mrm8488 commited on
Commit
9c0a241
1 Parent(s): 6e4b4f5

Initial commit from mrm8488

Browse files
README.md ADDED
@@ -0,0 +1,79 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - en
4
+ license: mit
5
+ tags:
6
+ - generated_from_trainer
7
+ datasets:
8
+ - glue
9
+ metrics:
10
+ - accuracy
11
+ model-index:
12
+ - name: deberta-v3-large-mnli-2
13
+ results:
14
+ - task:
15
+ name: Text Classification
16
+ type: text-classification
17
+ dataset:
18
+ name: GLUE MNLI
19
+ type: glue
20
+ args: mnli
21
+ metrics:
22
+ - name: Accuracy
23
+ type: accuracy
24
+ value: 0.8949349064279902
25
+ ---
26
+
27
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
28
+ should probably proofread and complete it, then remove this comment. -->
29
+
30
+ # deberta-v3-large-mnli-2
31
+
32
+ This model is a fine-tuned version of [microsoft/deberta-v3-large](https://huggingface.co/microsoft/deberta-v3-large) on the GLUE MNLI dataset.
33
+ It achieves the following results on the evaluation set:
34
+ - Loss: 0.6763
35
+ - Accuracy: 0.8949
36
+
37
+ ## Model description
38
+
39
+ More information needed
40
+
41
+ ## Intended uses & limitations
42
+
43
+ More information needed
44
+
45
+ ## Training and evaluation data
46
+
47
+ More information needed
48
+
49
+ ## Training procedure
50
+
51
+ ### Training hyperparameters
52
+
53
+ The following hyperparameters were used during training:
54
+ - learning_rate: 3e-05
55
+ - train_batch_size: 16
56
+ - eval_batch_size: 8
57
+ - seed: 42
58
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
59
+ - lr_scheduler_type: linear
60
+ - num_epochs: 5.0
61
+ - mixed_precision_training: Native AMP
62
+
63
+ ### Training results
64
+
65
+ | Training Loss | Epoch | Step | Validation Loss | Accuracy |
66
+ |:-------------:|:-----:|:------:|:---------------:|:--------:|
67
+ | 0.3676 | 1.0 | 24544 | 0.3761 | 0.8681 |
68
+ | 0.2782 | 2.0 | 49088 | 0.3605 | 0.8881 |
69
+ | 0.1986 | 3.0 | 73632 | 0.4672 | 0.8894 |
70
+ | 0.1299 | 4.0 | 98176 | 0.5248 | 0.8967 |
71
+ | 0.0643 | 5.0 | 122720 | 0.6489 | 0.8999 |
72
+
73
+
74
+ ### Framework versions
75
+
76
+ - Transformers 4.13.0.dev0
77
+ - Pytorch 1.10.0+cu111
78
+ - Datasets 1.16.1
79
+ - Tokenizers 0.10.3
added_tokens.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"[MASK]": 128000}
all_results.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 5.0,
3
+ "eval_accuracy": 0.8949349064279902,
4
+ "eval_loss": 0.6762520670890808,
5
+ "eval_runtime": 67.7704,
6
+ "eval_samples": 9832,
7
+ "eval_samples_per_second": 145.078,
8
+ "eval_steps_per_second": 18.135,
9
+ "train_loss": 0.2137734965519613,
10
+ "train_runtime": 32769.5489,
11
+ "train_samples": 392702,
12
+ "train_samples_per_second": 59.919,
13
+ "train_steps_per_second": 3.745
14
+ }
config.json ADDED
@@ -0,0 +1,46 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "microsoft/deberta-v3-large",
3
+ "architectures": [
4
+ "DebertaV2ForSequenceClassification"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.1,
7
+ "finetuning_task": "mnli",
8
+ "hidden_act": "gelu",
9
+ "hidden_dropout_prob": 0.1,
10
+ "hidden_size": 1024,
11
+ "id2label": {
12
+ "0": "entailment",
13
+ "1": "neutral",
14
+ "2": "contradiction"
15
+ },
16
+ "initializer_range": 0.02,
17
+ "intermediate_size": 4096,
18
+ "label2id": {
19
+ "contradiction": 2,
20
+ "entailment": 0,
21
+ "neutral": 1
22
+ },
23
+ "layer_norm_eps": 1e-07,
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": 16,
29
+ "num_hidden_layers": 24,
30
+ "pad_token_id": 0,
31
+ "pooler_dropout": 0,
32
+ "pooler_hidden_act": "gelu",
33
+ "pooler_hidden_size": 1024,
34
+ "pos_att_type": [
35
+ "p2c",
36
+ "c2p"
37
+ ],
38
+ "position_biased_input": false,
39
+ "position_buckets": 256,
40
+ "relative_attention": true,
41
+ "share_att_key": true,
42
+ "torch_dtype": "float32",
43
+ "transformers_version": "4.13.0.dev0",
44
+ "type_vocab_size": 0,
45
+ "vocab_size": 128100
46
+ }
eval_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 5.0,
3
+ "eval_accuracy": 0.8949349064279902,
4
+ "eval_loss": 0.6762520670890808,
5
+ "eval_runtime": 67.7704,
6
+ "eval_samples": 9832,
7
+ "eval_samples_per_second": 145.078,
8
+ "eval_steps_per_second": 18.135
9
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:862fb1da1706e5f73d3b814d8791d9f1c62132f1baab9e9a9cadfc01abf04bbf
3
+ size 1740426539
special_tokens_map.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"bos_token": "[CLS]", "eos_token": "[SEP]", "unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]"}
spm.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c679fbf93643d19aab7ee10c0b99e460bdbc02fedf34b92b05af343b4af586fd
3
+ size 2464616
tokenizer_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"do_lower_case": false, "bos_token": "[CLS]", "eos_token": "[SEP]", "unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]", "split_by_punct": false, "sp_model_kwargs": {}, "vocab_type": "spm", "special_tokens_map_file": null, "tokenizer_file": null, "name_or_path": "microsoft/deberta-v3-large", "tokenizer_class": "DebertaV2Tokenizer"}
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 5.0,
3
+ "train_loss": 0.2137734965519613,
4
+ "train_runtime": 32769.5489,
5
+ "train_samples": 392702,
6
+ "train_samples_per_second": 59.919,
7
+ "train_steps_per_second": 3.745
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,802 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 5.0,
5
+ "global_step": 122720,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.04,
12
+ "learning_rate": 2.975700782268579e-05,
13
+ "loss": 0.5382,
14
+ "step": 1000
15
+ },
16
+ {
17
+ "epoch": 0.08,
18
+ "learning_rate": 2.951254889178618e-05,
19
+ "loss": 0.4368,
20
+ "step": 2000
21
+ },
22
+ {
23
+ "epoch": 0.12,
24
+ "learning_rate": 2.9268334419817473e-05,
25
+ "loss": 0.4277,
26
+ "step": 3000
27
+ },
28
+ {
29
+ "epoch": 0.16,
30
+ "learning_rate": 2.9023875488917863e-05,
31
+ "loss": 0.4026,
32
+ "step": 4000
33
+ },
34
+ {
35
+ "epoch": 0.2,
36
+ "learning_rate": 2.877966101694915e-05,
37
+ "loss": 0.4137,
38
+ "step": 5000
39
+ },
40
+ {
41
+ "epoch": 0.24,
42
+ "learning_rate": 2.8535202086049545e-05,
43
+ "loss": 0.4048,
44
+ "step": 6000
45
+ },
46
+ {
47
+ "epoch": 0.29,
48
+ "learning_rate": 2.8290743155149935e-05,
49
+ "loss": 0.3949,
50
+ "step": 7000
51
+ },
52
+ {
53
+ "epoch": 0.33,
54
+ "learning_rate": 2.804628422425033e-05,
55
+ "loss": 0.3883,
56
+ "step": 8000
57
+ },
58
+ {
59
+ "epoch": 0.37,
60
+ "learning_rate": 2.780182529335072e-05,
61
+ "loss": 0.3759,
62
+ "step": 9000
63
+ },
64
+ {
65
+ "epoch": 0.41,
66
+ "learning_rate": 2.755736636245111e-05,
67
+ "loss": 0.3868,
68
+ "step": 10000
69
+ },
70
+ {
71
+ "epoch": 0.45,
72
+ "learning_rate": 2.73129074315515e-05,
73
+ "loss": 0.3809,
74
+ "step": 11000
75
+ },
76
+ {
77
+ "epoch": 0.49,
78
+ "learning_rate": 2.7068448500651893e-05,
79
+ "loss": 0.3754,
80
+ "step": 12000
81
+ },
82
+ {
83
+ "epoch": 0.53,
84
+ "learning_rate": 2.6823989569752283e-05,
85
+ "loss": 0.3733,
86
+ "step": 13000
87
+ },
88
+ {
89
+ "epoch": 0.57,
90
+ "learning_rate": 2.6579530638852673e-05,
91
+ "loss": 0.3825,
92
+ "step": 14000
93
+ },
94
+ {
95
+ "epoch": 0.61,
96
+ "learning_rate": 2.6335071707953063e-05,
97
+ "loss": 0.3719,
98
+ "step": 15000
99
+ },
100
+ {
101
+ "epoch": 0.65,
102
+ "learning_rate": 2.6091101694915253e-05,
103
+ "loss": 0.3725,
104
+ "step": 16000
105
+ },
106
+ {
107
+ "epoch": 0.69,
108
+ "learning_rate": 2.5846642764015644e-05,
109
+ "loss": 0.3694,
110
+ "step": 17000
111
+ },
112
+ {
113
+ "epoch": 0.73,
114
+ "learning_rate": 2.5602428292046935e-05,
115
+ "loss": 0.3714,
116
+ "step": 18000
117
+ },
118
+ {
119
+ "epoch": 0.77,
120
+ "learning_rate": 2.5358213820078227e-05,
121
+ "loss": 0.3818,
122
+ "step": 19000
123
+ },
124
+ {
125
+ "epoch": 0.81,
126
+ "learning_rate": 2.511399934810952e-05,
127
+ "loss": 0.374,
128
+ "step": 20000
129
+ },
130
+ {
131
+ "epoch": 0.86,
132
+ "learning_rate": 2.486954041720991e-05,
133
+ "loss": 0.4574,
134
+ "step": 21000
135
+ },
136
+ {
137
+ "epoch": 0.9,
138
+ "learning_rate": 2.46250814863103e-05,
139
+ "loss": 0.426,
140
+ "step": 22000
141
+ },
142
+ {
143
+ "epoch": 0.94,
144
+ "learning_rate": 2.4380622555410693e-05,
145
+ "loss": 0.3714,
146
+ "step": 23000
147
+ },
148
+ {
149
+ "epoch": 0.98,
150
+ "learning_rate": 2.4136163624511083e-05,
151
+ "loss": 0.3676,
152
+ "step": 24000
153
+ },
154
+ {
155
+ "epoch": 1.0,
156
+ "eval_accuracy": 0.8680590932246561,
157
+ "eval_loss": 0.3760605454444885,
158
+ "eval_runtime": 67.1189,
159
+ "eval_samples_per_second": 146.233,
160
+ "eval_steps_per_second": 18.281,
161
+ "step": 24544
162
+ },
163
+ {
164
+ "epoch": 1.02,
165
+ "learning_rate": 2.3892193611473273e-05,
166
+ "loss": 0.3246,
167
+ "step": 25000
168
+ },
169
+ {
170
+ "epoch": 1.06,
171
+ "learning_rate": 2.3647734680573663e-05,
172
+ "loss": 0.2698,
173
+ "step": 26000
174
+ },
175
+ {
176
+ "epoch": 1.1,
177
+ "learning_rate": 2.3403275749674057e-05,
178
+ "loss": 0.2783,
179
+ "step": 27000
180
+ },
181
+ {
182
+ "epoch": 1.14,
183
+ "learning_rate": 2.3158816818774447e-05,
184
+ "loss": 0.2761,
185
+ "step": 28000
186
+ },
187
+ {
188
+ "epoch": 1.18,
189
+ "learning_rate": 2.2914357887874837e-05,
190
+ "loss": 0.2831,
191
+ "step": 29000
192
+ },
193
+ {
194
+ "epoch": 1.22,
195
+ "learning_rate": 2.2669898956975228e-05,
196
+ "loss": 0.2819,
197
+ "step": 30000
198
+ },
199
+ {
200
+ "epoch": 1.26,
201
+ "learning_rate": 2.2425440026075618e-05,
202
+ "loss": 0.2822,
203
+ "step": 31000
204
+ },
205
+ {
206
+ "epoch": 1.3,
207
+ "learning_rate": 2.2180981095176008e-05,
208
+ "loss": 0.2769,
209
+ "step": 32000
210
+ },
211
+ {
212
+ "epoch": 1.34,
213
+ "learning_rate": 2.1936522164276405e-05,
214
+ "loss": 0.2853,
215
+ "step": 33000
216
+ },
217
+ {
218
+ "epoch": 1.39,
219
+ "learning_rate": 2.1692063233376795e-05,
220
+ "loss": 0.2891,
221
+ "step": 34000
222
+ },
223
+ {
224
+ "epoch": 1.43,
225
+ "learning_rate": 2.1447848761408084e-05,
226
+ "loss": 0.281,
227
+ "step": 35000
228
+ },
229
+ {
230
+ "epoch": 1.47,
231
+ "learning_rate": 2.1203389830508474e-05,
232
+ "loss": 0.2733,
233
+ "step": 36000
234
+ },
235
+ {
236
+ "epoch": 1.51,
237
+ "learning_rate": 2.0959175358539766e-05,
238
+ "loss": 0.2779,
239
+ "step": 37000
240
+ },
241
+ {
242
+ "epoch": 1.55,
243
+ "learning_rate": 2.0714716427640156e-05,
244
+ "loss": 0.2799,
245
+ "step": 38000
246
+ },
247
+ {
248
+ "epoch": 1.59,
249
+ "learning_rate": 2.0470257496740546e-05,
250
+ "loss": 0.2824,
251
+ "step": 39000
252
+ },
253
+ {
254
+ "epoch": 1.63,
255
+ "learning_rate": 2.022628748370274e-05,
256
+ "loss": 0.2823,
257
+ "step": 40000
258
+ },
259
+ {
260
+ "epoch": 1.67,
261
+ "learning_rate": 1.998182855280313e-05,
262
+ "loss": 0.2789,
263
+ "step": 41000
264
+ },
265
+ {
266
+ "epoch": 1.71,
267
+ "learning_rate": 1.973736962190352e-05,
268
+ "loss": 0.2806,
269
+ "step": 42000
270
+ },
271
+ {
272
+ "epoch": 1.75,
273
+ "learning_rate": 1.9492910691003913e-05,
274
+ "loss": 0.2808,
275
+ "step": 43000
276
+ },
277
+ {
278
+ "epoch": 1.79,
279
+ "learning_rate": 1.9248696219035202e-05,
280
+ "loss": 0.2745,
281
+ "step": 44000
282
+ },
283
+ {
284
+ "epoch": 1.83,
285
+ "learning_rate": 1.9004481747066494e-05,
286
+ "loss": 0.2918,
287
+ "step": 45000
288
+ },
289
+ {
290
+ "epoch": 1.87,
291
+ "learning_rate": 1.8760511734028684e-05,
292
+ "loss": 0.2789,
293
+ "step": 46000
294
+ },
295
+ {
296
+ "epoch": 1.91,
297
+ "learning_rate": 1.8516052803129074e-05,
298
+ "loss": 0.2754,
299
+ "step": 47000
300
+ },
301
+ {
302
+ "epoch": 1.96,
303
+ "learning_rate": 1.8271593872229467e-05,
304
+ "loss": 0.2709,
305
+ "step": 48000
306
+ },
307
+ {
308
+ "epoch": 2.0,
309
+ "learning_rate": 1.8027134941329858e-05,
310
+ "loss": 0.2782,
311
+ "step": 49000
312
+ },
313
+ {
314
+ "epoch": 2.0,
315
+ "eval_accuracy": 0.8881304126337239,
316
+ "eval_loss": 0.36052629351615906,
317
+ "eval_runtime": 68.3607,
318
+ "eval_samples_per_second": 143.577,
319
+ "eval_steps_per_second": 17.949,
320
+ "step": 49088
321
+ },
322
+ {
323
+ "epoch": 2.04,
324
+ "learning_rate": 1.7782676010430248e-05,
325
+ "loss": 0.2017,
326
+ "step": 50000
327
+ },
328
+ {
329
+ "epoch": 2.08,
330
+ "learning_rate": 1.7538217079530638e-05,
331
+ "loss": 0.1921,
332
+ "step": 51000
333
+ },
334
+ {
335
+ "epoch": 2.12,
336
+ "learning_rate": 1.7293758148631028e-05,
337
+ "loss": 0.2017,
338
+ "step": 52000
339
+ },
340
+ {
341
+ "epoch": 2.16,
342
+ "learning_rate": 1.704929921773142e-05,
343
+ "loss": 0.1894,
344
+ "step": 53000
345
+ },
346
+ {
347
+ "epoch": 2.2,
348
+ "learning_rate": 1.6805084745762714e-05,
349
+ "loss": 0.1978,
350
+ "step": 54000
351
+ },
352
+ {
353
+ "epoch": 2.24,
354
+ "learning_rate": 1.6560625814863104e-05,
355
+ "loss": 0.1983,
356
+ "step": 55000
357
+ },
358
+ {
359
+ "epoch": 2.28,
360
+ "learning_rate": 1.6316166883963494e-05,
361
+ "loss": 0.2003,
362
+ "step": 56000
363
+ },
364
+ {
365
+ "epoch": 2.32,
366
+ "learning_rate": 1.6071707953063884e-05,
367
+ "loss": 0.2059,
368
+ "step": 57000
369
+ },
370
+ {
371
+ "epoch": 2.36,
372
+ "learning_rate": 1.5827249022164274e-05,
373
+ "loss": 0.1971,
374
+ "step": 58000
375
+ },
376
+ {
377
+ "epoch": 2.4,
378
+ "learning_rate": 1.5582790091264668e-05,
379
+ "loss": 0.2014,
380
+ "step": 59000
381
+ },
382
+ {
383
+ "epoch": 2.44,
384
+ "learning_rate": 1.533833116036506e-05,
385
+ "loss": 0.1922,
386
+ "step": 60000
387
+ },
388
+ {
389
+ "epoch": 2.49,
390
+ "learning_rate": 1.5094116688396348e-05,
391
+ "loss": 0.1927,
392
+ "step": 61000
393
+ },
394
+ {
395
+ "epoch": 2.53,
396
+ "learning_rate": 1.4849657757496742e-05,
397
+ "loss": 0.1965,
398
+ "step": 62000
399
+ },
400
+ {
401
+ "epoch": 2.57,
402
+ "learning_rate": 1.4605443285528032e-05,
403
+ "loss": 0.1971,
404
+ "step": 63000
405
+ },
406
+ {
407
+ "epoch": 2.61,
408
+ "learning_rate": 1.4360984354628422e-05,
409
+ "loss": 0.2095,
410
+ "step": 64000
411
+ },
412
+ {
413
+ "epoch": 2.65,
414
+ "learning_rate": 1.4116769882659712e-05,
415
+ "loss": 0.199,
416
+ "step": 65000
417
+ },
418
+ {
419
+ "epoch": 2.69,
420
+ "learning_rate": 1.3872555410691004e-05,
421
+ "loss": 0.2072,
422
+ "step": 66000
423
+ },
424
+ {
425
+ "epoch": 2.73,
426
+ "learning_rate": 1.3628096479791396e-05,
427
+ "loss": 0.1955,
428
+ "step": 67000
429
+ },
430
+ {
431
+ "epoch": 2.77,
432
+ "learning_rate": 1.3383637548891786e-05,
433
+ "loss": 0.1924,
434
+ "step": 68000
435
+ },
436
+ {
437
+ "epoch": 2.81,
438
+ "learning_rate": 1.3139178617992178e-05,
439
+ "loss": 0.2055,
440
+ "step": 69000
441
+ },
442
+ {
443
+ "epoch": 2.85,
444
+ "learning_rate": 1.2894719687092568e-05,
445
+ "loss": 0.2016,
446
+ "step": 70000
447
+ },
448
+ {
449
+ "epoch": 2.89,
450
+ "learning_rate": 1.265026075619296e-05,
451
+ "loss": 0.2011,
452
+ "step": 71000
453
+ },
454
+ {
455
+ "epoch": 2.93,
456
+ "learning_rate": 1.2405801825293352e-05,
457
+ "loss": 0.1915,
458
+ "step": 72000
459
+ },
460
+ {
461
+ "epoch": 2.97,
462
+ "learning_rate": 1.2161587353324642e-05,
463
+ "loss": 0.1986,
464
+ "step": 73000
465
+ },
466
+ {
467
+ "epoch": 3.0,
468
+ "eval_accuracy": 0.8893530310748854,
469
+ "eval_loss": 0.4672204256057739,
470
+ "eval_runtime": 67.0424,
471
+ "eval_samples_per_second": 146.4,
472
+ "eval_steps_per_second": 18.302,
473
+ "step": 73632
474
+ },
475
+ {
476
+ "epoch": 3.01,
477
+ "learning_rate": 1.1917128422425032e-05,
478
+ "loss": 0.1697,
479
+ "step": 74000
480
+ },
481
+ {
482
+ "epoch": 3.06,
483
+ "learning_rate": 1.1672669491525423e-05,
484
+ "loss": 0.1246,
485
+ "step": 75000
486
+ },
487
+ {
488
+ "epoch": 3.1,
489
+ "learning_rate": 1.1428455019556716e-05,
490
+ "loss": 0.1231,
491
+ "step": 76000
492
+ },
493
+ {
494
+ "epoch": 3.14,
495
+ "learning_rate": 1.1183996088657106e-05,
496
+ "loss": 0.1314,
497
+ "step": 77000
498
+ },
499
+ {
500
+ "epoch": 3.18,
501
+ "learning_rate": 1.0939537157757497e-05,
502
+ "loss": 0.1272,
503
+ "step": 78000
504
+ },
505
+ {
506
+ "epoch": 3.22,
507
+ "learning_rate": 1.0695567144719687e-05,
508
+ "loss": 0.1257,
509
+ "step": 79000
510
+ },
511
+ {
512
+ "epoch": 3.26,
513
+ "learning_rate": 1.0451108213820078e-05,
514
+ "loss": 0.1278,
515
+ "step": 80000
516
+ },
517
+ {
518
+ "epoch": 3.3,
519
+ "learning_rate": 1.0206893741851369e-05,
520
+ "loss": 0.1365,
521
+ "step": 81000
522
+ },
523
+ {
524
+ "epoch": 3.34,
525
+ "learning_rate": 9.96243481095176e-06,
526
+ "loss": 0.1284,
527
+ "step": 82000
528
+ },
529
+ {
530
+ "epoch": 3.38,
531
+ "learning_rate": 9.717975880052152e-06,
532
+ "loss": 0.1435,
533
+ "step": 83000
534
+ },
535
+ {
536
+ "epoch": 3.42,
537
+ "learning_rate": 9.473761408083443e-06,
538
+ "loss": 0.1386,
539
+ "step": 84000
540
+ },
541
+ {
542
+ "epoch": 3.46,
543
+ "learning_rate": 9.229546936114733e-06,
544
+ "loss": 0.1301,
545
+ "step": 85000
546
+ },
547
+ {
548
+ "epoch": 3.5,
549
+ "learning_rate": 8.985088005215123e-06,
550
+ "loss": 0.1261,
551
+ "step": 86000
552
+ },
553
+ {
554
+ "epoch": 3.54,
555
+ "learning_rate": 8.740629074315516e-06,
556
+ "loss": 0.1291,
557
+ "step": 87000
558
+ },
559
+ {
560
+ "epoch": 3.59,
561
+ "learning_rate": 8.496170143415907e-06,
562
+ "loss": 0.1281,
563
+ "step": 88000
564
+ },
565
+ {
566
+ "epoch": 3.63,
567
+ "learning_rate": 8.251711212516297e-06,
568
+ "loss": 0.1166,
569
+ "step": 89000
570
+ },
571
+ {
572
+ "epoch": 3.67,
573
+ "learning_rate": 8.007252281616689e-06,
574
+ "loss": 0.124,
575
+ "step": 90000
576
+ },
577
+ {
578
+ "epoch": 3.71,
579
+ "learning_rate": 7.762793350717079e-06,
580
+ "loss": 0.1257,
581
+ "step": 91000
582
+ },
583
+ {
584
+ "epoch": 3.75,
585
+ "learning_rate": 7.518334419817472e-06,
586
+ "loss": 0.1232,
587
+ "step": 92000
588
+ },
589
+ {
590
+ "epoch": 3.79,
591
+ "learning_rate": 7.274119947848762e-06,
592
+ "loss": 0.1271,
593
+ "step": 93000
594
+ },
595
+ {
596
+ "epoch": 3.83,
597
+ "learning_rate": 7.029905475880053e-06,
598
+ "loss": 0.128,
599
+ "step": 94000
600
+ },
601
+ {
602
+ "epoch": 3.87,
603
+ "learning_rate": 6.785446544980443e-06,
604
+ "loss": 0.1212,
605
+ "step": 95000
606
+ },
607
+ {
608
+ "epoch": 3.91,
609
+ "learning_rate": 6.540987614080835e-06,
610
+ "loss": 0.1193,
611
+ "step": 96000
612
+ },
613
+ {
614
+ "epoch": 3.95,
615
+ "learning_rate": 6.296773142112125e-06,
616
+ "loss": 0.1199,
617
+ "step": 97000
618
+ },
619
+ {
620
+ "epoch": 3.99,
621
+ "learning_rate": 6.052314211212517e-06,
622
+ "loss": 0.1299,
623
+ "step": 98000
624
+ },
625
+ {
626
+ "epoch": 4.0,
627
+ "eval_accuracy": 0.8966887417218543,
628
+ "eval_loss": 0.5247582197189331,
629
+ "eval_runtime": 66.9471,
630
+ "eval_samples_per_second": 146.608,
631
+ "eval_steps_per_second": 18.328,
632
+ "step": 98176
633
+ },
634
+ {
635
+ "epoch": 4.03,
636
+ "learning_rate": 5.807855280312908e-06,
637
+ "loss": 0.0766,
638
+ "step": 99000
639
+ },
640
+ {
641
+ "epoch": 4.07,
642
+ "learning_rate": 5.563640808344198e-06,
643
+ "loss": 0.0642,
644
+ "step": 100000
645
+ },
646
+ {
647
+ "epoch": 4.12,
648
+ "learning_rate": 5.31918187744459e-06,
649
+ "loss": 0.0744,
650
+ "step": 101000
651
+ },
652
+ {
653
+ "epoch": 4.16,
654
+ "learning_rate": 5.07496740547588e-06,
655
+ "loss": 0.0684,
656
+ "step": 102000
657
+ },
658
+ {
659
+ "epoch": 4.2,
660
+ "learning_rate": 4.830508474576271e-06,
661
+ "loss": 0.0732,
662
+ "step": 103000
663
+ },
664
+ {
665
+ "epoch": 4.24,
666
+ "learning_rate": 4.586294002607562e-06,
667
+ "loss": 0.0637,
668
+ "step": 104000
669
+ },
670
+ {
671
+ "epoch": 4.28,
672
+ "learning_rate": 4.342323989569753e-06,
673
+ "loss": 0.071,
674
+ "step": 105000
675
+ },
676
+ {
677
+ "epoch": 4.32,
678
+ "learning_rate": 4.097865058670143e-06,
679
+ "loss": 0.066,
680
+ "step": 106000
681
+ },
682
+ {
683
+ "epoch": 4.36,
684
+ "learning_rate": 3.853406127770535e-06,
685
+ "loss": 0.0715,
686
+ "step": 107000
687
+ },
688
+ {
689
+ "epoch": 4.4,
690
+ "learning_rate": 3.6089471968709256e-06,
691
+ "loss": 0.069,
692
+ "step": 108000
693
+ },
694
+ {
695
+ "epoch": 4.44,
696
+ "learning_rate": 3.3647327249022166e-06,
697
+ "loss": 0.0707,
698
+ "step": 109000
699
+ },
700
+ {
701
+ "epoch": 4.48,
702
+ "learning_rate": 3.1202737940026076e-06,
703
+ "loss": 0.0702,
704
+ "step": 110000
705
+ },
706
+ {
707
+ "epoch": 4.52,
708
+ "learning_rate": 2.875814863102999e-06,
709
+ "loss": 0.0632,
710
+ "step": 111000
711
+ },
712
+ {
713
+ "epoch": 4.56,
714
+ "learning_rate": 2.63135593220339e-06,
715
+ "loss": 0.0672,
716
+ "step": 112000
717
+ },
718
+ {
719
+ "epoch": 4.6,
720
+ "learning_rate": 2.3871414602346807e-06,
721
+ "loss": 0.0665,
722
+ "step": 113000
723
+ },
724
+ {
725
+ "epoch": 4.64,
726
+ "learning_rate": 2.1426825293350717e-06,
727
+ "loss": 0.0637,
728
+ "step": 114000
729
+ },
730
+ {
731
+ "epoch": 4.69,
732
+ "learning_rate": 1.898223598435463e-06,
733
+ "loss": 0.0712,
734
+ "step": 115000
735
+ },
736
+ {
737
+ "epoch": 4.73,
738
+ "learning_rate": 1.653764667535854e-06,
739
+ "loss": 0.064,
740
+ "step": 116000
741
+ },
742
+ {
743
+ "epoch": 4.77,
744
+ "learning_rate": 1.4093057366362453e-06,
745
+ "loss": 0.0644,
746
+ "step": 117000
747
+ },
748
+ {
749
+ "epoch": 4.81,
750
+ "learning_rate": 1.1653357235984355e-06,
751
+ "loss": 0.0683,
752
+ "step": 118000
753
+ },
754
+ {
755
+ "epoch": 4.85,
756
+ "learning_rate": 9.208767926988267e-07,
757
+ "loss": 0.0622,
758
+ "step": 119000
759
+ },
760
+ {
761
+ "epoch": 4.89,
762
+ "learning_rate": 6.766623207301173e-07,
763
+ "loss": 0.0678,
764
+ "step": 120000
765
+ },
766
+ {
767
+ "epoch": 4.93,
768
+ "learning_rate": 4.322033898305085e-07,
769
+ "loss": 0.0578,
770
+ "step": 121000
771
+ },
772
+ {
773
+ "epoch": 4.97,
774
+ "learning_rate": 1.877444589308996e-07,
775
+ "loss": 0.0643,
776
+ "step": 122000
777
+ },
778
+ {
779
+ "epoch": 5.0,
780
+ "eval_accuracy": 0.8999490575649516,
781
+ "eval_loss": 0.6488531231880188,
782
+ "eval_runtime": 67.1753,
783
+ "eval_samples_per_second": 146.11,
784
+ "eval_steps_per_second": 18.266,
785
+ "step": 122720
786
+ },
787
+ {
788
+ "epoch": 5.0,
789
+ "step": 122720,
790
+ "total_flos": 4.5746877181130496e+17,
791
+ "train_loss": 0.2137734965519613,
792
+ "train_runtime": 32769.5489,
793
+ "train_samples_per_second": 59.919,
794
+ "train_steps_per_second": 3.745
795
+ }
796
+ ],
797
+ "max_steps": 122720,
798
+ "num_train_epochs": 5,
799
+ "total_flos": 4.5746877181130496e+17,
800
+ "trial_name": null,
801
+ "trial_params": null
802
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:15c3a40d9609cb3972f15030c64ec42efb005c428519957b0c8e74edf8840568
3
+ size 2927