parmarm commited on
Commit
dba13e7
·
verified ·
1 Parent(s): c133330

Initial commit with SafeTensors

Browse files
added_tokens.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ {
2
+ "[MASK]": 128000
3
+ }
config.json ADDED
@@ -0,0 +1,47 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "microsoft/deberta-v3-base",
3
+ "architectures": [
4
+ "DebertaV2ForSequenceClassification"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.15,
7
+ "hidden_act": "gelu",
8
+ "hidden_dropout_prob": 0.15,
9
+ "hidden_size": 768,
10
+ "id2label": {
11
+ "0": "NSFW",
12
+ "1": "GenHarm",
13
+ "2": "Toxic"
14
+ },
15
+ "initializer_range": 0.02,
16
+ "intermediate_size": 3072,
17
+ "label2id": {
18
+ "GenHarm": 1,
19
+ "NSFW": 0,
20
+ "Toxic": 2
21
+ },
22
+ "layer_norm_eps": 1e-07,
23
+ "legacy": true,
24
+ "max_position_embeddings": 512,
25
+ "max_relative_positions": -1,
26
+ "model_type": "deberta-v2",
27
+ "norm_rel_ebd": "layer_norm",
28
+ "num_attention_heads": 12,
29
+ "num_hidden_layers": 12,
30
+ "pad_token_id": 0,
31
+ "pooler_dropout": 0,
32
+ "pooler_hidden_act": "gelu",
33
+ "pooler_hidden_size": 768,
34
+ "pos_att_type": [
35
+ "p2c",
36
+ "c2p"
37
+ ],
38
+ "position_biased_input": false,
39
+ "position_buckets": 256,
40
+ "problem_type": "multi_label_classification",
41
+ "relative_attention": true,
42
+ "share_att_key": true,
43
+ "torch_dtype": "float32",
44
+ "transformers_version": "4.47.1",
45
+ "type_vocab_size": 0,
46
+ "vocab_size": 128100
47
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3877e4f52647da895b45e0ef8a8319d5dae63239658526d63914198ec4dfde49
3
+ size 737722356
optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cbed5b34b770cbf9a596148bc1c464bcd566a3de4ffe3f9be659059666a9cad6
3
+ size 1475564538
rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2ac7bb245edf5533edf15cad547a88cd0bac36f929a2c61a90d304cfe83ae6c8
3
+ size 14244
scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a2d564b724ae4c0f87828589c44b761aa0cde92642fe48b03a74d1d049a87d90
3
+ size 1064
special_tokens_map.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "[CLS]",
3
+ "cls_token": "[CLS]",
4
+ "eos_token": "[SEP]",
5
+ "mask_token": "[MASK]",
6
+ "pad_token": "[PAD]",
7
+ "sep_token": "[SEP]",
8
+ "unk_token": {
9
+ "content": "[UNK]",
10
+ "lstrip": false,
11
+ "normalized": true,
12
+ "rstrip": false,
13
+ "single_word": false
14
+ }
15
+ }
spm.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c679fbf93643d19aab7ee10c0b99e460bdbc02fedf34b92b05af343b4af586fd
3
+ size 2464616
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,59 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "[PAD]",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "[CLS]",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "[SEP]",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "[UNK]",
29
+ "lstrip": false,
30
+ "normalized": true,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "128000": {
36
+ "content": "[MASK]",
37
+ "lstrip": false,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ }
43
+ },
44
+ "bos_token": "[CLS]",
45
+ "clean_up_tokenization_spaces": false,
46
+ "cls_token": "[CLS]",
47
+ "do_lower_case": false,
48
+ "eos_token": "[SEP]",
49
+ "extra_special_tokens": {},
50
+ "mask_token": "[MASK]",
51
+ "model_max_length": 1000000000000000019884624838656,
52
+ "pad_token": "[PAD]",
53
+ "sep_token": "[SEP]",
54
+ "sp_model_kwargs": {},
55
+ "split_by_punct": false,
56
+ "tokenizer_class": "DebertaV2Tokenizer",
57
+ "unk_token": "[UNK]",
58
+ "vocab_type": "spm"
59
+ }
trainer_state.json ADDED
@@ -0,0 +1,1158 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.32019519805908203,
3
+ "best_model_checkpoint": "./deberta_multilabel_safetyattribution_run1/checkpoint-2172",
4
+ "epoch": 8.0,
5
+ "eval_steps": 500,
6
+ "global_step": 5792,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.06906077348066299,
13
+ "grad_norm": 17.054950714111328,
14
+ "learning_rate": 5.753739930955121e-07,
15
+ "loss": 2.7202,
16
+ "step": 50
17
+ },
18
+ {
19
+ "epoch": 0.13812154696132597,
20
+ "grad_norm": 15.804380416870117,
21
+ "learning_rate": 1.1507479861910242e-06,
22
+ "loss": 2.4784,
23
+ "step": 100
24
+ },
25
+ {
26
+ "epoch": 0.20718232044198895,
27
+ "grad_norm": 4.005581378936768,
28
+ "learning_rate": 1.7261219792865363e-06,
29
+ "loss": 1.9331,
30
+ "step": 150
31
+ },
32
+ {
33
+ "epoch": 0.27624309392265195,
34
+ "grad_norm": 3.5140247344970703,
35
+ "learning_rate": 2.3014959723820484e-06,
36
+ "loss": 1.3647,
37
+ "step": 200
38
+ },
39
+ {
40
+ "epoch": 0.3453038674033149,
41
+ "grad_norm": 10.396773338317871,
42
+ "learning_rate": 2.8768699654775607e-06,
43
+ "loss": 1.1814,
44
+ "step": 250
45
+ },
46
+ {
47
+ "epoch": 0.4143646408839779,
48
+ "grad_norm": 4.15683126449585,
49
+ "learning_rate": 3.4522439585730726e-06,
50
+ "loss": 1.1349,
51
+ "step": 300
52
+ },
53
+ {
54
+ "epoch": 0.48342541436464087,
55
+ "grad_norm": 13.827107429504395,
56
+ "learning_rate": 4.027617951668585e-06,
57
+ "loss": 1.0992,
58
+ "step": 350
59
+ },
60
+ {
61
+ "epoch": 0.5524861878453039,
62
+ "grad_norm": 5.466891765594482,
63
+ "learning_rate": 4.602991944764097e-06,
64
+ "loss": 1.0756,
65
+ "step": 400
66
+ },
67
+ {
68
+ "epoch": 0.6215469613259669,
69
+ "grad_norm": 5.3988118171691895,
70
+ "learning_rate": 5.1783659378596095e-06,
71
+ "loss": 1.0385,
72
+ "step": 450
73
+ },
74
+ {
75
+ "epoch": 0.6906077348066298,
76
+ "grad_norm": 6.582638740539551,
77
+ "learning_rate": 5.753739930955121e-06,
78
+ "loss": 0.9732,
79
+ "step": 500
80
+ },
81
+ {
82
+ "epoch": 0.7596685082872928,
83
+ "grad_norm": 4.769979953765869,
84
+ "learning_rate": 6.329113924050634e-06,
85
+ "loss": 0.9272,
86
+ "step": 550
87
+ },
88
+ {
89
+ "epoch": 0.8287292817679558,
90
+ "grad_norm": 7.325048923492432,
91
+ "learning_rate": 6.904487917146145e-06,
92
+ "loss": 0.9391,
93
+ "step": 600
94
+ },
95
+ {
96
+ "epoch": 0.8977900552486188,
97
+ "grad_norm": 5.211851596832275,
98
+ "learning_rate": 7.479861910241658e-06,
99
+ "loss": 0.9031,
100
+ "step": 650
101
+ },
102
+ {
103
+ "epoch": 0.9668508287292817,
104
+ "grad_norm": 8.523693084716797,
105
+ "learning_rate": 8.05523590333717e-06,
106
+ "loss": 0.8428,
107
+ "step": 700
108
+ },
109
+ {
110
+ "epoch": 1.0,
111
+ "eval_class_accuracy": {
112
+ "label_0": 0.7745931503521982,
113
+ "label_1": 0.9664804469273743,
114
+ "label_2": 0.8299732815156667
115
+ },
116
+ "eval_class_f1": {
117
+ "label_0": 0.5691736304549675,
118
+ "label_1": 0.7523923444976076,
119
+ "label_2": 0.6285815351963212
120
+ },
121
+ "eval_class_false_positive_rate": {
122
+ "label_0": 0.06384009691095531,
123
+ "label_1": 0.018147086914993647,
124
+ "label_2": 0.0660200595172417
125
+ },
126
+ "eval_class_precision": {
127
+ "label_0": 0.7772612003381234,
128
+ "label_1": 0.7505966587112172,
129
+ "label_2": 0.747895622895623
130
+ },
131
+ "eval_class_recall": {
132
+ "label_0": 0.448974609375,
133
+ "label_1": 0.7541966426858513,
134
+ "label_2": 0.5420988407565589
135
+ },
136
+ "eval_loss": 0.4103058874607086,
137
+ "eval_macro_f1": 0.6500491700496321,
138
+ "eval_macro_precision": 0.7585844939816545,
139
+ "eval_macro_recall": 0.5817566976058034,
140
+ "eval_micro_f1": 0.6157528285465622,
141
+ "eval_micro_precision": 0.760752688172043,
142
+ "eval_micro_recall": 0.5171783625730995,
143
+ "eval_runtime": 132.9806,
144
+ "eval_samples_per_second": 92.878,
145
+ "eval_steps_per_second": 1.451,
146
+ "eval_subset_accuracy": 0.6513642620030766,
147
+ "step": 724
148
+ },
149
+ {
150
+ "epoch": 1.0359116022099448,
151
+ "grad_norm": 5.507997989654541,
152
+ "learning_rate": 8.630609896432683e-06,
153
+ "loss": 0.8297,
154
+ "step": 750
155
+ },
156
+ {
157
+ "epoch": 1.1049723756906078,
158
+ "grad_norm": 6.73417329788208,
159
+ "learning_rate": 9.205983889528194e-06,
160
+ "loss": 0.7722,
161
+ "step": 800
162
+ },
163
+ {
164
+ "epoch": 1.1740331491712708,
165
+ "grad_norm": 5.604831218719482,
166
+ "learning_rate": 9.781357882623706e-06,
167
+ "loss": 0.7892,
168
+ "step": 850
169
+ },
170
+ {
171
+ "epoch": 1.2430939226519337,
172
+ "grad_norm": 6.269744873046875,
173
+ "learning_rate": 9.93703026609791e-06,
174
+ "loss": 0.7489,
175
+ "step": 900
176
+ },
177
+ {
178
+ "epoch": 1.3121546961325967,
179
+ "grad_norm": 5.223318576812744,
180
+ "learning_rate": 9.83546617915905e-06,
181
+ "loss": 0.756,
182
+ "step": 950
183
+ },
184
+ {
185
+ "epoch": 1.3812154696132597,
186
+ "grad_norm": 6.4933180809021,
187
+ "learning_rate": 9.733902092220193e-06,
188
+ "loss": 0.706,
189
+ "step": 1000
190
+ },
191
+ {
192
+ "epoch": 1.4502762430939227,
193
+ "grad_norm": 4.81841516494751,
194
+ "learning_rate": 9.632338005281333e-06,
195
+ "loss": 0.7007,
196
+ "step": 1050
197
+ },
198
+ {
199
+ "epoch": 1.5193370165745856,
200
+ "grad_norm": 4.895750999450684,
201
+ "learning_rate": 9.530773918342476e-06,
202
+ "loss": 0.7067,
203
+ "step": 1100
204
+ },
205
+ {
206
+ "epoch": 1.5883977900552486,
207
+ "grad_norm": 15.617860794067383,
208
+ "learning_rate": 9.429209831403616e-06,
209
+ "loss": 0.6879,
210
+ "step": 1150
211
+ },
212
+ {
213
+ "epoch": 1.6574585635359116,
214
+ "grad_norm": 5.868403911590576,
215
+ "learning_rate": 9.327645744464759e-06,
216
+ "loss": 0.6599,
217
+ "step": 1200
218
+ },
219
+ {
220
+ "epoch": 1.7265193370165746,
221
+ "grad_norm": 5.1376423835754395,
222
+ "learning_rate": 9.2260816575259e-06,
223
+ "loss": 0.6524,
224
+ "step": 1250
225
+ },
226
+ {
227
+ "epoch": 1.7955801104972375,
228
+ "grad_norm": 8.1726713180542,
229
+ "learning_rate": 9.124517570587042e-06,
230
+ "loss": 0.6301,
231
+ "step": 1300
232
+ },
233
+ {
234
+ "epoch": 1.8646408839779005,
235
+ "grad_norm": 5.737952709197998,
236
+ "learning_rate": 9.022953483648182e-06,
237
+ "loss": 0.6554,
238
+ "step": 1350
239
+ },
240
+ {
241
+ "epoch": 1.9337016574585635,
242
+ "grad_norm": 8.225056648254395,
243
+ "learning_rate": 8.921389396709325e-06,
244
+ "loss": 0.6457,
245
+ "step": 1400
246
+ },
247
+ {
248
+ "epoch": 2.0,
249
+ "eval_class_accuracy": {
250
+ "label_0": 0.8472998137802608,
251
+ "label_1": 0.9697190510889806,
252
+ "label_2": 0.8784713788357218
253
+ },
254
+ "eval_class_f1": {
255
+ "label_0": 0.7341415280518748,
256
+ "label_1": 0.7713936430317848,
257
+ "label_2": 0.7481966113068277
258
+ },
259
+ "eval_class_false_positive_rate": {
260
+ "label_0": 0.04772864930344667,
261
+ "label_1": 0.014847616566812985,
262
+ "label_2": 0.049928358866962425
263
+ },
264
+ "eval_class_precision": {
265
+ "label_0": 0.8685790527018012,
266
+ "label_1": 0.786783042394015,
267
+ "label_2": 0.8311591502049944
268
+ },
269
+ "eval_class_recall": {
270
+ "label_0": 0.6357421875,
271
+ "label_1": 0.7565947242206235,
272
+ "label_2": 0.6802928615009152
273
+ },
274
+ "eval_loss": 0.3337920904159546,
275
+ "eval_macro_f1": 0.7512439274634959,
276
+ "eval_macro_precision": 0.8288404151002702,
277
+ "eval_macro_recall": 0.6908765910738461,
278
+ "eval_micro_f1": 0.7439929208358859,
279
+ "eval_micro_precision": 0.8429739318216874,
280
+ "eval_micro_recall": 0.6658138401559455,
281
+ "eval_runtime": 132.9461,
282
+ "eval_samples_per_second": 92.902,
283
+ "eval_steps_per_second": 1.452,
284
+ "eval_subset_accuracy": 0.7538660837179175,
285
+ "step": 1448
286
+ },
287
+ {
288
+ "epoch": 2.0027624309392267,
289
+ "grad_norm": 4.911665439605713,
290
+ "learning_rate": 8.819825309770465e-06,
291
+ "loss": 0.6339,
292
+ "step": 1450
293
+ },
294
+ {
295
+ "epoch": 2.0718232044198897,
296
+ "grad_norm": 5.494454383850098,
297
+ "learning_rate": 8.718261222831608e-06,
298
+ "loss": 0.6068,
299
+ "step": 1500
300
+ },
301
+ {
302
+ "epoch": 2.1408839779005526,
303
+ "grad_norm": 7.909505844116211,
304
+ "learning_rate": 8.616697135892748e-06,
305
+ "loss": 0.6289,
306
+ "step": 1550
307
+ },
308
+ {
309
+ "epoch": 2.2099447513812156,
310
+ "grad_norm": 6.768340587615967,
311
+ "learning_rate": 8.515133048953891e-06,
312
+ "loss": 0.6021,
313
+ "step": 1600
314
+ },
315
+ {
316
+ "epoch": 2.2790055248618786,
317
+ "grad_norm": 5.231354236602783,
318
+ "learning_rate": 8.413568962015032e-06,
319
+ "loss": 0.585,
320
+ "step": 1650
321
+ },
322
+ {
323
+ "epoch": 2.3480662983425415,
324
+ "grad_norm": 3.701436758041382,
325
+ "learning_rate": 8.312004875076174e-06,
326
+ "loss": 0.5728,
327
+ "step": 1700
328
+ },
329
+ {
330
+ "epoch": 2.4171270718232045,
331
+ "grad_norm": 3.403343915939331,
332
+ "learning_rate": 8.210440788137315e-06,
333
+ "loss": 0.5746,
334
+ "step": 1750
335
+ },
336
+ {
337
+ "epoch": 2.4861878453038675,
338
+ "grad_norm": 4.147812843322754,
339
+ "learning_rate": 8.108876701198457e-06,
340
+ "loss": 0.6028,
341
+ "step": 1800
342
+ },
343
+ {
344
+ "epoch": 2.5552486187845305,
345
+ "grad_norm": 8.234676361083984,
346
+ "learning_rate": 8.007312614259598e-06,
347
+ "loss": 0.5451,
348
+ "step": 1850
349
+ },
350
+ {
351
+ "epoch": 2.6243093922651934,
352
+ "grad_norm": 3.46964955329895,
353
+ "learning_rate": 7.90574852732074e-06,
354
+ "loss": 0.591,
355
+ "step": 1900
356
+ },
357
+ {
358
+ "epoch": 2.6933701657458564,
359
+ "grad_norm": 5.345615386962891,
360
+ "learning_rate": 7.80418444038188e-06,
361
+ "loss": 0.5685,
362
+ "step": 1950
363
+ },
364
+ {
365
+ "epoch": 2.7624309392265194,
366
+ "grad_norm": 5.062536239624023,
367
+ "learning_rate": 7.702620353443023e-06,
368
+ "loss": 0.5786,
369
+ "step": 2000
370
+ },
371
+ {
372
+ "epoch": 2.8314917127071824,
373
+ "grad_norm": 5.448482036590576,
374
+ "learning_rate": 7.6010562665041645e-06,
375
+ "loss": 0.5773,
376
+ "step": 2050
377
+ },
378
+ {
379
+ "epoch": 2.9005524861878453,
380
+ "grad_norm": 7.135024547576904,
381
+ "learning_rate": 7.499492179565306e-06,
382
+ "loss": 0.5849,
383
+ "step": 2100
384
+ },
385
+ {
386
+ "epoch": 2.9696132596685083,
387
+ "grad_norm": 4.267426490783691,
388
+ "learning_rate": 7.3979280926264475e-06,
389
+ "loss": 0.5663,
390
+ "step": 2150
391
+ },
392
+ {
393
+ "epoch": 3.0,
394
+ "eval_class_accuracy": {
395
+ "label_0": 0.8591207189701239,
396
+ "label_1": 0.9690713302566594,
397
+ "label_2": 0.8903732491296251
398
+ },
399
+ "eval_class_f1": {
400
+ "label_0": 0.7654986522911051,
401
+ "label_1": 0.7768691588785047,
402
+ "label_2": 0.7841198979591837
403
+ },
404
+ "eval_class_false_positive_rate": {
405
+ "label_0": 0.05863113264687357,
406
+ "label_1": 0.01849439958322319,
407
+ "label_2": 0.05896616334177681
408
+ },
409
+ "eval_class_precision": {
410
+ "label_0": 0.8543922984356197,
411
+ "label_1": 0.7574031890660592,
412
+ "label_2": 0.821309285237141
413
+ },
414
+ "eval_class_recall": {
415
+ "label_0": 0.693359375,
416
+ "label_1": 0.7973621103117506,
417
+ "label_2": 0.7501525320317267
418
+ },
419
+ "eval_loss": 0.32019519805908203,
420
+ "eval_macro_f1": 0.775495903042931,
421
+ "eval_macro_precision": 0.8110349242462732,
422
+ "eval_macro_recall": 0.746958005781159,
423
+ "eval_micro_f1": 0.7743443261490522,
424
+ "eval_micro_precision": 0.8287937743190662,
425
+ "eval_micro_recall": 0.7266081871345029,
426
+ "eval_runtime": 132.8667,
427
+ "eval_samples_per_second": 92.958,
428
+ "eval_steps_per_second": 1.453,
429
+ "eval_subset_accuracy": 0.7761314873289612,
430
+ "step": 2172
431
+ },
432
+ {
433
+ "epoch": 3.0386740331491713,
434
+ "grad_norm": 3.6615378856658936,
435
+ "learning_rate": 7.29636400568759e-06,
436
+ "loss": 0.5629,
437
+ "step": 2200
438
+ },
439
+ {
440
+ "epoch": 3.1077348066298343,
441
+ "grad_norm": 7.720168590545654,
442
+ "learning_rate": 7.194799918748731e-06,
443
+ "loss": 0.5422,
444
+ "step": 2250
445
+ },
446
+ {
447
+ "epoch": 3.1767955801104972,
448
+ "grad_norm": 5.942800521850586,
449
+ "learning_rate": 7.093235831809873e-06,
450
+ "loss": 0.5198,
451
+ "step": 2300
452
+ },
453
+ {
454
+ "epoch": 3.24585635359116,
455
+ "grad_norm": 4.374157428741455,
456
+ "learning_rate": 6.991671744871014e-06,
457
+ "loss": 0.528,
458
+ "step": 2350
459
+ },
460
+ {
461
+ "epoch": 3.314917127071823,
462
+ "grad_norm": 4.527683734893799,
463
+ "learning_rate": 6.890107657932156e-06,
464
+ "loss": 0.5481,
465
+ "step": 2400
466
+ },
467
+ {
468
+ "epoch": 3.383977900552486,
469
+ "grad_norm": 3.410897731781006,
470
+ "learning_rate": 6.7885435709932975e-06,
471
+ "loss": 0.531,
472
+ "step": 2450
473
+ },
474
+ {
475
+ "epoch": 3.453038674033149,
476
+ "grad_norm": 3.105001449584961,
477
+ "learning_rate": 6.686979484054439e-06,
478
+ "loss": 0.5193,
479
+ "step": 2500
480
+ },
481
+ {
482
+ "epoch": 3.522099447513812,
483
+ "grad_norm": 4.019840717315674,
484
+ "learning_rate": 6.585415397115581e-06,
485
+ "loss": 0.5435,
486
+ "step": 2550
487
+ },
488
+ {
489
+ "epoch": 3.591160220994475,
490
+ "grad_norm": 2.8091235160827637,
491
+ "learning_rate": 6.483851310176722e-06,
492
+ "loss": 0.5028,
493
+ "step": 2600
494
+ },
495
+ {
496
+ "epoch": 3.660220994475138,
497
+ "grad_norm": 6.829853057861328,
498
+ "learning_rate": 6.382287223237864e-06,
499
+ "loss": 0.5308,
500
+ "step": 2650
501
+ },
502
+ {
503
+ "epoch": 3.729281767955801,
504
+ "grad_norm": 6.366962432861328,
505
+ "learning_rate": 6.280723136299005e-06,
506
+ "loss": 0.5431,
507
+ "step": 2700
508
+ },
509
+ {
510
+ "epoch": 3.798342541436464,
511
+ "grad_norm": 7.169396877288818,
512
+ "learning_rate": 6.1791590493601475e-06,
513
+ "loss": 0.5297,
514
+ "step": 2750
515
+ },
516
+ {
517
+ "epoch": 3.867403314917127,
518
+ "grad_norm": 4.60045051574707,
519
+ "learning_rate": 6.077594962421288e-06,
520
+ "loss": 0.5226,
521
+ "step": 2800
522
+ },
523
+ {
524
+ "epoch": 3.93646408839779,
525
+ "grad_norm": 3.7710654735565186,
526
+ "learning_rate": 5.9760308754824305e-06,
527
+ "loss": 0.528,
528
+ "step": 2850
529
+ },
530
+ {
531
+ "epoch": 4.0,
532
+ "eval_class_accuracy": {
533
+ "label_0": 0.8668933689579791,
534
+ "label_1": 0.9708525625455429,
535
+ "label_2": 0.8956359808922354
536
+ },
537
+ "eval_class_f1": {
538
+ "label_0": 0.7924242424242425,
539
+ "label_1": 0.7904540162980209,
540
+ "label_2": 0.7929982334992773
541
+ },
542
+ "eval_class_false_positive_rate": {
543
+ "label_0": 0.08310115081767618,
544
+ "label_1": 0.017799774246764103,
545
+ "label_2": 0.05290422131598667
546
+ },
547
+ "eval_class_precision": {
548
+ "label_0": 0.8206066945606695,
549
+ "label_1": 0.7680995475113123,
550
+ "label_2": 0.8372329603255341
551
+ },
552
+ "eval_class_recall": {
553
+ "label_0": 0.76611328125,
554
+ "label_1": 0.8141486810551559,
555
+ "label_2": 0.7532031726662599
556
+ },
557
+ "eval_loss": 0.3253972828388214,
558
+ "eval_macro_f1": 0.7919588307405135,
559
+ "eval_macro_precision": 0.8086464007991719,
560
+ "eval_macro_recall": 0.7778217116571385,
561
+ "eval_micro_f1": 0.7924361802710369,
562
+ "eval_micro_precision": 0.8209481520177615,
563
+ "eval_micro_recall": 0.7658382066276803,
564
+ "eval_runtime": 132.8554,
565
+ "eval_samples_per_second": 92.966,
566
+ "eval_steps_per_second": 1.453,
567
+ "eval_subset_accuracy": 0.7909481013683103,
568
+ "step": 2896
569
+ },
570
+ {
571
+ "epoch": 4.005524861878453,
572
+ "grad_norm": 6.861307621002197,
573
+ "learning_rate": 5.874466788543571e-06,
574
+ "loss": 0.5269,
575
+ "step": 2900
576
+ },
577
+ {
578
+ "epoch": 4.074585635359116,
579
+ "grad_norm": 5.186521530151367,
580
+ "learning_rate": 5.7729027016047135e-06,
581
+ "loss": 0.5032,
582
+ "step": 2950
583
+ },
584
+ {
585
+ "epoch": 4.143646408839779,
586
+ "grad_norm": 2.319528579711914,
587
+ "learning_rate": 5.671338614665854e-06,
588
+ "loss": 0.4947,
589
+ "step": 3000
590
+ },
591
+ {
592
+ "epoch": 4.212707182320442,
593
+ "grad_norm": 7.0490403175354,
594
+ "learning_rate": 5.5697745277269966e-06,
595
+ "loss": 0.475,
596
+ "step": 3050
597
+ },
598
+ {
599
+ "epoch": 4.281767955801105,
600
+ "grad_norm": 6.733311653137207,
601
+ "learning_rate": 5.468210440788137e-06,
602
+ "loss": 0.4758,
603
+ "step": 3100
604
+ },
605
+ {
606
+ "epoch": 4.350828729281768,
607
+ "grad_norm": 5.118820667266846,
608
+ "learning_rate": 5.36664635384928e-06,
609
+ "loss": 0.5216,
610
+ "step": 3150
611
+ },
612
+ {
613
+ "epoch": 4.419889502762431,
614
+ "grad_norm": 4.245698928833008,
615
+ "learning_rate": 5.26508226691042e-06,
616
+ "loss": 0.4965,
617
+ "step": 3200
618
+ },
619
+ {
620
+ "epoch": 4.488950276243094,
621
+ "grad_norm": 6.310864448547363,
622
+ "learning_rate": 5.163518179971563e-06,
623
+ "loss": 0.4979,
624
+ "step": 3250
625
+ },
626
+ {
627
+ "epoch": 4.558011049723757,
628
+ "grad_norm": 10.312039375305176,
629
+ "learning_rate": 5.061954093032704e-06,
630
+ "loss": 0.4812,
631
+ "step": 3300
632
+ },
633
+ {
634
+ "epoch": 4.62707182320442,
635
+ "grad_norm": 5.925340175628662,
636
+ "learning_rate": 4.960390006093846e-06,
637
+ "loss": 0.4946,
638
+ "step": 3350
639
+ },
640
+ {
641
+ "epoch": 4.696132596685083,
642
+ "grad_norm": 2.886112689971924,
643
+ "learning_rate": 4.858825919154987e-06,
644
+ "loss": 0.488,
645
+ "step": 3400
646
+ },
647
+ {
648
+ "epoch": 4.765193370165746,
649
+ "grad_norm": 5.802464008331299,
650
+ "learning_rate": 4.757261832216129e-06,
651
+ "loss": 0.4914,
652
+ "step": 3450
653
+ },
654
+ {
655
+ "epoch": 4.834254143646409,
656
+ "grad_norm": 3.5704684257507324,
657
+ "learning_rate": 4.65569774527727e-06,
658
+ "loss": 0.4581,
659
+ "step": 3500
660
+ },
661
+ {
662
+ "epoch": 4.903314917127072,
663
+ "grad_norm": 4.317540168762207,
664
+ "learning_rate": 4.554133658338412e-06,
665
+ "loss": 0.4881,
666
+ "step": 3550
667
+ },
668
+ {
669
+ "epoch": 4.972375690607735,
670
+ "grad_norm": 2.523470401763916,
671
+ "learning_rate": 4.452569571399554e-06,
672
+ "loss": 0.4949,
673
+ "step": 3600
674
+ },
675
+ {
676
+ "epoch": 5.0,
677
+ "eval_class_accuracy": {
678
+ "label_0": 0.8691603918711035,
679
+ "label_1": 0.9713383531697838,
680
+ "label_2": 0.8945024694356732
681
+ },
682
+ "eval_class_f1": {
683
+ "label_0": 0.7941925624044829,
684
+ "label_1": 0.787515006002401,
685
+ "label_2": 0.7994458981068185
686
+ },
687
+ "eval_class_false_positive_rate": {
688
+ "label_0": 0.07728649303451517,
689
+ "label_1": 0.015281757402099914,
690
+ "label_2": 0.0685550534552994
691
+ },
692
+ "eval_class_precision": {
693
+ "label_0": 0.8301384451544196,
694
+ "label_1": 0.7884615384615384,
695
+ "label_2": 0.8067722895309102
696
+ },
697
+ "eval_class_recall": {
698
+ "label_0": 0.76123046875,
699
+ "label_1": 0.7865707434052758,
700
+ "label_2": 0.7922513727882855
701
+ },
702
+ "eval_loss": 0.3214770257472992,
703
+ "eval_macro_f1": 0.7937178221712342,
704
+ "eval_macro_precision": 0.8084574243822894,
705
+ "eval_macro_recall": 0.7800175283145204,
706
+ "eval_micro_f1": 0.7956290977208866,
707
+ "eval_micro_precision": 0.816062508005636,
708
+ "eval_micro_recall": 0.7761939571150097,
709
+ "eval_runtime": 133.0557,
710
+ "eval_samples_per_second": 92.826,
711
+ "eval_steps_per_second": 1.451,
712
+ "eval_subset_accuracy": 0.791838717512752,
713
+ "step": 3620
714
+ },
715
+ {
716
+ "epoch": 5.041436464088398,
717
+ "grad_norm": 5.695593357086182,
718
+ "learning_rate": 4.351005484460696e-06,
719
+ "loss": 0.4785,
720
+ "step": 3650
721
+ },
722
+ {
723
+ "epoch": 5.110497237569061,
724
+ "grad_norm": 6.8877692222595215,
725
+ "learning_rate": 4.249441397521837e-06,
726
+ "loss": 0.4648,
727
+ "step": 3700
728
+ },
729
+ {
730
+ "epoch": 5.179558011049724,
731
+ "grad_norm": 4.254743576049805,
732
+ "learning_rate": 4.147877310582979e-06,
733
+ "loss": 0.4543,
734
+ "step": 3750
735
+ },
736
+ {
737
+ "epoch": 5.248618784530387,
738
+ "grad_norm": 4.990494251251221,
739
+ "learning_rate": 4.04631322364412e-06,
740
+ "loss": 0.4654,
741
+ "step": 3800
742
+ },
743
+ {
744
+ "epoch": 5.31767955801105,
745
+ "grad_norm": 20.026336669921875,
746
+ "learning_rate": 3.944749136705262e-06,
747
+ "loss": 0.4667,
748
+ "step": 3850
749
+ },
750
+ {
751
+ "epoch": 5.386740331491713,
752
+ "grad_norm": 5.121794700622559,
753
+ "learning_rate": 3.843185049766403e-06,
754
+ "loss": 0.4763,
755
+ "step": 3900
756
+ },
757
+ {
758
+ "epoch": 5.455801104972376,
759
+ "grad_norm": 7.3256516456604,
760
+ "learning_rate": 3.7416209628275447e-06,
761
+ "loss": 0.4418,
762
+ "step": 3950
763
+ },
764
+ {
765
+ "epoch": 5.524861878453039,
766
+ "grad_norm": 7.178898811340332,
767
+ "learning_rate": 3.6400568758886863e-06,
768
+ "loss": 0.467,
769
+ "step": 4000
770
+ },
771
+ {
772
+ "epoch": 5.593922651933702,
773
+ "grad_norm": 7.906302452087402,
774
+ "learning_rate": 3.5384927889498278e-06,
775
+ "loss": 0.426,
776
+ "step": 4050
777
+ },
778
+ {
779
+ "epoch": 5.662983425414365,
780
+ "grad_norm": 4.6435546875,
781
+ "learning_rate": 3.4369287020109693e-06,
782
+ "loss": 0.4585,
783
+ "step": 4100
784
+ },
785
+ {
786
+ "epoch": 5.732044198895028,
787
+ "grad_norm": 6.0284104347229,
788
+ "learning_rate": 3.335364615072111e-06,
789
+ "loss": 0.441,
790
+ "step": 4150
791
+ },
792
+ {
793
+ "epoch": 5.801104972375691,
794
+ "grad_norm": 4.433581829071045,
795
+ "learning_rate": 3.2338005281332523e-06,
796
+ "loss": 0.4645,
797
+ "step": 4200
798
+ },
799
+ {
800
+ "epoch": 5.870165745856354,
801
+ "grad_norm": 4.757821083068848,
802
+ "learning_rate": 3.132236441194394e-06,
803
+ "loss": 0.4652,
804
+ "step": 4250
805
+ },
806
+ {
807
+ "epoch": 5.939226519337017,
808
+ "grad_norm": 6.152480602264404,
809
+ "learning_rate": 3.0306723542555354e-06,
810
+ "loss": 0.4484,
811
+ "step": 4300
812
+ },
813
+ {
814
+ "epoch": 6.0,
815
+ "eval_class_accuracy": {
816
+ "label_0": 0.8723989960327099,
817
+ "label_1": 0.9698000161930208,
818
+ "label_2": 0.8949882600599142
819
+ },
820
+ "eval_class_f1": {
821
+ "label_0": 0.7942021415513189,
822
+ "label_1": 0.7852619458837076,
823
+ "label_2": 0.7991949218145223
824
+ },
825
+ "eval_class_false_positive_rate": {
826
+ "label_0": 0.0631132646880602,
827
+ "label_1": 0.019189024919682277,
828
+ "label_2": 0.06613027664498335
829
+ },
830
+ "eval_class_precision": {
831
+ "label_0": 0.8537338573834924,
832
+ "label_1": 0.7552602436323367,
833
+ "label_2": 0.8113800691606413
834
+ },
835
+ "eval_class_recall": {
836
+ "label_0": 0.742431640625,
837
+ "label_1": 0.8177458033573142,
838
+ "label_2": 0.7873703477730324
839
+ },
840
+ "eval_loss": 0.3344503939151764,
841
+ "eval_macro_f1": 0.7928863364165163,
842
+ "eval_macro_precision": 0.8067913900588235,
843
+ "eval_macro_recall": 0.7825159305851156,
844
+ "eval_micro_f1": 0.7952567175476221,
845
+ "eval_micro_precision": 0.8244833900078472,
846
+ "eval_micro_recall": 0.7680311890838206,
847
+ "eval_runtime": 132.822,
848
+ "eval_samples_per_second": 92.989,
849
+ "eval_steps_per_second": 1.453,
850
+ "eval_subset_accuracy": 0.792324508136993,
851
+ "step": 4344
852
+ },
853
+ {
854
+ "epoch": 6.00828729281768,
855
+ "grad_norm": 7.724102973937988,
856
+ "learning_rate": 2.929108267316677e-06,
857
+ "loss": 0.4457,
858
+ "step": 4350
859
+ },
860
+ {
861
+ "epoch": 6.077348066298343,
862
+ "grad_norm": 7.960653305053711,
863
+ "learning_rate": 2.827544180377819e-06,
864
+ "loss": 0.4507,
865
+ "step": 4400
866
+ },
867
+ {
868
+ "epoch": 6.1464088397790055,
869
+ "grad_norm": 7.553245544433594,
870
+ "learning_rate": 2.7259800934389603e-06,
871
+ "loss": 0.4214,
872
+ "step": 4450
873
+ },
874
+ {
875
+ "epoch": 6.2154696132596685,
876
+ "grad_norm": 13.122285842895508,
877
+ "learning_rate": 2.624416006500102e-06,
878
+ "loss": 0.4561,
879
+ "step": 4500
880
+ },
881
+ {
882
+ "epoch": 6.2845303867403315,
883
+ "grad_norm": 5.380446910858154,
884
+ "learning_rate": 2.5228519195612434e-06,
885
+ "loss": 0.4195,
886
+ "step": 4550
887
+ },
888
+ {
889
+ "epoch": 6.3535911602209945,
890
+ "grad_norm": 8.281150817871094,
891
+ "learning_rate": 2.421287832622385e-06,
892
+ "loss": 0.4411,
893
+ "step": 4600
894
+ },
895
+ {
896
+ "epoch": 6.422651933701657,
897
+ "grad_norm": 5.860757350921631,
898
+ "learning_rate": 2.3197237456835264e-06,
899
+ "loss": 0.461,
900
+ "step": 4650
901
+ },
902
+ {
903
+ "epoch": 6.49171270718232,
904
+ "grad_norm": 5.274202346801758,
905
+ "learning_rate": 2.218159658744668e-06,
906
+ "loss": 0.4469,
907
+ "step": 4700
908
+ },
909
+ {
910
+ "epoch": 6.560773480662983,
911
+ "grad_norm": 3.1023948192596436,
912
+ "learning_rate": 2.1165955718058095e-06,
913
+ "loss": 0.4353,
914
+ "step": 4750
915
+ },
916
+ {
917
+ "epoch": 6.629834254143646,
918
+ "grad_norm": 3.906625270843506,
919
+ "learning_rate": 2.015031484866951e-06,
920
+ "loss": 0.413,
921
+ "step": 4800
922
+ },
923
+ {
924
+ "epoch": 6.698895027624309,
925
+ "grad_norm": 4.5051422119140625,
926
+ "learning_rate": 1.9134673979280925e-06,
927
+ "loss": 0.4255,
928
+ "step": 4850
929
+ },
930
+ {
931
+ "epoch": 6.767955801104972,
932
+ "grad_norm": 6.132110118865967,
933
+ "learning_rate": 1.8119033109892342e-06,
934
+ "loss": 0.4412,
935
+ "step": 4900
936
+ },
937
+ {
938
+ "epoch": 6.837016574585635,
939
+ "grad_norm": 3.9629147052764893,
940
+ "learning_rate": 1.710339224050376e-06,
941
+ "loss": 0.4081,
942
+ "step": 4950
943
+ },
944
+ {
945
+ "epoch": 6.906077348066298,
946
+ "grad_norm": 8.018949508666992,
947
+ "learning_rate": 1.6087751371115177e-06,
948
+ "loss": 0.4497,
949
+ "step": 5000
950
+ },
951
+ {
952
+ "epoch": 6.975138121546961,
953
+ "grad_norm": 4.3800177574157715,
954
+ "learning_rate": 1.5072110501726592e-06,
955
+ "loss": 0.4381,
956
+ "step": 5050
957
+ },
958
+ {
959
+ "epoch": 7.0,
960
+ "eval_class_accuracy": {
961
+ "label_0": 0.8718322403044287,
962
+ "label_1": 0.9686665047364585,
963
+ "label_2": 0.896445631932637
964
+ },
965
+ "eval_class_f1": {
966
+ "label_0": 0.7970772977823356,
967
+ "label_1": 0.7792355961209355,
968
+ "label_2": 0.8016746782446891
969
+ },
970
+ "eval_class_false_positive_rate": {
971
+ "label_0": 0.07219866747424927,
972
+ "label_1": 0.020491447425543066,
973
+ "label_2": 0.0645872368566004
974
+ },
975
+ "eval_class_precision": {
976
+ "label_0": 0.839136302294197,
977
+ "label_1": 0.7431991294885746,
978
+ "label_2": 0.815200252286345
979
+ },
980
+ "eval_class_recall": {
981
+ "label_0": 0.759033203125,
982
+ "label_1": 0.8189448441247003,
983
+ "label_2": 0.7885906040268457
984
+ },
985
+ "eval_loss": 0.34001022577285767,
986
+ "eval_macro_f1": 0.79266252404932,
987
+ "eval_macro_precision": 0.7991785613563721,
988
+ "eval_macro_recall": 0.788856217092182,
989
+ "eval_micro_f1": 0.7969755670811722,
990
+ "eval_micro_precision": 0.8180885182809493,
991
+ "eval_micro_recall": 0.7769249512670565,
992
+ "eval_runtime": 133.0263,
993
+ "eval_samples_per_second": 92.846,
994
+ "eval_steps_per_second": 1.451,
995
+ "eval_subset_accuracy": 0.7941867055299167,
996
+ "step": 5068
997
+ },
998
+ {
999
+ "epoch": 7.044198895027624,
1000
+ "grad_norm": 5.756816387176514,
1001
+ "learning_rate": 1.4056469632338007e-06,
1002
+ "loss": 0.4057,
1003
+ "step": 5100
1004
+ },
1005
+ {
1006
+ "epoch": 7.113259668508287,
1007
+ "grad_norm": 10.789379119873047,
1008
+ "learning_rate": 1.3040828762949422e-06,
1009
+ "loss": 0.4082,
1010
+ "step": 5150
1011
+ },
1012
+ {
1013
+ "epoch": 7.18232044198895,
1014
+ "grad_norm": 6.213741779327393,
1015
+ "learning_rate": 1.2025187893560838e-06,
1016
+ "loss": 0.4371,
1017
+ "step": 5200
1018
+ },
1019
+ {
1020
+ "epoch": 7.251381215469613,
1021
+ "grad_norm": 10.314269065856934,
1022
+ "learning_rate": 1.1009547024172255e-06,
1023
+ "loss": 0.414,
1024
+ "step": 5250
1025
+ },
1026
+ {
1027
+ "epoch": 7.320441988950276,
1028
+ "grad_norm": 4.766229152679443,
1029
+ "learning_rate": 9.99390615478367e-07,
1030
+ "loss": 0.4262,
1031
+ "step": 5300
1032
+ },
1033
+ {
1034
+ "epoch": 7.389502762430939,
1035
+ "grad_norm": 10.421786308288574,
1036
+ "learning_rate": 8.978265285395085e-07,
1037
+ "loss": 0.4089,
1038
+ "step": 5350
1039
+ },
1040
+ {
1041
+ "epoch": 7.458563535911602,
1042
+ "grad_norm": 7.079195976257324,
1043
+ "learning_rate": 7.9626244160065e-07,
1044
+ "loss": 0.4176,
1045
+ "step": 5400
1046
+ },
1047
+ {
1048
+ "epoch": 7.527624309392265,
1049
+ "grad_norm": 11.755300521850586,
1050
+ "learning_rate": 6.946983546617917e-07,
1051
+ "loss": 0.4175,
1052
+ "step": 5450
1053
+ },
1054
+ {
1055
+ "epoch": 7.596685082872928,
1056
+ "grad_norm": 5.82271146774292,
1057
+ "learning_rate": 5.931342677229332e-07,
1058
+ "loss": 0.4028,
1059
+ "step": 5500
1060
+ },
1061
+ {
1062
+ "epoch": 7.665745856353591,
1063
+ "grad_norm": 5.967957019805908,
1064
+ "learning_rate": 4.915701807840748e-07,
1065
+ "loss": 0.4248,
1066
+ "step": 5550
1067
+ },
1068
+ {
1069
+ "epoch": 7.734806629834254,
1070
+ "grad_norm": 8.050609588623047,
1071
+ "learning_rate": 3.900060938452164e-07,
1072
+ "loss": 0.4203,
1073
+ "step": 5600
1074
+ },
1075
+ {
1076
+ "epoch": 7.803867403314917,
1077
+ "grad_norm": 5.390321254730225,
1078
+ "learning_rate": 2.8844200690635795e-07,
1079
+ "loss": 0.4054,
1080
+ "step": 5650
1081
+ },
1082
+ {
1083
+ "epoch": 7.87292817679558,
1084
+ "grad_norm": 4.956524848937988,
1085
+ "learning_rate": 1.868779199674995e-07,
1086
+ "loss": 0.4327,
1087
+ "step": 5700
1088
+ },
1089
+ {
1090
+ "epoch": 7.941988950276243,
1091
+ "grad_norm": 3.822096347808838,
1092
+ "learning_rate": 8.531383302864108e-08,
1093
+ "loss": 0.4183,
1094
+ "step": 5750
1095
+ },
1096
+ {
1097
+ "epoch": 8.0,
1098
+ "eval_class_accuracy": {
1099
+ "label_0": 0.8722370658246296,
1100
+ "label_1": 0.9690713302566594,
1101
+ "label_2": 0.8970123876609182
1102
+ },
1103
+ "eval_class_f1": {
1104
+ "label_0": 0.797640420620672,
1105
+ "label_1": 0.7791907514450868,
1106
+ "label_2": 0.8029129222187791
1107
+ },
1108
+ "eval_class_false_positive_rate": {
1109
+ "label_0": 0.07171411265898586,
1110
+ "label_1": 0.019275853086739663,
1111
+ "label_2": 0.06447701972885876
1112
+ },
1113
+ "eval_class_precision": {
1114
+ "label_0": 0.8400864397622907,
1115
+ "label_1": 0.7522321428571429,
1116
+ "label_2": 0.8158060453400504
1117
+ },
1118
+ "eval_class_recall": {
1119
+ "label_0": 0.75927734375,
1120
+ "label_1": 0.8081534772182254,
1121
+ "label_2": 0.7904209884075656
1122
+ },
1123
+ "eval_loss": 0.34776991605758667,
1124
+ "eval_macro_f1": 0.7932480314281793,
1125
+ "eval_macro_precision": 0.8027082093198281,
1126
+ "eval_macro_recall": 0.7859506031252637,
1127
+ "eval_micro_f1": 0.7977724940558127,
1128
+ "eval_micro_precision": 0.8200411628505274,
1129
+ "eval_micro_recall": 0.7766812865497076,
1130
+ "eval_runtime": 132.8386,
1131
+ "eval_samples_per_second": 92.977,
1132
+ "eval_steps_per_second": 1.453,
1133
+ "eval_subset_accuracy": 0.7951582867783985,
1134
+ "step": 5792
1135
+ }
1136
+ ],
1137
+ "logging_steps": 50,
1138
+ "max_steps": 5792,
1139
+ "num_input_tokens_seen": 0,
1140
+ "num_train_epochs": 8,
1141
+ "save_steps": 500,
1142
+ "stateful_callbacks": {
1143
+ "TrainerControl": {
1144
+ "args": {
1145
+ "should_epoch_stop": false,
1146
+ "should_evaluate": false,
1147
+ "should_log": false,
1148
+ "should_save": true,
1149
+ "should_training_stop": true
1150
+ },
1151
+ "attributes": {}
1152
+ }
1153
+ },
1154
+ "total_flos": 1.9496632912293888e+17,
1155
+ "train_batch_size": 64,
1156
+ "trial_name": null,
1157
+ "trial_params": null
1158
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:091514c36d0fd173ba63baa2ea42184f6fd20ba27cfe14f5732e4ddbf5b83c16
3
+ size 5304