gagan3012 commited on
Commit
83d33d6
1 Parent(s): 7a39ec1

Initial commit

Browse files
all_results.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 5.0,
3
+ "init_mem_cpu_alloc_delta": 10329645,
4
+ "init_mem_cpu_peaked_delta": 18306,
5
+ "init_mem_gpu_alloc_delta": 334744576,
6
+ "init_mem_gpu_peaked_delta": 0,
7
+ "perplexity": 2.7666771850966008,
8
+ "train_mem_cpu_alloc_delta": 1355433,
9
+ "train_mem_cpu_peaked_delta": 1484888,
10
+ "train_mem_gpu_alloc_delta": 986922496,
11
+ "train_mem_gpu_peaked_delta": 2066717696,
12
+ "train_runtime": 9052.4232,
13
+ "train_samples_per_second": 6.641
14
+ }
config.json ADDED
@@ -0,0 +1,43 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "distilgpt2",
3
+ "_num_labels": 1,
4
+ "activation_function": "gelu_new",
5
+ "architectures": [
6
+ "GPT2LMHeadModel"
7
+ ],
8
+ "attn_pdrop": 0.1,
9
+ "bos_token_id": 50256,
10
+ "embd_pdrop": 0.1,
11
+ "eos_token_id": 50256,
12
+ "gradient_checkpointing": false,
13
+ "id2label": {
14
+ "0": "LABEL_0"
15
+ },
16
+ "initializer_range": 0.02,
17
+ "label2id": {
18
+ "LABEL_0": 0
19
+ },
20
+ "layer_norm_epsilon": 1e-05,
21
+ "model_type": "gpt2",
22
+ "n_ctx": 1024,
23
+ "n_embd": 768,
24
+ "n_head": 12,
25
+ "n_inner": null,
26
+ "n_layer": 6,
27
+ "n_positions": 1024,
28
+ "resid_pdrop": 0.1,
29
+ "summary_activation": null,
30
+ "summary_first_dropout": 0.1,
31
+ "summary_proj_to_labels": true,
32
+ "summary_type": "cls_index",
33
+ "summary_use_proj": true,
34
+ "task_specific_params": {
35
+ "text-generation": {
36
+ "do_sample": true,
37
+ "max_length": 50
38
+ }
39
+ },
40
+ "transformers_version": "4.4.0.dev0",
41
+ "use_cache": true,
42
+ "vocab_size": 50257
43
+ }
eval_results.json ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ {
2
+ "perplexity": 2.7666771850966008
3
+ }
merges.txt ADDED
The diff for this file is too large to render. See raw diff
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e848cf348d7e6d28db7b3de2f83770222992e40c762e58669ad6ef939b2c3f1d
3
+ size 333975623
special_tokens_map.json ADDED
@@ -0,0 +1 @@
 
1
+ {"bos_token": "<|endoftext|>", "eos_token": "<|endoftext|>", "unk_token": "<|endoftext|>"}
tokenizer_config.json ADDED
@@ -0,0 +1 @@
 
1
+ {"unk_token": "<|endoftext|>", "bos_token": "<|endoftext|>", "eos_token": "<|endoftext|>", "add_prefix_space": false, "model_max_length": 1024, "name_or_path": "distilgpt2"}
train_results.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 5.0,
3
+ "init_mem_cpu_alloc_delta": 10329645,
4
+ "init_mem_cpu_peaked_delta": 18306,
5
+ "init_mem_gpu_alloc_delta": 334744576,
6
+ "init_mem_gpu_peaked_delta": 0,
7
+ "train_mem_cpu_alloc_delta": 1355433,
8
+ "train_mem_cpu_peaked_delta": 1484888,
9
+ "train_mem_gpu_alloc_delta": 986922496,
10
+ "train_mem_gpu_peaked_delta": 2066717696,
11
+ "train_runtime": 9052.4232,
12
+ "train_samples_per_second": 6.641
13
+ }
trainer_state.json ADDED
@@ -0,0 +1,743 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 5.0,
5
+ "global_step": 60120,
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": 4.958416500332668e-05,
13
+ "loss": 2.6853,
14
+ "step": 500
15
+ },
16
+ {
17
+ "epoch": 0.08,
18
+ "learning_rate": 4.916833000665336e-05,
19
+ "loss": 2.3888,
20
+ "step": 1000
21
+ },
22
+ {
23
+ "epoch": 0.12,
24
+ "learning_rate": 4.8752495009980045e-05,
25
+ "loss": 2.4518,
26
+ "step": 1500
27
+ },
28
+ {
29
+ "epoch": 0.17,
30
+ "learning_rate": 4.833666001330672e-05,
31
+ "loss": 2.4134,
32
+ "step": 2000
33
+ },
34
+ {
35
+ "epoch": 0.21,
36
+ "learning_rate": 4.79208250166334e-05,
37
+ "loss": 2.2574,
38
+ "step": 2500
39
+ },
40
+ {
41
+ "epoch": 0.25,
42
+ "learning_rate": 4.750499001996008e-05,
43
+ "loss": 2.1327,
44
+ "step": 3000
45
+ },
46
+ {
47
+ "epoch": 0.29,
48
+ "learning_rate": 4.7089155023286765e-05,
49
+ "loss": 2.2362,
50
+ "step": 3500
51
+ },
52
+ {
53
+ "epoch": 0.33,
54
+ "learning_rate": 4.6673320026613444e-05,
55
+ "loss": 2.1094,
56
+ "step": 4000
57
+ },
58
+ {
59
+ "epoch": 0.37,
60
+ "learning_rate": 4.625748502994012e-05,
61
+ "loss": 2.1622,
62
+ "step": 4500
63
+ },
64
+ {
65
+ "epoch": 0.42,
66
+ "learning_rate": 4.58416500332668e-05,
67
+ "loss": 2.0353,
68
+ "step": 5000
69
+ },
70
+ {
71
+ "epoch": 0.46,
72
+ "learning_rate": 4.5425815036593486e-05,
73
+ "loss": 2.0771,
74
+ "step": 5500
75
+ },
76
+ {
77
+ "epoch": 0.5,
78
+ "learning_rate": 4.5009980039920164e-05,
79
+ "loss": 2.1499,
80
+ "step": 6000
81
+ },
82
+ {
83
+ "epoch": 0.54,
84
+ "learning_rate": 4.459414504324684e-05,
85
+ "loss": 2.1235,
86
+ "step": 6500
87
+ },
88
+ {
89
+ "epoch": 0.58,
90
+ "learning_rate": 4.417831004657352e-05,
91
+ "loss": 2.061,
92
+ "step": 7000
93
+ },
94
+ {
95
+ "epoch": 0.62,
96
+ "learning_rate": 4.3762475049900207e-05,
97
+ "loss": 2.066,
98
+ "step": 7500
99
+ },
100
+ {
101
+ "epoch": 0.67,
102
+ "learning_rate": 4.3346640053226885e-05,
103
+ "loss": 2.055,
104
+ "step": 8000
105
+ },
106
+ {
107
+ "epoch": 0.71,
108
+ "learning_rate": 4.2930805056553564e-05,
109
+ "loss": 2.012,
110
+ "step": 8500
111
+ },
112
+ {
113
+ "epoch": 0.75,
114
+ "learning_rate": 4.251497005988024e-05,
115
+ "loss": 2.0656,
116
+ "step": 9000
117
+ },
118
+ {
119
+ "epoch": 0.79,
120
+ "learning_rate": 4.209913506320692e-05,
121
+ "loss": 2.1063,
122
+ "step": 9500
123
+ },
124
+ {
125
+ "epoch": 0.83,
126
+ "learning_rate": 4.1683300066533606e-05,
127
+ "loss": 2.007,
128
+ "step": 10000
129
+ },
130
+ {
131
+ "epoch": 0.87,
132
+ "learning_rate": 4.1267465069860284e-05,
133
+ "loss": 2.0566,
134
+ "step": 10500
135
+ },
136
+ {
137
+ "epoch": 0.91,
138
+ "learning_rate": 4.085163007318696e-05,
139
+ "loss": 2.0895,
140
+ "step": 11000
141
+ },
142
+ {
143
+ "epoch": 0.96,
144
+ "learning_rate": 4.043579507651364e-05,
145
+ "loss": 1.9675,
146
+ "step": 11500
147
+ },
148
+ {
149
+ "epoch": 1.0,
150
+ "learning_rate": 4.0019960079840326e-05,
151
+ "loss": 2.0837,
152
+ "step": 12000
153
+ },
154
+ {
155
+ "epoch": 1.04,
156
+ "learning_rate": 3.9604125083167e-05,
157
+ "loss": 1.964,
158
+ "step": 12500
159
+ },
160
+ {
161
+ "epoch": 1.08,
162
+ "learning_rate": 3.9188290086493677e-05,
163
+ "loss": 2.024,
164
+ "step": 13000
165
+ },
166
+ {
167
+ "epoch": 1.12,
168
+ "learning_rate": 3.877245508982036e-05,
169
+ "loss": 1.9672,
170
+ "step": 13500
171
+ },
172
+ {
173
+ "epoch": 1.16,
174
+ "learning_rate": 3.835662009314704e-05,
175
+ "loss": 1.9733,
176
+ "step": 14000
177
+ },
178
+ {
179
+ "epoch": 1.21,
180
+ "learning_rate": 3.794078509647372e-05,
181
+ "loss": 2.0202,
182
+ "step": 14500
183
+ },
184
+ {
185
+ "epoch": 1.25,
186
+ "learning_rate": 3.75249500998004e-05,
187
+ "loss": 1.9932,
188
+ "step": 15000
189
+ },
190
+ {
191
+ "epoch": 1.29,
192
+ "learning_rate": 3.7109115103127076e-05,
193
+ "loss": 1.956,
194
+ "step": 15500
195
+ },
196
+ {
197
+ "epoch": 1.33,
198
+ "learning_rate": 3.669328010645376e-05,
199
+ "loss": 1.9439,
200
+ "step": 16000
201
+ },
202
+ {
203
+ "epoch": 1.37,
204
+ "learning_rate": 3.627744510978044e-05,
205
+ "loss": 2.024,
206
+ "step": 16500
207
+ },
208
+ {
209
+ "epoch": 1.41,
210
+ "learning_rate": 3.586161011310712e-05,
211
+ "loss": 1.9913,
212
+ "step": 17000
213
+ },
214
+ {
215
+ "epoch": 1.46,
216
+ "learning_rate": 3.5445775116433796e-05,
217
+ "loss": 1.8802,
218
+ "step": 17500
219
+ },
220
+ {
221
+ "epoch": 1.5,
222
+ "learning_rate": 3.502994011976048e-05,
223
+ "loss": 1.9544,
224
+ "step": 18000
225
+ },
226
+ {
227
+ "epoch": 1.54,
228
+ "learning_rate": 3.461410512308716e-05,
229
+ "loss": 1.9787,
230
+ "step": 18500
231
+ },
232
+ {
233
+ "epoch": 1.58,
234
+ "learning_rate": 3.419827012641384e-05,
235
+ "loss": 1.81,
236
+ "step": 19000
237
+ },
238
+ {
239
+ "epoch": 1.62,
240
+ "learning_rate": 3.378243512974052e-05,
241
+ "loss": 1.8795,
242
+ "step": 19500
243
+ },
244
+ {
245
+ "epoch": 1.66,
246
+ "learning_rate": 3.33666001330672e-05,
247
+ "loss": 1.9943,
248
+ "step": 20000
249
+ },
250
+ {
251
+ "epoch": 1.7,
252
+ "learning_rate": 3.295076513639388e-05,
253
+ "loss": 1.9593,
254
+ "step": 20500
255
+ },
256
+ {
257
+ "epoch": 1.75,
258
+ "learning_rate": 3.253493013972056e-05,
259
+ "loss": 1.8693,
260
+ "step": 21000
261
+ },
262
+ {
263
+ "epoch": 1.79,
264
+ "learning_rate": 3.211909514304724e-05,
265
+ "loss": 1.8688,
266
+ "step": 21500
267
+ },
268
+ {
269
+ "epoch": 1.83,
270
+ "learning_rate": 3.170326014637392e-05,
271
+ "loss": 1.9299,
272
+ "step": 22000
273
+ },
274
+ {
275
+ "epoch": 1.87,
276
+ "learning_rate": 3.12874251497006e-05,
277
+ "loss": 1.9862,
278
+ "step": 22500
279
+ },
280
+ {
281
+ "epoch": 1.91,
282
+ "learning_rate": 3.087159015302728e-05,
283
+ "loss": 1.9493,
284
+ "step": 23000
285
+ },
286
+ {
287
+ "epoch": 1.95,
288
+ "learning_rate": 3.045575515635396e-05,
289
+ "loss": 1.8195,
290
+ "step": 23500
291
+ },
292
+ {
293
+ "epoch": 2.0,
294
+ "learning_rate": 3.003992015968064e-05,
295
+ "loss": 2.0447,
296
+ "step": 24000
297
+ },
298
+ {
299
+ "epoch": 2.04,
300
+ "learning_rate": 2.962408516300732e-05,
301
+ "loss": 1.9297,
302
+ "step": 24500
303
+ },
304
+ {
305
+ "epoch": 2.08,
306
+ "learning_rate": 2.9208250166334e-05,
307
+ "loss": 1.8736,
308
+ "step": 25000
309
+ },
310
+ {
311
+ "epoch": 2.12,
312
+ "learning_rate": 2.879241516966068e-05,
313
+ "loss": 1.7594,
314
+ "step": 25500
315
+ },
316
+ {
317
+ "epoch": 2.16,
318
+ "learning_rate": 2.837658017298736e-05,
319
+ "loss": 1.913,
320
+ "step": 26000
321
+ },
322
+ {
323
+ "epoch": 2.2,
324
+ "learning_rate": 2.796074517631404e-05,
325
+ "loss": 1.8812,
326
+ "step": 26500
327
+ },
328
+ {
329
+ "epoch": 2.25,
330
+ "learning_rate": 2.754491017964072e-05,
331
+ "loss": 1.7825,
332
+ "step": 27000
333
+ },
334
+ {
335
+ "epoch": 2.29,
336
+ "learning_rate": 2.71290751829674e-05,
337
+ "loss": 1.9509,
338
+ "step": 27500
339
+ },
340
+ {
341
+ "epoch": 2.33,
342
+ "learning_rate": 2.671324018629408e-05,
343
+ "loss": 1.8853,
344
+ "step": 28000
345
+ },
346
+ {
347
+ "epoch": 2.37,
348
+ "learning_rate": 2.629740518962076e-05,
349
+ "loss": 1.9285,
350
+ "step": 28500
351
+ },
352
+ {
353
+ "epoch": 2.41,
354
+ "learning_rate": 2.5881570192947442e-05,
355
+ "loss": 1.9212,
356
+ "step": 29000
357
+ },
358
+ {
359
+ "epoch": 2.45,
360
+ "learning_rate": 2.546573519627412e-05,
361
+ "loss": 1.8891,
362
+ "step": 29500
363
+ },
364
+ {
365
+ "epoch": 2.5,
366
+ "learning_rate": 2.5049900199600802e-05,
367
+ "loss": 1.9006,
368
+ "step": 30000
369
+ },
370
+ {
371
+ "epoch": 2.54,
372
+ "learning_rate": 2.4634065202927477e-05,
373
+ "loss": 1.9119,
374
+ "step": 30500
375
+ },
376
+ {
377
+ "epoch": 2.58,
378
+ "learning_rate": 2.421823020625416e-05,
379
+ "loss": 1.8479,
380
+ "step": 31000
381
+ },
382
+ {
383
+ "epoch": 2.62,
384
+ "learning_rate": 2.3802395209580838e-05,
385
+ "loss": 1.9032,
386
+ "step": 31500
387
+ },
388
+ {
389
+ "epoch": 2.66,
390
+ "learning_rate": 2.338656021290752e-05,
391
+ "loss": 1.9279,
392
+ "step": 32000
393
+ },
394
+ {
395
+ "epoch": 2.7,
396
+ "learning_rate": 2.2970725216234198e-05,
397
+ "loss": 1.8265,
398
+ "step": 32500
399
+ },
400
+ {
401
+ "epoch": 2.74,
402
+ "learning_rate": 2.255489021956088e-05,
403
+ "loss": 1.8167,
404
+ "step": 33000
405
+ },
406
+ {
407
+ "epoch": 2.79,
408
+ "learning_rate": 2.2139055222887558e-05,
409
+ "loss": 1.9173,
410
+ "step": 33500
411
+ },
412
+ {
413
+ "epoch": 2.83,
414
+ "learning_rate": 2.172322022621424e-05,
415
+ "loss": 1.7094,
416
+ "step": 34000
417
+ },
418
+ {
419
+ "epoch": 2.87,
420
+ "learning_rate": 2.130738522954092e-05,
421
+ "loss": 1.8966,
422
+ "step": 34500
423
+ },
424
+ {
425
+ "epoch": 2.91,
426
+ "learning_rate": 2.08915502328676e-05,
427
+ "loss": 1.9069,
428
+ "step": 35000
429
+ },
430
+ {
431
+ "epoch": 2.95,
432
+ "learning_rate": 2.047571523619428e-05,
433
+ "loss": 1.8751,
434
+ "step": 35500
435
+ },
436
+ {
437
+ "epoch": 2.99,
438
+ "learning_rate": 2.0059880239520957e-05,
439
+ "loss": 1.8776,
440
+ "step": 36000
441
+ },
442
+ {
443
+ "epoch": 3.04,
444
+ "learning_rate": 1.964404524284764e-05,
445
+ "loss": 1.907,
446
+ "step": 36500
447
+ },
448
+ {
449
+ "epoch": 3.08,
450
+ "learning_rate": 1.9228210246174318e-05,
451
+ "loss": 1.8639,
452
+ "step": 37000
453
+ },
454
+ {
455
+ "epoch": 3.12,
456
+ "learning_rate": 1.8812375249501e-05,
457
+ "loss": 1.8729,
458
+ "step": 37500
459
+ },
460
+ {
461
+ "epoch": 3.16,
462
+ "learning_rate": 1.8396540252827678e-05,
463
+ "loss": 1.8241,
464
+ "step": 38000
465
+ },
466
+ {
467
+ "epoch": 3.2,
468
+ "learning_rate": 1.798070525615436e-05,
469
+ "loss": 1.7862,
470
+ "step": 38500
471
+ },
472
+ {
473
+ "epoch": 3.24,
474
+ "learning_rate": 1.756487025948104e-05,
475
+ "loss": 1.8483,
476
+ "step": 39000
477
+ },
478
+ {
479
+ "epoch": 3.29,
480
+ "learning_rate": 1.714903526280772e-05,
481
+ "loss": 1.9772,
482
+ "step": 39500
483
+ },
484
+ {
485
+ "epoch": 3.33,
486
+ "learning_rate": 1.67332002661344e-05,
487
+ "loss": 1.825,
488
+ "step": 40000
489
+ },
490
+ {
491
+ "epoch": 3.37,
492
+ "learning_rate": 1.631736526946108e-05,
493
+ "loss": 1.9203,
494
+ "step": 40500
495
+ },
496
+ {
497
+ "epoch": 3.41,
498
+ "learning_rate": 1.590153027278776e-05,
499
+ "loss": 1.8294,
500
+ "step": 41000
501
+ },
502
+ {
503
+ "epoch": 3.45,
504
+ "learning_rate": 1.548569527611444e-05,
505
+ "loss": 1.7971,
506
+ "step": 41500
507
+ },
508
+ {
509
+ "epoch": 3.49,
510
+ "learning_rate": 1.506986027944112e-05,
511
+ "loss": 1.7856,
512
+ "step": 42000
513
+ },
514
+ {
515
+ "epoch": 3.53,
516
+ "learning_rate": 1.4654025282767798e-05,
517
+ "loss": 1.834,
518
+ "step": 42500
519
+ },
520
+ {
521
+ "epoch": 3.58,
522
+ "learning_rate": 1.4238190286094476e-05,
523
+ "loss": 1.7484,
524
+ "step": 43000
525
+ },
526
+ {
527
+ "epoch": 3.62,
528
+ "learning_rate": 1.3822355289421157e-05,
529
+ "loss": 1.7906,
530
+ "step": 43500
531
+ },
532
+ {
533
+ "epoch": 3.66,
534
+ "learning_rate": 1.3406520292747837e-05,
535
+ "loss": 1.8818,
536
+ "step": 44000
537
+ },
538
+ {
539
+ "epoch": 3.7,
540
+ "learning_rate": 1.2990685296074517e-05,
541
+ "loss": 1.9182,
542
+ "step": 44500
543
+ },
544
+ {
545
+ "epoch": 3.74,
546
+ "learning_rate": 1.2574850299401197e-05,
547
+ "loss": 1.8451,
548
+ "step": 45000
549
+ },
550
+ {
551
+ "epoch": 3.78,
552
+ "learning_rate": 1.2159015302727877e-05,
553
+ "loss": 1.8165,
554
+ "step": 45500
555
+ },
556
+ {
557
+ "epoch": 3.83,
558
+ "learning_rate": 1.1743180306054557e-05,
559
+ "loss": 1.8486,
560
+ "step": 46000
561
+ },
562
+ {
563
+ "epoch": 3.87,
564
+ "learning_rate": 1.1327345309381238e-05,
565
+ "loss": 1.6533,
566
+ "step": 46500
567
+ },
568
+ {
569
+ "epoch": 3.91,
570
+ "learning_rate": 1.0911510312707918e-05,
571
+ "loss": 1.7231,
572
+ "step": 47000
573
+ },
574
+ {
575
+ "epoch": 3.95,
576
+ "learning_rate": 1.0495675316034598e-05,
577
+ "loss": 1.884,
578
+ "step": 47500
579
+ },
580
+ {
581
+ "epoch": 3.99,
582
+ "learning_rate": 1.0079840319361278e-05,
583
+ "loss": 1.7961,
584
+ "step": 48000
585
+ },
586
+ {
587
+ "epoch": 4.03,
588
+ "learning_rate": 9.664005322687958e-06,
589
+ "loss": 1.8359,
590
+ "step": 48500
591
+ },
592
+ {
593
+ "epoch": 4.08,
594
+ "learning_rate": 9.248170326014638e-06,
595
+ "loss": 1.7858,
596
+ "step": 49000
597
+ },
598
+ {
599
+ "epoch": 4.12,
600
+ "learning_rate": 8.832335329341319e-06,
601
+ "loss": 1.8043,
602
+ "step": 49500
603
+ },
604
+ {
605
+ "epoch": 4.16,
606
+ "learning_rate": 8.416500332667999e-06,
607
+ "loss": 1.9039,
608
+ "step": 50000
609
+ },
610
+ {
611
+ "epoch": 4.2,
612
+ "learning_rate": 8.000665335994679e-06,
613
+ "loss": 1.7695,
614
+ "step": 50500
615
+ },
616
+ {
617
+ "epoch": 4.24,
618
+ "learning_rate": 7.584830339321358e-06,
619
+ "loss": 1.8137,
620
+ "step": 51000
621
+ },
622
+ {
623
+ "epoch": 4.28,
624
+ "learning_rate": 7.168995342648038e-06,
625
+ "loss": 1.7972,
626
+ "step": 51500
627
+ },
628
+ {
629
+ "epoch": 4.32,
630
+ "learning_rate": 6.753160345974717e-06,
631
+ "loss": 1.8118,
632
+ "step": 52000
633
+ },
634
+ {
635
+ "epoch": 4.37,
636
+ "learning_rate": 6.337325349301397e-06,
637
+ "loss": 1.8511,
638
+ "step": 52500
639
+ },
640
+ {
641
+ "epoch": 4.41,
642
+ "learning_rate": 5.921490352628078e-06,
643
+ "loss": 1.8975,
644
+ "step": 53000
645
+ },
646
+ {
647
+ "epoch": 4.45,
648
+ "learning_rate": 5.505655355954757e-06,
649
+ "loss": 1.7506,
650
+ "step": 53500
651
+ },
652
+ {
653
+ "epoch": 4.49,
654
+ "learning_rate": 5.0898203592814375e-06,
655
+ "loss": 1.7564,
656
+ "step": 54000
657
+ },
658
+ {
659
+ "epoch": 4.53,
660
+ "learning_rate": 4.673985362608118e-06,
661
+ "loss": 1.9306,
662
+ "step": 54500
663
+ },
664
+ {
665
+ "epoch": 4.57,
666
+ "learning_rate": 4.258150365934797e-06,
667
+ "loss": 1.7743,
668
+ "step": 55000
669
+ },
670
+ {
671
+ "epoch": 4.62,
672
+ "learning_rate": 3.842315369261477e-06,
673
+ "loss": 1.8154,
674
+ "step": 55500
675
+ },
676
+ {
677
+ "epoch": 4.66,
678
+ "learning_rate": 3.4264803725881573e-06,
679
+ "loss": 1.7763,
680
+ "step": 56000
681
+ },
682
+ {
683
+ "epoch": 4.7,
684
+ "learning_rate": 3.010645375914837e-06,
685
+ "loss": 1.7074,
686
+ "step": 56500
687
+ },
688
+ {
689
+ "epoch": 4.74,
690
+ "learning_rate": 2.594810379241517e-06,
691
+ "loss": 1.8081,
692
+ "step": 57000
693
+ },
694
+ {
695
+ "epoch": 4.78,
696
+ "learning_rate": 2.178975382568197e-06,
697
+ "loss": 1.7563,
698
+ "step": 57500
699
+ },
700
+ {
701
+ "epoch": 4.82,
702
+ "learning_rate": 1.7631403858948771e-06,
703
+ "loss": 1.818,
704
+ "step": 58000
705
+ },
706
+ {
707
+ "epoch": 4.87,
708
+ "learning_rate": 1.3473053892215569e-06,
709
+ "loss": 1.7706,
710
+ "step": 58500
711
+ },
712
+ {
713
+ "epoch": 4.91,
714
+ "learning_rate": 9.314703925482368e-07,
715
+ "loss": 1.8389,
716
+ "step": 59000
717
+ },
718
+ {
719
+ "epoch": 4.95,
720
+ "learning_rate": 5.156353958749169e-07,
721
+ "loss": 1.8149,
722
+ "step": 59500
723
+ },
724
+ {
725
+ "epoch": 4.99,
726
+ "learning_rate": 9.98003992015968e-08,
727
+ "loss": 1.7843,
728
+ "step": 60000
729
+ },
730
+ {
731
+ "epoch": 5.0,
732
+ "step": 60120,
733
+ "total_flos": 3.025664452067328e+16,
734
+ "train_runtime": 9052.4232,
735
+ "train_samples_per_second": 6.641
736
+ }
737
+ ],
738
+ "max_steps": 60120,
739
+ "num_train_epochs": 5,
740
+ "total_flos": 3.025664452067328e+16,
741
+ "trial_name": null,
742
+ "trial_params": null
743
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:872709e973d39d6d742c76a9b06ad4af3c839ada5c70807675010bfb8aecceee
3
+ size 2287
vocab.json ADDED
The diff for this file is too large to render. See raw diff