tmnam20 commited on
Commit
04b382f
1 Parent(s): 8f265f8

Training in progress, step 23000, checkpoint

Browse files
checkpoint-23000/added_tokens.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "<bos>": 40478,
3
+ "<eos>": 40479,
4
+ "<pad>": 40480
5
+ }
checkpoint-23000/config.json ADDED
@@ -0,0 +1,46 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "openai-gpt",
3
+ "afn": "gelu",
4
+ "architectures": [
5
+ "OpenAIGPTForSequenceClassification"
6
+ ],
7
+ "attn_pdrop": 0.1,
8
+ "embd_pdrop": 0.1,
9
+ "id2label": {
10
+ "0": "entailment",
11
+ "1": "neutral",
12
+ "2": "contradiction"
13
+ },
14
+ "initializer_range": 0.02,
15
+ "label2id": {
16
+ "contradiction": 2,
17
+ "entailment": 0,
18
+ "neutral": 1
19
+ },
20
+ "layer_norm_epsilon": 1e-05,
21
+ "model_type": "openai-gpt",
22
+ "n_ctx": 512,
23
+ "n_embd": 768,
24
+ "n_head": 12,
25
+ "n_layer": 4,
26
+ "n_positions": 512,
27
+ "n_special": 0,
28
+ "pad_token_id": 40480,
29
+ "predict_special_tokens": true,
30
+ "problem_type": "single_label_classification",
31
+ "resid_pdrop": 0.1,
32
+ "summary_activation": null,
33
+ "summary_first_dropout": 0.1,
34
+ "summary_proj_to_labels": true,
35
+ "summary_type": "last",
36
+ "summary_use_proj": true,
37
+ "task_specific_params": {
38
+ "text-generation": {
39
+ "do_sample": true,
40
+ "max_length": 50
41
+ }
42
+ },
43
+ "torch_dtype": "float32",
44
+ "transformers_version": "4.33.0.dev0",
45
+ "vocab_size": 40481
46
+ }
checkpoint-23000/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-23000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8763dbd137971c52419809faa993cd6a92ae8e41cf4e9ec7ee6a6f941cfde6aa
3
+ size 478721221
checkpoint-23000/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:927daa70c80f93259f6a78734ffc4e66fe576e624c9b578feb3cc188d035644e
3
+ size 239361994
checkpoint-23000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3d7c569b31ac64d19f13f0571c9a324df2ac2ce49244b198c7065ae0d3108535
3
+ size 14575
checkpoint-23000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:12056657fb01aa91fa2bc3df57968ac5273971ff7ca5bf1d33cf2c2196f7f334
3
+ size 627
checkpoint-23000/special_tokens_map.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<bos>",
3
+ "eos_token": "<eos>",
4
+ "pad_token": "<pad>",
5
+ "unk_token": "<unk>"
6
+ }
checkpoint-23000/tokenizer_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "clean_up_tokenization_spaces": true,
3
+ "model_max_length": 512,
4
+ "padding_side": "left",
5
+ "tokenizer_class": "OpenAIGPTTokenizer",
6
+ "unk_token": "<unk>"
7
+ }
checkpoint-23000/trainer_state.json ADDED
@@ -0,0 +1,370 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 1.0591468811035156,
3
+ "best_model_checkpoint": "/kaggle/output/checkpoint-23000",
4
+ "epoch": 0.9370925684485006,
5
+ "eval_steps": 1000,
6
+ "global_step": 23000,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.0,
13
+ "learning_rate": 2.7777777777777777e-11,
14
+ "loss": 1.2184,
15
+ "step": 1
16
+ },
17
+ {
18
+ "epoch": 0.04,
19
+ "learning_rate": 2.7750000000000004e-08,
20
+ "loss": 1.1394,
21
+ "step": 1000
22
+ },
23
+ {
24
+ "epoch": 0.04,
25
+ "eval_accuracy": 0.3327345309381238,
26
+ "eval_loss": 1.1149410009384155,
27
+ "eval_runtime": 20.6803,
28
+ "eval_samples_per_second": 242.26,
29
+ "eval_steps_per_second": 30.319,
30
+ "step": 1000
31
+ },
32
+ {
33
+ "epoch": 0.08,
34
+ "learning_rate": 5.5527777777777784e-08,
35
+ "loss": 1.1141,
36
+ "step": 2000
37
+ },
38
+ {
39
+ "epoch": 0.08,
40
+ "eval_accuracy": 0.3401197604790419,
41
+ "eval_loss": 1.104099988937378,
42
+ "eval_runtime": 20.8477,
43
+ "eval_samples_per_second": 240.314,
44
+ "eval_steps_per_second": 30.075,
45
+ "step": 2000
46
+ },
47
+ {
48
+ "epoch": 0.12,
49
+ "learning_rate": 8.330555555555556e-08,
50
+ "loss": 1.116,
51
+ "step": 3000
52
+ },
53
+ {
54
+ "epoch": 0.12,
55
+ "eval_accuracy": 0.3407185628742515,
56
+ "eval_loss": 1.1040862798690796,
57
+ "eval_runtime": 20.6818,
58
+ "eval_samples_per_second": 242.242,
59
+ "eval_steps_per_second": 30.317,
60
+ "step": 3000
61
+ },
62
+ {
63
+ "epoch": 0.16,
64
+ "learning_rate": 1.1108333333333333e-07,
65
+ "loss": 1.1158,
66
+ "step": 4000
67
+ },
68
+ {
69
+ "epoch": 0.16,
70
+ "eval_accuracy": 0.32894211576846305,
71
+ "eval_loss": 1.1020556688308716,
72
+ "eval_runtime": 20.8541,
73
+ "eval_samples_per_second": 240.241,
74
+ "eval_steps_per_second": 30.066,
75
+ "step": 4000
76
+ },
77
+ {
78
+ "epoch": 0.2,
79
+ "learning_rate": 1.3883333333333335e-07,
80
+ "loss": 1.1135,
81
+ "step": 5000
82
+ },
83
+ {
84
+ "epoch": 0.2,
85
+ "eval_accuracy": 0.34271457085828344,
86
+ "eval_loss": 1.1008552312850952,
87
+ "eval_runtime": 20.8055,
88
+ "eval_samples_per_second": 240.802,
89
+ "eval_steps_per_second": 30.136,
90
+ "step": 5000
91
+ },
92
+ {
93
+ "epoch": 0.24,
94
+ "learning_rate": 1.6658333333333335e-07,
95
+ "loss": 1.1121,
96
+ "step": 6000
97
+ },
98
+ {
99
+ "epoch": 0.24,
100
+ "eval_accuracy": 0.3395209580838323,
101
+ "eval_loss": 1.1004050970077515,
102
+ "eval_runtime": 20.8985,
103
+ "eval_samples_per_second": 239.731,
104
+ "eval_steps_per_second": 30.002,
105
+ "step": 6000
106
+ },
107
+ {
108
+ "epoch": 0.29,
109
+ "learning_rate": 1.9436111111111112e-07,
110
+ "loss": 1.1089,
111
+ "step": 7000
112
+ },
113
+ {
114
+ "epoch": 0.29,
115
+ "eval_accuracy": 0.35788423153692617,
116
+ "eval_loss": 1.0985721349716187,
117
+ "eval_runtime": 20.84,
118
+ "eval_samples_per_second": 240.403,
119
+ "eval_steps_per_second": 30.086,
120
+ "step": 7000
121
+ },
122
+ {
123
+ "epoch": 0.33,
124
+ "learning_rate": 2.2213888888888891e-07,
125
+ "loss": 1.1079,
126
+ "step": 8000
127
+ },
128
+ {
129
+ "epoch": 0.33,
130
+ "eval_accuracy": 0.3331337325349301,
131
+ "eval_loss": 1.098374843597412,
132
+ "eval_runtime": 20.7886,
133
+ "eval_samples_per_second": 240.998,
134
+ "eval_steps_per_second": 30.161,
135
+ "step": 8000
136
+ },
137
+ {
138
+ "epoch": 0.37,
139
+ "learning_rate": 2.4988888888888893e-07,
140
+ "loss": 1.1087,
141
+ "step": 9000
142
+ },
143
+ {
144
+ "epoch": 0.37,
145
+ "eval_accuracy": 0.34510978043912177,
146
+ "eval_loss": 1.0993521213531494,
147
+ "eval_runtime": 20.782,
148
+ "eval_samples_per_second": 241.074,
149
+ "eval_steps_per_second": 30.17,
150
+ "step": 9000
151
+ },
152
+ {
153
+ "epoch": 0.41,
154
+ "learning_rate": 2.776666666666667e-07,
155
+ "loss": 1.109,
156
+ "step": 10000
157
+ },
158
+ {
159
+ "epoch": 0.41,
160
+ "eval_accuracy": 0.3475049900199601,
161
+ "eval_loss": 1.0967597961425781,
162
+ "eval_runtime": 20.6798,
163
+ "eval_samples_per_second": 242.265,
164
+ "eval_steps_per_second": 30.319,
165
+ "step": 10000
166
+ },
167
+ {
168
+ "epoch": 0.45,
169
+ "learning_rate": 3.054444444444444e-07,
170
+ "loss": 1.1052,
171
+ "step": 11000
172
+ },
173
+ {
174
+ "epoch": 0.45,
175
+ "eval_accuracy": 0.37544910179640717,
176
+ "eval_loss": 1.0941349267959595,
177
+ "eval_runtime": 20.8641,
178
+ "eval_samples_per_second": 240.126,
179
+ "eval_steps_per_second": 30.052,
180
+ "step": 11000
181
+ },
182
+ {
183
+ "epoch": 0.49,
184
+ "learning_rate": 3.3322222222222225e-07,
185
+ "loss": 1.105,
186
+ "step": 12000
187
+ },
188
+ {
189
+ "epoch": 0.49,
190
+ "eval_accuracy": 0.3834331337325349,
191
+ "eval_loss": 1.0927647352218628,
192
+ "eval_runtime": 20.6541,
193
+ "eval_samples_per_second": 242.567,
194
+ "eval_steps_per_second": 30.357,
195
+ "step": 12000
196
+ },
197
+ {
198
+ "epoch": 0.53,
199
+ "learning_rate": 3.609722222222222e-07,
200
+ "loss": 1.1016,
201
+ "step": 13000
202
+ },
203
+ {
204
+ "epoch": 0.53,
205
+ "eval_accuracy": 0.3457085828343313,
206
+ "eval_loss": 1.0942081212997437,
207
+ "eval_runtime": 21.0733,
208
+ "eval_samples_per_second": 237.742,
209
+ "eval_steps_per_second": 29.753,
210
+ "step": 13000
211
+ },
212
+ {
213
+ "epoch": 0.57,
214
+ "learning_rate": 3.8875e-07,
215
+ "loss": 1.1031,
216
+ "step": 14000
217
+ },
218
+ {
219
+ "epoch": 0.57,
220
+ "eval_accuracy": 0.37005988023952097,
221
+ "eval_loss": 1.0918152332305908,
222
+ "eval_runtime": 20.9151,
223
+ "eval_samples_per_second": 239.54,
224
+ "eval_steps_per_second": 29.978,
225
+ "step": 14000
226
+ },
227
+ {
228
+ "epoch": 0.61,
229
+ "learning_rate": 4.1652777777777786e-07,
230
+ "loss": 1.1026,
231
+ "step": 15000
232
+ },
233
+ {
234
+ "epoch": 0.61,
235
+ "eval_accuracy": 0.3790419161676647,
236
+ "eval_loss": 1.0895211696624756,
237
+ "eval_runtime": 21.0591,
238
+ "eval_samples_per_second": 237.902,
239
+ "eval_steps_per_second": 29.773,
240
+ "step": 15000
241
+ },
242
+ {
243
+ "epoch": 0.65,
244
+ "learning_rate": 4.4427777777777783e-07,
245
+ "loss": 1.0988,
246
+ "step": 16000
247
+ },
248
+ {
249
+ "epoch": 0.65,
250
+ "eval_accuracy": 0.4101796407185629,
251
+ "eval_loss": 1.0852997303009033,
252
+ "eval_runtime": 20.9509,
253
+ "eval_samples_per_second": 239.131,
254
+ "eval_steps_per_second": 29.927,
255
+ "step": 16000
256
+ },
257
+ {
258
+ "epoch": 0.69,
259
+ "learning_rate": 4.720555555555556e-07,
260
+ "loss": 1.0974,
261
+ "step": 17000
262
+ },
263
+ {
264
+ "epoch": 0.69,
265
+ "eval_accuracy": 0.43213572854291415,
266
+ "eval_loss": 1.0791982412338257,
267
+ "eval_runtime": 20.7526,
268
+ "eval_samples_per_second": 241.415,
269
+ "eval_steps_per_second": 30.213,
270
+ "step": 17000
271
+ },
272
+ {
273
+ "epoch": 0.73,
274
+ "learning_rate": 4.998055555555556e-07,
275
+ "loss": 1.0932,
276
+ "step": 18000
277
+ },
278
+ {
279
+ "epoch": 0.73,
280
+ "eval_accuracy": 0.4275449101796407,
281
+ "eval_loss": 1.072191596031189,
282
+ "eval_runtime": 21.2435,
283
+ "eval_samples_per_second": 235.837,
284
+ "eval_steps_per_second": 29.515,
285
+ "step": 18000
286
+ },
287
+ {
288
+ "epoch": 0.77,
289
+ "learning_rate": 5.275833333333334e-07,
290
+ "loss": 1.0833,
291
+ "step": 19000
292
+ },
293
+ {
294
+ "epoch": 0.77,
295
+ "eval_accuracy": 0.43233532934131735,
296
+ "eval_loss": 1.06425940990448,
297
+ "eval_runtime": 20.7923,
298
+ "eval_samples_per_second": 240.955,
299
+ "eval_steps_per_second": 30.155,
300
+ "step": 19000
301
+ },
302
+ {
303
+ "epoch": 0.81,
304
+ "learning_rate": 5.553333333333334e-07,
305
+ "loss": 1.0787,
306
+ "step": 20000
307
+ },
308
+ {
309
+ "epoch": 0.81,
310
+ "eval_accuracy": 0.4295409181636727,
311
+ "eval_loss": 1.0638529062271118,
312
+ "eval_runtime": 21.0018,
313
+ "eval_samples_per_second": 238.551,
314
+ "eval_steps_per_second": 29.855,
315
+ "step": 20000
316
+ },
317
+ {
318
+ "epoch": 0.86,
319
+ "learning_rate": 5.831111111111111e-07,
320
+ "loss": 1.0779,
321
+ "step": 21000
322
+ },
323
+ {
324
+ "epoch": 0.86,
325
+ "eval_accuracy": 0.4243512974051896,
326
+ "eval_loss": 1.0603673458099365,
327
+ "eval_runtime": 20.9689,
328
+ "eval_samples_per_second": 238.926,
329
+ "eval_steps_per_second": 29.901,
330
+ "step": 21000
331
+ },
332
+ {
333
+ "epoch": 0.9,
334
+ "learning_rate": 6.108888888888888e-07,
335
+ "loss": 1.0751,
336
+ "step": 22000
337
+ },
338
+ {
339
+ "epoch": 0.9,
340
+ "eval_accuracy": 0.43233532934131735,
341
+ "eval_loss": 1.0603009462356567,
342
+ "eval_runtime": 20.8897,
343
+ "eval_samples_per_second": 239.831,
344
+ "eval_steps_per_second": 30.015,
345
+ "step": 22000
346
+ },
347
+ {
348
+ "epoch": 0.94,
349
+ "learning_rate": 6.386388888888889e-07,
350
+ "loss": 1.0776,
351
+ "step": 23000
352
+ },
353
+ {
354
+ "epoch": 0.94,
355
+ "eval_accuracy": 0.42734530938123755,
356
+ "eval_loss": 1.0591468811035156,
357
+ "eval_runtime": 20.964,
358
+ "eval_samples_per_second": 238.981,
359
+ "eval_steps_per_second": 29.908,
360
+ "step": 23000
361
+ }
362
+ ],
363
+ "logging_steps": 1000,
364
+ "max_steps": 10000000,
365
+ "num_train_epochs": 408,
366
+ "save_steps": 1000,
367
+ "total_flos": 1.6026924220416e+16,
368
+ "trial_name": null,
369
+ "trial_params": null
370
+ }
checkpoint-23000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d42437e01c6a8e691de275f2b2ef4e9ee68d60c829bca2508adb832bd55170f9
3
+ size 4027
checkpoint-23000/vocab.json ADDED
The diff for this file is too large to render. See raw diff