Hugging Face's logo

zylandy
/
mae-intent-classifier

Text Classification
Transformers
Safetensors
roberta
intent-classification
Model card Files
xet
 Community
zylandy commited on
Commit
968b95a
·
verified ·
1 Parent(s): 32297cb

Upload 33 files

Browse files
Files changed (33) hide show
  1. balanced_finetuned_model/.DS_Store +0 -0
  2. balanced_finetuned_model/checkpoint-1848/config.json +42 -0
  3. balanced_finetuned_model/checkpoint-1848/merges.txt +0 -0
  4. balanced_finetuned_model/checkpoint-1848/model.safetensors +3 -0
  5. balanced_finetuned_model/checkpoint-1848/optimizer.pt +3 -0
  6. balanced_finetuned_model/checkpoint-1848/rng_state.pth +3 -0
  7. balanced_finetuned_model/checkpoint-1848/scheduler.pt +3 -0
  8. balanced_finetuned_model/checkpoint-1848/special_tokens_map.json +51 -0
  9. balanced_finetuned_model/checkpoint-1848/tokenizer_config.json +56 -0
  10. balanced_finetuned_model/checkpoint-1848/trainer_state.json +1171 -0
  11. balanced_finetuned_model/checkpoint-1848/training_args.bin +3 -0
  12. balanced_finetuned_model/checkpoint-1848/vocab.json +0 -0
  13. balanced_finetuned_model/checkpoint-2310/config.json +42 -0
  14. balanced_finetuned_model/checkpoint-2310/merges.txt +0 -0
  15. balanced_finetuned_model/checkpoint-2310/model.safetensors +3 -0
  16. balanced_finetuned_model/checkpoint-2310/optimizer.pt +3 -0
  17. balanced_finetuned_model/checkpoint-2310/rng_state.pth +3 -0
  18. balanced_finetuned_model/checkpoint-2310/scheduler.pt +3 -0
  19. balanced_finetuned_model/checkpoint-2310/special_tokens_map.json +51 -0
  20. balanced_finetuned_model/checkpoint-2310/tokenizer_config.json +56 -0
  21. balanced_finetuned_model/checkpoint-2310/trainer_state.json +1465 -0
  22. balanced_finetuned_model/checkpoint-2310/training_args.bin +3 -0
  23. balanced_finetuned_model/checkpoint-2310/vocab.json +0 -0
  24. balanced_finetuned_model/config.json +42 -0
  25. balanced_finetuned_model/label_mapping.json +23 -0
  26. balanced_finetuned_model/merges.txt +0 -0
  27. balanced_finetuned_model/model.safetensors +3 -0
  28. balanced_finetuned_model/runs/Jul03_19-02-50_hayashis-MacBook-Pro.local/events.out.tfevents.1751583770.hayashis-MacBook-Pro.local.69542.0 +3 -0
  29. balanced_finetuned_model/runs/Jul03_19-09-19_hayashis-MacBook-Pro.local/events.out.tfevents.1751584159.hayashis-MacBook-Pro.local.69763.0 +3 -0
  30. balanced_finetuned_model/runs/Jul03_19-09-19_hayashis-MacBook-Pro.local/events.out.tfevents.1751588684.hayashis-MacBook-Pro.local.69763.1 +3 -0
  31. balanced_finetuned_model/special_tokens_map.json +51 -0
  32. balanced_finetuned_model/tokenizer_config.json +56 -0
  33. balanced_finetuned_model/vocab.json +0 -0
balanced_finetuned_model/.DS_Store ADDED
Binary file (6.15 kB). View file
 
balanced_finetuned_model/checkpoint-1848/config.json ADDED
@@ -0,0 +1,42 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "./checkpoint-18750",
3
+ "architectures": [
4
+ "RobertaForSequenceClassification"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.1,
7
+ "bos_token_id": 0,
8
+ "classifier_dropout": null,
9
+ "eos_token_id": 2,
10
+ "hidden_act": "gelu",
11
+ "hidden_dropout_prob": 0.1,
12
+ "hidden_size": 768,
13
+ "id2label": {
14
+ "0": "Exploration and Reflection",
15
+ "1": "Feedback and Support",
16
+ "2": "Goal Setting and Planning",
17
+ "3": "Problem Solving and Critical Thinking",
18
+ "4": "Understanding and Clarification"
19
+ },
20
+ "initializer_range": 0.02,
21
+ "intermediate_size": 3072,
22
+ "label2id": {
23
+ "Exploration and Reflection": 0,
24
+ "Feedback and Support": 1,
25
+ "Goal Setting and Planning": 2,
26
+ "Problem Solving and Critical Thinking": 3,
27
+ "Understanding and Clarification": 4
28
+ },
29
+ "layer_norm_eps": 1e-05,
30
+ "max_position_embeddings": 514,
31
+ "model_type": "roberta",
32
+ "num_attention_heads": 12,
33
+ "num_hidden_layers": 12,
34
+ "pad_token_id": 1,
35
+ "position_embedding_type": "absolute",
36
+ "problem_type": "single_label_classification",
37
+ "torch_dtype": "float32",
38
+ "transformers_version": "4.35.2",
39
+ "type_vocab_size": 1,
40
+ "use_cache": true,
41
+ "vocab_size": 50265
42
+ }
balanced_finetuned_model/checkpoint-1848/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
balanced_finetuned_model/checkpoint-1848/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:239de4f4efa9aeddc0ac83e902ac064dc9c883476297d3a3ce567df855b5a314
3
+ size 498622052
balanced_finetuned_model/checkpoint-1848/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:60364537e883aa971de6c81d0b1603689280881beb84da34fd85fb74b04fc1ea
3
+ size 997357893
balanced_finetuned_model/checkpoint-1848/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:13938901139a11f0d571585eb3909fb49a6aa0224cc1127f30d98c04b87a4bab
3
+ size 13553
balanced_finetuned_model/checkpoint-1848/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c53e8b668917f65f06efbe8160243f51112567235723e8262333b16642ba9fea
3
+ size 627
balanced_finetuned_model/checkpoint-1848/special_tokens_map.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": true,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "cls_token": {
10
+ "content": "<s>",
11
+ "lstrip": false,
12
+ "normalized": true,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "eos_token": {
17
+ "content": "</s>",
18
+ "lstrip": false,
19
+ "normalized": true,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ },
23
+ "mask_token": {
24
+ "content": "<mask>",
25
+ "lstrip": true,
26
+ "normalized": false,
27
+ "rstrip": false,
28
+ "single_word": false
29
+ },
30
+ "pad_token": {
31
+ "content": "<pad>",
32
+ "lstrip": false,
33
+ "normalized": true,
34
+ "rstrip": false,
35
+ "single_word": false
36
+ },
37
+ "sep_token": {
38
+ "content": "</s>",
39
+ "lstrip": false,
40
+ "normalized": true,
41
+ "rstrip": false,
42
+ "single_word": false
43
+ },
44
+ "unk_token": {
45
+ "content": "<unk>",
46
+ "lstrip": false,
47
+ "normalized": true,
48
+ "rstrip": false,
49
+ "single_word": false
50
+ }
51
+ }
balanced_finetuned_model/checkpoint-1848/tokenizer_config.json ADDED
@@ -0,0 +1,56 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "added_tokens_decoder": {
4
+ "0": {
5
+ "content": "<s>",
6
+ "lstrip": false,
7
+ "normalized": true,
8
+ "rstrip": false,
9
+ "single_word": false,
10
+ "special": true
11
+ },
12
+ "1": {
13
+ "content": "<pad>",
14
+ "lstrip": false,
15
+ "normalized": true,
16
+ "rstrip": false,
17
+ "single_word": false,
18
+ "special": true
19
+ },
20
+ "2": {
21
+ "content": "</s>",
22
+ "lstrip": false,
23
+ "normalized": true,
24
+ "rstrip": false,
25
+ "single_word": false,
26
+ "special": true
27
+ },
28
+ "3": {
29
+ "content": "<unk>",
30
+ "lstrip": false,
31
+ "normalized": true,
32
+ "rstrip": false,
33
+ "single_word": false,
34
+ "special": true
35
+ },
36
+ "50264": {
37
+ "content": "<mask>",
38
+ "lstrip": true,
39
+ "normalized": false,
40
+ "rstrip": false,
41
+ "single_word": false,
42
+ "special": true
43
+ }
44
+ },
45
+ "bos_token": "<s>",
46
+ "clean_up_tokenization_spaces": true,
47
+ "cls_token": "<s>",
48
+ "eos_token": "</s>",
49
+ "errors": "replace",
50
+ "mask_token": "<mask>",
51
+ "model_max_length": 512,
52
+ "pad_token": "<pad>",
53
+ "sep_token": "</s>",
54
+ "tokenizer_class": "RobertaTokenizer",
55
+ "unk_token": "<unk>"
56
+ }
balanced_finetuned_model/checkpoint-1848/trainer_state.json ADDED
@@ -0,0 +1,1171 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.8039960230056994,
3
+ "best_model_checkpoint": "./balanced_finetuned_model/checkpoint-1848",
4
+ "epoch": 4.0,
5
+ "eval_steps": 500,
6
+ "global_step": 1848,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.02,
13
+ "learning_rate": 4.329004329004329e-07,
14
+ "loss": 10.8262,
15
+ "step": 10
16
+ },
17
+ {
18
+ "epoch": 0.04,
19
+ "learning_rate": 8.658008658008658e-07,
20
+ "loss": 12.2906,
21
+ "step": 20
22
+ },
23
+ {
24
+ "epoch": 0.06,
25
+ "learning_rate": 1.2987012987012986e-06,
26
+ "loss": 11.018,
27
+ "step": 30
28
+ },
29
+ {
30
+ "epoch": 0.09,
31
+ "learning_rate": 1.7316017316017317e-06,
32
+ "loss": 10.4001,
33
+ "step": 40
34
+ },
35
+ {
36
+ "epoch": 0.11,
37
+ "learning_rate": 2.1645021645021648e-06,
38
+ "loss": 10.7695,
39
+ "step": 50
40
+ },
41
+ {
42
+ "epoch": 0.13,
43
+ "learning_rate": 2.597402597402597e-06,
44
+ "loss": 9.9439,
45
+ "step": 60
46
+ },
47
+ {
48
+ "epoch": 0.15,
49
+ "learning_rate": 3.0303030303030305e-06,
50
+ "loss": 10.3621,
51
+ "step": 70
52
+ },
53
+ {
54
+ "epoch": 0.17,
55
+ "learning_rate": 3.4632034632034634e-06,
56
+ "loss": 8.5374,
57
+ "step": 80
58
+ },
59
+ {
60
+ "epoch": 0.19,
61
+ "learning_rate": 3.896103896103897e-06,
62
+ "loss": 6.6053,
63
+ "step": 90
64
+ },
65
+ {
66
+ "epoch": 0.22,
67
+ "learning_rate": 4.3290043290043295e-06,
68
+ "loss": 3.9921,
69
+ "step": 100
70
+ },
71
+ {
72
+ "epoch": 0.24,
73
+ "learning_rate": 4.761904761904762e-06,
74
+ "loss": 3.2861,
75
+ "step": 110
76
+ },
77
+ {
78
+ "epoch": 0.26,
79
+ "learning_rate": 5.194805194805194e-06,
80
+ "loss": 2.7387,
81
+ "step": 120
82
+ },
83
+ {
84
+ "epoch": 0.28,
85
+ "learning_rate": 5.627705627705629e-06,
86
+ "loss": 1.5201,
87
+ "step": 130
88
+ },
89
+ {
90
+ "epoch": 0.3,
91
+ "learning_rate": 6.060606060606061e-06,
92
+ "loss": 2.2568,
93
+ "step": 140
94
+ },
95
+ {
96
+ "epoch": 0.32,
97
+ "learning_rate": 6.493506493506494e-06,
98
+ "loss": 1.7721,
99
+ "step": 150
100
+ },
101
+ {
102
+ "epoch": 0.35,
103
+ "learning_rate": 6.926406926406927e-06,
104
+ "loss": 1.9366,
105
+ "step": 160
106
+ },
107
+ {
108
+ "epoch": 0.37,
109
+ "learning_rate": 7.3593073593073596e-06,
110
+ "loss": 1.9072,
111
+ "step": 170
112
+ },
113
+ {
114
+ "epoch": 0.39,
115
+ "learning_rate": 7.792207792207793e-06,
116
+ "loss": 2.1488,
117
+ "step": 180
118
+ },
119
+ {
120
+ "epoch": 0.41,
121
+ "learning_rate": 8.225108225108225e-06,
122
+ "loss": 1.6105,
123
+ "step": 190
124
+ },
125
+ {
126
+ "epoch": 0.43,
127
+ "learning_rate": 8.658008658008659e-06,
128
+ "loss": 2.0,
129
+ "step": 200
130
+ },
131
+ {
132
+ "epoch": 0.45,
133
+ "learning_rate": 9.090909090909091e-06,
134
+ "loss": 1.572,
135
+ "step": 210
136
+ },
137
+ {
138
+ "epoch": 0.48,
139
+ "learning_rate": 9.523809523809525e-06,
140
+ "loss": 1.5928,
141
+ "step": 220
142
+ },
143
+ {
144
+ "epoch": 0.5,
145
+ "learning_rate": 9.956709956709958e-06,
146
+ "loss": 1.3095,
147
+ "step": 230
148
+ },
149
+ {
150
+ "epoch": 0.52,
151
+ "learning_rate": 9.956709956709958e-06,
152
+ "loss": 1.513,
153
+ "step": 240
154
+ },
155
+ {
156
+ "epoch": 0.54,
157
+ "learning_rate": 9.90860990860991e-06,
158
+ "loss": 1.1392,
159
+ "step": 250
160
+ },
161
+ {
162
+ "epoch": 0.56,
163
+ "learning_rate": 9.860509860509861e-06,
164
+ "loss": 1.2247,
165
+ "step": 260
166
+ },
167
+ {
168
+ "epoch": 0.58,
169
+ "learning_rate": 9.812409812409814e-06,
170
+ "loss": 0.8886,
171
+ "step": 270
172
+ },
173
+ {
174
+ "epoch": 0.61,
175
+ "learning_rate": 9.764309764309765e-06,
176
+ "loss": 0.9848,
177
+ "step": 280
178
+ },
179
+ {
180
+ "epoch": 0.63,
181
+ "learning_rate": 9.716209716209716e-06,
182
+ "loss": 1.1762,
183
+ "step": 290
184
+ },
185
+ {
186
+ "epoch": 0.65,
187
+ "learning_rate": 9.66810966810967e-06,
188
+ "loss": 1.1722,
189
+ "step": 300
190
+ },
191
+ {
192
+ "epoch": 0.67,
193
+ "learning_rate": 9.62000962000962e-06,
194
+ "loss": 1.0029,
195
+ "step": 310
196
+ },
197
+ {
198
+ "epoch": 0.69,
199
+ "learning_rate": 9.571909571909572e-06,
200
+ "loss": 0.9868,
201
+ "step": 320
202
+ },
203
+ {
204
+ "epoch": 0.71,
205
+ "learning_rate": 9.523809523809525e-06,
206
+ "loss": 0.9915,
207
+ "step": 330
208
+ },
209
+ {
210
+ "epoch": 0.74,
211
+ "learning_rate": 9.475709475709478e-06,
212
+ "loss": 0.8529,
213
+ "step": 340
214
+ },
215
+ {
216
+ "epoch": 0.76,
217
+ "learning_rate": 9.427609427609429e-06,
218
+ "loss": 0.9799,
219
+ "step": 350
220
+ },
221
+ {
222
+ "epoch": 0.78,
223
+ "learning_rate": 9.37950937950938e-06,
224
+ "loss": 1.1586,
225
+ "step": 360
226
+ },
227
+ {
228
+ "epoch": 0.8,
229
+ "learning_rate": 9.331409331409333e-06,
230
+ "loss": 1.0889,
231
+ "step": 370
232
+ },
233
+ {
234
+ "epoch": 0.82,
235
+ "learning_rate": 9.283309283309284e-06,
236
+ "loss": 1.2502,
237
+ "step": 380
238
+ },
239
+ {
240
+ "epoch": 0.84,
241
+ "learning_rate": 9.235209235209236e-06,
242
+ "loss": 0.9826,
243
+ "step": 390
244
+ },
245
+ {
246
+ "epoch": 0.87,
247
+ "learning_rate": 9.187109187109189e-06,
248
+ "loss": 0.9052,
249
+ "step": 400
250
+ },
251
+ {
252
+ "epoch": 0.89,
253
+ "learning_rate": 9.13900913900914e-06,
254
+ "loss": 0.9651,
255
+ "step": 410
256
+ },
257
+ {
258
+ "epoch": 0.91,
259
+ "learning_rate": 9.090909090909091e-06,
260
+ "loss": 0.8358,
261
+ "step": 420
262
+ },
263
+ {
264
+ "epoch": 0.93,
265
+ "learning_rate": 9.042809042809044e-06,
266
+ "loss": 1.2029,
267
+ "step": 430
268
+ },
269
+ {
270
+ "epoch": 0.95,
271
+ "learning_rate": 8.994708994708995e-06,
272
+ "loss": 0.9126,
273
+ "step": 440
274
+ },
275
+ {
276
+ "epoch": 0.97,
277
+ "learning_rate": 8.946608946608948e-06,
278
+ "loss": 0.9049,
279
+ "step": 450
280
+ },
281
+ {
282
+ "epoch": 1.0,
283
+ "learning_rate": 8.8985088985089e-06,
284
+ "loss": 0.8561,
285
+ "step": 460
286
+ },
287
+ {
288
+ "epoch": 1.0,
289
+ "eval_accuracy": 0.6951351351351351,
290
+ "eval_f1": 0.6941963163251571,
291
+ "eval_loss": 0.8276578187942505,
292
+ "eval_precision": 0.7097058096914333,
293
+ "eval_recall": 0.6951351351351351,
294
+ "eval_runtime": 41.1958,
295
+ "eval_samples_per_second": 22.454,
296
+ "eval_steps_per_second": 2.816,
297
+ "step": 462
298
+ },
299
+ {
300
+ "epoch": 1.02,
301
+ "learning_rate": 8.85040885040885e-06,
302
+ "loss": 0.9094,
303
+ "step": 470
304
+ },
305
+ {
306
+ "epoch": 1.04,
307
+ "learning_rate": 8.802308802308804e-06,
308
+ "loss": 0.7898,
309
+ "step": 480
310
+ },
311
+ {
312
+ "epoch": 1.06,
313
+ "learning_rate": 8.754208754208755e-06,
314
+ "loss": 0.6191,
315
+ "step": 490
316
+ },
317
+ {
318
+ "epoch": 1.08,
319
+ "learning_rate": 8.706108706108706e-06,
320
+ "loss": 0.9794,
321
+ "step": 500
322
+ },
323
+ {
324
+ "epoch": 1.1,
325
+ "learning_rate": 8.658008658008659e-06,
326
+ "loss": 0.7676,
327
+ "step": 510
328
+ },
329
+ {
330
+ "epoch": 1.13,
331
+ "learning_rate": 8.60990860990861e-06,
332
+ "loss": 0.7618,
333
+ "step": 520
334
+ },
335
+ {
336
+ "epoch": 1.15,
337
+ "learning_rate": 8.561808561808562e-06,
338
+ "loss": 0.7243,
339
+ "step": 530
340
+ },
341
+ {
342
+ "epoch": 1.17,
343
+ "learning_rate": 8.513708513708514e-06,
344
+ "loss": 1.0373,
345
+ "step": 540
346
+ },
347
+ {
348
+ "epoch": 1.19,
349
+ "learning_rate": 8.465608465608466e-06,
350
+ "loss": 0.8194,
351
+ "step": 550
352
+ },
353
+ {
354
+ "epoch": 1.21,
355
+ "learning_rate": 8.417508417508419e-06,
356
+ "loss": 0.8167,
357
+ "step": 560
358
+ },
359
+ {
360
+ "epoch": 1.23,
361
+ "learning_rate": 8.36940836940837e-06,
362
+ "loss": 0.6841,
363
+ "step": 570
364
+ },
365
+ {
366
+ "epoch": 1.26,
367
+ "learning_rate": 8.321308321308323e-06,
368
+ "loss": 0.5651,
369
+ "step": 580
370
+ },
371
+ {
372
+ "epoch": 1.28,
373
+ "learning_rate": 8.273208273208274e-06,
374
+ "loss": 0.9127,
375
+ "step": 590
376
+ },
377
+ {
378
+ "epoch": 1.3,
379
+ "learning_rate": 8.225108225108225e-06,
380
+ "loss": 0.8384,
381
+ "step": 600
382
+ },
383
+ {
384
+ "epoch": 1.32,
385
+ "learning_rate": 8.177008177008178e-06,
386
+ "loss": 0.6718,
387
+ "step": 610
388
+ },
389
+ {
390
+ "epoch": 1.34,
391
+ "learning_rate": 8.12890812890813e-06,
392
+ "loss": 0.5633,
393
+ "step": 620
394
+ },
395
+ {
396
+ "epoch": 1.36,
397
+ "learning_rate": 8.08080808080808e-06,
398
+ "loss": 0.6944,
399
+ "step": 630
400
+ },
401
+ {
402
+ "epoch": 1.39,
403
+ "learning_rate": 8.032708032708034e-06,
404
+ "loss": 0.4755,
405
+ "step": 640
406
+ },
407
+ {
408
+ "epoch": 1.41,
409
+ "learning_rate": 7.984607984607985e-06,
410
+ "loss": 0.9228,
411
+ "step": 650
412
+ },
413
+ {
414
+ "epoch": 1.43,
415
+ "learning_rate": 7.936507936507936e-06,
416
+ "loss": 0.5992,
417
+ "step": 660
418
+ },
419
+ {
420
+ "epoch": 1.45,
421
+ "learning_rate": 7.888407888407889e-06,
422
+ "loss": 0.7077,
423
+ "step": 670
424
+ },
425
+ {
426
+ "epoch": 1.47,
427
+ "learning_rate": 7.840307840307842e-06,
428
+ "loss": 0.7204,
429
+ "step": 680
430
+ },
431
+ {
432
+ "epoch": 1.49,
433
+ "learning_rate": 7.792207792207793e-06,
434
+ "loss": 0.7907,
435
+ "step": 690
436
+ },
437
+ {
438
+ "epoch": 1.52,
439
+ "learning_rate": 7.744107744107745e-06,
440
+ "loss": 0.531,
441
+ "step": 700
442
+ },
443
+ {
444
+ "epoch": 1.54,
445
+ "learning_rate": 7.696007696007697e-06,
446
+ "loss": 0.889,
447
+ "step": 710
448
+ },
449
+ {
450
+ "epoch": 1.56,
451
+ "learning_rate": 7.647907647907649e-06,
452
+ "loss": 0.6319,
453
+ "step": 720
454
+ },
455
+ {
456
+ "epoch": 1.58,
457
+ "learning_rate": 7.599807599807601e-06,
458
+ "loss": 0.5859,
459
+ "step": 730
460
+ },
461
+ {
462
+ "epoch": 1.6,
463
+ "learning_rate": 7.551707551707552e-06,
464
+ "loss": 0.6143,
465
+ "step": 740
466
+ },
467
+ {
468
+ "epoch": 1.62,
469
+ "learning_rate": 7.503607503607504e-06,
470
+ "loss": 0.5408,
471
+ "step": 750
472
+ },
473
+ {
474
+ "epoch": 1.65,
475
+ "learning_rate": 7.455507455507455e-06,
476
+ "loss": 0.8811,
477
+ "step": 760
478
+ },
479
+ {
480
+ "epoch": 1.67,
481
+ "learning_rate": 7.4074074074074075e-06,
482
+ "loss": 0.6971,
483
+ "step": 770
484
+ },
485
+ {
486
+ "epoch": 1.69,
487
+ "learning_rate": 7.3593073593073596e-06,
488
+ "loss": 0.4133,
489
+ "step": 780
490
+ },
491
+ {
492
+ "epoch": 1.71,
493
+ "learning_rate": 7.3112073112073125e-06,
494
+ "loss": 0.4327,
495
+ "step": 790
496
+ },
497
+ {
498
+ "epoch": 1.73,
499
+ "learning_rate": 7.263107263107264e-06,
500
+ "loss": 0.4936,
501
+ "step": 800
502
+ },
503
+ {
504
+ "epoch": 1.75,
505
+ "learning_rate": 7.215007215007216e-06,
506
+ "loss": 0.6713,
507
+ "step": 810
508
+ },
509
+ {
510
+ "epoch": 1.77,
511
+ "learning_rate": 7.166907166907168e-06,
512
+ "loss": 0.5208,
513
+ "step": 820
514
+ },
515
+ {
516
+ "epoch": 1.8,
517
+ "learning_rate": 7.118807118807119e-06,
518
+ "loss": 0.81,
519
+ "step": 830
520
+ },
521
+ {
522
+ "epoch": 1.82,
523
+ "learning_rate": 7.070707070707071e-06,
524
+ "loss": 0.5224,
525
+ "step": 840
526
+ },
527
+ {
528
+ "epoch": 1.84,
529
+ "learning_rate": 7.022607022607023e-06,
530
+ "loss": 0.7043,
531
+ "step": 850
532
+ },
533
+ {
534
+ "epoch": 1.86,
535
+ "learning_rate": 6.974506974506975e-06,
536
+ "loss": 0.7131,
537
+ "step": 860
538
+ },
539
+ {
540
+ "epoch": 1.88,
541
+ "learning_rate": 6.926406926406927e-06,
542
+ "loss": 0.8933,
543
+ "step": 870
544
+ },
545
+ {
546
+ "epoch": 1.9,
547
+ "learning_rate": 6.878306878306879e-06,
548
+ "loss": 0.7053,
549
+ "step": 880
550
+ },
551
+ {
552
+ "epoch": 1.93,
553
+ "learning_rate": 6.83020683020683e-06,
554
+ "loss": 0.6044,
555
+ "step": 890
556
+ },
557
+ {
558
+ "epoch": 1.95,
559
+ "learning_rate": 6.782106782106783e-06,
560
+ "loss": 0.5334,
561
+ "step": 900
562
+ },
563
+ {
564
+ "epoch": 1.97,
565
+ "learning_rate": 6.734006734006735e-06,
566
+ "loss": 0.5943,
567
+ "step": 910
568
+ },
569
+ {
570
+ "epoch": 1.99,
571
+ "learning_rate": 6.685906685906687e-06,
572
+ "loss": 0.4249,
573
+ "step": 920
574
+ },
575
+ {
576
+ "epoch": 2.0,
577
+ "eval_accuracy": 0.7881081081081082,
578
+ "eval_f1": 0.7883030241253077,
579
+ "eval_loss": 0.6265861392021179,
580
+ "eval_precision": 0.7896965590352171,
581
+ "eval_recall": 0.7881081081081082,
582
+ "eval_runtime": 42.8724,
583
+ "eval_samples_per_second": 21.576,
584
+ "eval_steps_per_second": 2.706,
585
+ "step": 924
586
+ },
587
+ {
588
+ "epoch": 2.01,
589
+ "learning_rate": 6.637806637806638e-06,
590
+ "loss": 0.527,
591
+ "step": 930
592
+ },
593
+ {
594
+ "epoch": 2.03,
595
+ "learning_rate": 6.5897065897065905e-06,
596
+ "loss": 0.3392,
597
+ "step": 940
598
+ },
599
+ {
600
+ "epoch": 2.06,
601
+ "learning_rate": 6.541606541606543e-06,
602
+ "loss": 0.6126,
603
+ "step": 950
604
+ },
605
+ {
606
+ "epoch": 2.08,
607
+ "learning_rate": 6.493506493506494e-06,
608
+ "loss": 0.4029,
609
+ "step": 960
610
+ },
611
+ {
612
+ "epoch": 2.1,
613
+ "learning_rate": 6.445406445406446e-06,
614
+ "loss": 0.5203,
615
+ "step": 970
616
+ },
617
+ {
618
+ "epoch": 2.12,
619
+ "learning_rate": 6.397306397306397e-06,
620
+ "loss": 0.49,
621
+ "step": 980
622
+ },
623
+ {
624
+ "epoch": 2.14,
625
+ "learning_rate": 6.349206349206349e-06,
626
+ "loss": 0.4803,
627
+ "step": 990
628
+ },
629
+ {
630
+ "epoch": 2.16,
631
+ "learning_rate": 6.301106301106301e-06,
632
+ "loss": 0.4338,
633
+ "step": 1000
634
+ },
635
+ {
636
+ "epoch": 2.19,
637
+ "learning_rate": 6.253006253006253e-06,
638
+ "loss": 0.3084,
639
+ "step": 1010
640
+ },
641
+ {
642
+ "epoch": 2.21,
643
+ "learning_rate": 6.2049062049062055e-06,
644
+ "loss": 0.3061,
645
+ "step": 1020
646
+ },
647
+ {
648
+ "epoch": 2.23,
649
+ "learning_rate": 6.156806156806158e-06,
650
+ "loss": 0.6207,
651
+ "step": 1030
652
+ },
653
+ {
654
+ "epoch": 2.25,
655
+ "learning_rate": 6.10870610870611e-06,
656
+ "loss": 0.3268,
657
+ "step": 1040
658
+ },
659
+ {
660
+ "epoch": 2.27,
661
+ "learning_rate": 6.060606060606061e-06,
662
+ "loss": 0.6113,
663
+ "step": 1050
664
+ },
665
+ {
666
+ "epoch": 2.29,
667
+ "learning_rate": 6.012506012506013e-06,
668
+ "loss": 0.3333,
669
+ "step": 1060
670
+ },
671
+ {
672
+ "epoch": 2.32,
673
+ "learning_rate": 5.964405964405965e-06,
674
+ "loss": 0.4616,
675
+ "step": 1070
676
+ },
677
+ {
678
+ "epoch": 2.34,
679
+ "learning_rate": 5.916305916305916e-06,
680
+ "loss": 0.7176,
681
+ "step": 1080
682
+ },
683
+ {
684
+ "epoch": 2.36,
685
+ "learning_rate": 5.8682058682058685e-06,
686
+ "loss": 0.6433,
687
+ "step": 1090
688
+ },
689
+ {
690
+ "epoch": 2.38,
691
+ "learning_rate": 5.820105820105821e-06,
692
+ "loss": 0.2657,
693
+ "step": 1100
694
+ },
695
+ {
696
+ "epoch": 2.4,
697
+ "learning_rate": 5.772005772005772e-06,
698
+ "loss": 0.7557,
699
+ "step": 1110
700
+ },
701
+ {
702
+ "epoch": 2.42,
703
+ "learning_rate": 5.723905723905724e-06,
704
+ "loss": 0.499,
705
+ "step": 1120
706
+ },
707
+ {
708
+ "epoch": 2.45,
709
+ "learning_rate": 5.675805675805677e-06,
710
+ "loss": 0.6084,
711
+ "step": 1130
712
+ },
713
+ {
714
+ "epoch": 2.47,
715
+ "learning_rate": 5.627705627705629e-06,
716
+ "loss": 0.4959,
717
+ "step": 1140
718
+ },
719
+ {
720
+ "epoch": 2.49,
721
+ "learning_rate": 5.57960557960558e-06,
722
+ "loss": 0.3092,
723
+ "step": 1150
724
+ },
725
+ {
726
+ "epoch": 2.51,
727
+ "learning_rate": 5.531505531505532e-06,
728
+ "loss": 0.4155,
729
+ "step": 1160
730
+ },
731
+ {
732
+ "epoch": 2.53,
733
+ "learning_rate": 5.4834054834054835e-06,
734
+ "loss": 0.448,
735
+ "step": 1170
736
+ },
737
+ {
738
+ "epoch": 2.55,
739
+ "learning_rate": 5.435305435305436e-06,
740
+ "loss": 0.4538,
741
+ "step": 1180
742
+ },
743
+ {
744
+ "epoch": 2.58,
745
+ "learning_rate": 5.387205387205388e-06,
746
+ "loss": 0.4018,
747
+ "step": 1190
748
+ },
749
+ {
750
+ "epoch": 2.6,
751
+ "learning_rate": 5.339105339105339e-06,
752
+ "loss": 0.6252,
753
+ "step": 1200
754
+ },
755
+ {
756
+ "epoch": 2.62,
757
+ "learning_rate": 5.291005291005291e-06,
758
+ "loss": 0.4832,
759
+ "step": 1210
760
+ },
761
+ {
762
+ "epoch": 2.64,
763
+ "learning_rate": 5.242905242905243e-06,
764
+ "loss": 0.3838,
765
+ "step": 1220
766
+ },
767
+ {
768
+ "epoch": 2.66,
769
+ "learning_rate": 5.194805194805194e-06,
770
+ "loss": 0.3117,
771
+ "step": 1230
772
+ },
773
+ {
774
+ "epoch": 2.68,
775
+ "learning_rate": 5.146705146705147e-06,
776
+ "loss": 0.3003,
777
+ "step": 1240
778
+ },
779
+ {
780
+ "epoch": 2.71,
781
+ "learning_rate": 5.0986050986050994e-06,
782
+ "loss": 0.2706,
783
+ "step": 1250
784
+ },
785
+ {
786
+ "epoch": 2.73,
787
+ "learning_rate": 5.0505050505050515e-06,
788
+ "loss": 0.7442,
789
+ "step": 1260
790
+ },
791
+ {
792
+ "epoch": 2.75,
793
+ "learning_rate": 5.002405002405003e-06,
794
+ "loss": 0.2976,
795
+ "step": 1270
796
+ },
797
+ {
798
+ "epoch": 2.77,
799
+ "learning_rate": 4.954304954304955e-06,
800
+ "loss": 0.3882,
801
+ "step": 1280
802
+ },
803
+ {
804
+ "epoch": 2.79,
805
+ "learning_rate": 4.906204906204907e-06,
806
+ "loss": 0.3143,
807
+ "step": 1290
808
+ },
809
+ {
810
+ "epoch": 2.81,
811
+ "learning_rate": 4.858104858104858e-06,
812
+ "loss": 0.8753,
813
+ "step": 1300
814
+ },
815
+ {
816
+ "epoch": 2.84,
817
+ "learning_rate": 4.81000481000481e-06,
818
+ "loss": 0.706,
819
+ "step": 1310
820
+ },
821
+ {
822
+ "epoch": 2.86,
823
+ "learning_rate": 4.761904761904762e-06,
824
+ "loss": 0.478,
825
+ "step": 1320
826
+ },
827
+ {
828
+ "epoch": 2.88,
829
+ "learning_rate": 4.7138047138047145e-06,
830
+ "loss": 0.6678,
831
+ "step": 1330
832
+ },
833
+ {
834
+ "epoch": 2.9,
835
+ "learning_rate": 4.6657046657046666e-06,
836
+ "loss": 0.3574,
837
+ "step": 1340
838
+ },
839
+ {
840
+ "epoch": 2.92,
841
+ "learning_rate": 4.617604617604618e-06,
842
+ "loss": 0.3732,
843
+ "step": 1350
844
+ },
845
+ {
846
+ "epoch": 2.94,
847
+ "learning_rate": 4.56950456950457e-06,
848
+ "loss": 0.2018,
849
+ "step": 1360
850
+ },
851
+ {
852
+ "epoch": 2.97,
853
+ "learning_rate": 4.521404521404522e-06,
854
+ "loss": 0.4718,
855
+ "step": 1370
856
+ },
857
+ {
858
+ "epoch": 2.99,
859
+ "learning_rate": 4.473304473304474e-06,
860
+ "loss": 0.404,
861
+ "step": 1380
862
+ },
863
+ {
864
+ "epoch": 3.0,
865
+ "eval_accuracy": 0.7967567567567567,
866
+ "eval_f1": 0.7969295115959946,
867
+ "eval_loss": 0.662196695804596,
868
+ "eval_precision": 0.8024330408412987,
869
+ "eval_recall": 0.7967567567567567,
870
+ "eval_runtime": 42.3812,
871
+ "eval_samples_per_second": 21.826,
872
+ "eval_steps_per_second": 2.737,
873
+ "step": 1386
874
+ },
875
+ {
876
+ "epoch": 3.01,
877
+ "learning_rate": 4.425204425204425e-06,
878
+ "loss": 0.4288,
879
+ "step": 1390
880
+ },
881
+ {
882
+ "epoch": 3.03,
883
+ "learning_rate": 4.377104377104377e-06,
884
+ "loss": 0.1295,
885
+ "step": 1400
886
+ },
887
+ {
888
+ "epoch": 3.05,
889
+ "learning_rate": 4.3290043290043295e-06,
890
+ "loss": 0.3256,
891
+ "step": 1410
892
+ },
893
+ {
894
+ "epoch": 3.07,
895
+ "learning_rate": 4.280904280904281e-06,
896
+ "loss": 0.4232,
897
+ "step": 1420
898
+ },
899
+ {
900
+ "epoch": 3.1,
901
+ "learning_rate": 4.232804232804233e-06,
902
+ "loss": 0.4204,
903
+ "step": 1430
904
+ },
905
+ {
906
+ "epoch": 3.12,
907
+ "learning_rate": 4.184704184704185e-06,
908
+ "loss": 0.1602,
909
+ "step": 1440
910
+ },
911
+ {
912
+ "epoch": 3.14,
913
+ "learning_rate": 4.136604136604137e-06,
914
+ "loss": 0.344,
915
+ "step": 1450
916
+ },
917
+ {
918
+ "epoch": 3.16,
919
+ "learning_rate": 4.088504088504089e-06,
920
+ "loss": 0.2476,
921
+ "step": 1460
922
+ },
923
+ {
924
+ "epoch": 3.18,
925
+ "learning_rate": 4.04040404040404e-06,
926
+ "loss": 0.2518,
927
+ "step": 1470
928
+ },
929
+ {
930
+ "epoch": 3.2,
931
+ "learning_rate": 3.9923039923039925e-06,
932
+ "loss": 0.4476,
933
+ "step": 1480
934
+ },
935
+ {
936
+ "epoch": 3.23,
937
+ "learning_rate": 3.9442039442039446e-06,
938
+ "loss": 0.4147,
939
+ "step": 1490
940
+ },
941
+ {
942
+ "epoch": 3.25,
943
+ "learning_rate": 3.896103896103897e-06,
944
+ "loss": 0.3457,
945
+ "step": 1500
946
+ },
947
+ {
948
+ "epoch": 3.27,
949
+ "learning_rate": 3.848003848003849e-06,
950
+ "loss": 0.4206,
951
+ "step": 1510
952
+ },
953
+ {
954
+ "epoch": 3.29,
955
+ "learning_rate": 3.7999037999038004e-06,
956
+ "loss": 0.2476,
957
+ "step": 1520
958
+ },
959
+ {
960
+ "epoch": 3.31,
961
+ "learning_rate": 3.751803751803752e-06,
962
+ "loss": 0.3427,
963
+ "step": 1530
964
+ },
965
+ {
966
+ "epoch": 3.33,
967
+ "learning_rate": 3.7037037037037037e-06,
968
+ "loss": 0.3846,
969
+ "step": 1540
970
+ },
971
+ {
972
+ "epoch": 3.35,
973
+ "learning_rate": 3.6556036556036563e-06,
974
+ "loss": 0.3671,
975
+ "step": 1550
976
+ },
977
+ {
978
+ "epoch": 3.38,
979
+ "learning_rate": 3.607503607503608e-06,
980
+ "loss": 0.3038,
981
+ "step": 1560
982
+ },
983
+ {
984
+ "epoch": 3.4,
985
+ "learning_rate": 3.5594035594035596e-06,
986
+ "loss": 0.307,
987
+ "step": 1570
988
+ },
989
+ {
990
+ "epoch": 3.42,
991
+ "learning_rate": 3.5113035113035117e-06,
992
+ "loss": 0.2364,
993
+ "step": 1580
994
+ },
995
+ {
996
+ "epoch": 3.44,
997
+ "learning_rate": 3.4632034632034634e-06,
998
+ "loss": 0.3357,
999
+ "step": 1590
1000
+ },
1001
+ {
1002
+ "epoch": 3.46,
1003
+ "learning_rate": 3.415103415103415e-06,
1004
+ "loss": 0.2825,
1005
+ "step": 1600
1006
+ },
1007
+ {
1008
+ "epoch": 3.48,
1009
+ "learning_rate": 3.3670033670033675e-06,
1010
+ "loss": 0.3363,
1011
+ "step": 1610
1012
+ },
1013
+ {
1014
+ "epoch": 3.51,
1015
+ "learning_rate": 3.318903318903319e-06,
1016
+ "loss": 0.469,
1017
+ "step": 1620
1018
+ },
1019
+ {
1020
+ "epoch": 3.53,
1021
+ "learning_rate": 3.2708032708032713e-06,
1022
+ "loss": 0.5258,
1023
+ "step": 1630
1024
+ },
1025
+ {
1026
+ "epoch": 3.55,
1027
+ "learning_rate": 3.222703222703223e-06,
1028
+ "loss": 0.2061,
1029
+ "step": 1640
1030
+ },
1031
+ {
1032
+ "epoch": 3.57,
1033
+ "learning_rate": 3.1746031746031746e-06,
1034
+ "loss": 0.1948,
1035
+ "step": 1650
1036
+ },
1037
+ {
1038
+ "epoch": 3.59,
1039
+ "learning_rate": 3.1265031265031263e-06,
1040
+ "loss": 0.4708,
1041
+ "step": 1660
1042
+ },
1043
+ {
1044
+ "epoch": 3.61,
1045
+ "learning_rate": 3.078403078403079e-06,
1046
+ "loss": 0.4367,
1047
+ "step": 1670
1048
+ },
1049
+ {
1050
+ "epoch": 3.64,
1051
+ "learning_rate": 3.0303030303030305e-06,
1052
+ "loss": 0.4284,
1053
+ "step": 1680
1054
+ },
1055
+ {
1056
+ "epoch": 3.66,
1057
+ "learning_rate": 2.9822029822029826e-06,
1058
+ "loss": 0.3583,
1059
+ "step": 1690
1060
+ },
1061
+ {
1062
+ "epoch": 3.68,
1063
+ "learning_rate": 2.9341029341029342e-06,
1064
+ "loss": 0.3829,
1065
+ "step": 1700
1066
+ },
1067
+ {
1068
+ "epoch": 3.7,
1069
+ "learning_rate": 2.886002886002886e-06,
1070
+ "loss": 0.3757,
1071
+ "step": 1710
1072
+ },
1073
+ {
1074
+ "epoch": 3.72,
1075
+ "learning_rate": 2.8379028379028384e-06,
1076
+ "loss": 0.2352,
1077
+ "step": 1720
1078
+ },
1079
+ {
1080
+ "epoch": 3.74,
1081
+ "learning_rate": 2.78980278980279e-06,
1082
+ "loss": 0.3246,
1083
+ "step": 1730
1084
+ },
1085
+ {
1086
+ "epoch": 3.77,
1087
+ "learning_rate": 2.7417027417027418e-06,
1088
+ "loss": 0.2555,
1089
+ "step": 1740
1090
+ },
1091
+ {
1092
+ "epoch": 3.79,
1093
+ "learning_rate": 2.693602693602694e-06,
1094
+ "loss": 0.1718,
1095
+ "step": 1750
1096
+ },
1097
+ {
1098
+ "epoch": 3.81,
1099
+ "learning_rate": 2.6455026455026455e-06,
1100
+ "loss": 0.3823,
1101
+ "step": 1760
1102
+ },
1103
+ {
1104
+ "epoch": 3.83,
1105
+ "learning_rate": 2.597402597402597e-06,
1106
+ "loss": 0.3836,
1107
+ "step": 1770
1108
+ },
1109
+ {
1110
+ "epoch": 3.85,
1111
+ "learning_rate": 2.5493025493025497e-06,
1112
+ "loss": 0.0989,
1113
+ "step": 1780
1114
+ },
1115
+ {
1116
+ "epoch": 3.87,
1117
+ "learning_rate": 2.5012025012025014e-06,
1118
+ "loss": 0.3505,
1119
+ "step": 1790
1120
+ },
1121
+ {
1122
+ "epoch": 3.9,
1123
+ "learning_rate": 2.4531024531024535e-06,
1124
+ "loss": 0.4152,
1125
+ "step": 1800
1126
+ },
1127
+ {
1128
+ "epoch": 3.92,
1129
+ "learning_rate": 2.405002405002405e-06,
1130
+ "loss": 0.1573,
1131
+ "step": 1810
1132
+ },
1133
+ {
1134
+ "epoch": 3.94,
1135
+ "learning_rate": 2.3569023569023572e-06,
1136
+ "loss": 0.4271,
1137
+ "step": 1820
1138
+ },
1139
+ {
1140
+ "epoch": 3.96,
1141
+ "learning_rate": 2.308802308802309e-06,
1142
+ "loss": 0.5782,
1143
+ "step": 1830
1144
+ },
1145
+ {
1146
+ "epoch": 3.98,
1147
+ "learning_rate": 2.260702260702261e-06,
1148
+ "loss": 0.4115,
1149
+ "step": 1840
1150
+ },
1151
+ {
1152
+ "epoch": 4.0,
1153
+ "eval_accuracy": 0.8043243243243243,
1154
+ "eval_f1": 0.8039960230056994,
1155
+ "eval_loss": 0.6978992819786072,
1156
+ "eval_precision": 0.8072324249099836,
1157
+ "eval_recall": 0.8043243243243243,
1158
+ "eval_runtime": 42.9375,
1159
+ "eval_samples_per_second": 21.543,
1160
+ "eval_steps_per_second": 2.702,
1161
+ "step": 1848
1162
+ }
1163
+ ],
1164
+ "logging_steps": 10,
1165
+ "max_steps": 2310,
1166
+ "num_train_epochs": 5,
1167
+ "save_steps": 500,
1168
+ "total_flos": 1679906559108240.0,
1169
+ "trial_name": null,
1170
+ "trial_params": null
1171
+ }
balanced_finetuned_model/checkpoint-1848/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2e8c11de7aff56066100df364f86da392520ed4522bc7c376ec702df94a5ace7
3
+ size 4155
balanced_finetuned_model/checkpoint-1848/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
balanced_finetuned_model/checkpoint-2310/config.json ADDED
@@ -0,0 +1,42 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "./checkpoint-18750",
3
+ "architectures": [
4
+ "RobertaForSequenceClassification"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.1,
7
+ "bos_token_id": 0,
8
+ "classifier_dropout": null,
9
+ "eos_token_id": 2,
10
+ "hidden_act": "gelu",
11
+ "hidden_dropout_prob": 0.1,
12
+ "hidden_size": 768,
13
+ "id2label": {
14
+ "0": "Exploration and Reflection",
15
+ "1": "Feedback and Support",
16
+ "2": "Goal Setting and Planning",
17
+ "3": "Problem Solving and Critical Thinking",
18
+ "4": "Understanding and Clarification"
19
+ },
20
+ "initializer_range": 0.02,
21
+ "intermediate_size": 3072,
22
+ "label2id": {
23
+ "Exploration and Reflection": 0,
24
+ "Feedback and Support": 1,
25
+ "Goal Setting and Planning": 2,
26
+ "Problem Solving and Critical Thinking": 3,
27
+ "Understanding and Clarification": 4
28
+ },
29
+ "layer_norm_eps": 1e-05,
30
+ "max_position_embeddings": 514,
31
+ "model_type": "roberta",
32
+ "num_attention_heads": 12,
33
+ "num_hidden_layers": 12,
34
+ "pad_token_id": 1,
35
+ "position_embedding_type": "absolute",
36
+ "problem_type": "single_label_classification",
37
+ "torch_dtype": "float32",
38
+ "transformers_version": "4.35.2",
39
+ "type_vocab_size": 1,
40
+ "use_cache": true,
41
+ "vocab_size": 50265
42
+ }
balanced_finetuned_model/checkpoint-2310/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
balanced_finetuned_model/checkpoint-2310/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4ec1dac41d174534829a2ac7f7456cd92ed4cba58731a0ba25409849f53e2437
3
+ size 498622052
balanced_finetuned_model/checkpoint-2310/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2abc7211189997815b080aea3d93df4b075b39bcec02c658692c8120fd848029
3
+ size 997357893
balanced_finetuned_model/checkpoint-2310/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a19faecd5e6f7b529b18bd8f38eba789c55a57d7b544e49e9c631b4d1da863de
3
+ size 13553
balanced_finetuned_model/checkpoint-2310/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6a4eb7d6e5e1a5182be250ef9b1d56953dcc1b26de7b6fbc3c972b4692756b1d
3
+ size 627
balanced_finetuned_model/checkpoint-2310/special_tokens_map.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": true,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "cls_token": {
10
+ "content": "<s>",
11
+ "lstrip": false,
12
+ "normalized": true,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "eos_token": {
17
+ "content": "</s>",
18
+ "lstrip": false,
19
+ "normalized": true,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ },
23
+ "mask_token": {
24
+ "content": "<mask>",
25
+ "lstrip": true,
26
+ "normalized": false,
27
+ "rstrip": false,
28
+ "single_word": false
29
+ },
30
+ "pad_token": {
31
+ "content": "<pad>",
32
+ "lstrip": false,
33
+ "normalized": true,
34
+ "rstrip": false,
35
+ "single_word": false
36
+ },
37
+ "sep_token": {
38
+ "content": "</s>",
39
+ "lstrip": false,
40
+ "normalized": true,
41
+ "rstrip": false,
42
+ "single_word": false
43
+ },
44
+ "unk_token": {
45
+ "content": "<unk>",
46
+ "lstrip": false,
47
+ "normalized": true,
48
+ "rstrip": false,
49
+ "single_word": false
50
+ }
51
+ }
balanced_finetuned_model/checkpoint-2310/tokenizer_config.json ADDED
@@ -0,0 +1,56 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "added_tokens_decoder": {
4
+ "0": {
5
+ "content": "<s>",
6
+ "lstrip": false,
7
+ "normalized": true,
8
+ "rstrip": false,
9
+ "single_word": false,
10
+ "special": true
11
+ },
12
+ "1": {
13
+ "content": "<pad>",
14
+ "lstrip": false,
15
+ "normalized": true,
16
+ "rstrip": false,
17
+ "single_word": false,
18
+ "special": true
19
+ },
20
+ "2": {
21
+ "content": "</s>",
22
+ "lstrip": false,
23
+ "normalized": true,
24
+ "rstrip": false,
25
+ "single_word": false,
26
+ "special": true
27
+ },
28
+ "3": {
29
+ "content": "<unk>",
30
+ "lstrip": false,
31
+ "normalized": true,
32
+ "rstrip": false,
33
+ "single_word": false,
34
+ "special": true
35
+ },
36
+ "50264": {
37
+ "content": "<mask>",
38
+ "lstrip": true,
39
+ "normalized": false,
40
+ "rstrip": false,
41
+ "single_word": false,
42
+ "special": true
43
+ }
44
+ },
45
+ "bos_token": "<s>",
46
+ "clean_up_tokenization_spaces": true,
47
+ "cls_token": "<s>",
48
+ "eos_token": "</s>",
49
+ "errors": "replace",
50
+ "mask_token": "<mask>",
51
+ "model_max_length": 512,
52
+ "pad_token": "<pad>",
53
+ "sep_token": "</s>",
54
+ "tokenizer_class": "RobertaTokenizer",
55
+ "unk_token": "<unk>"
56
+ }
balanced_finetuned_model/checkpoint-2310/trainer_state.json ADDED
@@ -0,0 +1,1465 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.8079253061520617,
3
+ "best_model_checkpoint": "./balanced_finetuned_model/checkpoint-2310",
4
+ "epoch": 5.0,
5
+ "eval_steps": 500,
6
+ "global_step": 2310,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.02,
13
+ "learning_rate": 4.329004329004329e-07,
14
+ "loss": 10.8262,
15
+ "step": 10
16
+ },
17
+ {
18
+ "epoch": 0.04,
19
+ "learning_rate": 8.658008658008658e-07,
20
+ "loss": 12.2906,
21
+ "step": 20
22
+ },
23
+ {
24
+ "epoch": 0.06,
25
+ "learning_rate": 1.2987012987012986e-06,
26
+ "loss": 11.018,
27
+ "step": 30
28
+ },
29
+ {
30
+ "epoch": 0.09,
31
+ "learning_rate": 1.7316017316017317e-06,
32
+ "loss": 10.4001,
33
+ "step": 40
34
+ },
35
+ {
36
+ "epoch": 0.11,
37
+ "learning_rate": 2.1645021645021648e-06,
38
+ "loss": 10.7695,
39
+ "step": 50
40
+ },
41
+ {
42
+ "epoch": 0.13,
43
+ "learning_rate": 2.597402597402597e-06,
44
+ "loss": 9.9439,
45
+ "step": 60
46
+ },
47
+ {
48
+ "epoch": 0.15,
49
+ "learning_rate": 3.0303030303030305e-06,
50
+ "loss": 10.3621,
51
+ "step": 70
52
+ },
53
+ {
54
+ "epoch": 0.17,
55
+ "learning_rate": 3.4632034632034634e-06,
56
+ "loss": 8.5374,
57
+ "step": 80
58
+ },
59
+ {
60
+ "epoch": 0.19,
61
+ "learning_rate": 3.896103896103897e-06,
62
+ "loss": 6.6053,
63
+ "step": 90
64
+ },
65
+ {
66
+ "epoch": 0.22,
67
+ "learning_rate": 4.3290043290043295e-06,
68
+ "loss": 3.9921,
69
+ "step": 100
70
+ },
71
+ {
72
+ "epoch": 0.24,
73
+ "learning_rate": 4.761904761904762e-06,
74
+ "loss": 3.2861,
75
+ "step": 110
76
+ },
77
+ {
78
+ "epoch": 0.26,
79
+ "learning_rate": 5.194805194805194e-06,
80
+ "loss": 2.7387,
81
+ "step": 120
82
+ },
83
+ {
84
+ "epoch": 0.28,
85
+ "learning_rate": 5.627705627705629e-06,
86
+ "loss": 1.5201,
87
+ "step": 130
88
+ },
89
+ {
90
+ "epoch": 0.3,
91
+ "learning_rate": 6.060606060606061e-06,
92
+ "loss": 2.2568,
93
+ "step": 140
94
+ },
95
+ {
96
+ "epoch": 0.32,
97
+ "learning_rate": 6.493506493506494e-06,
98
+ "loss": 1.7721,
99
+ "step": 150
100
+ },
101
+ {
102
+ "epoch": 0.35,
103
+ "learning_rate": 6.926406926406927e-06,
104
+ "loss": 1.9366,
105
+ "step": 160
106
+ },
107
+ {
108
+ "epoch": 0.37,
109
+ "learning_rate": 7.3593073593073596e-06,
110
+ "loss": 1.9072,
111
+ "step": 170
112
+ },
113
+ {
114
+ "epoch": 0.39,
115
+ "learning_rate": 7.792207792207793e-06,
116
+ "loss": 2.1488,
117
+ "step": 180
118
+ },
119
+ {
120
+ "epoch": 0.41,
121
+ "learning_rate": 8.225108225108225e-06,
122
+ "loss": 1.6105,
123
+ "step": 190
124
+ },
125
+ {
126
+ "epoch": 0.43,
127
+ "learning_rate": 8.658008658008659e-06,
128
+ "loss": 2.0,
129
+ "step": 200
130
+ },
131
+ {
132
+ "epoch": 0.45,
133
+ "learning_rate": 9.090909090909091e-06,
134
+ "loss": 1.572,
135
+ "step": 210
136
+ },
137
+ {
138
+ "epoch": 0.48,
139
+ "learning_rate": 9.523809523809525e-06,
140
+ "loss": 1.5928,
141
+ "step": 220
142
+ },
143
+ {
144
+ "epoch": 0.5,
145
+ "learning_rate": 9.956709956709958e-06,
146
+ "loss": 1.3095,
147
+ "step": 230
148
+ },
149
+ {
150
+ "epoch": 0.52,
151
+ "learning_rate": 9.956709956709958e-06,
152
+ "loss": 1.513,
153
+ "step": 240
154
+ },
155
+ {
156
+ "epoch": 0.54,
157
+ "learning_rate": 9.90860990860991e-06,
158
+ "loss": 1.1392,
159
+ "step": 250
160
+ },
161
+ {
162
+ "epoch": 0.56,
163
+ "learning_rate": 9.860509860509861e-06,
164
+ "loss": 1.2247,
165
+ "step": 260
166
+ },
167
+ {
168
+ "epoch": 0.58,
169
+ "learning_rate": 9.812409812409814e-06,
170
+ "loss": 0.8886,
171
+ "step": 270
172
+ },
173
+ {
174
+ "epoch": 0.61,
175
+ "learning_rate": 9.764309764309765e-06,
176
+ "loss": 0.9848,
177
+ "step": 280
178
+ },
179
+ {
180
+ "epoch": 0.63,
181
+ "learning_rate": 9.716209716209716e-06,
182
+ "loss": 1.1762,
183
+ "step": 290
184
+ },
185
+ {
186
+ "epoch": 0.65,
187
+ "learning_rate": 9.66810966810967e-06,
188
+ "loss": 1.1722,
189
+ "step": 300
190
+ },
191
+ {
192
+ "epoch": 0.67,
193
+ "learning_rate": 9.62000962000962e-06,
194
+ "loss": 1.0029,
195
+ "step": 310
196
+ },
197
+ {
198
+ "epoch": 0.69,
199
+ "learning_rate": 9.571909571909572e-06,
200
+ "loss": 0.9868,
201
+ "step": 320
202
+ },
203
+ {
204
+ "epoch": 0.71,
205
+ "learning_rate": 9.523809523809525e-06,
206
+ "loss": 0.9915,
207
+ "step": 330
208
+ },
209
+ {
210
+ "epoch": 0.74,
211
+ "learning_rate": 9.475709475709478e-06,
212
+ "loss": 0.8529,
213
+ "step": 340
214
+ },
215
+ {
216
+ "epoch": 0.76,
217
+ "learning_rate": 9.427609427609429e-06,
218
+ "loss": 0.9799,
219
+ "step": 350
220
+ },
221
+ {
222
+ "epoch": 0.78,
223
+ "learning_rate": 9.37950937950938e-06,
224
+ "loss": 1.1586,
225
+ "step": 360
226
+ },
227
+ {
228
+ "epoch": 0.8,
229
+ "learning_rate": 9.331409331409333e-06,
230
+ "loss": 1.0889,
231
+ "step": 370
232
+ },
233
+ {
234
+ "epoch": 0.82,
235
+ "learning_rate": 9.283309283309284e-06,
236
+ "loss": 1.2502,
237
+ "step": 380
238
+ },
239
+ {
240
+ "epoch": 0.84,
241
+ "learning_rate": 9.235209235209236e-06,
242
+ "loss": 0.9826,
243
+ "step": 390
244
+ },
245
+ {
246
+ "epoch": 0.87,
247
+ "learning_rate": 9.187109187109189e-06,
248
+ "loss": 0.9052,
249
+ "step": 400
250
+ },
251
+ {
252
+ "epoch": 0.89,
253
+ "learning_rate": 9.13900913900914e-06,
254
+ "loss": 0.9651,
255
+ "step": 410
256
+ },
257
+ {
258
+ "epoch": 0.91,
259
+ "learning_rate": 9.090909090909091e-06,
260
+ "loss": 0.8358,
261
+ "step": 420
262
+ },
263
+ {
264
+ "epoch": 0.93,
265
+ "learning_rate": 9.042809042809044e-06,
266
+ "loss": 1.2029,
267
+ "step": 430
268
+ },
269
+ {
270
+ "epoch": 0.95,
271
+ "learning_rate": 8.994708994708995e-06,
272
+ "loss": 0.9126,
273
+ "step": 440
274
+ },
275
+ {
276
+ "epoch": 0.97,
277
+ "learning_rate": 8.946608946608948e-06,
278
+ "loss": 0.9049,
279
+ "step": 450
280
+ },
281
+ {
282
+ "epoch": 1.0,
283
+ "learning_rate": 8.8985088985089e-06,
284
+ "loss": 0.8561,
285
+ "step": 460
286
+ },
287
+ {
288
+ "epoch": 1.0,
289
+ "eval_accuracy": 0.6951351351351351,
290
+ "eval_f1": 0.6941963163251571,
291
+ "eval_loss": 0.8276578187942505,
292
+ "eval_precision": 0.7097058096914333,
293
+ "eval_recall": 0.6951351351351351,
294
+ "eval_runtime": 41.1958,
295
+ "eval_samples_per_second": 22.454,
296
+ "eval_steps_per_second": 2.816,
297
+ "step": 462
298
+ },
299
+ {
300
+ "epoch": 1.02,
301
+ "learning_rate": 8.85040885040885e-06,
302
+ "loss": 0.9094,
303
+ "step": 470
304
+ },
305
+ {
306
+ "epoch": 1.04,
307
+ "learning_rate": 8.802308802308804e-06,
308
+ "loss": 0.7898,
309
+ "step": 480
310
+ },
311
+ {
312
+ "epoch": 1.06,
313
+ "learning_rate": 8.754208754208755e-06,
314
+ "loss": 0.6191,
315
+ "step": 490
316
+ },
317
+ {
318
+ "epoch": 1.08,
319
+ "learning_rate": 8.706108706108706e-06,
320
+ "loss": 0.9794,
321
+ "step": 500
322
+ },
323
+ {
324
+ "epoch": 1.1,
325
+ "learning_rate": 8.658008658008659e-06,
326
+ "loss": 0.7676,
327
+ "step": 510
328
+ },
329
+ {
330
+ "epoch": 1.13,
331
+ "learning_rate": 8.60990860990861e-06,
332
+ "loss": 0.7618,
333
+ "step": 520
334
+ },
335
+ {
336
+ "epoch": 1.15,
337
+ "learning_rate": 8.561808561808562e-06,
338
+ "loss": 0.7243,
339
+ "step": 530
340
+ },
341
+ {
342
+ "epoch": 1.17,
343
+ "learning_rate": 8.513708513708514e-06,
344
+ "loss": 1.0373,
345
+ "step": 540
346
+ },
347
+ {
348
+ "epoch": 1.19,
349
+ "learning_rate": 8.465608465608466e-06,
350
+ "loss": 0.8194,
351
+ "step": 550
352
+ },
353
+ {
354
+ "epoch": 1.21,
355
+ "learning_rate": 8.417508417508419e-06,
356
+ "loss": 0.8167,
357
+ "step": 560
358
+ },
359
+ {
360
+ "epoch": 1.23,
361
+ "learning_rate": 8.36940836940837e-06,
362
+ "loss": 0.6841,
363
+ "step": 570
364
+ },
365
+ {
366
+ "epoch": 1.26,
367
+ "learning_rate": 8.321308321308323e-06,
368
+ "loss": 0.5651,
369
+ "step": 580
370
+ },
371
+ {
372
+ "epoch": 1.28,
373
+ "learning_rate": 8.273208273208274e-06,
374
+ "loss": 0.9127,
375
+ "step": 590
376
+ },
377
+ {
378
+ "epoch": 1.3,
379
+ "learning_rate": 8.225108225108225e-06,
380
+ "loss": 0.8384,
381
+ "step": 600
382
+ },
383
+ {
384
+ "epoch": 1.32,
385
+ "learning_rate": 8.177008177008178e-06,
386
+ "loss": 0.6718,
387
+ "step": 610
388
+ },
389
+ {
390
+ "epoch": 1.34,
391
+ "learning_rate": 8.12890812890813e-06,
392
+ "loss": 0.5633,
393
+ "step": 620
394
+ },
395
+ {
396
+ "epoch": 1.36,
397
+ "learning_rate": 8.08080808080808e-06,
398
+ "loss": 0.6944,
399
+ "step": 630
400
+ },
401
+ {
402
+ "epoch": 1.39,
403
+ "learning_rate": 8.032708032708034e-06,
404
+ "loss": 0.4755,
405
+ "step": 640
406
+ },
407
+ {
408
+ "epoch": 1.41,
409
+ "learning_rate": 7.984607984607985e-06,
410
+ "loss": 0.9228,
411
+ "step": 650
412
+ },
413
+ {
414
+ "epoch": 1.43,
415
+ "learning_rate": 7.936507936507936e-06,
416
+ "loss": 0.5992,
417
+ "step": 660
418
+ },
419
+ {
420
+ "epoch": 1.45,
421
+ "learning_rate": 7.888407888407889e-06,
422
+ "loss": 0.7077,
423
+ "step": 670
424
+ },
425
+ {
426
+ "epoch": 1.47,
427
+ "learning_rate": 7.840307840307842e-06,
428
+ "loss": 0.7204,
429
+ "step": 680
430
+ },
431
+ {
432
+ "epoch": 1.49,
433
+ "learning_rate": 7.792207792207793e-06,
434
+ "loss": 0.7907,
435
+ "step": 690
436
+ },
437
+ {
438
+ "epoch": 1.52,
439
+ "learning_rate": 7.744107744107745e-06,
440
+ "loss": 0.531,
441
+ "step": 700
442
+ },
443
+ {
444
+ "epoch": 1.54,
445
+ "learning_rate": 7.696007696007697e-06,
446
+ "loss": 0.889,
447
+ "step": 710
448
+ },
449
+ {
450
+ "epoch": 1.56,
451
+ "learning_rate": 7.647907647907649e-06,
452
+ "loss": 0.6319,
453
+ "step": 720
454
+ },
455
+ {
456
+ "epoch": 1.58,
457
+ "learning_rate": 7.599807599807601e-06,
458
+ "loss": 0.5859,
459
+ "step": 730
460
+ },
461
+ {
462
+ "epoch": 1.6,
463
+ "learning_rate": 7.551707551707552e-06,
464
+ "loss": 0.6143,
465
+ "step": 740
466
+ },
467
+ {
468
+ "epoch": 1.62,
469
+ "learning_rate": 7.503607503607504e-06,
470
+ "loss": 0.5408,
471
+ "step": 750
472
+ },
473
+ {
474
+ "epoch": 1.65,
475
+ "learning_rate": 7.455507455507455e-06,
476
+ "loss": 0.8811,
477
+ "step": 760
478
+ },
479
+ {
480
+ "epoch": 1.67,
481
+ "learning_rate": 7.4074074074074075e-06,
482
+ "loss": 0.6971,
483
+ "step": 770
484
+ },
485
+ {
486
+ "epoch": 1.69,
487
+ "learning_rate": 7.3593073593073596e-06,
488
+ "loss": 0.4133,
489
+ "step": 780
490
+ },
491
+ {
492
+ "epoch": 1.71,
493
+ "learning_rate": 7.3112073112073125e-06,
494
+ "loss": 0.4327,
495
+ "step": 790
496
+ },
497
+ {
498
+ "epoch": 1.73,
499
+ "learning_rate": 7.263107263107264e-06,
500
+ "loss": 0.4936,
501
+ "step": 800
502
+ },
503
+ {
504
+ "epoch": 1.75,
505
+ "learning_rate": 7.215007215007216e-06,
506
+ "loss": 0.6713,
507
+ "step": 810
508
+ },
509
+ {
510
+ "epoch": 1.77,
511
+ "learning_rate": 7.166907166907168e-06,
512
+ "loss": 0.5208,
513
+ "step": 820
514
+ },
515
+ {
516
+ "epoch": 1.8,
517
+ "learning_rate": 7.118807118807119e-06,
518
+ "loss": 0.81,
519
+ "step": 830
520
+ },
521
+ {
522
+ "epoch": 1.82,
523
+ "learning_rate": 7.070707070707071e-06,
524
+ "loss": 0.5224,
525
+ "step": 840
526
+ },
527
+ {
528
+ "epoch": 1.84,
529
+ "learning_rate": 7.022607022607023e-06,
530
+ "loss": 0.7043,
531
+ "step": 850
532
+ },
533
+ {
534
+ "epoch": 1.86,
535
+ "learning_rate": 6.974506974506975e-06,
536
+ "loss": 0.7131,
537
+ "step": 860
538
+ },
539
+ {
540
+ "epoch": 1.88,
541
+ "learning_rate": 6.926406926406927e-06,
542
+ "loss": 0.8933,
543
+ "step": 870
544
+ },
545
+ {
546
+ "epoch": 1.9,
547
+ "learning_rate": 6.878306878306879e-06,
548
+ "loss": 0.7053,
549
+ "step": 880
550
+ },
551
+ {
552
+ "epoch": 1.93,
553
+ "learning_rate": 6.83020683020683e-06,
554
+ "loss": 0.6044,
555
+ "step": 890
556
+ },
557
+ {
558
+ "epoch": 1.95,
559
+ "learning_rate": 6.782106782106783e-06,
560
+ "loss": 0.5334,
561
+ "step": 900
562
+ },
563
+ {
564
+ "epoch": 1.97,
565
+ "learning_rate": 6.734006734006735e-06,
566
+ "loss": 0.5943,
567
+ "step": 910
568
+ },
569
+ {
570
+ "epoch": 1.99,
571
+ "learning_rate": 6.685906685906687e-06,
572
+ "loss": 0.4249,
573
+ "step": 920
574
+ },
575
+ {
576
+ "epoch": 2.0,
577
+ "eval_accuracy": 0.7881081081081082,
578
+ "eval_f1": 0.7883030241253077,
579
+ "eval_loss": 0.6265861392021179,
580
+ "eval_precision": 0.7896965590352171,
581
+ "eval_recall": 0.7881081081081082,
582
+ "eval_runtime": 42.8724,
583
+ "eval_samples_per_second": 21.576,
584
+ "eval_steps_per_second": 2.706,
585
+ "step": 924
586
+ },
587
+ {
588
+ "epoch": 2.01,
589
+ "learning_rate": 6.637806637806638e-06,
590
+ "loss": 0.527,
591
+ "step": 930
592
+ },
593
+ {
594
+ "epoch": 2.03,
595
+ "learning_rate": 6.5897065897065905e-06,
596
+ "loss": 0.3392,
597
+ "step": 940
598
+ },
599
+ {
600
+ "epoch": 2.06,
601
+ "learning_rate": 6.541606541606543e-06,
602
+ "loss": 0.6126,
603
+ "step": 950
604
+ },
605
+ {
606
+ "epoch": 2.08,
607
+ "learning_rate": 6.493506493506494e-06,
608
+ "loss": 0.4029,
609
+ "step": 960
610
+ },
611
+ {
612
+ "epoch": 2.1,
613
+ "learning_rate": 6.445406445406446e-06,
614
+ "loss": 0.5203,
615
+ "step": 970
616
+ },
617
+ {
618
+ "epoch": 2.12,
619
+ "learning_rate": 6.397306397306397e-06,
620
+ "loss": 0.49,
621
+ "step": 980
622
+ },
623
+ {
624
+ "epoch": 2.14,
625
+ "learning_rate": 6.349206349206349e-06,
626
+ "loss": 0.4803,
627
+ "step": 990
628
+ },
629
+ {
630
+ "epoch": 2.16,
631
+ "learning_rate": 6.301106301106301e-06,
632
+ "loss": 0.4338,
633
+ "step": 1000
634
+ },
635
+ {
636
+ "epoch": 2.19,
637
+ "learning_rate": 6.253006253006253e-06,
638
+ "loss": 0.3084,
639
+ "step": 1010
640
+ },
641
+ {
642
+ "epoch": 2.21,
643
+ "learning_rate": 6.2049062049062055e-06,
644
+ "loss": 0.3061,
645
+ "step": 1020
646
+ },
647
+ {
648
+ "epoch": 2.23,
649
+ "learning_rate": 6.156806156806158e-06,
650
+ "loss": 0.6207,
651
+ "step": 1030
652
+ },
653
+ {
654
+ "epoch": 2.25,
655
+ "learning_rate": 6.10870610870611e-06,
656
+ "loss": 0.3268,
657
+ "step": 1040
658
+ },
659
+ {
660
+ "epoch": 2.27,
661
+ "learning_rate": 6.060606060606061e-06,
662
+ "loss": 0.6113,
663
+ "step": 1050
664
+ },
665
+ {
666
+ "epoch": 2.29,
667
+ "learning_rate": 6.012506012506013e-06,
668
+ "loss": 0.3333,
669
+ "step": 1060
670
+ },
671
+ {
672
+ "epoch": 2.32,
673
+ "learning_rate": 5.964405964405965e-06,
674
+ "loss": 0.4616,
675
+ "step": 1070
676
+ },
677
+ {
678
+ "epoch": 2.34,
679
+ "learning_rate": 5.916305916305916e-06,
680
+ "loss": 0.7176,
681
+ "step": 1080
682
+ },
683
+ {
684
+ "epoch": 2.36,
685
+ "learning_rate": 5.8682058682058685e-06,
686
+ "loss": 0.6433,
687
+ "step": 1090
688
+ },
689
+ {
690
+ "epoch": 2.38,
691
+ "learning_rate": 5.820105820105821e-06,
692
+ "loss": 0.2657,
693
+ "step": 1100
694
+ },
695
+ {
696
+ "epoch": 2.4,
697
+ "learning_rate": 5.772005772005772e-06,
698
+ "loss": 0.7557,
699
+ "step": 1110
700
+ },
701
+ {
702
+ "epoch": 2.42,
703
+ "learning_rate": 5.723905723905724e-06,
704
+ "loss": 0.499,
705
+ "step": 1120
706
+ },
707
+ {
708
+ "epoch": 2.45,
709
+ "learning_rate": 5.675805675805677e-06,
710
+ "loss": 0.6084,
711
+ "step": 1130
712
+ },
713
+ {
714
+ "epoch": 2.47,
715
+ "learning_rate": 5.627705627705629e-06,
716
+ "loss": 0.4959,
717
+ "step": 1140
718
+ },
719
+ {
720
+ "epoch": 2.49,
721
+ "learning_rate": 5.57960557960558e-06,
722
+ "loss": 0.3092,
723
+ "step": 1150
724
+ },
725
+ {
726
+ "epoch": 2.51,
727
+ "learning_rate": 5.531505531505532e-06,
728
+ "loss": 0.4155,
729
+ "step": 1160
730
+ },
731
+ {
732
+ "epoch": 2.53,
733
+ "learning_rate": 5.4834054834054835e-06,
734
+ "loss": 0.448,
735
+ "step": 1170
736
+ },
737
+ {
738
+ "epoch": 2.55,
739
+ "learning_rate": 5.435305435305436e-06,
740
+ "loss": 0.4538,
741
+ "step": 1180
742
+ },
743
+ {
744
+ "epoch": 2.58,
745
+ "learning_rate": 5.387205387205388e-06,
746
+ "loss": 0.4018,
747
+ "step": 1190
748
+ },
749
+ {
750
+ "epoch": 2.6,
751
+ "learning_rate": 5.339105339105339e-06,
752
+ "loss": 0.6252,
753
+ "step": 1200
754
+ },
755
+ {
756
+ "epoch": 2.62,
757
+ "learning_rate": 5.291005291005291e-06,
758
+ "loss": 0.4832,
759
+ "step": 1210
760
+ },
761
+ {
762
+ "epoch": 2.64,
763
+ "learning_rate": 5.242905242905243e-06,
764
+ "loss": 0.3838,
765
+ "step": 1220
766
+ },
767
+ {
768
+ "epoch": 2.66,
769
+ "learning_rate": 5.194805194805194e-06,
770
+ "loss": 0.3117,
771
+ "step": 1230
772
+ },
773
+ {
774
+ "epoch": 2.68,
775
+ "learning_rate": 5.146705146705147e-06,
776
+ "loss": 0.3003,
777
+ "step": 1240
778
+ },
779
+ {
780
+ "epoch": 2.71,
781
+ "learning_rate": 5.0986050986050994e-06,
782
+ "loss": 0.2706,
783
+ "step": 1250
784
+ },
785
+ {
786
+ "epoch": 2.73,
787
+ "learning_rate": 5.0505050505050515e-06,
788
+ "loss": 0.7442,
789
+ "step": 1260
790
+ },
791
+ {
792
+ "epoch": 2.75,
793
+ "learning_rate": 5.002405002405003e-06,
794
+ "loss": 0.2976,
795
+ "step": 1270
796
+ },
797
+ {
798
+ "epoch": 2.77,
799
+ "learning_rate": 4.954304954304955e-06,
800
+ "loss": 0.3882,
801
+ "step": 1280
802
+ },
803
+ {
804
+ "epoch": 2.79,
805
+ "learning_rate": 4.906204906204907e-06,
806
+ "loss": 0.3143,
807
+ "step": 1290
808
+ },
809
+ {
810
+ "epoch": 2.81,
811
+ "learning_rate": 4.858104858104858e-06,
812
+ "loss": 0.8753,
813
+ "step": 1300
814
+ },
815
+ {
816
+ "epoch": 2.84,
817
+ "learning_rate": 4.81000481000481e-06,
818
+ "loss": 0.706,
819
+ "step": 1310
820
+ },
821
+ {
822
+ "epoch": 2.86,
823
+ "learning_rate": 4.761904761904762e-06,
824
+ "loss": 0.478,
825
+ "step": 1320
826
+ },
827
+ {
828
+ "epoch": 2.88,
829
+ "learning_rate": 4.7138047138047145e-06,
830
+ "loss": 0.6678,
831
+ "step": 1330
832
+ },
833
+ {
834
+ "epoch": 2.9,
835
+ "learning_rate": 4.6657046657046666e-06,
836
+ "loss": 0.3574,
837
+ "step": 1340
838
+ },
839
+ {
840
+ "epoch": 2.92,
841
+ "learning_rate": 4.617604617604618e-06,
842
+ "loss": 0.3732,
843
+ "step": 1350
844
+ },
845
+ {
846
+ "epoch": 2.94,
847
+ "learning_rate": 4.56950456950457e-06,
848
+ "loss": 0.2018,
849
+ "step": 1360
850
+ },
851
+ {
852
+ "epoch": 2.97,
853
+ "learning_rate": 4.521404521404522e-06,
854
+ "loss": 0.4718,
855
+ "step": 1370
856
+ },
857
+ {
858
+ "epoch": 2.99,
859
+ "learning_rate": 4.473304473304474e-06,
860
+ "loss": 0.404,
861
+ "step": 1380
862
+ },
863
+ {
864
+ "epoch": 3.0,
865
+ "eval_accuracy": 0.7967567567567567,
866
+ "eval_f1": 0.7969295115959946,
867
+ "eval_loss": 0.662196695804596,
868
+ "eval_precision": 0.8024330408412987,
869
+ "eval_recall": 0.7967567567567567,
870
+ "eval_runtime": 42.3812,
871
+ "eval_samples_per_second": 21.826,
872
+ "eval_steps_per_second": 2.737,
873
+ "step": 1386
874
+ },
875
+ {
876
+ "epoch": 3.01,
877
+ "learning_rate": 4.425204425204425e-06,
878
+ "loss": 0.4288,
879
+ "step": 1390
880
+ },
881
+ {
882
+ "epoch": 3.03,
883
+ "learning_rate": 4.377104377104377e-06,
884
+ "loss": 0.1295,
885
+ "step": 1400
886
+ },
887
+ {
888
+ "epoch": 3.05,
889
+ "learning_rate": 4.3290043290043295e-06,
890
+ "loss": 0.3256,
891
+ "step": 1410
892
+ },
893
+ {
894
+ "epoch": 3.07,
895
+ "learning_rate": 4.280904280904281e-06,
896
+ "loss": 0.4232,
897
+ "step": 1420
898
+ },
899
+ {
900
+ "epoch": 3.1,
901
+ "learning_rate": 4.232804232804233e-06,
902
+ "loss": 0.4204,
903
+ "step": 1430
904
+ },
905
+ {
906
+ "epoch": 3.12,
907
+ "learning_rate": 4.184704184704185e-06,
908
+ "loss": 0.1602,
909
+ "step": 1440
910
+ },
911
+ {
912
+ "epoch": 3.14,
913
+ "learning_rate": 4.136604136604137e-06,
914
+ "loss": 0.344,
915
+ "step": 1450
916
+ },
917
+ {
918
+ "epoch": 3.16,
919
+ "learning_rate": 4.088504088504089e-06,
920
+ "loss": 0.2476,
921
+ "step": 1460
922
+ },
923
+ {
924
+ "epoch": 3.18,
925
+ "learning_rate": 4.04040404040404e-06,
926
+ "loss": 0.2518,
927
+ "step": 1470
928
+ },
929
+ {
930
+ "epoch": 3.2,
931
+ "learning_rate": 3.9923039923039925e-06,
932
+ "loss": 0.4476,
933
+ "step": 1480
934
+ },
935
+ {
936
+ "epoch": 3.23,
937
+ "learning_rate": 3.9442039442039446e-06,
938
+ "loss": 0.4147,
939
+ "step": 1490
940
+ },
941
+ {
942
+ "epoch": 3.25,
943
+ "learning_rate": 3.896103896103897e-06,
944
+ "loss": 0.3457,
945
+ "step": 1500
946
+ },
947
+ {
948
+ "epoch": 3.27,
949
+ "learning_rate": 3.848003848003849e-06,
950
+ "loss": 0.4206,
951
+ "step": 1510
952
+ },
953
+ {
954
+ "epoch": 3.29,
955
+ "learning_rate": 3.7999037999038004e-06,
956
+ "loss": 0.2476,
957
+ "step": 1520
958
+ },
959
+ {
960
+ "epoch": 3.31,
961
+ "learning_rate": 3.751803751803752e-06,
962
+ "loss": 0.3427,
963
+ "step": 1530
964
+ },
965
+ {
966
+ "epoch": 3.33,
967
+ "learning_rate": 3.7037037037037037e-06,
968
+ "loss": 0.3846,
969
+ "step": 1540
970
+ },
971
+ {
972
+ "epoch": 3.35,
973
+ "learning_rate": 3.6556036556036563e-06,
974
+ "loss": 0.3671,
975
+ "step": 1550
976
+ },
977
+ {
978
+ "epoch": 3.38,
979
+ "learning_rate": 3.607503607503608e-06,
980
+ "loss": 0.3038,
981
+ "step": 1560
982
+ },
983
+ {
984
+ "epoch": 3.4,
985
+ "learning_rate": 3.5594035594035596e-06,
986
+ "loss": 0.307,
987
+ "step": 1570
988
+ },
989
+ {
990
+ "epoch": 3.42,
991
+ "learning_rate": 3.5113035113035117e-06,
992
+ "loss": 0.2364,
993
+ "step": 1580
994
+ },
995
+ {
996
+ "epoch": 3.44,
997
+ "learning_rate": 3.4632034632034634e-06,
998
+ "loss": 0.3357,
999
+ "step": 1590
1000
+ },
1001
+ {
1002
+ "epoch": 3.46,
1003
+ "learning_rate": 3.415103415103415e-06,
1004
+ "loss": 0.2825,
1005
+ "step": 1600
1006
+ },
1007
+ {
1008
+ "epoch": 3.48,
1009
+ "learning_rate": 3.3670033670033675e-06,
1010
+ "loss": 0.3363,
1011
+ "step": 1610
1012
+ },
1013
+ {
1014
+ "epoch": 3.51,
1015
+ "learning_rate": 3.318903318903319e-06,
1016
+ "loss": 0.469,
1017
+ "step": 1620
1018
+ },
1019
+ {
1020
+ "epoch": 3.53,
1021
+ "learning_rate": 3.2708032708032713e-06,
1022
+ "loss": 0.5258,
1023
+ "step": 1630
1024
+ },
1025
+ {
1026
+ "epoch": 3.55,
1027
+ "learning_rate": 3.222703222703223e-06,
1028
+ "loss": 0.2061,
1029
+ "step": 1640
1030
+ },
1031
+ {
1032
+ "epoch": 3.57,
1033
+ "learning_rate": 3.1746031746031746e-06,
1034
+ "loss": 0.1948,
1035
+ "step": 1650
1036
+ },
1037
+ {
1038
+ "epoch": 3.59,
1039
+ "learning_rate": 3.1265031265031263e-06,
1040
+ "loss": 0.4708,
1041
+ "step": 1660
1042
+ },
1043
+ {
1044
+ "epoch": 3.61,
1045
+ "learning_rate": 3.078403078403079e-06,
1046
+ "loss": 0.4367,
1047
+ "step": 1670
1048
+ },
1049
+ {
1050
+ "epoch": 3.64,
1051
+ "learning_rate": 3.0303030303030305e-06,
1052
+ "loss": 0.4284,
1053
+ "step": 1680
1054
+ },
1055
+ {
1056
+ "epoch": 3.66,
1057
+ "learning_rate": 2.9822029822029826e-06,
1058
+ "loss": 0.3583,
1059
+ "step": 1690
1060
+ },
1061
+ {
1062
+ "epoch": 3.68,
1063
+ "learning_rate": 2.9341029341029342e-06,
1064
+ "loss": 0.3829,
1065
+ "step": 1700
1066
+ },
1067
+ {
1068
+ "epoch": 3.7,
1069
+ "learning_rate": 2.886002886002886e-06,
1070
+ "loss": 0.3757,
1071
+ "step": 1710
1072
+ },
1073
+ {
1074
+ "epoch": 3.72,
1075
+ "learning_rate": 2.8379028379028384e-06,
1076
+ "loss": 0.2352,
1077
+ "step": 1720
1078
+ },
1079
+ {
1080
+ "epoch": 3.74,
1081
+ "learning_rate": 2.78980278980279e-06,
1082
+ "loss": 0.3246,
1083
+ "step": 1730
1084
+ },
1085
+ {
1086
+ "epoch": 3.77,
1087
+ "learning_rate": 2.7417027417027418e-06,
1088
+ "loss": 0.2555,
1089
+ "step": 1740
1090
+ },
1091
+ {
1092
+ "epoch": 3.79,
1093
+ "learning_rate": 2.693602693602694e-06,
1094
+ "loss": 0.1718,
1095
+ "step": 1750
1096
+ },
1097
+ {
1098
+ "epoch": 3.81,
1099
+ "learning_rate": 2.6455026455026455e-06,
1100
+ "loss": 0.3823,
1101
+ "step": 1760
1102
+ },
1103
+ {
1104
+ "epoch": 3.83,
1105
+ "learning_rate": 2.597402597402597e-06,
1106
+ "loss": 0.3836,
1107
+ "step": 1770
1108
+ },
1109
+ {
1110
+ "epoch": 3.85,
1111
+ "learning_rate": 2.5493025493025497e-06,
1112
+ "loss": 0.0989,
1113
+ "step": 1780
1114
+ },
1115
+ {
1116
+ "epoch": 3.87,
1117
+ "learning_rate": 2.5012025012025014e-06,
1118
+ "loss": 0.3505,
1119
+ "step": 1790
1120
+ },
1121
+ {
1122
+ "epoch": 3.9,
1123
+ "learning_rate": 2.4531024531024535e-06,
1124
+ "loss": 0.4152,
1125
+ "step": 1800
1126
+ },
1127
+ {
1128
+ "epoch": 3.92,
1129
+ "learning_rate": 2.405002405002405e-06,
1130
+ "loss": 0.1573,
1131
+ "step": 1810
1132
+ },
1133
+ {
1134
+ "epoch": 3.94,
1135
+ "learning_rate": 2.3569023569023572e-06,
1136
+ "loss": 0.4271,
1137
+ "step": 1820
1138
+ },
1139
+ {
1140
+ "epoch": 3.96,
1141
+ "learning_rate": 2.308802308802309e-06,
1142
+ "loss": 0.5782,
1143
+ "step": 1830
1144
+ },
1145
+ {
1146
+ "epoch": 3.98,
1147
+ "learning_rate": 2.260702260702261e-06,
1148
+ "loss": 0.4115,
1149
+ "step": 1840
1150
+ },
1151
+ {
1152
+ "epoch": 4.0,
1153
+ "eval_accuracy": 0.8043243243243243,
1154
+ "eval_f1": 0.8039960230056994,
1155
+ "eval_loss": 0.6978992819786072,
1156
+ "eval_precision": 0.8072324249099836,
1157
+ "eval_recall": 0.8043243243243243,
1158
+ "eval_runtime": 42.9375,
1159
+ "eval_samples_per_second": 21.543,
1160
+ "eval_steps_per_second": 2.702,
1161
+ "step": 1848
1162
+ },
1163
+ {
1164
+ "epoch": 4.0,
1165
+ "learning_rate": 2.2126022126022127e-06,
1166
+ "loss": 0.3399,
1167
+ "step": 1850
1168
+ },
1169
+ {
1170
+ "epoch": 4.03,
1171
+ "learning_rate": 2.1645021645021648e-06,
1172
+ "loss": 0.2122,
1173
+ "step": 1860
1174
+ },
1175
+ {
1176
+ "epoch": 4.05,
1177
+ "learning_rate": 2.1164021164021164e-06,
1178
+ "loss": 0.227,
1179
+ "step": 1870
1180
+ },
1181
+ {
1182
+ "epoch": 4.07,
1183
+ "learning_rate": 2.0683020683020685e-06,
1184
+ "loss": 0.3432,
1185
+ "step": 1880
1186
+ },
1187
+ {
1188
+ "epoch": 4.09,
1189
+ "learning_rate": 2.02020202020202e-06,
1190
+ "loss": 0.316,
1191
+ "step": 1890
1192
+ },
1193
+ {
1194
+ "epoch": 4.11,
1195
+ "learning_rate": 1.9721019721019723e-06,
1196
+ "loss": 0.4086,
1197
+ "step": 1900
1198
+ },
1199
+ {
1200
+ "epoch": 4.13,
1201
+ "learning_rate": 1.9240019240019244e-06,
1202
+ "loss": 0.1236,
1203
+ "step": 1910
1204
+ },
1205
+ {
1206
+ "epoch": 4.16,
1207
+ "learning_rate": 1.875901875901876e-06,
1208
+ "loss": 0.2039,
1209
+ "step": 1920
1210
+ },
1211
+ {
1212
+ "epoch": 4.18,
1213
+ "learning_rate": 1.8278018278018281e-06,
1214
+ "loss": 0.2467,
1215
+ "step": 1930
1216
+ },
1217
+ {
1218
+ "epoch": 4.2,
1219
+ "learning_rate": 1.7797017797017798e-06,
1220
+ "loss": 0.1929,
1221
+ "step": 1940
1222
+ },
1223
+ {
1224
+ "epoch": 4.22,
1225
+ "learning_rate": 1.7316017316017317e-06,
1226
+ "loss": 0.1663,
1227
+ "step": 1950
1228
+ },
1229
+ {
1230
+ "epoch": 4.24,
1231
+ "learning_rate": 1.6835016835016838e-06,
1232
+ "loss": 0.3821,
1233
+ "step": 1960
1234
+ },
1235
+ {
1236
+ "epoch": 4.26,
1237
+ "learning_rate": 1.6354016354016356e-06,
1238
+ "loss": 0.2922,
1239
+ "step": 1970
1240
+ },
1241
+ {
1242
+ "epoch": 4.29,
1243
+ "learning_rate": 1.5873015873015873e-06,
1244
+ "loss": 0.1047,
1245
+ "step": 1980
1246
+ },
1247
+ {
1248
+ "epoch": 4.31,
1249
+ "learning_rate": 1.5392015392015394e-06,
1250
+ "loss": 0.1757,
1251
+ "step": 1990
1252
+ },
1253
+ {
1254
+ "epoch": 4.33,
1255
+ "learning_rate": 1.4911014911014913e-06,
1256
+ "loss": 0.2511,
1257
+ "step": 2000
1258
+ },
1259
+ {
1260
+ "epoch": 4.35,
1261
+ "learning_rate": 1.443001443001443e-06,
1262
+ "loss": 0.0429,
1263
+ "step": 2010
1264
+ },
1265
+ {
1266
+ "epoch": 4.37,
1267
+ "learning_rate": 1.394901394901395e-06,
1268
+ "loss": 0.2221,
1269
+ "step": 2020
1270
+ },
1271
+ {
1272
+ "epoch": 4.39,
1273
+ "learning_rate": 1.346801346801347e-06,
1274
+ "loss": 0.2715,
1275
+ "step": 2030
1276
+ },
1277
+ {
1278
+ "epoch": 4.42,
1279
+ "learning_rate": 1.2987012987012986e-06,
1280
+ "loss": 0.2049,
1281
+ "step": 2040
1282
+ },
1283
+ {
1284
+ "epoch": 4.44,
1285
+ "learning_rate": 1.2506012506012507e-06,
1286
+ "loss": 0.1993,
1287
+ "step": 2050
1288
+ },
1289
+ {
1290
+ "epoch": 4.46,
1291
+ "learning_rate": 1.2025012025012026e-06,
1292
+ "loss": 0.2417,
1293
+ "step": 2060
1294
+ },
1295
+ {
1296
+ "epoch": 4.48,
1297
+ "learning_rate": 1.1544011544011545e-06,
1298
+ "loss": 0.2595,
1299
+ "step": 2070
1300
+ },
1301
+ {
1302
+ "epoch": 4.5,
1303
+ "learning_rate": 1.1063011063011063e-06,
1304
+ "loss": 0.3325,
1305
+ "step": 2080
1306
+ },
1307
+ {
1308
+ "epoch": 4.52,
1309
+ "learning_rate": 1.0582010582010582e-06,
1310
+ "loss": 0.2834,
1311
+ "step": 2090
1312
+ },
1313
+ {
1314
+ "epoch": 4.55,
1315
+ "learning_rate": 1.01010101010101e-06,
1316
+ "loss": 0.1687,
1317
+ "step": 2100
1318
+ },
1319
+ {
1320
+ "epoch": 4.57,
1321
+ "learning_rate": 9.620009620009622e-07,
1322
+ "loss": 0.2747,
1323
+ "step": 2110
1324
+ },
1325
+ {
1326
+ "epoch": 4.59,
1327
+ "learning_rate": 9.139009139009141e-07,
1328
+ "loss": 0.2611,
1329
+ "step": 2120
1330
+ },
1331
+ {
1332
+ "epoch": 4.61,
1333
+ "learning_rate": 8.658008658008658e-07,
1334
+ "loss": 0.1968,
1335
+ "step": 2130
1336
+ },
1337
+ {
1338
+ "epoch": 4.63,
1339
+ "learning_rate": 8.177008177008178e-07,
1340
+ "loss": 0.151,
1341
+ "step": 2140
1342
+ },
1343
+ {
1344
+ "epoch": 4.65,
1345
+ "learning_rate": 7.696007696007697e-07,
1346
+ "loss": 0.2216,
1347
+ "step": 2150
1348
+ },
1349
+ {
1350
+ "epoch": 4.68,
1351
+ "learning_rate": 7.215007215007215e-07,
1352
+ "loss": 0.27,
1353
+ "step": 2160
1354
+ },
1355
+ {
1356
+ "epoch": 4.7,
1357
+ "learning_rate": 6.734006734006735e-07,
1358
+ "loss": 0.1358,
1359
+ "step": 2170
1360
+ },
1361
+ {
1362
+ "epoch": 4.72,
1363
+ "learning_rate": 6.253006253006253e-07,
1364
+ "loss": 0.4274,
1365
+ "step": 2180
1366
+ },
1367
+ {
1368
+ "epoch": 4.74,
1369
+ "learning_rate": 5.772005772005772e-07,
1370
+ "loss": 0.2386,
1371
+ "step": 2190
1372
+ },
1373
+ {
1374
+ "epoch": 4.76,
1375
+ "learning_rate": 5.291005291005291e-07,
1376
+ "loss": 0.2219,
1377
+ "step": 2200
1378
+ },
1379
+ {
1380
+ "epoch": 4.78,
1381
+ "learning_rate": 4.810004810004811e-07,
1382
+ "loss": 0.2619,
1383
+ "step": 2210
1384
+ },
1385
+ {
1386
+ "epoch": 4.81,
1387
+ "learning_rate": 4.329004329004329e-07,
1388
+ "loss": 0.307,
1389
+ "step": 2220
1390
+ },
1391
+ {
1392
+ "epoch": 4.83,
1393
+ "learning_rate": 3.8480038480038485e-07,
1394
+ "loss": 0.2637,
1395
+ "step": 2230
1396
+ },
1397
+ {
1398
+ "epoch": 4.85,
1399
+ "learning_rate": 3.3670033670033673e-07,
1400
+ "loss": 0.1671,
1401
+ "step": 2240
1402
+ },
1403
+ {
1404
+ "epoch": 4.87,
1405
+ "learning_rate": 2.886002886002886e-07,
1406
+ "loss": 0.432,
1407
+ "step": 2250
1408
+ },
1409
+ {
1410
+ "epoch": 4.89,
1411
+ "learning_rate": 2.4050024050024055e-07,
1412
+ "loss": 0.3521,
1413
+ "step": 2260
1414
+ },
1415
+ {
1416
+ "epoch": 4.91,
1417
+ "learning_rate": 1.9240019240019243e-07,
1418
+ "loss": 0.3122,
1419
+ "step": 2270
1420
+ },
1421
+ {
1422
+ "epoch": 4.94,
1423
+ "learning_rate": 1.443001443001443e-07,
1424
+ "loss": 0.2866,
1425
+ "step": 2280
1426
+ },
1427
+ {
1428
+ "epoch": 4.96,
1429
+ "learning_rate": 9.620009620009621e-08,
1430
+ "loss": 0.3384,
1431
+ "step": 2290
1432
+ },
1433
+ {
1434
+ "epoch": 4.98,
1435
+ "learning_rate": 4.8100048100048107e-08,
1436
+ "loss": 0.1356,
1437
+ "step": 2300
1438
+ },
1439
+ {
1440
+ "epoch": 5.0,
1441
+ "learning_rate": 0.0,
1442
+ "loss": 0.2733,
1443
+ "step": 2310
1444
+ },
1445
+ {
1446
+ "epoch": 5.0,
1447
+ "eval_accuracy": 0.8086486486486486,
1448
+ "eval_f1": 0.8079253061520617,
1449
+ "eval_loss": 0.7432375550270081,
1450
+ "eval_precision": 0.8093600008268919,
1451
+ "eval_recall": 0.8086486486486486,
1452
+ "eval_runtime": 42.4495,
1453
+ "eval_samples_per_second": 21.791,
1454
+ "eval_steps_per_second": 2.733,
1455
+ "step": 2310
1456
+ }
1457
+ ],
1458
+ "logging_steps": 10,
1459
+ "max_steps": 2310,
1460
+ "num_train_epochs": 5,
1461
+ "save_steps": 500,
1462
+ "total_flos": 2099403213836352.0,
1463
+ "trial_name": null,
1464
+ "trial_params": null
1465
+ }
balanced_finetuned_model/checkpoint-2310/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2e8c11de7aff56066100df364f86da392520ed4522bc7c376ec702df94a5ace7
3
+ size 4155
balanced_finetuned_model/checkpoint-2310/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
balanced_finetuned_model/config.json ADDED
@@ -0,0 +1,42 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "./checkpoint-18750",
3
+ "architectures": [
4
+ "RobertaForSequenceClassification"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.1,
7
+ "bos_token_id": 0,
8
+ "classifier_dropout": null,
9
+ "eos_token_id": 2,
10
+ "hidden_act": "gelu",
11
+ "hidden_dropout_prob": 0.1,
12
+ "hidden_size": 768,
13
+ "id2label": {
14
+ "0": "Exploration and Reflection",
15
+ "1": "Feedback and Support",
16
+ "2": "Goal Setting and Planning",
17
+ "3": "Problem Solving and Critical Thinking",
18
+ "4": "Understanding and Clarification"
19
+ },
20
+ "initializer_range": 0.02,
21
+ "intermediate_size": 3072,
22
+ "label2id": {
23
+ "Exploration and Reflection": 0,
24
+ "Feedback and Support": 1,
25
+ "Goal Setting and Planning": 2,
26
+ "Problem Solving and Critical Thinking": 3,
27
+ "Understanding and Clarification": 4
28
+ },
29
+ "layer_norm_eps": 1e-05,
30
+ "max_position_embeddings": 514,
31
+ "model_type": "roberta",
32
+ "num_attention_heads": 12,
33
+ "num_hidden_layers": 12,
34
+ "pad_token_id": 1,
35
+ "position_embedding_type": "absolute",
36
+ "problem_type": "single_label_classification",
37
+ "torch_dtype": "float32",
38
+ "transformers_version": "4.35.2",
39
+ "type_vocab_size": 1,
40
+ "use_cache": true,
41
+ "vocab_size": 50265
42
+ }
balanced_finetuned_model/label_mapping.json ADDED
@@ -0,0 +1,23 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "label_to_id": {
3
+ "Exploration and Reflection": 0,
4
+ "Feedback and Support": 1,
5
+ "Goal Setting and Planning": 2,
6
+ "Problem Solving and Critical Thinking": 3,
7
+ "Understanding and Clarification": 4
8
+ },
9
+ "id_to_label": {
10
+ "0": "Exploration and Reflection",
11
+ "1": "Feedback and Support",
12
+ "2": "Goal Setting and Planning",
13
+ "3": "Problem Solving and Critical Thinking",
14
+ "4": "Understanding and Clarification"
15
+ },
16
+ "all_labels": [
17
+ "Exploration and Reflection",
18
+ "Feedback and Support",
19
+ "Goal Setting and Planning",
20
+ "Problem Solving and Critical Thinking",
21
+ "Understanding and Clarification"
22
+ ]
23
+ }
balanced_finetuned_model/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
balanced_finetuned_model/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4ec1dac41d174534829a2ac7f7456cd92ed4cba58731a0ba25409849f53e2437
3
+ size 498622052
balanced_finetuned_model/runs/Jul03_19-02-50_hayashis-MacBook-Pro.local/events.out.tfevents.1751583770.hayashis-MacBook-Pro.local.69542.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:07bb3e55c507d9ba320873edd6a73acd4d0f99e18b6e84572113f4e8e3c4437a
3
+ size 6098
balanced_finetuned_model/runs/Jul03_19-09-19_hayashis-MacBook-Pro.local/events.out.tfevents.1751584159.hayashis-MacBook-Pro.local.69763.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2b0310cffb5e56242cf2110efa82f9ef9d543b0dffd22f1304838c43c049aafd
3
+ size 43727
balanced_finetuned_model/runs/Jul03_19-09-19_hayashis-MacBook-Pro.local/events.out.tfevents.1751588684.hayashis-MacBook-Pro.local.69763.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dce9b11ee60bf54e1cf9e4f2728c866dae1a3c1bc2cee267a4317aaff4370e87
3
+ size 560
balanced_finetuned_model/special_tokens_map.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": true,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "cls_token": {
10
+ "content": "<s>",
11
+ "lstrip": false,
12
+ "normalized": true,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "eos_token": {
17
+ "content": "</s>",
18
+ "lstrip": false,
19
+ "normalized": true,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ },
23
+ "mask_token": {
24
+ "content": "<mask>",
25
+ "lstrip": true,
26
+ "normalized": false,
27
+ "rstrip": false,
28
+ "single_word": false
29
+ },
30
+ "pad_token": {
31
+ "content": "<pad>",
32
+ "lstrip": false,
33
+ "normalized": true,
34
+ "rstrip": false,
35
+ "single_word": false
36
+ },
37
+ "sep_token": {
38
+ "content": "</s>",
39
+ "lstrip": false,
40
+ "normalized": true,
41
+ "rstrip": false,
42
+ "single_word": false
43
+ },
44
+ "unk_token": {
45
+ "content": "<unk>",
46
+ "lstrip": false,
47
+ "normalized": true,
48
+ "rstrip": false,
49
+ "single_word": false
50
+ }
51
+ }
balanced_finetuned_model/tokenizer_config.json ADDED
@@ -0,0 +1,56 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "added_tokens_decoder": {
4
+ "0": {
5
+ "content": "<s>",
6
+ "lstrip": false,
7
+ "normalized": true,
8
+ "rstrip": false,
9
+ "single_word": false,
10
+ "special": true
11
+ },
12
+ "1": {
13
+ "content": "<pad>",
14
+ "lstrip": false,
15
+ "normalized": true,
16
+ "rstrip": false,
17
+ "single_word": false,
18
+ "special": true
19
+ },
20
+ "2": {
21
+ "content": "</s>",
22
+ "lstrip": false,
23
+ "normalized": true,
24
+ "rstrip": false,
25
+ "single_word": false,
26
+ "special": true
27
+ },
28
+ "3": {
29
+ "content": "<unk>",
30
+ "lstrip": false,
31
+ "normalized": true,
32
+ "rstrip": false,
33
+ "single_word": false,
34
+ "special": true
35
+ },
36
+ "50264": {
37
+ "content": "<mask>",
38
+ "lstrip": true,
39
+ "normalized": false,
40
+ "rstrip": false,
41
+ "single_word": false,
42
+ "special": true
43
+ }
44
+ },
45
+ "bos_token": "<s>",
46
+ "clean_up_tokenization_spaces": true,
47
+ "cls_token": "<s>",
48
+ "eos_token": "</s>",
49
+ "errors": "replace",
50
+ "mask_token": "<mask>",
51
+ "model_max_length": 512,
52
+ "pad_token": "<pad>",
53
+ "sep_token": "</s>",
54
+ "tokenizer_class": "RobertaTokenizer",
55
+ "unk_token": "<unk>"
56
+ }
balanced_finetuned_model/vocab.json ADDED
The diff for this file is too large to render. See raw diff