tmnam20 commited on
Commit
87cf2f4
1 Parent(s): 66ca2d1

Training in progress, step 53000, checkpoint

Browse files
checkpoint-53000/added_tokens.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "<bos>": 40478,
3
+ "<eos>": 40479,
4
+ "<pad>": 40480
5
+ }
checkpoint-53000/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": 2,
25
+ "n_layer": 12,
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": "cls_index",
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-53000/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-53000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:31459ef00ea1fdf0368e1650bf29bf7cca9ee77d6afe69ae8a7b36d8b7b105f5
3
+ size 932401285
checkpoint-53000/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4429b1980d74d96bfbffb0b498b082351bee8f07d729d3271b38c06f69469a89
3
+ size 466205145
checkpoint-53000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9b781b5597b30365e3ebcbe3255fe411b183697ec5434eca00f5edc669e074ef
3
+ size 14575
checkpoint-53000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0fd8357075e75ccefbc060d5c904c1e2561d150187c128e33efb385f0e2f7ca0
3
+ size 627
checkpoint-53000/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-53000/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-53000/trainer_state.json ADDED
@@ -0,0 +1,820 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.9880945086479187,
3
+ "best_model_checkpoint": "/kaggle/output/checkpoint-53000",
4
+ "epoch": 2.159387222946545,
5
+ "eval_steps": 1000,
6
+ "global_step": 53000,
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": 1.0788,
15
+ "step": 1
16
+ },
17
+ {
18
+ "epoch": 0.04,
19
+ "learning_rate": 2.7750000000000004e-08,
20
+ "loss": 1.2063,
21
+ "step": 1000
22
+ },
23
+ {
24
+ "epoch": 0.04,
25
+ "eval_accuracy": 0.32894211576846305,
26
+ "eval_loss": 1.109653115272522,
27
+ "eval_runtime": 50.6057,
28
+ "eval_samples_per_second": 99.001,
29
+ "eval_steps_per_second": 12.39,
30
+ "step": 1000
31
+ },
32
+ {
33
+ "epoch": 0.08,
34
+ "learning_rate": 5.5527777777777784e-08,
35
+ "loss": 1.1293,
36
+ "step": 2000
37
+ },
38
+ {
39
+ "epoch": 0.08,
40
+ "eval_accuracy": 0.32994011976047904,
41
+ "eval_loss": 1.1056402921676636,
42
+ "eval_runtime": 50.7169,
43
+ "eval_samples_per_second": 98.784,
44
+ "eval_steps_per_second": 12.363,
45
+ "step": 2000
46
+ },
47
+ {
48
+ "epoch": 0.12,
49
+ "learning_rate": 8.327777777777778e-08,
50
+ "loss": 1.1263,
51
+ "step": 3000
52
+ },
53
+ {
54
+ "epoch": 0.12,
55
+ "eval_accuracy": 0.3347305389221557,
56
+ "eval_loss": 1.1063188314437866,
57
+ "eval_runtime": 50.8081,
58
+ "eval_samples_per_second": 98.606,
59
+ "eval_steps_per_second": 12.341,
60
+ "step": 3000
61
+ },
62
+ {
63
+ "epoch": 0.16,
64
+ "learning_rate": 1.1105555555555557e-07,
65
+ "loss": 1.1211,
66
+ "step": 4000
67
+ },
68
+ {
69
+ "epoch": 0.16,
70
+ "eval_accuracy": 0.34291417165668664,
71
+ "eval_loss": 1.1010773181915283,
72
+ "eval_runtime": 50.6469,
73
+ "eval_samples_per_second": 98.92,
74
+ "eval_steps_per_second": 12.38,
75
+ "step": 4000
76
+ },
77
+ {
78
+ "epoch": 0.2,
79
+ "learning_rate": 1.3883333333333335e-07,
80
+ "loss": 1.1176,
81
+ "step": 5000
82
+ },
83
+ {
84
+ "epoch": 0.2,
85
+ "eval_accuracy": 0.3395209580838323,
86
+ "eval_loss": 1.099412441253662,
87
+ "eval_runtime": 50.7266,
88
+ "eval_samples_per_second": 98.765,
89
+ "eval_steps_per_second": 12.36,
90
+ "step": 5000
91
+ },
92
+ {
93
+ "epoch": 0.24,
94
+ "learning_rate": 1.6658333333333335e-07,
95
+ "loss": 1.1165,
96
+ "step": 6000
97
+ },
98
+ {
99
+ "epoch": 0.24,
100
+ "eval_accuracy": 0.3596806387225549,
101
+ "eval_loss": 1.0983413457870483,
102
+ "eval_runtime": 50.7493,
103
+ "eval_samples_per_second": 98.721,
104
+ "eval_steps_per_second": 12.355,
105
+ "step": 6000
106
+ },
107
+ {
108
+ "epoch": 0.29,
109
+ "learning_rate": 1.9436111111111112e-07,
110
+ "loss": 1.1127,
111
+ "step": 7000
112
+ },
113
+ {
114
+ "epoch": 0.29,
115
+ "eval_accuracy": 0.35768463073852297,
116
+ "eval_loss": 1.094804048538208,
117
+ "eval_runtime": 50.8322,
118
+ "eval_samples_per_second": 98.56,
119
+ "eval_steps_per_second": 12.335,
120
+ "step": 7000
121
+ },
122
+ {
123
+ "epoch": 0.33,
124
+ "learning_rate": 2.2213888888888891e-07,
125
+ "loss": 1.1095,
126
+ "step": 8000
127
+ },
128
+ {
129
+ "epoch": 0.33,
130
+ "eval_accuracy": 0.36407185628742517,
131
+ "eval_loss": 1.0937609672546387,
132
+ "eval_runtime": 50.8609,
133
+ "eval_samples_per_second": 98.504,
134
+ "eval_steps_per_second": 12.328,
135
+ "step": 8000
136
+ },
137
+ {
138
+ "epoch": 0.37,
139
+ "learning_rate": 2.499166666666667e-07,
140
+ "loss": 1.1074,
141
+ "step": 9000
142
+ },
143
+ {
144
+ "epoch": 0.37,
145
+ "eval_accuracy": 0.34770459081836325,
146
+ "eval_loss": 1.1025313138961792,
147
+ "eval_runtime": 50.8052,
148
+ "eval_samples_per_second": 98.612,
149
+ "eval_steps_per_second": 12.341,
150
+ "step": 9000
151
+ },
152
+ {
153
+ "epoch": 0.41,
154
+ "learning_rate": 2.776666666666667e-07,
155
+ "loss": 1.1065,
156
+ "step": 10000
157
+ },
158
+ {
159
+ "epoch": 0.41,
160
+ "eval_accuracy": 0.38163672654690617,
161
+ "eval_loss": 1.0920748710632324,
162
+ "eval_runtime": 50.8964,
163
+ "eval_samples_per_second": 98.435,
164
+ "eval_steps_per_second": 12.319,
165
+ "step": 10000
166
+ },
167
+ {
168
+ "epoch": 0.45,
169
+ "learning_rate": 3.054444444444444e-07,
170
+ "loss": 1.1026,
171
+ "step": 11000
172
+ },
173
+ {
174
+ "epoch": 0.45,
175
+ "eval_accuracy": 0.37944111776447104,
176
+ "eval_loss": 1.0899525880813599,
177
+ "eval_runtime": 50.6905,
178
+ "eval_samples_per_second": 98.835,
179
+ "eval_steps_per_second": 12.369,
180
+ "step": 11000
181
+ },
182
+ {
183
+ "epoch": 0.49,
184
+ "learning_rate": 3.3319444444444444e-07,
185
+ "loss": 1.1016,
186
+ "step": 12000
187
+ },
188
+ {
189
+ "epoch": 0.49,
190
+ "eval_accuracy": 0.4,
191
+ "eval_loss": 1.0867687463760376,
192
+ "eval_runtime": 50.8282,
193
+ "eval_samples_per_second": 98.567,
194
+ "eval_steps_per_second": 12.336,
195
+ "step": 12000
196
+ },
197
+ {
198
+ "epoch": 0.53,
199
+ "learning_rate": 3.6094444444444446e-07,
200
+ "loss": 1.0999,
201
+ "step": 13000
202
+ },
203
+ {
204
+ "epoch": 0.53,
205
+ "eval_accuracy": 0.3467065868263473,
206
+ "eval_loss": 1.0919947624206543,
207
+ "eval_runtime": 50.7953,
208
+ "eval_samples_per_second": 98.631,
209
+ "eval_steps_per_second": 12.344,
210
+ "step": 13000
211
+ },
212
+ {
213
+ "epoch": 0.57,
214
+ "learning_rate": 3.8872222222222223e-07,
215
+ "loss": 1.0976,
216
+ "step": 14000
217
+ },
218
+ {
219
+ "epoch": 0.57,
220
+ "eval_accuracy": 0.38003992015968063,
221
+ "eval_loss": 1.0882444381713867,
222
+ "eval_runtime": 50.8866,
223
+ "eval_samples_per_second": 98.454,
224
+ "eval_steps_per_second": 12.322,
225
+ "step": 14000
226
+ },
227
+ {
228
+ "epoch": 0.61,
229
+ "learning_rate": 4.1650000000000006e-07,
230
+ "loss": 1.0951,
231
+ "step": 15000
232
+ },
233
+ {
234
+ "epoch": 0.61,
235
+ "eval_accuracy": 0.4229540918163673,
236
+ "eval_loss": 1.0766451358795166,
237
+ "eval_runtime": 50.7597,
238
+ "eval_samples_per_second": 98.7,
239
+ "eval_steps_per_second": 12.352,
240
+ "step": 15000
241
+ },
242
+ {
243
+ "epoch": 0.65,
244
+ "learning_rate": 4.4425e-07,
245
+ "loss": 1.087,
246
+ "step": 16000
247
+ },
248
+ {
249
+ "epoch": 0.65,
250
+ "eval_accuracy": 0.4339321357285429,
251
+ "eval_loss": 1.0621711015701294,
252
+ "eval_runtime": 50.7559,
253
+ "eval_samples_per_second": 98.708,
254
+ "eval_steps_per_second": 12.353,
255
+ "step": 16000
256
+ },
257
+ {
258
+ "epoch": 0.69,
259
+ "learning_rate": 4.7202777777777785e-07,
260
+ "loss": 1.0769,
261
+ "step": 17000
262
+ },
263
+ {
264
+ "epoch": 0.69,
265
+ "eval_accuracy": 0.4357285429141717,
266
+ "eval_loss": 1.0624364614486694,
267
+ "eval_runtime": 50.8608,
268
+ "eval_samples_per_second": 98.504,
269
+ "eval_steps_per_second": 12.328,
270
+ "step": 17000
271
+ },
272
+ {
273
+ "epoch": 0.73,
274
+ "learning_rate": 4.998055555555556e-07,
275
+ "loss": 1.0798,
276
+ "step": 18000
277
+ },
278
+ {
279
+ "epoch": 0.73,
280
+ "eval_accuracy": 0.4311377245508982,
281
+ "eval_loss": 1.0611094236373901,
282
+ "eval_runtime": 50.7255,
283
+ "eval_samples_per_second": 98.767,
284
+ "eval_steps_per_second": 12.361,
285
+ "step": 18000
286
+ },
287
+ {
288
+ "epoch": 0.77,
289
+ "learning_rate": 5.275555555555556e-07,
290
+ "loss": 1.0717,
291
+ "step": 19000
292
+ },
293
+ {
294
+ "epoch": 0.77,
295
+ "eval_accuracy": 0.43852295409181635,
296
+ "eval_loss": 1.056412935256958,
297
+ "eval_runtime": 50.8955,
298
+ "eval_samples_per_second": 98.437,
299
+ "eval_steps_per_second": 12.319,
300
+ "step": 19000
301
+ },
302
+ {
303
+ "epoch": 0.81,
304
+ "learning_rate": 5.553333333333334e-07,
305
+ "loss": 1.0682,
306
+ "step": 20000
307
+ },
308
+ {
309
+ "epoch": 0.81,
310
+ "eval_accuracy": 0.43313373253493015,
311
+ "eval_loss": 1.0576339960098267,
312
+ "eval_runtime": 50.6675,
313
+ "eval_samples_per_second": 98.88,
314
+ "eval_steps_per_second": 12.375,
315
+ "step": 20000
316
+ },
317
+ {
318
+ "epoch": 0.86,
319
+ "learning_rate": 5.830833333333334e-07,
320
+ "loss": 1.0674,
321
+ "step": 21000
322
+ },
323
+ {
324
+ "epoch": 0.86,
325
+ "eval_accuracy": 0.4305389221556886,
326
+ "eval_loss": 1.057605504989624,
327
+ "eval_runtime": 50.8082,
328
+ "eval_samples_per_second": 98.606,
329
+ "eval_steps_per_second": 12.341,
330
+ "step": 21000
331
+ },
332
+ {
333
+ "epoch": 0.9,
334
+ "learning_rate": 6.108611111111111e-07,
335
+ "loss": 1.0689,
336
+ "step": 22000
337
+ },
338
+ {
339
+ "epoch": 0.9,
340
+ "eval_accuracy": 0.4351297405189621,
341
+ "eval_loss": 1.0509696006774902,
342
+ "eval_runtime": 50.7595,
343
+ "eval_samples_per_second": 98.701,
344
+ "eval_steps_per_second": 12.352,
345
+ "step": 22000
346
+ },
347
+ {
348
+ "epoch": 0.94,
349
+ "learning_rate": 6.386388888888889e-07,
350
+ "loss": 1.065,
351
+ "step": 23000
352
+ },
353
+ {
354
+ "epoch": 0.94,
355
+ "eval_accuracy": 0.4363273453093812,
356
+ "eval_loss": 1.0513038635253906,
357
+ "eval_runtime": 50.685,
358
+ "eval_samples_per_second": 98.846,
359
+ "eval_steps_per_second": 12.371,
360
+ "step": 23000
361
+ },
362
+ {
363
+ "epoch": 0.98,
364
+ "learning_rate": 6.663888888888889e-07,
365
+ "loss": 1.0638,
366
+ "step": 24000
367
+ },
368
+ {
369
+ "epoch": 0.98,
370
+ "eval_accuracy": 0.4429141716566866,
371
+ "eval_loss": 1.04515540599823,
372
+ "eval_runtime": 50.9215,
373
+ "eval_samples_per_second": 98.387,
374
+ "eval_steps_per_second": 12.313,
375
+ "step": 24000
376
+ },
377
+ {
378
+ "epoch": 1.02,
379
+ "learning_rate": 6.941666666666667e-07,
380
+ "loss": 1.0607,
381
+ "step": 25000
382
+ },
383
+ {
384
+ "epoch": 1.02,
385
+ "eval_accuracy": 0.4471057884231537,
386
+ "eval_loss": 1.0432593822479248,
387
+ "eval_runtime": 50.8338,
388
+ "eval_samples_per_second": 98.556,
389
+ "eval_steps_per_second": 12.334,
390
+ "step": 25000
391
+ },
392
+ {
393
+ "epoch": 1.06,
394
+ "learning_rate": 7.219444444444444e-07,
395
+ "loss": 1.0578,
396
+ "step": 26000
397
+ },
398
+ {
399
+ "epoch": 1.06,
400
+ "eval_accuracy": 0.4407185628742515,
401
+ "eval_loss": 1.044198751449585,
402
+ "eval_runtime": 50.8791,
403
+ "eval_samples_per_second": 98.469,
404
+ "eval_steps_per_second": 12.323,
405
+ "step": 26000
406
+ },
407
+ {
408
+ "epoch": 1.1,
409
+ "learning_rate": 7.496944444444444e-07,
410
+ "loss": 1.0602,
411
+ "step": 27000
412
+ },
413
+ {
414
+ "epoch": 1.1,
415
+ "eval_accuracy": 0.4363273453093812,
416
+ "eval_loss": 1.045542597770691,
417
+ "eval_runtime": 50.8933,
418
+ "eval_samples_per_second": 98.441,
419
+ "eval_steps_per_second": 12.32,
420
+ "step": 27000
421
+ },
422
+ {
423
+ "epoch": 1.14,
424
+ "learning_rate": 7.774722222222223e-07,
425
+ "loss": 1.0566,
426
+ "step": 28000
427
+ },
428
+ {
429
+ "epoch": 1.14,
430
+ "eval_accuracy": 0.437125748502994,
431
+ "eval_loss": 1.038167953491211,
432
+ "eval_runtime": 50.8894,
433
+ "eval_samples_per_second": 98.449,
434
+ "eval_steps_per_second": 12.321,
435
+ "step": 28000
436
+ },
437
+ {
438
+ "epoch": 1.18,
439
+ "learning_rate": 8.052222222222223e-07,
440
+ "loss": 1.0528,
441
+ "step": 29000
442
+ },
443
+ {
444
+ "epoch": 1.18,
445
+ "eval_accuracy": 0.45069860279441115,
446
+ "eval_loss": 1.0337190628051758,
447
+ "eval_runtime": 50.8393,
448
+ "eval_samples_per_second": 98.546,
449
+ "eval_steps_per_second": 12.333,
450
+ "step": 29000
451
+ },
452
+ {
453
+ "epoch": 1.22,
454
+ "learning_rate": 8.330000000000001e-07,
455
+ "loss": 1.0507,
456
+ "step": 30000
457
+ },
458
+ {
459
+ "epoch": 1.22,
460
+ "eval_accuracy": 0.4497005988023952,
461
+ "eval_loss": 1.0287295579910278,
462
+ "eval_runtime": 50.8361,
463
+ "eval_samples_per_second": 98.552,
464
+ "eval_steps_per_second": 12.334,
465
+ "step": 30000
466
+ },
467
+ {
468
+ "epoch": 1.26,
469
+ "learning_rate": 8.607777777777779e-07,
470
+ "loss": 1.0553,
471
+ "step": 31000
472
+ },
473
+ {
474
+ "epoch": 1.26,
475
+ "eval_accuracy": 0.45209580838323354,
476
+ "eval_loss": 1.0318636894226074,
477
+ "eval_runtime": 50.8224,
478
+ "eval_samples_per_second": 98.578,
479
+ "eval_steps_per_second": 12.337,
480
+ "step": 31000
481
+ },
482
+ {
483
+ "epoch": 1.3,
484
+ "learning_rate": 8.885277777777779e-07,
485
+ "loss": 1.0509,
486
+ "step": 32000
487
+ },
488
+ {
489
+ "epoch": 1.3,
490
+ "eval_accuracy": 0.4530938123752495,
491
+ "eval_loss": 1.0215647220611572,
492
+ "eval_runtime": 50.8903,
493
+ "eval_samples_per_second": 98.447,
494
+ "eval_steps_per_second": 12.321,
495
+ "step": 32000
496
+ },
497
+ {
498
+ "epoch": 1.34,
499
+ "learning_rate": 9.163055555555556e-07,
500
+ "loss": 1.0444,
501
+ "step": 33000
502
+ },
503
+ {
504
+ "epoch": 1.34,
505
+ "eval_accuracy": 0.47025948103792414,
506
+ "eval_loss": 1.0156230926513672,
507
+ "eval_runtime": 50.8879,
508
+ "eval_samples_per_second": 98.452,
509
+ "eval_steps_per_second": 12.321,
510
+ "step": 33000
511
+ },
512
+ {
513
+ "epoch": 1.39,
514
+ "learning_rate": 9.440833333333335e-07,
515
+ "loss": 1.0383,
516
+ "step": 34000
517
+ },
518
+ {
519
+ "epoch": 1.39,
520
+ "eval_accuracy": 0.4686626746506986,
521
+ "eval_loss": 1.0187312364578247,
522
+ "eval_runtime": 50.7712,
523
+ "eval_samples_per_second": 98.678,
524
+ "eval_steps_per_second": 12.35,
525
+ "step": 34000
526
+ },
527
+ {
528
+ "epoch": 1.43,
529
+ "learning_rate": 9.718333333333334e-07,
530
+ "loss": 1.043,
531
+ "step": 35000
532
+ },
533
+ {
534
+ "epoch": 1.43,
535
+ "eval_accuracy": 0.4748502994011976,
536
+ "eval_loss": 1.0167384147644043,
537
+ "eval_runtime": 50.6223,
538
+ "eval_samples_per_second": 98.968,
539
+ "eval_steps_per_second": 12.386,
540
+ "step": 35000
541
+ },
542
+ {
543
+ "epoch": 1.47,
544
+ "learning_rate": 9.996111111111111e-07,
545
+ "loss": 1.0396,
546
+ "step": 36000
547
+ },
548
+ {
549
+ "epoch": 1.47,
550
+ "eval_accuracy": 0.47964071856287427,
551
+ "eval_loss": 1.0082147121429443,
552
+ "eval_runtime": 50.671,
553
+ "eval_samples_per_second": 98.873,
554
+ "eval_steps_per_second": 12.374,
555
+ "step": 36000
556
+ },
557
+ {
558
+ "epoch": 1.51,
559
+ "learning_rate": 1.0273888888888891e-06,
560
+ "loss": 1.0373,
561
+ "step": 37000
562
+ },
563
+ {
564
+ "epoch": 1.51,
565
+ "eval_accuracy": 0.4802395209580838,
566
+ "eval_loss": 1.0071511268615723,
567
+ "eval_runtime": 50.5903,
568
+ "eval_samples_per_second": 99.031,
569
+ "eval_steps_per_second": 12.394,
570
+ "step": 37000
571
+ },
572
+ {
573
+ "epoch": 1.55,
574
+ "learning_rate": 1.0551666666666669e-06,
575
+ "loss": 1.0371,
576
+ "step": 38000
577
+ },
578
+ {
579
+ "epoch": 1.55,
580
+ "eval_accuracy": 0.49481037924151694,
581
+ "eval_loss": 1.0086288452148438,
582
+ "eval_runtime": 50.7157,
583
+ "eval_samples_per_second": 98.786,
584
+ "eval_steps_per_second": 12.363,
585
+ "step": 38000
586
+ },
587
+ {
588
+ "epoch": 1.59,
589
+ "learning_rate": 1.082888888888889e-06,
590
+ "loss": 1.0367,
591
+ "step": 39000
592
+ },
593
+ {
594
+ "epoch": 1.59,
595
+ "eval_accuracy": 0.4884231536926148,
596
+ "eval_loss": 1.0103219747543335,
597
+ "eval_runtime": 50.6401,
598
+ "eval_samples_per_second": 98.934,
599
+ "eval_steps_per_second": 12.382,
600
+ "step": 39000
601
+ },
602
+ {
603
+ "epoch": 1.63,
604
+ "learning_rate": 1.1106666666666668e-06,
605
+ "loss": 1.0311,
606
+ "step": 40000
607
+ },
608
+ {
609
+ "epoch": 1.63,
610
+ "eval_accuracy": 0.4846307385229541,
611
+ "eval_loss": 1.0027852058410645,
612
+ "eval_runtime": 50.6939,
613
+ "eval_samples_per_second": 98.828,
614
+ "eval_steps_per_second": 12.368,
615
+ "step": 40000
616
+ },
617
+ {
618
+ "epoch": 1.67,
619
+ "learning_rate": 1.1384444444444446e-06,
620
+ "loss": 1.032,
621
+ "step": 41000
622
+ },
623
+ {
624
+ "epoch": 1.67,
625
+ "eval_accuracy": 0.48862275449101794,
626
+ "eval_loss": 1.0152778625488281,
627
+ "eval_runtime": 50.7815,
628
+ "eval_samples_per_second": 98.658,
629
+ "eval_steps_per_second": 12.347,
630
+ "step": 41000
631
+ },
632
+ {
633
+ "epoch": 1.71,
634
+ "learning_rate": 1.1661944444444447e-06,
635
+ "loss": 1.0311,
636
+ "step": 42000
637
+ },
638
+ {
639
+ "epoch": 1.71,
640
+ "eval_accuracy": 0.4812375249500998,
641
+ "eval_loss": 1.0021992921829224,
642
+ "eval_runtime": 50.7722,
643
+ "eval_samples_per_second": 98.676,
644
+ "eval_steps_per_second": 12.349,
645
+ "step": 42000
646
+ },
647
+ {
648
+ "epoch": 1.75,
649
+ "learning_rate": 1.1939722222222222e-06,
650
+ "loss": 1.0325,
651
+ "step": 43000
652
+ },
653
+ {
654
+ "epoch": 1.75,
655
+ "eval_accuracy": 0.47305389221556887,
656
+ "eval_loss": 1.0248425006866455,
657
+ "eval_runtime": 50.7202,
658
+ "eval_samples_per_second": 98.777,
659
+ "eval_steps_per_second": 12.362,
660
+ "step": 43000
661
+ },
662
+ {
663
+ "epoch": 1.79,
664
+ "learning_rate": 1.22175e-06,
665
+ "loss": 1.0365,
666
+ "step": 44000
667
+ },
668
+ {
669
+ "epoch": 1.79,
670
+ "eval_accuracy": 0.4906187624750499,
671
+ "eval_loss": 1.0018315315246582,
672
+ "eval_runtime": 50.7515,
673
+ "eval_samples_per_second": 98.716,
674
+ "eval_steps_per_second": 12.354,
675
+ "step": 44000
676
+ },
677
+ {
678
+ "epoch": 1.83,
679
+ "learning_rate": 1.2495e-06,
680
+ "loss": 1.0292,
681
+ "step": 45000
682
+ },
683
+ {
684
+ "epoch": 1.83,
685
+ "eval_accuracy": 0.4946107784431138,
686
+ "eval_loss": 1.0031094551086426,
687
+ "eval_runtime": 50.8865,
688
+ "eval_samples_per_second": 98.454,
689
+ "eval_steps_per_second": 12.322,
690
+ "step": 45000
691
+ },
692
+ {
693
+ "epoch": 1.87,
694
+ "learning_rate": 1.2772777777777778e-06,
695
+ "loss": 1.0312,
696
+ "step": 46000
697
+ },
698
+ {
699
+ "epoch": 1.87,
700
+ "eval_accuracy": 0.48882235528942114,
701
+ "eval_loss": 1.0050362348556519,
702
+ "eval_runtime": 50.9403,
703
+ "eval_samples_per_second": 98.35,
704
+ "eval_steps_per_second": 12.309,
705
+ "step": 46000
706
+ },
707
+ {
708
+ "epoch": 1.91,
709
+ "learning_rate": 1.3050277777777777e-06,
710
+ "loss": 1.0257,
711
+ "step": 47000
712
+ },
713
+ {
714
+ "epoch": 1.91,
715
+ "eval_accuracy": 0.49560878243512974,
716
+ "eval_loss": 0.9955324530601501,
717
+ "eval_runtime": 50.9191,
718
+ "eval_samples_per_second": 98.391,
719
+ "eval_steps_per_second": 12.314,
720
+ "step": 47000
721
+ },
722
+ {
723
+ "epoch": 1.96,
724
+ "learning_rate": 1.3328055555555555e-06,
725
+ "loss": 1.03,
726
+ "step": 48000
727
+ },
728
+ {
729
+ "epoch": 1.96,
730
+ "eval_accuracy": 0.48882235528942114,
731
+ "eval_loss": 1.0171200037002563,
732
+ "eval_runtime": 50.9431,
733
+ "eval_samples_per_second": 98.345,
734
+ "eval_steps_per_second": 12.308,
735
+ "step": 48000
736
+ },
737
+ {
738
+ "epoch": 2.0,
739
+ "learning_rate": 1.3605555555555555e-06,
740
+ "loss": 1.0241,
741
+ "step": 49000
742
+ },
743
+ {
744
+ "epoch": 2.0,
745
+ "eval_accuracy": 0.4906187624750499,
746
+ "eval_loss": 0.993681013584137,
747
+ "eval_runtime": 50.7775,
748
+ "eval_samples_per_second": 98.666,
749
+ "eval_steps_per_second": 12.348,
750
+ "step": 49000
751
+ },
752
+ {
753
+ "epoch": 2.04,
754
+ "learning_rate": 1.3883333333333333e-06,
755
+ "loss": 1.0252,
756
+ "step": 50000
757
+ },
758
+ {
759
+ "epoch": 2.04,
760
+ "eval_accuracy": 0.49161676646706587,
761
+ "eval_loss": 1.000351071357727,
762
+ "eval_runtime": 50.7897,
763
+ "eval_samples_per_second": 98.642,
764
+ "eval_steps_per_second": 12.345,
765
+ "step": 50000
766
+ },
767
+ {
768
+ "epoch": 2.08,
769
+ "learning_rate": 1.4160833333333334e-06,
770
+ "loss": 1.0227,
771
+ "step": 51000
772
+ },
773
+ {
774
+ "epoch": 2.08,
775
+ "eval_accuracy": 0.5021956087824351,
776
+ "eval_loss": 0.9907731413841248,
777
+ "eval_runtime": 50.9267,
778
+ "eval_samples_per_second": 98.377,
779
+ "eval_steps_per_second": 12.312,
780
+ "step": 51000
781
+ },
782
+ {
783
+ "epoch": 2.12,
784
+ "learning_rate": 1.4438611111111112e-06,
785
+ "loss": 1.019,
786
+ "step": 52000
787
+ },
788
+ {
789
+ "epoch": 2.12,
790
+ "eval_accuracy": 0.49001996007984033,
791
+ "eval_loss": 1.0103719234466553,
792
+ "eval_runtime": 50.9179,
793
+ "eval_samples_per_second": 98.394,
794
+ "eval_steps_per_second": 12.314,
795
+ "step": 52000
796
+ },
797
+ {
798
+ "epoch": 2.16,
799
+ "learning_rate": 1.471611111111111e-06,
800
+ "loss": 1.0268,
801
+ "step": 53000
802
+ },
803
+ {
804
+ "epoch": 2.16,
805
+ "eval_accuracy": 0.5001996007984032,
806
+ "eval_loss": 0.9880945086479187,
807
+ "eval_runtime": 50.9661,
808
+ "eval_samples_per_second": 98.301,
809
+ "eval_steps_per_second": 12.302,
810
+ "step": 53000
811
+ }
812
+ ],
813
+ "logging_steps": 1000,
814
+ "max_steps": 10000000,
815
+ "num_train_epochs": 408,
816
+ "save_steps": 1000,
817
+ "total_flos": 1.107882993207214e+17,
818
+ "trial_name": null,
819
+ "trial_params": null
820
+ }
checkpoint-53000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0ba6621b829e7b7a24837f1c8da621351c891fe426b370399b04766c22090faa
3
+ size 4091
checkpoint-53000/vocab.json ADDED
The diff for this file is too large to render. See raw diff