pgfeldman commited on
Commit
5ae1cd4
1 Parent(s): 4241ac9

Upload 12 files

Browse files
all_results.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 6.0,
3
+ "eval_loss": 2.896003007888794,
4
+ "eval_runtime": 170.7284,
5
+ "eval_samples": 2479,
6
+ "eval_samples_per_second": 14.52,
7
+ "eval_steps_per_second": 1.816,
8
+ "perplexity": 18.10164843512799,
9
+ "train_loss": 2.887446778400044,
10
+ "train_runtime": 18728.1376,
11
+ "train_samples": 12168,
12
+ "train_samples_per_second": 3.898,
13
+ "train_steps_per_second": 3.898
14
+ }
config.json ADDED
@@ -0,0 +1,38 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "gpt2",
3
+ "activation_function": "gelu_new",
4
+ "architectures": [
5
+ "GPT2LMHeadModel"
6
+ ],
7
+ "attn_pdrop": 0.1,
8
+ "bos_token_id": 50256,
9
+ "embd_pdrop": 0.1,
10
+ "eos_token_id": 50256,
11
+ "gradient_checkpointing": false,
12
+ "initializer_range": 0.02,
13
+ "layer_norm_epsilon": 1e-05,
14
+ "model_type": "gpt2",
15
+ "n_ctx": 1024,
16
+ "n_embd": 768,
17
+ "n_head": 12,
18
+ "n_inner": null,
19
+ "n_layer": 12,
20
+ "n_positions": 1024,
21
+ "resid_pdrop": 0.1,
22
+ "scale_attn_weights": true,
23
+ "summary_activation": null,
24
+ "summary_first_dropout": 0.1,
25
+ "summary_proj_to_labels": true,
26
+ "summary_type": "cls_index",
27
+ "summary_use_proj": true,
28
+ "task_specific_params": {
29
+ "text-generation": {
30
+ "do_sample": true,
31
+ "max_length": 50
32
+ }
33
+ },
34
+ "torch_dtype": "float32",
35
+ "transformers_version": "4.10.0.dev0",
36
+ "use_cache": true,
37
+ "vocab_size": 50257
38
+ }
eval_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 6.0,
3
+ "eval_loss": 2.896003007888794,
4
+ "eval_runtime": 170.7284,
5
+ "eval_samples": 2479,
6
+ "eval_samples_per_second": 14.52,
7
+ "eval_steps_per_second": 1.816,
8
+ "perplexity": 18.10164843512799
9
+ }
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 sha256:36866cd96e3f3da8af987efb25c5ffb0cc7932896e4612bba93091c7a066ff22
3
+ size 510407951
special_tokens_map.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"bos_token": "<|endoftext|>", "eos_token": "<|endoftext|>", "unk_token": "<|endoftext|>"}
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
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, "special_tokens_map_file": null, "name_or_path": "gpt2", "tokenizer_class": "GPT2Tokenizer"}
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 6.0,
3
+ "train_loss": 2.887446778400044,
4
+ "train_runtime": 18728.1376,
5
+ "train_samples": 12168,
6
+ "train_samples_per_second": 3.898,
7
+ "train_steps_per_second": 3.898
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,901 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 6.0,
5
+ "global_step": 73008,
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": 4.965757177295639e-05,
13
+ "loss": 3.431,
14
+ "step": 500
15
+ },
16
+ {
17
+ "epoch": 0.08,
18
+ "learning_rate": 4.931514354591278e-05,
19
+ "loss": 3.3117,
20
+ "step": 1000
21
+ },
22
+ {
23
+ "epoch": 0.12,
24
+ "learning_rate": 4.8972715318869165e-05,
25
+ "loss": 3.2778,
26
+ "step": 1500
27
+ },
28
+ {
29
+ "epoch": 0.16,
30
+ "learning_rate": 4.863028709182555e-05,
31
+ "loss": 3.2331,
32
+ "step": 2000
33
+ },
34
+ {
35
+ "epoch": 0.21,
36
+ "learning_rate": 4.828785886478194e-05,
37
+ "loss": 3.2122,
38
+ "step": 2500
39
+ },
40
+ {
41
+ "epoch": 0.25,
42
+ "learning_rate": 4.794543063773833e-05,
43
+ "loss": 3.1961,
44
+ "step": 3000
45
+ },
46
+ {
47
+ "epoch": 0.29,
48
+ "learning_rate": 4.7603002410694716e-05,
49
+ "loss": 3.1842,
50
+ "step": 3500
51
+ },
52
+ {
53
+ "epoch": 0.33,
54
+ "learning_rate": 4.726057418365111e-05,
55
+ "loss": 3.1603,
56
+ "step": 4000
57
+ },
58
+ {
59
+ "epoch": 0.37,
60
+ "learning_rate": 4.69181459566075e-05,
61
+ "loss": 3.147,
62
+ "step": 4500
63
+ },
64
+ {
65
+ "epoch": 0.41,
66
+ "learning_rate": 4.6575717729563886e-05,
67
+ "loss": 3.1572,
68
+ "step": 5000
69
+ },
70
+ {
71
+ "epoch": 0.45,
72
+ "learning_rate": 4.623328950252027e-05,
73
+ "loss": 3.1329,
74
+ "step": 5500
75
+ },
76
+ {
77
+ "epoch": 0.49,
78
+ "learning_rate": 4.589086127547666e-05,
79
+ "loss": 3.1132,
80
+ "step": 6000
81
+ },
82
+ {
83
+ "epoch": 0.53,
84
+ "learning_rate": 4.554843304843305e-05,
85
+ "loss": 3.1188,
86
+ "step": 6500
87
+ },
88
+ {
89
+ "epoch": 0.58,
90
+ "learning_rate": 4.520600482138944e-05,
91
+ "loss": 3.1141,
92
+ "step": 7000
93
+ },
94
+ {
95
+ "epoch": 0.62,
96
+ "learning_rate": 4.486357659434583e-05,
97
+ "loss": 3.1044,
98
+ "step": 7500
99
+ },
100
+ {
101
+ "epoch": 0.66,
102
+ "learning_rate": 4.452114836730222e-05,
103
+ "loss": 3.0998,
104
+ "step": 8000
105
+ },
106
+ {
107
+ "epoch": 0.7,
108
+ "learning_rate": 4.4178720140258606e-05,
109
+ "loss": 3.0963,
110
+ "step": 8500
111
+ },
112
+ {
113
+ "epoch": 0.74,
114
+ "learning_rate": 4.3836291913214994e-05,
115
+ "loss": 3.0978,
116
+ "step": 9000
117
+ },
118
+ {
119
+ "epoch": 0.78,
120
+ "learning_rate": 4.349386368617138e-05,
121
+ "loss": 3.0674,
122
+ "step": 9500
123
+ },
124
+ {
125
+ "epoch": 0.82,
126
+ "learning_rate": 4.315143545912777e-05,
127
+ "loss": 3.0855,
128
+ "step": 10000
129
+ },
130
+ {
131
+ "epoch": 0.86,
132
+ "learning_rate": 4.2809007232084157e-05,
133
+ "loss": 3.0664,
134
+ "step": 10500
135
+ },
136
+ {
137
+ "epoch": 0.9,
138
+ "learning_rate": 4.2466579005040544e-05,
139
+ "loss": 3.0678,
140
+ "step": 11000
141
+ },
142
+ {
143
+ "epoch": 0.95,
144
+ "learning_rate": 4.212415077799693e-05,
145
+ "loss": 3.0681,
146
+ "step": 11500
147
+ },
148
+ {
149
+ "epoch": 0.99,
150
+ "learning_rate": 4.178172255095332e-05,
151
+ "loss": 3.0565,
152
+ "step": 12000
153
+ },
154
+ {
155
+ "epoch": 1.03,
156
+ "learning_rate": 4.1439294323909714e-05,
157
+ "loss": 3.0114,
158
+ "step": 12500
159
+ },
160
+ {
161
+ "epoch": 1.07,
162
+ "learning_rate": 4.10968660968661e-05,
163
+ "loss": 2.9869,
164
+ "step": 13000
165
+ },
166
+ {
167
+ "epoch": 1.11,
168
+ "learning_rate": 4.075443786982249e-05,
169
+ "loss": 2.9743,
170
+ "step": 13500
171
+ },
172
+ {
173
+ "epoch": 1.15,
174
+ "learning_rate": 4.041200964277888e-05,
175
+ "loss": 2.9838,
176
+ "step": 14000
177
+ },
178
+ {
179
+ "epoch": 1.19,
180
+ "learning_rate": 4.0069581415735264e-05,
181
+ "loss": 2.9806,
182
+ "step": 14500
183
+ },
184
+ {
185
+ "epoch": 1.23,
186
+ "learning_rate": 3.972715318869165e-05,
187
+ "loss": 2.9715,
188
+ "step": 15000
189
+ },
190
+ {
191
+ "epoch": 1.27,
192
+ "learning_rate": 3.938472496164804e-05,
193
+ "loss": 2.9702,
194
+ "step": 15500
195
+ },
196
+ {
197
+ "epoch": 1.31,
198
+ "learning_rate": 3.904229673460443e-05,
199
+ "loss": 2.9667,
200
+ "step": 16000
201
+ },
202
+ {
203
+ "epoch": 1.36,
204
+ "learning_rate": 3.8699868507560815e-05,
205
+ "loss": 2.9737,
206
+ "step": 16500
207
+ },
208
+ {
209
+ "epoch": 1.4,
210
+ "learning_rate": 3.83574402805172e-05,
211
+ "loss": 2.9569,
212
+ "step": 17000
213
+ },
214
+ {
215
+ "epoch": 1.44,
216
+ "learning_rate": 3.801501205347359e-05,
217
+ "loss": 2.9759,
218
+ "step": 17500
219
+ },
220
+ {
221
+ "epoch": 1.48,
222
+ "learning_rate": 3.767258382642998e-05,
223
+ "loss": 2.9749,
224
+ "step": 18000
225
+ },
226
+ {
227
+ "epoch": 1.52,
228
+ "learning_rate": 3.7330155599386366e-05,
229
+ "loss": 2.9742,
230
+ "step": 18500
231
+ },
232
+ {
233
+ "epoch": 1.56,
234
+ "learning_rate": 3.698772737234275e-05,
235
+ "loss": 2.97,
236
+ "step": 19000
237
+ },
238
+ {
239
+ "epoch": 1.6,
240
+ "learning_rate": 3.664529914529915e-05,
241
+ "loss": 2.9551,
242
+ "step": 19500
243
+ },
244
+ {
245
+ "epoch": 1.64,
246
+ "learning_rate": 3.6302870918255535e-05,
247
+ "loss": 2.9464,
248
+ "step": 20000
249
+ },
250
+ {
251
+ "epoch": 1.68,
252
+ "learning_rate": 3.596044269121192e-05,
253
+ "loss": 2.9549,
254
+ "step": 20500
255
+ },
256
+ {
257
+ "epoch": 1.73,
258
+ "learning_rate": 3.561801446416832e-05,
259
+ "loss": 2.969,
260
+ "step": 21000
261
+ },
262
+ {
263
+ "epoch": 1.77,
264
+ "learning_rate": 3.5275586237124705e-05,
265
+ "loss": 2.9478,
266
+ "step": 21500
267
+ },
268
+ {
269
+ "epoch": 1.81,
270
+ "learning_rate": 3.493315801008109e-05,
271
+ "loss": 2.9568,
272
+ "step": 22000
273
+ },
274
+ {
275
+ "epoch": 1.85,
276
+ "learning_rate": 3.459072978303748e-05,
277
+ "loss": 2.9509,
278
+ "step": 22500
279
+ },
280
+ {
281
+ "epoch": 1.89,
282
+ "learning_rate": 3.424830155599387e-05,
283
+ "loss": 2.9484,
284
+ "step": 23000
285
+ },
286
+ {
287
+ "epoch": 1.93,
288
+ "learning_rate": 3.3905873328950256e-05,
289
+ "loss": 2.9646,
290
+ "step": 23500
291
+ },
292
+ {
293
+ "epoch": 1.97,
294
+ "learning_rate": 3.356344510190664e-05,
295
+ "loss": 2.9476,
296
+ "step": 24000
297
+ },
298
+ {
299
+ "epoch": 2.01,
300
+ "learning_rate": 3.322101687486303e-05,
301
+ "loss": 2.9165,
302
+ "step": 24500
303
+ },
304
+ {
305
+ "epoch": 2.05,
306
+ "learning_rate": 3.287858864781942e-05,
307
+ "loss": 2.8741,
308
+ "step": 25000
309
+ },
310
+ {
311
+ "epoch": 2.1,
312
+ "learning_rate": 3.2536160420775806e-05,
313
+ "loss": 2.8902,
314
+ "step": 25500
315
+ },
316
+ {
317
+ "epoch": 2.14,
318
+ "learning_rate": 3.2193732193732194e-05,
319
+ "loss": 2.8788,
320
+ "step": 26000
321
+ },
322
+ {
323
+ "epoch": 2.18,
324
+ "learning_rate": 3.185130396668858e-05,
325
+ "loss": 2.8775,
326
+ "step": 26500
327
+ },
328
+ {
329
+ "epoch": 2.22,
330
+ "learning_rate": 3.150887573964497e-05,
331
+ "loss": 2.885,
332
+ "step": 27000
333
+ },
334
+ {
335
+ "epoch": 2.26,
336
+ "learning_rate": 3.116644751260136e-05,
337
+ "loss": 2.8678,
338
+ "step": 27500
339
+ },
340
+ {
341
+ "epoch": 2.3,
342
+ "learning_rate": 3.082401928555775e-05,
343
+ "loss": 2.8771,
344
+ "step": 28000
345
+ },
346
+ {
347
+ "epoch": 2.34,
348
+ "learning_rate": 3.0481591058514136e-05,
349
+ "loss": 2.8873,
350
+ "step": 28500
351
+ },
352
+ {
353
+ "epoch": 2.38,
354
+ "learning_rate": 3.0139162831470523e-05,
355
+ "loss": 2.8768,
356
+ "step": 29000
357
+ },
358
+ {
359
+ "epoch": 2.42,
360
+ "learning_rate": 2.979673460442691e-05,
361
+ "loss": 2.8938,
362
+ "step": 29500
363
+ },
364
+ {
365
+ "epoch": 2.47,
366
+ "learning_rate": 2.9454306377383302e-05,
367
+ "loss": 2.8774,
368
+ "step": 30000
369
+ },
370
+ {
371
+ "epoch": 2.51,
372
+ "learning_rate": 2.911187815033969e-05,
373
+ "loss": 2.8761,
374
+ "step": 30500
375
+ },
376
+ {
377
+ "epoch": 2.55,
378
+ "learning_rate": 2.8769449923296077e-05,
379
+ "loss": 2.8824,
380
+ "step": 31000
381
+ },
382
+ {
383
+ "epoch": 2.59,
384
+ "learning_rate": 2.8427021696252465e-05,
385
+ "loss": 2.8737,
386
+ "step": 31500
387
+ },
388
+ {
389
+ "epoch": 2.63,
390
+ "learning_rate": 2.8084593469208852e-05,
391
+ "loss": 2.8802,
392
+ "step": 32000
393
+ },
394
+ {
395
+ "epoch": 2.67,
396
+ "learning_rate": 2.774216524216524e-05,
397
+ "loss": 2.8744,
398
+ "step": 32500
399
+ },
400
+ {
401
+ "epoch": 2.71,
402
+ "learning_rate": 2.7399737015121628e-05,
403
+ "loss": 2.8713,
404
+ "step": 33000
405
+ },
406
+ {
407
+ "epoch": 2.75,
408
+ "learning_rate": 2.705730878807802e-05,
409
+ "loss": 2.8822,
410
+ "step": 33500
411
+ },
412
+ {
413
+ "epoch": 2.79,
414
+ "learning_rate": 2.671488056103441e-05,
415
+ "loss": 2.8809,
416
+ "step": 34000
417
+ },
418
+ {
419
+ "epoch": 2.84,
420
+ "learning_rate": 2.63724523339908e-05,
421
+ "loss": 2.8694,
422
+ "step": 34500
423
+ },
424
+ {
425
+ "epoch": 2.88,
426
+ "learning_rate": 2.603002410694719e-05,
427
+ "loss": 2.8723,
428
+ "step": 35000
429
+ },
430
+ {
431
+ "epoch": 2.92,
432
+ "learning_rate": 2.5687595879903576e-05,
433
+ "loss": 2.8734,
434
+ "step": 35500
435
+ },
436
+ {
437
+ "epoch": 2.96,
438
+ "learning_rate": 2.5345167652859964e-05,
439
+ "loss": 2.8757,
440
+ "step": 36000
441
+ },
442
+ {
443
+ "epoch": 3.0,
444
+ "learning_rate": 2.500273942581635e-05,
445
+ "loss": 2.8737,
446
+ "step": 36500
447
+ },
448
+ {
449
+ "epoch": 3.04,
450
+ "learning_rate": 2.466031119877274e-05,
451
+ "loss": 2.8095,
452
+ "step": 37000
453
+ },
454
+ {
455
+ "epoch": 3.08,
456
+ "learning_rate": 2.4317882971729127e-05,
457
+ "loss": 2.8181,
458
+ "step": 37500
459
+ },
460
+ {
461
+ "epoch": 3.12,
462
+ "learning_rate": 2.3975454744685514e-05,
463
+ "loss": 2.8102,
464
+ "step": 38000
465
+ },
466
+ {
467
+ "epoch": 3.16,
468
+ "learning_rate": 2.3633026517641905e-05,
469
+ "loss": 2.8261,
470
+ "step": 38500
471
+ },
472
+ {
473
+ "epoch": 3.21,
474
+ "learning_rate": 2.3290598290598293e-05,
475
+ "loss": 2.8121,
476
+ "step": 39000
477
+ },
478
+ {
479
+ "epoch": 3.25,
480
+ "learning_rate": 2.294817006355468e-05,
481
+ "loss": 2.8127,
482
+ "step": 39500
483
+ },
484
+ {
485
+ "epoch": 3.29,
486
+ "learning_rate": 2.260574183651107e-05,
487
+ "loss": 2.8246,
488
+ "step": 40000
489
+ },
490
+ {
491
+ "epoch": 3.33,
492
+ "learning_rate": 2.2263313609467456e-05,
493
+ "loss": 2.8188,
494
+ "step": 40500
495
+ },
496
+ {
497
+ "epoch": 3.37,
498
+ "learning_rate": 2.1920885382423844e-05,
499
+ "loss": 2.8247,
500
+ "step": 41000
501
+ },
502
+ {
503
+ "epoch": 3.41,
504
+ "learning_rate": 2.157845715538023e-05,
505
+ "loss": 2.8087,
506
+ "step": 41500
507
+ },
508
+ {
509
+ "epoch": 3.45,
510
+ "learning_rate": 2.1236028928336622e-05,
511
+ "loss": 2.8135,
512
+ "step": 42000
513
+ },
514
+ {
515
+ "epoch": 3.49,
516
+ "learning_rate": 2.089360070129301e-05,
517
+ "loss": 2.8208,
518
+ "step": 42500
519
+ },
520
+ {
521
+ "epoch": 3.53,
522
+ "learning_rate": 2.0551172474249398e-05,
523
+ "loss": 2.8227,
524
+ "step": 43000
525
+ },
526
+ {
527
+ "epoch": 3.57,
528
+ "learning_rate": 2.0208744247205785e-05,
529
+ "loss": 2.8272,
530
+ "step": 43500
531
+ },
532
+ {
533
+ "epoch": 3.62,
534
+ "learning_rate": 1.9866316020162176e-05,
535
+ "loss": 2.8168,
536
+ "step": 44000
537
+ },
538
+ {
539
+ "epoch": 3.66,
540
+ "learning_rate": 1.9523887793118564e-05,
541
+ "loss": 2.8304,
542
+ "step": 44500
543
+ },
544
+ {
545
+ "epoch": 3.7,
546
+ "learning_rate": 1.918145956607495e-05,
547
+ "loss": 2.8102,
548
+ "step": 45000
549
+ },
550
+ {
551
+ "epoch": 3.74,
552
+ "learning_rate": 1.883903133903134e-05,
553
+ "loss": 2.8139,
554
+ "step": 45500
555
+ },
556
+ {
557
+ "epoch": 3.78,
558
+ "learning_rate": 1.849660311198773e-05,
559
+ "loss": 2.806,
560
+ "step": 46000
561
+ },
562
+ {
563
+ "epoch": 3.82,
564
+ "learning_rate": 1.8154174884944118e-05,
565
+ "loss": 2.8253,
566
+ "step": 46500
567
+ },
568
+ {
569
+ "epoch": 3.86,
570
+ "learning_rate": 1.7811746657900506e-05,
571
+ "loss": 2.8237,
572
+ "step": 47000
573
+ },
574
+ {
575
+ "epoch": 3.9,
576
+ "learning_rate": 1.7469318430856893e-05,
577
+ "loss": 2.8118,
578
+ "step": 47500
579
+ },
580
+ {
581
+ "epoch": 3.94,
582
+ "learning_rate": 1.712689020381328e-05,
583
+ "loss": 2.8203,
584
+ "step": 48000
585
+ },
586
+ {
587
+ "epoch": 3.99,
588
+ "learning_rate": 1.678446197676967e-05,
589
+ "loss": 2.8142,
590
+ "step": 48500
591
+ },
592
+ {
593
+ "epoch": 4.03,
594
+ "learning_rate": 1.6442033749726056e-05,
595
+ "loss": 2.7918,
596
+ "step": 49000
597
+ },
598
+ {
599
+ "epoch": 4.07,
600
+ "learning_rate": 1.6099605522682447e-05,
601
+ "loss": 2.7683,
602
+ "step": 49500
603
+ },
604
+ {
605
+ "epoch": 4.11,
606
+ "learning_rate": 1.5757177295638835e-05,
607
+ "loss": 2.7645,
608
+ "step": 50000
609
+ },
610
+ {
611
+ "epoch": 4.15,
612
+ "learning_rate": 1.5414749068595226e-05,
613
+ "loss": 2.7562,
614
+ "step": 50500
615
+ },
616
+ {
617
+ "epoch": 4.19,
618
+ "learning_rate": 1.5072320841551612e-05,
619
+ "loss": 2.7706,
620
+ "step": 51000
621
+ },
622
+ {
623
+ "epoch": 4.23,
624
+ "learning_rate": 1.4729892614508001e-05,
625
+ "loss": 2.779,
626
+ "step": 51500
627
+ },
628
+ {
629
+ "epoch": 4.27,
630
+ "learning_rate": 1.4387464387464389e-05,
631
+ "loss": 2.7665,
632
+ "step": 52000
633
+ },
634
+ {
635
+ "epoch": 4.31,
636
+ "learning_rate": 1.4045036160420776e-05,
637
+ "loss": 2.7696,
638
+ "step": 52500
639
+ },
640
+ {
641
+ "epoch": 4.36,
642
+ "learning_rate": 1.3702607933377166e-05,
643
+ "loss": 2.7725,
644
+ "step": 53000
645
+ },
646
+ {
647
+ "epoch": 4.4,
648
+ "learning_rate": 1.3360179706333553e-05,
649
+ "loss": 2.7724,
650
+ "step": 53500
651
+ },
652
+ {
653
+ "epoch": 4.44,
654
+ "learning_rate": 1.3017751479289941e-05,
655
+ "loss": 2.7725,
656
+ "step": 54000
657
+ },
658
+ {
659
+ "epoch": 4.48,
660
+ "learning_rate": 1.2675323252246329e-05,
661
+ "loss": 2.7759,
662
+ "step": 54500
663
+ },
664
+ {
665
+ "epoch": 4.52,
666
+ "learning_rate": 1.2332895025202718e-05,
667
+ "loss": 2.7655,
668
+ "step": 55000
669
+ },
670
+ {
671
+ "epoch": 4.56,
672
+ "learning_rate": 1.1990466798159107e-05,
673
+ "loss": 2.7836,
674
+ "step": 55500
675
+ },
676
+ {
677
+ "epoch": 4.6,
678
+ "learning_rate": 1.1648038571115495e-05,
679
+ "loss": 2.7709,
680
+ "step": 56000
681
+ },
682
+ {
683
+ "epoch": 4.64,
684
+ "learning_rate": 1.1305610344071883e-05,
685
+ "loss": 2.7758,
686
+ "step": 56500
687
+ },
688
+ {
689
+ "epoch": 4.68,
690
+ "learning_rate": 1.0963182117028272e-05,
691
+ "loss": 2.7825,
692
+ "step": 57000
693
+ },
694
+ {
695
+ "epoch": 4.73,
696
+ "learning_rate": 1.062075388998466e-05,
697
+ "loss": 2.7641,
698
+ "step": 57500
699
+ },
700
+ {
701
+ "epoch": 4.77,
702
+ "learning_rate": 1.0278325662941047e-05,
703
+ "loss": 2.7886,
704
+ "step": 58000
705
+ },
706
+ {
707
+ "epoch": 4.81,
708
+ "learning_rate": 9.935897435897435e-06,
709
+ "loss": 2.7784,
710
+ "step": 58500
711
+ },
712
+ {
713
+ "epoch": 4.85,
714
+ "learning_rate": 9.593469208853826e-06,
715
+ "loss": 2.7692,
716
+ "step": 59000
717
+ },
718
+ {
719
+ "epoch": 4.89,
720
+ "learning_rate": 9.251040981810214e-06,
721
+ "loss": 2.7633,
722
+ "step": 59500
723
+ },
724
+ {
725
+ "epoch": 4.93,
726
+ "learning_rate": 8.908612754766601e-06,
727
+ "loss": 2.7832,
728
+ "step": 60000
729
+ },
730
+ {
731
+ "epoch": 4.97,
732
+ "learning_rate": 8.566184527722989e-06,
733
+ "loss": 2.7778,
734
+ "step": 60500
735
+ },
736
+ {
737
+ "epoch": 5.01,
738
+ "learning_rate": 8.223756300679378e-06,
739
+ "loss": 2.7643,
740
+ "step": 61000
741
+ },
742
+ {
743
+ "epoch": 5.05,
744
+ "learning_rate": 7.881328073635766e-06,
745
+ "loss": 2.7539,
746
+ "step": 61500
747
+ },
748
+ {
749
+ "epoch": 5.1,
750
+ "learning_rate": 7.538899846592154e-06,
751
+ "loss": 2.7482,
752
+ "step": 62000
753
+ },
754
+ {
755
+ "epoch": 5.14,
756
+ "learning_rate": 7.196471619548544e-06,
757
+ "loss": 2.7536,
758
+ "step": 62500
759
+ },
760
+ {
761
+ "epoch": 5.18,
762
+ "learning_rate": 6.854043392504931e-06,
763
+ "loss": 2.7419,
764
+ "step": 63000
765
+ },
766
+ {
767
+ "epoch": 5.22,
768
+ "learning_rate": 6.51161516546132e-06,
769
+ "loss": 2.731,
770
+ "step": 63500
771
+ },
772
+ {
773
+ "epoch": 5.26,
774
+ "learning_rate": 6.1691869384177075e-06,
775
+ "loss": 2.7355,
776
+ "step": 64000
777
+ },
778
+ {
779
+ "epoch": 5.3,
780
+ "learning_rate": 5.826758711374096e-06,
781
+ "loss": 2.7378,
782
+ "step": 64500
783
+ },
784
+ {
785
+ "epoch": 5.34,
786
+ "learning_rate": 5.4843304843304845e-06,
787
+ "loss": 2.7447,
788
+ "step": 65000
789
+ },
790
+ {
791
+ "epoch": 5.38,
792
+ "learning_rate": 5.141902257286873e-06,
793
+ "loss": 2.7548,
794
+ "step": 65500
795
+ },
796
+ {
797
+ "epoch": 5.42,
798
+ "learning_rate": 4.799474030243261e-06,
799
+ "loss": 2.7362,
800
+ "step": 66000
801
+ },
802
+ {
803
+ "epoch": 5.47,
804
+ "learning_rate": 4.45704580319965e-06,
805
+ "loss": 2.7434,
806
+ "step": 66500
807
+ },
808
+ {
809
+ "epoch": 5.51,
810
+ "learning_rate": 4.114617576156038e-06,
811
+ "loss": 2.7424,
812
+ "step": 67000
813
+ },
814
+ {
815
+ "epoch": 5.55,
816
+ "learning_rate": 3.772189349112426e-06,
817
+ "loss": 2.7468,
818
+ "step": 67500
819
+ },
820
+ {
821
+ "epoch": 5.59,
822
+ "learning_rate": 3.429761122068815e-06,
823
+ "loss": 2.7451,
824
+ "step": 68000
825
+ },
826
+ {
827
+ "epoch": 5.63,
828
+ "learning_rate": 3.087332895025203e-06,
829
+ "loss": 2.7504,
830
+ "step": 68500
831
+ },
832
+ {
833
+ "epoch": 5.67,
834
+ "learning_rate": 2.744904667981591e-06,
835
+ "loss": 2.7372,
836
+ "step": 69000
837
+ },
838
+ {
839
+ "epoch": 5.71,
840
+ "learning_rate": 2.4024764409379797e-06,
841
+ "loss": 2.7349,
842
+ "step": 69500
843
+ },
844
+ {
845
+ "epoch": 5.75,
846
+ "learning_rate": 2.0600482138943677e-06,
847
+ "loss": 2.7477,
848
+ "step": 70000
849
+ },
850
+ {
851
+ "epoch": 5.79,
852
+ "learning_rate": 1.7176199868507562e-06,
853
+ "loss": 2.7394,
854
+ "step": 70500
855
+ },
856
+ {
857
+ "epoch": 5.83,
858
+ "learning_rate": 1.3751917598071445e-06,
859
+ "loss": 2.7303,
860
+ "step": 71000
861
+ },
862
+ {
863
+ "epoch": 5.88,
864
+ "learning_rate": 1.0327635327635328e-06,
865
+ "loss": 2.7408,
866
+ "step": 71500
867
+ },
868
+ {
869
+ "epoch": 5.92,
870
+ "learning_rate": 6.903353057199211e-07,
871
+ "loss": 2.7455,
872
+ "step": 72000
873
+ },
874
+ {
875
+ "epoch": 5.96,
876
+ "learning_rate": 3.479070786763094e-07,
877
+ "loss": 2.7378,
878
+ "step": 72500
879
+ },
880
+ {
881
+ "epoch": 6.0,
882
+ "learning_rate": 5.478851632697787e-09,
883
+ "loss": 2.7414,
884
+ "step": 73000
885
+ },
886
+ {
887
+ "epoch": 6.0,
888
+ "step": 73008,
889
+ "total_flos": 3.8152817344512e+16,
890
+ "train_loss": 2.887446778400044,
891
+ "train_runtime": 18728.1376,
892
+ "train_samples_per_second": 3.898,
893
+ "train_steps_per_second": 3.898
894
+ }
895
+ ],
896
+ "max_steps": 73008,
897
+ "num_train_epochs": 6,
898
+ "total_flos": 3.8152817344512e+16,
899
+ "trial_name": null,
900
+ "trial_params": null
901
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:35d2e854d9ce89a987257c734e0e0a275d7bb12008817e12741f6efa1d772afe
3
+ size 2607
vocab.json ADDED
The diff for this file is too large to render. See raw diff