
1441
+ },
1442
+ {
1443
+ "epoch": 2.45,
1444
+ "learning_rate": 5.108695652173914e-05,
1445
+ "loss": 0.0293,
1446
+ "step": 450
1447
+ },
1448
+ {
1449
+ "epoch": 2.45,
1450
+ "eval_f1": 0.9251812311308268,
1451
+ "eval_loss": 0.024492362514138222,
1452
+ "eval_precision": 0.9251268214697702,
1453
+ "eval_recall": 0.9252456483686047,
1454
+ "eval_runtime": 13.5649,
1455
+ "eval_samples_per_second": 136.823,
1456
+ "eval_steps_per_second": 0.59,
1457
+ "step": 450
1458
+ },
1459
+ {
1460
+ "epoch": 2.46,
1461
+ "learning_rate": 5.08695652173913e-05,
1462
+ "loss": 0.0271,
1463
+ "step": 452
1464
+ },
1465
+ {
1466
+ "epoch": 2.47,
1467
+ "learning_rate": 5.0652173913043474e-05,
1468
+ "loss": 0.0246,
1469
+ "step": 454
1470
+ },
1471
+ {
1472
+ "epoch": 2.48,
1473
+ "learning_rate": 5.0434782608695655e-05,
1474
+ "loss": 0.0219,
1475
+ "step": 456
1476
+ },
1477
+ {
1478
+ "epoch": 2.49,
1479
+ "learning_rate": 5.021739130434783e-05,
1480
+ "loss": 0.0282,
1481
+ "step": 458
1482
+ },
1483
+ {
1484
+ "epoch": 2.5,
1485
+ "learning_rate": 5e-05,
1486
+ "loss": 0.0319,
1487
+ "step": 460
1488
+ },
1489
+ {
1490
+ "epoch": 2.51,
1491
+ "learning_rate": 4.9782608695652176e-05,
1492
+ "loss": 0.026,
1493
+ "step": 462
1494
+ },
1495
+ {
1496
+ "epoch": 2.52,
1497
+ "learning_rate": 4.956521739130435e-05,
1498
+ "loss": 0.0285,
1499
+ "step": 464
1500
+ },
1501
+ {
1502
+ "epoch": 2.53,
1503
+ "learning_rate": 4.9347826086956524e-05,
1504
+ "loss": 0.0299,
1505
+ "step": 466
1506
+ },
1507
+ {
1508
+ "epoch": 2.54,
1509
+ "learning_rate": 4.91304347826087e-05,
1510
+ "loss": 0.0311,
1511
+ "step": 468
1512
+ },
1513
+ {
1514
+ "epoch": 2.55,
1515
+ "learning_rate": 4.891304347826087e-05,
1516
+ "loss": 0.0283,
1517
+ "step": 470
1518
+ },
1519
+ {
1520
+ "epoch": 2.57,
1521
+ "learning_rate": 4.8695652173913046e-05,
1522
+ "loss": 0.0291,
1523
+ "step": 472
1524
+ },
1525
+ {
1526
+ "epoch": 2.58,
1527
+ "learning_rate": 4.847826086956522e-05,
1528
+ "loss": 0.0287,
1529
+ "step": 474
1530
+ },
1531
+ {
1532
+ "epoch": 2.59,
1533
+ "learning_rate": 4.8260869565217394e-05,
1534
+ "loss": 0.0262,
1535
+ "step": 476
1536
+ },
1537
+ {
1538
+ "epoch": 2.6,
1539
+ "learning_rate": 4.804347826086957e-05,
1540
+ "loss": 0.0308,
1541
+ "step": 478
1542
+ },
1543
+ {
1544
+ "epoch": 2.61,
1545
+ "learning_rate": 4.782608695652174e-05,
1546
+ "loss": 0.0252,
1547
+ "step": 480
1548
+ },
1549
+ {
1550
+ "epoch": 2.62,
1551
+ "learning_rate": 4.7608695652173916e-05,
1552
+ "loss": 0.029,
1553
+ "step": 482
1554
+ },
1555
+ {
1556
+ "epoch": 2.63,
1557
+ "learning_rate": 4.739130434782609e-05,
1558
+ "loss": 0.027,
1559
+ "step": 484
1560
+ },
1561
+ {
1562
+ "epoch": 2.64,
1563
+ "learning_rate": 4.7173913043478264e-05,
1564
+ "loss": 0.0253,
1565
+ "step": 486
1566
+ },
1567
+ {
1568
+ "epoch": 2.65,
1569
+ "learning_rate": 4.695652173913044e-05,
1570
+ "loss": 0.0271,
1571
+ "step": 488
1572
+ },
1573
+ {
1574
+ "epoch": 2.66,
1575
+ "learning_rate": 4.673913043478261e-05,
1576
+ "loss": 0.0264,
1577
+ "step": 490
1578
+ },
1579
+ {
1580
+ "epoch": 2.67,
1581
+ "learning_rate": 4.6521739130434785e-05,
1582
+ "loss": 0.0276,
1583
+ "step": 492
1584
+ },
1585
+ {
1586
+ "epoch": 2.68,
1587
+ "learning_rate": 4.630434782608696e-05,
1588
+ "loss": 0.0277,
1589
+ "step": 494
1590
+ },
1591
+ {
1592
+ "epoch": 2.7,
1593
+ "learning_rate": 4.608695652173913e-05,
1594
+ "loss": 0.0259,
1595
+ "step": 496
1596
+ },
1597
+ {
1598
+ "epoch": 2.71,
1599
+ "learning_rate": 4.586956521739131e-05,
1600
+ "loss": 0.0314,
1601
+ "step": 498
1602
+ },
1603
+ {
1604
+ "epoch": 2.72,
1605
+ "learning_rate": 4.565217391304348e-05,
1606
+ "loss": 0.0271,
1607
+ "step": 500
1608
+ },
1609
+ {
1610
+ "epoch": 2.72,
1611
+ "eval_f1": 0.9348680681377384,
1612
+ "eval_loss": 0.02203366905450821,
1613
+ "eval_precision": 0.9391445230465687,
1614
+ "eval_recall": 0.9306402922544215,
1615
+ "eval_runtime": 13.4883,
1616
+ "eval_samples_per_second": 137.601,
1617
+ "eval_steps_per_second": 0.593,
1618
+ "step": 500
1619
+ },
1620
+ {
1621
+ "epoch": 2.73,
1622
+ "learning_rate": 4.5434782608695655e-05,
1623
+ "loss": 0.0282,
1624
+ "step": 502
1625
+ },
1626
+ {
1627
+ "epoch": 2.74,
1628
+ "learning_rate": 4.521739130434783e-05,
1629
+ "loss": 0.0275,
1630
+ "step": 504
1631
+ },
1632
+ {
1633
+ "epoch": 2.75,
1634
+ "learning_rate": 4.5e-05,
1635
+ "loss": 0.0211,
1636
+ "step": 506
1637
+ },
1638
+ {
1639
+ "epoch": 2.76,
1640
+ "learning_rate": 4.478260869565218e-05,
1641
+ "loss": 0.0287,
1642
+ "step": 508
1643
+ },
1644
+ {
1645
+ "epoch": 2.77,
1646
+ "learning_rate": 4.456521739130435e-05,
1647
+ "loss": 0.0215,
1648
+ "step": 510
1649
+ },
1650
+ {
1651
+ "epoch": 2.78,
1652
+ "learning_rate": 4.4347826086956525e-05,
1653
+ "loss": 0.0257,
1654
+ "step": 512
1655
+ },
1656
+ {
1657
+ "epoch": 2.79,
1658
+ "learning_rate": 4.41304347826087e-05,
1659
+ "loss": 0.0333,
1660
+ "step": 514
1661
+ },
1662
+ {
1663
+ "epoch": 2.8,
1664
+ "learning_rate": 4.391304347826087e-05,
1665
+ "loss": 0.0285,
1666
+ "step": 516
1667
+ },
1668
+ {
1669
+ "epoch": 2.82,
1670
+ "learning_rate": 4.3695652173913046e-05,
1671
+ "loss": 0.0211,
1672
+ "step": 518
1673
+ },
1674
+ {
1675
+ "epoch": 2.83,
1676
+ "learning_rate": 4.347826086956522e-05,
1677
+ "loss": 0.0253,
1678
+ "step": 520
1679
+ },
1680
+ {
1681
+ "epoch": 2.84,
1682
+ "learning_rate": 4.3260869565217394e-05,
1683
+ "loss": 0.0222,
1684
+ "step": 522
1685
+ },
1686
+ {
1687
+ "epoch": 2.85,
1688
+ "learning_rate": 4.304347826086957e-05,
1689
+ "loss": 0.0274,
1690
+ "step": 524
1691
+ },
1692
+ {
1693
+ "epoch": 2.86,
1694
+ "learning_rate": 4.282608695652174e-05,
1695
+ "loss": 0.0244,
1696
+ "step": 526
1697
+ },
1698
+ {
1699
+ "epoch": 2.87,
1700
+ "learning_rate": 4.2608695652173916e-05,
1701
+ "loss": 0.0313,
1702
+ "step": 528
1703
+ },
1704
+ {
1705
+ "epoch": 2.88,
1706
+ "learning_rate": 4.239130434782609e-05,
1707
+ "loss": 0.0236,
1708
+ "step": 530
1709
+ },
1710
+ {
1711
+ "epoch": 2.89,
1712
+ "learning_rate": 4.2173913043478264e-05,
1713
+ "loss": 0.0241,
1714
+ "step": 532
1715
+ },
1716
+ {
1717
+ "epoch": 2.9,
1718
+ "learning_rate": 4.195652173913044e-05,
1719
+ "loss": 0.0228,
1720
+ "step": 534
1721
+ },
1722
+ {
1723
+ "epoch": 2.91,
1724
+ "learning_rate": 4.1739130434782605e-05,
1725
+ "loss": 0.03,
1726
+ "step": 536
1727
+ },
1728
+ {
1729
+ "epoch": 2.92,
1730
+ "learning_rate": 4.1521739130434786e-05,
1731
+ "loss": 0.0232,
1732
+ "step": 538
1733
+ },
1734
+ {
1735
+ "epoch": 2.93,
1736
+ "learning_rate": 4.130434782608696e-05,
1737
+ "loss": 0.0242,
1738
+ "step": 540
1739
+ },
1740
+ {
1741
+ "epoch": 2.95,
1742
+ "learning_rate": 4.1086956521739134e-05,
1743
+ "loss": 0.0235,
1744
+ "step": 542
1745
+ },
1746
+ {
1747
+ "epoch": 2.96,
1748
+ "learning_rate": 4.086956521739131e-05,
1749
+ "loss": 0.0235,
1750
+ "step": 544
1751
+ },
1752
+ {
1753
+ "epoch": 2.97,
1754
+ "learning_rate": 4.065217391304348e-05,
1755
+ "loss": 0.0262,
1756
+ "step": 546
1757
+ },
1758
+ {
1759
+ "epoch": 2.98,
1760
+ "learning_rate": 4.0434782608695655e-05,
1761
+ "loss": 0.0227,
1762
+ "step": 548
1763
+ },
1764
+ {
1765
+ "epoch": 2.99,
1766
+ "learning_rate": 4.021739130434783e-05,
1767
+ "loss": 0.0189,
1768
+ "step": 550
1769
+ },
1770
+ {
1771
+ "epoch": 2.99,
1772
+ "eval_f1": 0.9353923823099244,
1773
+ "eval_loss": 0.0210172887891531,
1774
+ "eval_precision": 0.9298178812552713,
1775
+ "eval_recall": 0.9410442483199254,
1776
+ "eval_runtime": 13.8124,
1777
+ "eval_samples_per_second": 134.372,
1778
+ "eval_steps_per_second": 0.579,
1779
+ "step": 550
1780
+ },
1781
+ {
1782
+ "epoch": 3.0,
1783
+ "learning_rate": 4e-05,
1784
+ "loss": 0.0232,
1785
+ "step": 552
1786
+ },
1787
+ {
1788
+ "epoch": 3.01,
1789
+ "learning_rate": 3.978260869565217e-05,
1790
+ "loss": 0.0251,
1791
+ "step": 554
1792
+ },
1793
+ {
1794
+ "epoch": 3.02,
1795
+ "learning_rate": 3.956521739130435e-05,
1796
+ "loss": 0.0249,
1797
+ "step": 556
1798
+ },
1799
+ {
1800
+ "epoch": 3.03,
1801
+ "learning_rate": 3.9347826086956525e-05,
1802
+ "loss": 0.0289,
1803
+ "step": 558
1804
+ },
1805
+ {
1806
+ "epoch": 3.04,
1807
+ "learning_rate": 3.91304347826087e-05,
1808
+ "loss": 0.0201,
1809
+ "step": 560
1810
+ },
1811
+ {
1812
+ "epoch": 3.05,
1813
+ "learning_rate": 3.8913043478260866e-05,
1814
+ "loss": 0.023,
1815
+ "step": 562
1816
+ },
1817
+ {
1818
+ "epoch": 3.07,
1819
+ "learning_rate": 3.869565217391305e-05,
1820
+ "loss": 0.0241,
1821
+ "step": 564
1822
+ },
1823
+ {
1824
+ "epoch": 3.08,
1825
+ "learning_rate": 3.847826086956522e-05,
1826
+ "loss": 0.0266,
1827
+ "step": 566
1828
+ },
1829
+ {
1830
+ "epoch": 3.09,
1831
+ "learning_rate": 3.8260869565217395e-05,
1832
+ "loss": 0.0212,
1833
+ "step": 568
1834
+ },
1835
+ {
1836
+ "epoch": 3.1,
1837
+ "learning_rate": 3.804347826086957e-05,
1838
+ "loss": 0.024,
1839
+ "step": 570
1840
+ },
1841
+ {
1842
+ "epoch": 3.11,
1843
+ "learning_rate": 3.7826086956521736e-05,
1844
+ "loss": 0.0214,
1845
+ "step": 572
1846
+ },
1847
+ {
1848
+ "epoch": 3.12,
1849
+ "learning_rate": 3.7608695652173917e-05,
1850
+ "loss": 0.0294,
1851
+ "step": 574
1852
+ },
1853
+ {
1854
+ "epoch": 3.13,
1855
+ "learning_rate": 3.739130434782609e-05,
1856
+ "loss": 0.0233,
1857
+ "step": 576
1858
+ },
1859
+ {
1860
+ "epoch": 3.14,
1861
+ "learning_rate": 3.7173913043478264e-05,
1862
+ "loss": 0.0268,
1863
+ "step": 578
1864
+ },
1865
+ {
1866
+ "epoch": 3.15,
1867
+ "learning_rate": 3.695652173913043e-05,
1868
+ "loss": 0.0189,
1869
+ "step": 580
1870
+ },
1871
+ {
1872
+ "epoch": 3.16,
1873
+ "learning_rate": 3.673913043478261e-05,
1874
+ "loss": 0.0217,
1875
+ "step": 582
1876
+ },
1877
+ {
1878
+ "epoch": 3.17,
1879
+ "learning_rate": 3.6521739130434786e-05,
1880
+ "loss": 0.0242,
1881
+ "step": 584
1882
+ },
1883
+ {
1884
+ "epoch": 3.18,
1885
+ "learning_rate": 3.630434782608696e-05,
1886
+ "loss": 0.0234,
1887
+ "step": 586
1888
+ },
1889
+ {
1890
+ "epoch": 3.2,
1891
+ "learning_rate": 3.6086956521739134e-05,
1892
+ "loss": 0.0209,
1893
+ "step": 588
1894
+ },
1895
+ {
1896
+ "epoch": 3.21,
1897
+ "learning_rate": 3.58695652173913e-05,
1898
+ "loss": 0.0257,
1899
+ "step": 590
1900
+ },
1901
+ {
1902
+ "epoch": 3.22,
1903
+ "learning_rate": 3.565217391304348e-05,
1904
+ "loss": 0.0243,
1905
+ "step": 592
1906
+ },
1907
+ {
1908
+ "epoch": 3.23,
1909
+ "learning_rate": 3.5434782608695656e-05,
1910
+ "loss": 0.0268,
1911
+ "step": 594
1912
+ },
1913
+ {
1914
+ "epoch": 3.24,
1915
+ "learning_rate": 3.521739130434783e-05,
1916
+ "loss": 0.0228,
1917
+ "step": 596
1918
+ },
1919
+ {
1920
+ "epoch": 3.25,
1921
+ "learning_rate": 3.5e-05,
1922
+ "loss": 0.0201,
1923
+ "step": 598
1924
+ },
1925
+ {
1926
+ "epoch": 3.26,
1927
+ "learning_rate": 3.478260869565218e-05,
1928
+ "loss": 0.0224,
1929
+ "step": 600
1930
+ },
1931
+ {
1932
+ "epoch": 3.26,
1933
+ "eval_f1": 0.9388881585680969,
1934
+ "eval_loss": 0.019923681393265724,
1935
+ "eval_precision": 0.9364339098342593,
1936
+ "eval_recall": 0.9413653580750335,
1937
+ "eval_runtime": 13.3913,
1938
+ "eval_samples_per_second": 138.598,
1939
+ "eval_steps_per_second": 0.597,
1940
+ "step": 600
1941
+ },
1942
+ {
1943
+ "epoch": 3.27,
1944
+ "learning_rate": 3.456521739130435e-05,
1945
+ "loss": 0.0245,
1946
+ "step": 602
1947
+ },
1948
+ {
1949
+ "epoch": 3.28,
1950
+ "learning_rate": 3.4347826086956526e-05,
1951
+ "loss": 0.0213,
1952
+ "step": 604
1953
+ },
1954
+ {
1955
+ "epoch": 3.29,
1956
+ "learning_rate": 3.413043478260869e-05,
1957
+ "loss": 0.0236,
1958
+ "step": 606
1959
+ },
1960
+ {
1961
+ "epoch": 3.3,
1962
+ "learning_rate": 3.3913043478260867e-05,
1963
+ "loss": 0.0189,
1964
+ "step": 608
1965
+ },
1966
+ {
1967
+ "epoch": 3.32,
1968
+ "learning_rate": 3.369565217391305e-05,
1969
+ "loss": 0.0242,
1970
+ "step": 610
1971
+ },
1972
+ {
1973
+ "epoch": 3.33,
1974
+ "learning_rate": 3.347826086956522e-05,
1975
+ "loss": 0.025,
1976
+ "step": 612
1977
+ },
1978
+ {
1979
+ "epoch": 3.34,
1980
+ "learning_rate": 3.3260869565217395e-05,
1981
+ "loss": 0.0255,
1982
+ "step": 614
1983
+ },
1984
+ {
1985
+ "epoch": 3.35,
1986
+ "learning_rate": 3.304347826086956e-05,
1987
+ "loss": 0.0203,
1988
+ "step": 616
1989
+ },
1990
+ {
1991
+ "epoch": 3.36,
1992
+ "learning_rate": 3.282608695652174e-05,
1993
+ "loss": 0.0263,
1994
+ "step": 618
1995
+ },
1996
+ {
1997
+ "epoch": 3.37,
1998
+ "learning_rate": 3.260869565217392e-05,
1999
+ "loss": 0.0227,
2000
+ "step": 620
2001
+ },
2002
+ {
2003
+ "epoch": 3.38,
2004
+ "learning_rate": 3.239130434782609e-05,
2005
+ "loss": 0.0225,
2006
+ "step": 622
2007
+ },
2008
+ {
2009
+ "epoch": 3.39,
2010
+ "learning_rate": 3.217391304347826e-05,
2011
+ "loss": 0.0172,
2012
+ "step": 624
2013
+ },
2014
+ {
2015
+ "epoch": 3.4,
2016
+ "learning_rate": 3.195652173913043e-05,
2017
+ "loss": 0.0176,
2018
+ "step": 626
2019
+ },
2020
+ {
2021
+ "epoch": 3.41,
2022
+ "learning_rate": 3.173913043478261e-05,
2023
+ "loss": 0.0202,
2024
+ "step": 628
2025
+ },
2026
+ {
2027
+ "epoch": 3.42,
2028
+ "learning_rate": 3.152173913043479e-05,
2029
+ "loss": 0.0223,
2030
+ "step": 630
2031
+ },
2032
+ {
2033
+ "epoch": 3.43,
2034
+ "learning_rate": 3.130434782608696e-05,
2035
+ "loss": 0.0248,
2036
+ "step": 632
2037
+ },
2038
+ {
2039
+ "epoch": 3.45,
2040
+ "learning_rate": 3.108695652173913e-05,
2041
+ "loss": 0.0213,
2042
+ "step": 634
2043
+ },
2044
+ {
2045
+ "epoch": 3.46,
2046
+ "learning_rate": 3.086956521739131e-05,
2047
+ "loss": 0.0189,
2048
+ "step": 636
2049
+ },
2050
+ {
2051
+ "epoch": 3.47,
2052
+ "learning_rate": 3.065217391304348e-05,
2053
+ "loss": 0.0221,
2054
+ "step": 638
2055
+ },
2056
+ {
2057
+ "epoch": 3.48,
2058
+ "learning_rate": 3.0434782608695656e-05,
2059
+ "loss": 0.0223,
2060
+ "step": 640
2061
+ },
2062
+ {
2063
+ "epoch": 3.49,
2064
+ "learning_rate": 3.0217391304347827e-05,
2065
+ "loss": 0.0242,
2066
+ "step": 642
2067
+ },
2068
+ {
2069
+ "epoch": 3.5,
2070
+ "learning_rate": 3e-05,
2071
+ "loss": 0.0228,
2072
+ "step": 644
2073
+ },
2074
+ {
2075
+ "epoch": 3.51,
2076
+ "learning_rate": 2.9782608695652175e-05,
2077
+ "loss": 0.0217,
2078
+ "step": 646
2079
+ },
2080
+ {
2081
+ "epoch": 3.52,
2082
+ "learning_rate": 2.9565217391304352e-05,
2083
+ "loss": 0.021,
2084
+ "step": 648
2085
+ },
2086
+ {
2087
+ "epoch": 3.53,
2088
+ "learning_rate": 2.9347826086956526e-05,
2089
+ "loss": 0.0247,
2090
+ "step": 650
2091
+ },
2092
+ {
2093
+ "epoch": 3.53,
2094
+ "eval_f1": 0.9423872456441195,
2095
+ "eval_loss": 0.018970053642988205,
2096
+ "eval_precision": 0.9418783673711327,
2097
+ "eval_recall": 0.9429066848995525,
2098
+ "eval_runtime": 14.0441,
2099
+ "eval_samples_per_second": 132.155,
2100
+ "eval_steps_per_second": 0.57,
2101
+ "step": 650
2102
+ },
2103
+ {
2104
+ "epoch": 3.54,
2105
+ "learning_rate": 2.9130434782608696e-05,
2106
+ "loss": 0.0254,
2107
+ "step": 652
2108
+ },
2109
+ {
2110
+ "epoch": 3.55,
2111
+ "learning_rate": 2.891304347826087e-05,
2112
+ "loss": 0.0215,
2113
+ "step": 654
2114
+ },
2115
+ {
2116
+ "epoch": 3.57,
2117
+ "learning_rate": 2.8695652173913044e-05,
2118
+ "loss": 0.0264,
2119
+ "step": 656
2120
+ },
2121
+ {
2122
+ "epoch": 3.58,
2123
+ "learning_rate": 2.847826086956522e-05,
2124
+ "loss": 0.0239,
2125
+ "step": 658
2126
+ },
2127
+ {
2128
+ "epoch": 3.59,
2129
+ "learning_rate": 2.826086956521739e-05,
2130
+ "loss": 0.018,
2131
+ "step": 660
2132
+ },
2133
+ {
2134
+ "epoch": 3.6,
2135
+ "learning_rate": 2.8043478260869566e-05,
2136
+ "loss": 0.0156,
2137
+ "step": 662
2138
+ },
2139
+ {
2140
+ "epoch": 3.61,
2141
+ "learning_rate": 2.782608695652174e-05,
2142
+ "loss": 0.0206,
2143
+ "step": 664
2144
+ },
2145
+ {
2146
+ "epoch": 3.62,
2147
+ "learning_rate": 2.7608695652173917e-05,
2148
+ "loss": 0.0208,
2149
+ "step": 666
2150
+ },
2151
+ {
2152
+ "epoch": 3.63,
2153
+ "learning_rate": 2.7391304347826085e-05,
2154
+ "loss": 0.0192,
2155
+ "step": 668
2156
+ },
2157
+ {
2158
+ "epoch": 3.64,
2159
+ "learning_rate": 2.7173913043478262e-05,
2160
+ "loss": 0.023,
2161
+ "step": 670
2162
+ },
2163
+ {
2164
+ "epoch": 3.65,
2165
+ "learning_rate": 2.6956521739130436e-05,
2166
+ "loss": 0.0203,
2167
+ "step": 672
2168
+ },
2169
+ {
2170
+ "epoch": 3.66,
2171
+ "learning_rate": 2.673913043478261e-05,
2172
+ "loss": 0.0217,
2173
+ "step": 674
2174
+ },
2175
+ {
2176
+ "epoch": 3.67,
2177
+ "learning_rate": 2.6521739130434787e-05,
2178
+ "loss": 0.0216,
2179
+ "step": 676
2180
+ },
2181
+ {
2182
+ "epoch": 3.68,
2183
+ "learning_rate": 2.6304347826086954e-05,
2184
+ "loss": 0.0182,
2185
+ "step": 678
2186
+ },
2187
+ {
2188
+ "epoch": 3.7,
2189
+ "learning_rate": 2.608695652173913e-05,
2190
+ "loss": 0.0231,
2191
+ "step": 680
2192
+ },
2193
+ {
2194
+ "epoch": 3.71,
2195
+ "learning_rate": 2.5869565217391305e-05,
2196
+ "loss": 0.0189,
2197
+ "step": 682
2198
+ },
2199
+ {
2200
+ "epoch": 3.72,
2201
+ "learning_rate": 2.5652173913043483e-05,
2202
+ "loss": 0.0215,
2203
+ "step": 684
2204
+ },
2205
+ {
2206
+ "epoch": 3.73,
2207
+ "learning_rate": 2.543478260869565e-05,
2208
+ "loss": 0.0257,
2209
+ "step": 686
2210
+ },
2211
+ {
2212
+ "epoch": 3.74,
2213
+ "learning_rate": 2.5217391304347827e-05,
2214
+ "loss": 0.0145,
2215
+ "step": 688
2216
+ },
2217
+ {
2218
+ "epoch": 3.75,
2219
+ "learning_rate": 2.5e-05,
2220
+ "loss": 0.0186,
2221
+ "step": 690
2222
+ },
2223
+ {
2224
+ "epoch": 3.76,
2225
+ "learning_rate": 2.4782608695652175e-05,
2226
+ "loss": 0.0174,
2227
+ "step": 692
2228
+ },
2229
+ {
2230
+ "epoch": 3.77,
2231
+ "learning_rate": 2.456521739130435e-05,
2232
+ "loss": 0.023,
2233
+ "step": 694
2234
+ },
2235
+ {
2236
+ "epoch": 3.78,
2237
+ "learning_rate": 2.4347826086956523e-05,
2238
+ "loss": 0.021,
2239
+ "step": 696
2240
+ },
2241
+ {
2242
+ "epoch": 3.79,
2243
+ "learning_rate": 2.4130434782608697e-05,
2244
+ "loss": 0.0202,
2245
+ "step": 698
2246
+ },
2247
+ {
2248
+ "epoch": 3.8,
2249
+ "learning_rate": 2.391304347826087e-05,
2250
+ "loss": 0.0184,
2251
+ "step": 700
2252
+ },
2253
+ {
2254
+ "epoch": 3.8,
2255
+ "eval_f1": 0.9427964190902829,
2256
+ "eval_loss": 0.01869898848235607,
2257
+ "eval_precision": 0.9387495218777859,
2258
+ "eval_recall": 0.9468884458628936,
2259
+ "eval_runtime": 13.5723,
2260
+ "eval_samples_per_second": 136.749,
2261
+ "eval_steps_per_second": 0.589,
2262
+ "step": 700
2263
+ },
2264
+ {
2265
+ "epoch": 3.82,
2266
+ "learning_rate": 2.3695652173913045e-05,
2267
+ "loss": 0.02,
2268
+ "step": 702
2269
+ },
2270
+ {
2271
+ "epoch": 3.83,
2272
+ "learning_rate": 2.347826086956522e-05,
2273
+ "loss": 0.0202,
2274
+ "step": 704
2275
+ },
2276
+ {
2277
+ "epoch": 3.84,
2278
+ "learning_rate": 2.3260869565217393e-05,
2279
+ "loss": 0.0207,
2280
+ "step": 706
2281
+ },
2282
+ {
2283
+ "epoch": 3.85,
2284
+ "learning_rate": 2.3043478260869567e-05,
2285
+ "loss": 0.0244,
2286
+ "step": 708
2287
+ },
2288
+ {
2289
+ "epoch": 3.86,
2290
+ "learning_rate": 2.282608695652174e-05,
2291
+ "loss": 0.0203,
2292
+ "step": 710
2293
+ },
2294
+ {
2295
+ "epoch": 3.87,
2296
+ "learning_rate": 2.2608695652173914e-05,
2297
+ "loss": 0.0195,
2298
+ "step": 712
2299
+ },
2300
+ {
2301
+ "epoch": 3.88,
2302
+ "learning_rate": 2.239130434782609e-05,
2303
+ "loss": 0.0164,
2304
+ "step": 714
2305
+ },
2306
+ {
2307
+ "epoch": 3.89,
2308
+ "learning_rate": 2.2173913043478262e-05,
2309
+ "loss": 0.0236,
2310
+ "step": 716
2311
+ },
2312
+ {
2313
+ "epoch": 3.9,
2314
+ "learning_rate": 2.1956521739130436e-05,
2315
+ "loss": 0.0209,
2316
+ "step": 718
2317
+ },
2318
+ {
2319
+ "epoch": 3.91,
2320
+ "learning_rate": 2.173913043478261e-05,
2321
+ "loss": 0.0218,
2322
+ "step": 720
2323
+ },
2324
+ {
2325
+ "epoch": 3.92,
2326
+ "learning_rate": 2.1521739130434784e-05,
2327
+ "loss": 0.0259,
2328
+ "step": 722
2329
+ },
2330
+ {
2331
+ "epoch": 3.93,
2332
+ "learning_rate": 2.1304347826086958e-05,
2333
+ "loss": 0.0207,
2334
+ "step": 724
2335
+ },
2336
+ {
2337
+ "epoch": 3.95,
2338
+ "learning_rate": 2.1086956521739132e-05,
2339
+ "loss": 0.0188,
2340
+ "step": 726
2341
+ },
2342
+ {
2343
+ "epoch": 3.96,
2344
+ "learning_rate": 2.0869565217391303e-05,
2345
+ "loss": 0.0231,
2346
+ "step": 728
2347
+ },
2348
+ {
2349
+ "epoch": 3.97,
2350
+ "learning_rate": 2.065217391304348e-05,
2351
+ "loss": 0.0222,
2352
+ "step": 730
2353
+ },
2354
+ {
2355
+ "epoch": 3.98,
2356
+ "learning_rate": 2.0434782608695654e-05,
2357
+ "loss": 0.0191,
2358
+ "step": 732
2359
+ },
2360
+ {
2361
+ "epoch": 3.99,
2362
+ "learning_rate": 2.0217391304347828e-05,
2363
+ "loss": 0.0221,
2364
+ "step": 734
2365
+ },
2366
+ {
2367
+ "epoch": 4.0,
2368
+ "learning_rate": 2e-05,
2369
+ "loss": 0.0211,
2370
+ "step": 736
2371
+ },
2372
+ {
2373
+ "epoch": 4.01,
2374
+ "learning_rate": 1.9782608695652176e-05,
2375
+ "loss": 0.02,
2376
+ "step": 738
2377
+ },
2378
+ {
2379
+ "epoch": 4.02,
2380
+ "learning_rate": 1.956521739130435e-05,
2381
+ "loss": 0.0197,
2382
+ "step": 740
2383
+ },
2384
+ {
2385
+ "epoch": 4.03,
2386
+ "learning_rate": 1.9347826086956523e-05,
2387
+ "loss": 0.0208,
2388
+ "step": 742
2389
+ },
2390
+ {
2391
+ "epoch": 4.04,
2392
+ "learning_rate": 1.9130434782608697e-05,
2393
+ "loss": 0.0181,
2394
+ "step": 744
2395
+ },
2396
+ {
2397
+ "epoch": 4.05,
2398
+ "learning_rate": 1.8913043478260868e-05,
2399
+ "loss": 0.0223,
2400
+ "step": 746
2401
+ },
2402
+ {
2403
+ "epoch": 4.07,
2404
+ "learning_rate": 1.8695652173913045e-05,
2405
+ "loss": 0.0212,
2406
+ "step": 748
2407
+ },
2408
+ {
2409
+ "epoch": 4.08,
2410
+ "learning_rate": 1.8478260869565216e-05,
2411
+ "loss": 0.0183,
2412
+ "step": 750
2413
+ },
2414
+ {
2415
+ "epoch": 4.08,
2416
+ "eval_f1": 0.9462391459396587,
2417
+ "eval_loss": 0.017822517082095146,
2418
+ "eval_precision": 0.9451528154384874,
2419
+ "eval_recall": 0.9473379995200449,
2420
+ "eval_runtime": 13.4572,
2421
+ "eval_samples_per_second": 137.919,
2422
+ "eval_steps_per_second": 0.594,
2423
+ "step": 750
2424
+ },
2425
+ {
2426
+ "epoch": 4.09,
2427
+ "learning_rate": 1.8260869565217393e-05,
2428
+ "loss": 0.0193,
2429
+ "step": 752
2430
+ },
2431
+ {
2432
+ "epoch": 4.1,
2433
+ "learning_rate": 1.8043478260869567e-05,
2434
+ "loss": 0.0159,
2435
+ "step": 754
2436
+ },
2437
+ {
2438
+ "epoch": 4.11,
2439
+ "learning_rate": 1.782608695652174e-05,
2440
+ "loss": 0.0212,
2441
+ "step": 756
2442
+ },
2443
+ {
2444
+ "epoch": 4.12,
2445
+ "learning_rate": 1.7608695652173915e-05,
2446
+ "loss": 0.0192,
2447
+ "step": 758
2448
+ },
2449
+ {
2450
+ "epoch": 4.13,
2451
+ "learning_rate": 1.739130434782609e-05,
2452
+ "loss": 0.0223,
2453
+ "step": 760
2454
+ },
2455
+ {
2456
+ "epoch": 4.14,
2457
+ "learning_rate": 1.7173913043478263e-05,
2458
+ "loss": 0.0166,
2459
+ "step": 762
2460
+ },
2461
+ {
2462
+ "epoch": 4.15,
2463
+ "learning_rate": 1.6956521739130433e-05,
2464
+ "loss": 0.0188,
2465
+ "step": 764
2466
+ },
2467
+ {
2468
+ "epoch": 4.16,
2469
+ "learning_rate": 1.673913043478261e-05,
2470
+ "loss": 0.0182,
2471
+ "step": 766
2472
+ },
2473
+ {
2474
+ "epoch": 4.17,
2475
+ "learning_rate": 1.652173913043478e-05,
2476
+ "loss": 0.0193,
2477
+ "step": 768
2478
+ },
2479
+ {
2480
+ "epoch": 4.18,
2481
+ "learning_rate": 1.630434782608696e-05,
2482
+ "loss": 0.0177,
2483
+ "step": 770
2484
+ },
2485
+ {
2486
+ "epoch": 4.2,
2487
+ "learning_rate": 1.608695652173913e-05,
2488
+ "loss": 0.0176,
2489
+ "step": 772
2490
+ },
2491
+ {
2492
+ "epoch": 4.21,
2493
+ "learning_rate": 1.5869565217391306e-05,
2494
+ "loss": 0.0202,
2495
+ "step": 774
2496
+ },
2497
+ {
2498
+ "epoch": 4.22,
2499
+ "learning_rate": 1.565217391304348e-05,
2500
+ "loss": 0.0199,
2501
+ "step": 776
2502
+ },
2503
+ {
2504
+ "epoch": 4.23,
2505
+ "learning_rate": 1.5434782608695654e-05,
2506
+ "loss": 0.0204,
2507
+ "step": 778
2508
+ },
2509
+ {
2510
+ "epoch": 4.24,
2511
+ "learning_rate": 1.5217391304347828e-05,
2512
+ "loss": 0.0199,
2513
+ "step": 780
2514
+ },
2515
+ {
2516
+ "epoch": 4.25,
2517
+ "learning_rate": 1.5e-05,
2518
+ "loss": 0.0205,
2519
+ "step": 782
2520
+ },
2521
+ {
2522
+ "epoch": 4.26,
2523
+ "learning_rate": 1.4782608695652176e-05,
2524
+ "loss": 0.0215,
2525
+ "step": 784
2526
+ },
2527
+ {
2528
+ "epoch": 4.27,
2529
+ "learning_rate": 1.4565217391304348e-05,
2530
+ "loss": 0.0179,
2531
+ "step": 786
2532
+ },
2533
+ {
2534
+ "epoch": 4.28,
2535
+ "learning_rate": 1.4347826086956522e-05,
2536
+ "loss": 0.0181,
2537
+ "step": 788
2538
+ },
2539
+ {
2540
+ "epoch": 4.29,
2541
+ "learning_rate": 1.4130434782608694e-05,
2542
+ "loss": 0.0217,
2543
+ "step": 790
2544
+ },
2545
+ {
2546
+ "epoch": 4.3,
2547
+ "learning_rate": 1.391304347826087e-05,
2548
+ "loss": 0.0212,
2549
+ "step": 792
2550
+ },
2551
+ {
2552
+ "epoch": 4.32,
2553
+ "learning_rate": 1.3695652173913042e-05,
2554
+ "loss": 0.0208,
2555
+ "step": 794
2556
+ },
2557
+ {
2558
+ "epoch": 4.33,
2559
+ "learning_rate": 1.3478260869565218e-05,
2560
+ "loss": 0.0157,
2561
+ "step": 796
2562
+ },
2563
+ {
2564
+ "epoch": 4.34,
2565
+ "learning_rate": 1.3260869565217394e-05,
2566
+ "loss": 0.0161,
2567
+ "step": 798
2568
+ },
2569
+ {
2570
+ "epoch": 4.35,
2571
+ "learning_rate": 1.3043478260869566e-05,
2572
+ "loss": 0.0175,
2573
+ "step": 800
2574
+ },
2575
+ {
2576
+ "epoch": 4.35,
2577
+ "eval_f1": 0.9480086057353725,
2578
+ "eval_loss": 0.017221176996827126,
2579
+ "eval_precision": 0.9474055538789009,
2580
+ "eval_recall": 0.9486224385404776,
2581
+ "eval_runtime": 13.6694,
2582
+ "eval_samples_per_second": 135.777,
2583
+ "eval_steps_per_second": 0.585,
2584
+ "step": 800
2585
+ }
2586
+ ],
2587
+ "max_steps": 920,
2588
+ "num_train_epochs": 5,
2589
+ "total_flos": 4.01525989965824e+16,
2590
+ "trial_name": null,
2591
+ "trial_params": null
2592
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3c53b114ee3695618d4d8a5a6aaf0b4cd5b626d9a81d765a87663ad619c8ec86
3
+ size 4015