tmnam20 commited on
Commit
6efe0dc
1 Parent(s): 74ae22c

Training in progress, step 68000, checkpoint

Browse files
checkpoint-68000/added_tokens.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "<bos>": 40478,
3
+ "<eos>": 40479,
4
+ "<pad>": 40480
5
+ }
checkpoint-68000/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": 24,
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-68000/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-68000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2dd77bf9320f4d759fc9ccabb82522ad28e400dd08f0079b745c3b8d4406607e
3
+ size 932401605
checkpoint-68000/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ce473f947a03683f2aa4703641565741ad6528846274f8516d9d0303242d3349
3
+ size 466205145
checkpoint-68000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1653bf5ffa880828c00f5e1b4a896d12cfb6f7743faac6b40b5516a52ef43e66
3
+ size 14575
checkpoint-68000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0a542b2666cfb6f09db9958702d09011950d5fbe8b175ffa4374646fdec7277f
3
+ size 627
checkpoint-68000/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-68000/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-68000/trainer_state.json ADDED
@@ -0,0 +1,1045 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.9776698350906372,
3
+ "best_model_checkpoint": "/kaggle/output/checkpoint-68000",
4
+ "epoch": 2.770534550195567,
5
+ "eval_steps": 1000,
6
+ "global_step": 68000,
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": 1.1506,
15
+ "step": 1
16
+ },
17
+ {
18
+ "epoch": 0.04,
19
+ "learning_rate": 2.7750000000000004e-08,
20
+ "loss": 1.1326,
21
+ "step": 1000
22
+ },
23
+ {
24
+ "epoch": 0.04,
25
+ "eval_accuracy": 0.3331337325349301,
26
+ "eval_loss": 1.1057127714157104,
27
+ "eval_runtime": 61.3538,
28
+ "eval_samples_per_second": 81.657,
29
+ "eval_steps_per_second": 10.219,
30
+ "step": 1000
31
+ },
32
+ {
33
+ "epoch": 0.08,
34
+ "learning_rate": 5.5527777777777784e-08,
35
+ "loss": 1.1238,
36
+ "step": 2000
37
+ },
38
+ {
39
+ "epoch": 0.08,
40
+ "eval_accuracy": 0.33812375249501,
41
+ "eval_loss": 1.1044437885284424,
42
+ "eval_runtime": 61.256,
43
+ "eval_samples_per_second": 81.788,
44
+ "eval_steps_per_second": 10.236,
45
+ "step": 2000
46
+ },
47
+ {
48
+ "epoch": 0.12,
49
+ "learning_rate": 8.327777777777778e-08,
50
+ "loss": 1.1251,
51
+ "step": 3000
52
+ },
53
+ {
54
+ "epoch": 0.12,
55
+ "eval_accuracy": 0.3407185628742515,
56
+ "eval_loss": 1.1043975353240967,
57
+ "eval_runtime": 61.2433,
58
+ "eval_samples_per_second": 81.805,
59
+ "eval_steps_per_second": 10.238,
60
+ "step": 3000
61
+ },
62
+ {
63
+ "epoch": 0.16,
64
+ "learning_rate": 1.1105555555555557e-07,
65
+ "loss": 1.1205,
66
+ "step": 4000
67
+ },
68
+ {
69
+ "epoch": 0.16,
70
+ "eval_accuracy": 0.3387225548902196,
71
+ "eval_loss": 1.100860357284546,
72
+ "eval_runtime": 61.4457,
73
+ "eval_samples_per_second": 81.535,
74
+ "eval_steps_per_second": 10.204,
75
+ "step": 4000
76
+ },
77
+ {
78
+ "epoch": 0.2,
79
+ "learning_rate": 1.3880555555555558e-07,
80
+ "loss": 1.1192,
81
+ "step": 5000
82
+ },
83
+ {
84
+ "epoch": 0.2,
85
+ "eval_accuracy": 0.34151696606786425,
86
+ "eval_loss": 1.1005476713180542,
87
+ "eval_runtime": 61.2568,
88
+ "eval_samples_per_second": 81.787,
89
+ "eval_steps_per_second": 10.236,
90
+ "step": 5000
91
+ },
92
+ {
93
+ "epoch": 0.24,
94
+ "learning_rate": 1.6658333333333335e-07,
95
+ "loss": 1.1145,
96
+ "step": 6000
97
+ },
98
+ {
99
+ "epoch": 0.24,
100
+ "eval_accuracy": 0.3508982035928144,
101
+ "eval_loss": 1.0979759693145752,
102
+ "eval_runtime": 61.2221,
103
+ "eval_samples_per_second": 81.833,
104
+ "eval_steps_per_second": 10.241,
105
+ "step": 6000
106
+ },
107
+ {
108
+ "epoch": 0.29,
109
+ "learning_rate": 1.9433333333333334e-07,
110
+ "loss": 1.1164,
111
+ "step": 7000
112
+ },
113
+ {
114
+ "epoch": 0.29,
115
+ "eval_accuracy": 0.35508982035928144,
116
+ "eval_loss": 1.0955135822296143,
117
+ "eval_runtime": 61.255,
118
+ "eval_samples_per_second": 81.789,
119
+ "eval_steps_per_second": 10.236,
120
+ "step": 7000
121
+ },
122
+ {
123
+ "epoch": 0.33,
124
+ "learning_rate": 2.2211111111111114e-07,
125
+ "loss": 1.1091,
126
+ "step": 8000
127
+ },
128
+ {
129
+ "epoch": 0.33,
130
+ "eval_accuracy": 0.36147704590818364,
131
+ "eval_loss": 1.0946425199508667,
132
+ "eval_runtime": 61.3011,
133
+ "eval_samples_per_second": 81.728,
134
+ "eval_steps_per_second": 10.228,
135
+ "step": 8000
136
+ },
137
+ {
138
+ "epoch": 0.37,
139
+ "learning_rate": 2.4986111111111113e-07,
140
+ "loss": 1.107,
141
+ "step": 9000
142
+ },
143
+ {
144
+ "epoch": 0.37,
145
+ "eval_accuracy": 0.3471057884231537,
146
+ "eval_loss": 1.10226309299469,
147
+ "eval_runtime": 61.3193,
148
+ "eval_samples_per_second": 81.703,
149
+ "eval_steps_per_second": 10.225,
150
+ "step": 9000
151
+ },
152
+ {
153
+ "epoch": 0.41,
154
+ "learning_rate": 2.776388888888889e-07,
155
+ "loss": 1.1055,
156
+ "step": 10000
157
+ },
158
+ {
159
+ "epoch": 0.41,
160
+ "eval_accuracy": 0.3774451097804391,
161
+ "eval_loss": 1.0922847986221313,
162
+ "eval_runtime": 61.1866,
163
+ "eval_samples_per_second": 81.881,
164
+ "eval_steps_per_second": 10.247,
165
+ "step": 10000
166
+ },
167
+ {
168
+ "epoch": 0.45,
169
+ "learning_rate": 3.0541666666666667e-07,
170
+ "loss": 1.1028,
171
+ "step": 11000
172
+ },
173
+ {
174
+ "epoch": 0.45,
175
+ "eval_accuracy": 0.37425149700598803,
176
+ "eval_loss": 1.0914534330368042,
177
+ "eval_runtime": 61.341,
178
+ "eval_samples_per_second": 81.675,
179
+ "eval_steps_per_second": 10.222,
180
+ "step": 11000
181
+ },
182
+ {
183
+ "epoch": 0.49,
184
+ "learning_rate": 3.3319444444444444e-07,
185
+ "loss": 1.1024,
186
+ "step": 12000
187
+ },
188
+ {
189
+ "epoch": 0.49,
190
+ "eval_accuracy": 0.3872255489021956,
191
+ "eval_loss": 1.0888582468032837,
192
+ "eval_runtime": 61.4245,
193
+ "eval_samples_per_second": 81.564,
194
+ "eval_steps_per_second": 10.208,
195
+ "step": 12000
196
+ },
197
+ {
198
+ "epoch": 0.53,
199
+ "learning_rate": 3.6094444444444446e-07,
200
+ "loss": 1.1015,
201
+ "step": 13000
202
+ },
203
+ {
204
+ "epoch": 0.53,
205
+ "eval_accuracy": 0.3397205588822355,
206
+ "eval_loss": 1.096097707748413,
207
+ "eval_runtime": 61.3257,
208
+ "eval_samples_per_second": 81.695,
209
+ "eval_steps_per_second": 10.224,
210
+ "step": 13000
211
+ },
212
+ {
213
+ "epoch": 0.57,
214
+ "learning_rate": 3.8872222222222223e-07,
215
+ "loss": 1.1,
216
+ "step": 14000
217
+ },
218
+ {
219
+ "epoch": 0.57,
220
+ "eval_accuracy": 0.36387225548902197,
221
+ "eval_loss": 1.0940359830856323,
222
+ "eval_runtime": 61.2945,
223
+ "eval_samples_per_second": 81.737,
224
+ "eval_steps_per_second": 10.229,
225
+ "step": 14000
226
+ },
227
+ {
228
+ "epoch": 0.61,
229
+ "learning_rate": 4.1650000000000006e-07,
230
+ "loss": 1.0997,
231
+ "step": 15000
232
+ },
233
+ {
234
+ "epoch": 0.61,
235
+ "eval_accuracy": 0.38602794411177643,
236
+ "eval_loss": 1.0880299806594849,
237
+ "eval_runtime": 61.2546,
238
+ "eval_samples_per_second": 81.79,
239
+ "eval_steps_per_second": 10.236,
240
+ "step": 15000
241
+ },
242
+ {
243
+ "epoch": 0.65,
244
+ "learning_rate": 4.4425e-07,
245
+ "loss": 1.0987,
246
+ "step": 16000
247
+ },
248
+ {
249
+ "epoch": 0.65,
250
+ "eval_accuracy": 0.38962075848303396,
251
+ "eval_loss": 1.0837221145629883,
252
+ "eval_runtime": 61.1947,
253
+ "eval_samples_per_second": 81.87,
254
+ "eval_steps_per_second": 10.246,
255
+ "step": 16000
256
+ },
257
+ {
258
+ "epoch": 0.69,
259
+ "learning_rate": 4.7202777777777785e-07,
260
+ "loss": 1.0943,
261
+ "step": 17000
262
+ },
263
+ {
264
+ "epoch": 0.69,
265
+ "eval_accuracy": 0.4033932135728543,
266
+ "eval_loss": 1.081850290298462,
267
+ "eval_runtime": 61.2293,
268
+ "eval_samples_per_second": 81.824,
269
+ "eval_steps_per_second": 10.24,
270
+ "step": 17000
271
+ },
272
+ {
273
+ "epoch": 0.73,
274
+ "learning_rate": 4.997777777777779e-07,
275
+ "loss": 1.0966,
276
+ "step": 18000
277
+ },
278
+ {
279
+ "epoch": 0.73,
280
+ "eval_accuracy": 0.4031936127744511,
281
+ "eval_loss": 1.0839308500289917,
282
+ "eval_runtime": 61.2192,
283
+ "eval_samples_per_second": 81.837,
284
+ "eval_steps_per_second": 10.242,
285
+ "step": 18000
286
+ },
287
+ {
288
+ "epoch": 0.77,
289
+ "learning_rate": 5.275555555555556e-07,
290
+ "loss": 1.0935,
291
+ "step": 19000
292
+ },
293
+ {
294
+ "epoch": 0.77,
295
+ "eval_accuracy": 0.40099800399201596,
296
+ "eval_loss": 1.0814112424850464,
297
+ "eval_runtime": 61.1921,
298
+ "eval_samples_per_second": 81.873,
299
+ "eval_steps_per_second": 10.246,
300
+ "step": 19000
301
+ },
302
+ {
303
+ "epoch": 0.81,
304
+ "learning_rate": 5.553055555555556e-07,
305
+ "loss": 1.0939,
306
+ "step": 20000
307
+ },
308
+ {
309
+ "epoch": 0.81,
310
+ "eval_accuracy": 0.39241516966067863,
311
+ "eval_loss": 1.0840386152267456,
312
+ "eval_runtime": 61.2683,
313
+ "eval_samples_per_second": 81.772,
314
+ "eval_steps_per_second": 10.234,
315
+ "step": 20000
316
+ },
317
+ {
318
+ "epoch": 0.86,
319
+ "learning_rate": 5.830833333333334e-07,
320
+ "loss": 1.0917,
321
+ "step": 21000
322
+ },
323
+ {
324
+ "epoch": 0.86,
325
+ "eval_accuracy": 0.38303393213572856,
326
+ "eval_loss": 1.083555817604065,
327
+ "eval_runtime": 61.2974,
328
+ "eval_samples_per_second": 81.733,
329
+ "eval_steps_per_second": 10.229,
330
+ "step": 21000
331
+ },
332
+ {
333
+ "epoch": 0.9,
334
+ "learning_rate": 6.108333333333333e-07,
335
+ "loss": 1.0926,
336
+ "step": 22000
337
+ },
338
+ {
339
+ "epoch": 0.9,
340
+ "eval_accuracy": 0.39481037924151696,
341
+ "eval_loss": 1.080513834953308,
342
+ "eval_runtime": 61.4078,
343
+ "eval_samples_per_second": 81.586,
344
+ "eval_steps_per_second": 10.21,
345
+ "step": 22000
346
+ },
347
+ {
348
+ "epoch": 0.94,
349
+ "learning_rate": 6.386111111111112e-07,
350
+ "loss": 1.09,
351
+ "step": 23000
352
+ },
353
+ {
354
+ "epoch": 0.94,
355
+ "eval_accuracy": 0.4115768463073852,
356
+ "eval_loss": 1.076084017753601,
357
+ "eval_runtime": 61.6647,
358
+ "eval_samples_per_second": 81.246,
359
+ "eval_steps_per_second": 10.168,
360
+ "step": 23000
361
+ },
362
+ {
363
+ "epoch": 0.98,
364
+ "learning_rate": 6.663888888888889e-07,
365
+ "loss": 1.0879,
366
+ "step": 24000
367
+ },
368
+ {
369
+ "epoch": 0.98,
370
+ "eval_accuracy": 0.42734530938123755,
371
+ "eval_loss": 1.0709619522094727,
372
+ "eval_runtime": 61.3291,
373
+ "eval_samples_per_second": 81.69,
374
+ "eval_steps_per_second": 10.224,
375
+ "step": 24000
376
+ },
377
+ {
378
+ "epoch": 1.02,
379
+ "learning_rate": 6.94138888888889e-07,
380
+ "loss": 1.0846,
381
+ "step": 25000
382
+ },
383
+ {
384
+ "epoch": 1.02,
385
+ "eval_accuracy": 0.417564870259481,
386
+ "eval_loss": 1.0698238611221313,
387
+ "eval_runtime": 61.3091,
388
+ "eval_samples_per_second": 81.717,
389
+ "eval_steps_per_second": 10.227,
390
+ "step": 25000
391
+ },
392
+ {
393
+ "epoch": 1.06,
394
+ "learning_rate": 7.219166666666666e-07,
395
+ "loss": 1.0821,
396
+ "step": 26000
397
+ },
398
+ {
399
+ "epoch": 1.06,
400
+ "eval_accuracy": 0.43313373253493015,
401
+ "eval_loss": 1.065280795097351,
402
+ "eval_runtime": 61.4486,
403
+ "eval_samples_per_second": 81.532,
404
+ "eval_steps_per_second": 10.204,
405
+ "step": 26000
406
+ },
407
+ {
408
+ "epoch": 1.1,
409
+ "learning_rate": 7.496666666666667e-07,
410
+ "loss": 1.078,
411
+ "step": 27000
412
+ },
413
+ {
414
+ "epoch": 1.1,
415
+ "eval_accuracy": 0.4367265469061876,
416
+ "eval_loss": 1.059377908706665,
417
+ "eval_runtime": 61.6257,
418
+ "eval_samples_per_second": 81.297,
419
+ "eval_steps_per_second": 10.174,
420
+ "step": 27000
421
+ },
422
+ {
423
+ "epoch": 1.14,
424
+ "learning_rate": 7.774444444444445e-07,
425
+ "loss": 1.0707,
426
+ "step": 28000
427
+ },
428
+ {
429
+ "epoch": 1.14,
430
+ "eval_accuracy": 0.437125748502994,
431
+ "eval_loss": 1.0536556243896484,
432
+ "eval_runtime": 61.8255,
433
+ "eval_samples_per_second": 81.035,
434
+ "eval_steps_per_second": 10.141,
435
+ "step": 28000
436
+ },
437
+ {
438
+ "epoch": 1.18,
439
+ "learning_rate": 8.051944444444445e-07,
440
+ "loss": 1.0658,
441
+ "step": 29000
442
+ },
443
+ {
444
+ "epoch": 1.18,
445
+ "eval_accuracy": 0.4425149700598802,
446
+ "eval_loss": 1.0465552806854248,
447
+ "eval_runtime": 61.6813,
448
+ "eval_samples_per_second": 81.224,
449
+ "eval_steps_per_second": 10.165,
450
+ "step": 29000
451
+ },
452
+ {
453
+ "epoch": 1.22,
454
+ "learning_rate": 8.329722222222223e-07,
455
+ "loss": 1.0596,
456
+ "step": 30000
457
+ },
458
+ {
459
+ "epoch": 1.22,
460
+ "eval_accuracy": 0.44011976047904194,
461
+ "eval_loss": 1.042420506477356,
462
+ "eval_runtime": 61.7796,
463
+ "eval_samples_per_second": 81.095,
464
+ "eval_steps_per_second": 10.149,
465
+ "step": 30000
466
+ },
467
+ {
468
+ "epoch": 1.26,
469
+ "learning_rate": 8.607222222222223e-07,
470
+ "loss": 1.0665,
471
+ "step": 31000
472
+ },
473
+ {
474
+ "epoch": 1.26,
475
+ "eval_accuracy": 0.44011976047904194,
476
+ "eval_loss": 1.0469056367874146,
477
+ "eval_runtime": 62.7365,
478
+ "eval_samples_per_second": 79.858,
479
+ "eval_steps_per_second": 9.994,
480
+ "step": 31000
481
+ },
482
+ {
483
+ "epoch": 1.3,
484
+ "learning_rate": 8.885e-07,
485
+ "loss": 1.0586,
486
+ "step": 32000
487
+ },
488
+ {
489
+ "epoch": 1.3,
490
+ "eval_accuracy": 0.4363273453093812,
491
+ "eval_loss": 1.041648507118225,
492
+ "eval_runtime": 61.7699,
493
+ "eval_samples_per_second": 81.107,
494
+ "eval_steps_per_second": 10.151,
495
+ "step": 32000
496
+ },
497
+ {
498
+ "epoch": 1.34,
499
+ "learning_rate": 9.162500000000001e-07,
500
+ "loss": 1.0579,
501
+ "step": 33000
502
+ },
503
+ {
504
+ "epoch": 1.34,
505
+ "eval_accuracy": 0.44550898203592815,
506
+ "eval_loss": 1.0353455543518066,
507
+ "eval_runtime": 61.7392,
508
+ "eval_samples_per_second": 81.148,
509
+ "eval_steps_per_second": 10.156,
510
+ "step": 33000
511
+ },
512
+ {
513
+ "epoch": 1.39,
514
+ "learning_rate": 9.440277777777779e-07,
515
+ "loss": 1.0499,
516
+ "step": 34000
517
+ },
518
+ {
519
+ "epoch": 1.39,
520
+ "eval_accuracy": 0.45109780439121755,
521
+ "eval_loss": 1.0313138961791992,
522
+ "eval_runtime": 61.661,
523
+ "eval_samples_per_second": 81.251,
524
+ "eval_steps_per_second": 10.169,
525
+ "step": 34000
526
+ },
527
+ {
528
+ "epoch": 1.43,
529
+ "learning_rate": 9.718055555555557e-07,
530
+ "loss": 1.0551,
531
+ "step": 35000
532
+ },
533
+ {
534
+ "epoch": 1.43,
535
+ "eval_accuracy": 0.4481037924151697,
536
+ "eval_loss": 1.0282059907913208,
537
+ "eval_runtime": 61.7111,
538
+ "eval_samples_per_second": 81.185,
539
+ "eval_steps_per_second": 10.16,
540
+ "step": 35000
541
+ },
542
+ {
543
+ "epoch": 1.47,
544
+ "learning_rate": 9.995555555555557e-07,
545
+ "loss": 1.0509,
546
+ "step": 36000
547
+ },
548
+ {
549
+ "epoch": 1.47,
550
+ "eval_accuracy": 0.4568862275449102,
551
+ "eval_loss": 1.0246349573135376,
552
+ "eval_runtime": 61.4536,
553
+ "eval_samples_per_second": 81.525,
554
+ "eval_steps_per_second": 10.203,
555
+ "step": 36000
556
+ },
557
+ {
558
+ "epoch": 1.51,
559
+ "learning_rate": 1.0273333333333335e-06,
560
+ "loss": 1.0486,
561
+ "step": 37000
562
+ },
563
+ {
564
+ "epoch": 1.51,
565
+ "eval_accuracy": 0.45868263473053894,
566
+ "eval_loss": 1.0200064182281494,
567
+ "eval_runtime": 61.3644,
568
+ "eval_samples_per_second": 81.643,
569
+ "eval_steps_per_second": 10.218,
570
+ "step": 37000
571
+ },
572
+ {
573
+ "epoch": 1.55,
574
+ "learning_rate": 1.0550833333333334e-06,
575
+ "loss": 1.0468,
576
+ "step": 38000
577
+ },
578
+ {
579
+ "epoch": 1.55,
580
+ "eval_accuracy": 0.4564870259481038,
581
+ "eval_loss": 1.0205750465393066,
582
+ "eval_runtime": 61.6145,
583
+ "eval_samples_per_second": 81.312,
584
+ "eval_steps_per_second": 10.176,
585
+ "step": 38000
586
+ },
587
+ {
588
+ "epoch": 1.59,
589
+ "learning_rate": 1.0828611111111111e-06,
590
+ "loss": 1.0464,
591
+ "step": 39000
592
+ },
593
+ {
594
+ "epoch": 1.59,
595
+ "eval_accuracy": 0.4554890219560878,
596
+ "eval_loss": 1.021471619606018,
597
+ "eval_runtime": 61.6216,
598
+ "eval_samples_per_second": 81.303,
599
+ "eval_steps_per_second": 10.175,
600
+ "step": 39000
601
+ },
602
+ {
603
+ "epoch": 1.63,
604
+ "learning_rate": 1.1106111111111112e-06,
605
+ "loss": 1.0374,
606
+ "step": 40000
607
+ },
608
+ {
609
+ "epoch": 1.63,
610
+ "eval_accuracy": 0.45728542914171655,
611
+ "eval_loss": 1.0143259763717651,
612
+ "eval_runtime": 61.6469,
613
+ "eval_samples_per_second": 81.269,
614
+ "eval_steps_per_second": 10.171,
615
+ "step": 40000
616
+ },
617
+ {
618
+ "epoch": 1.67,
619
+ "learning_rate": 1.138388888888889e-06,
620
+ "loss": 1.0427,
621
+ "step": 41000
622
+ },
623
+ {
624
+ "epoch": 1.67,
625
+ "eval_accuracy": 0.45848303393213574,
626
+ "eval_loss": 1.0229008197784424,
627
+ "eval_runtime": 61.4018,
628
+ "eval_samples_per_second": 81.594,
629
+ "eval_steps_per_second": 10.211,
630
+ "step": 41000
631
+ },
632
+ {
633
+ "epoch": 1.71,
634
+ "learning_rate": 1.166138888888889e-06,
635
+ "loss": 1.0359,
636
+ "step": 42000
637
+ },
638
+ {
639
+ "epoch": 1.71,
640
+ "eval_accuracy": 0.4674650698602794,
641
+ "eval_loss": 1.0088660717010498,
642
+ "eval_runtime": 61.4607,
643
+ "eval_samples_per_second": 81.515,
644
+ "eval_steps_per_second": 10.202,
645
+ "step": 42000
646
+ },
647
+ {
648
+ "epoch": 1.75,
649
+ "learning_rate": 1.1939166666666668e-06,
650
+ "loss": 1.0382,
651
+ "step": 43000
652
+ },
653
+ {
654
+ "epoch": 1.75,
655
+ "eval_accuracy": 0.4489021956087824,
656
+ "eval_loss": 1.0299837589263916,
657
+ "eval_runtime": 61.6434,
658
+ "eval_samples_per_second": 81.274,
659
+ "eval_steps_per_second": 10.171,
660
+ "step": 43000
661
+ },
662
+ {
663
+ "epoch": 1.79,
664
+ "learning_rate": 1.2216944444444446e-06,
665
+ "loss": 1.0381,
666
+ "step": 44000
667
+ },
668
+ {
669
+ "epoch": 1.79,
670
+ "eval_accuracy": 0.4786427145708583,
671
+ "eval_loss": 1.0064666271209717,
672
+ "eval_runtime": 61.6779,
673
+ "eval_samples_per_second": 81.228,
674
+ "eval_steps_per_second": 10.166,
675
+ "step": 44000
676
+ },
677
+ {
678
+ "epoch": 1.83,
679
+ "learning_rate": 1.2494444444444447e-06,
680
+ "loss": 1.0369,
681
+ "step": 45000
682
+ },
683
+ {
684
+ "epoch": 1.83,
685
+ "eval_accuracy": 0.4716566866267465,
686
+ "eval_loss": 1.0102956295013428,
687
+ "eval_runtime": 61.6864,
688
+ "eval_samples_per_second": 81.217,
689
+ "eval_steps_per_second": 10.164,
690
+ "step": 45000
691
+ },
692
+ {
693
+ "epoch": 1.87,
694
+ "learning_rate": 1.2772222222222224e-06,
695
+ "loss": 1.0359,
696
+ "step": 46000
697
+ },
698
+ {
699
+ "epoch": 1.87,
700
+ "eval_accuracy": 0.4686626746506986,
701
+ "eval_loss": 1.011850118637085,
702
+ "eval_runtime": 61.3581,
703
+ "eval_samples_per_second": 81.652,
704
+ "eval_steps_per_second": 10.219,
705
+ "step": 46000
706
+ },
707
+ {
708
+ "epoch": 1.91,
709
+ "learning_rate": 1.3049722222222223e-06,
710
+ "loss": 1.0305,
711
+ "step": 47000
712
+ },
713
+ {
714
+ "epoch": 1.91,
715
+ "eval_accuracy": 0.48283433133732534,
716
+ "eval_loss": 1.0011930465698242,
717
+ "eval_runtime": 61.3161,
718
+ "eval_samples_per_second": 81.708,
719
+ "eval_steps_per_second": 10.226,
720
+ "step": 47000
721
+ },
722
+ {
723
+ "epoch": 1.96,
724
+ "learning_rate": 1.33275e-06,
725
+ "loss": 1.035,
726
+ "step": 48000
727
+ },
728
+ {
729
+ "epoch": 1.96,
730
+ "eval_accuracy": 0.4467065868263473,
731
+ "eval_loss": 1.0333038568496704,
732
+ "eval_runtime": 61.3247,
733
+ "eval_samples_per_second": 81.696,
734
+ "eval_steps_per_second": 10.224,
735
+ "step": 48000
736
+ },
737
+ {
738
+ "epoch": 2.0,
739
+ "learning_rate": 1.3605277777777778e-06,
740
+ "loss": 1.027,
741
+ "step": 49000
742
+ },
743
+ {
744
+ "epoch": 2.0,
745
+ "eval_accuracy": 0.48862275449101794,
746
+ "eval_loss": 0.9966129064559937,
747
+ "eval_runtime": 61.479,
748
+ "eval_samples_per_second": 81.491,
749
+ "eval_steps_per_second": 10.199,
750
+ "step": 49000
751
+ },
752
+ {
753
+ "epoch": 2.04,
754
+ "learning_rate": 1.388277777777778e-06,
755
+ "loss": 1.0271,
756
+ "step": 50000
757
+ },
758
+ {
759
+ "epoch": 2.04,
760
+ "eval_accuracy": 0.4744510978043912,
761
+ "eval_loss": 1.006515622138977,
762
+ "eval_runtime": 61.3701,
763
+ "eval_samples_per_second": 81.636,
764
+ "eval_steps_per_second": 10.217,
765
+ "step": 50000
766
+ },
767
+ {
768
+ "epoch": 2.08,
769
+ "learning_rate": 1.4160555555555555e-06,
770
+ "loss": 1.0271,
771
+ "step": 51000
772
+ },
773
+ {
774
+ "epoch": 2.08,
775
+ "eval_accuracy": 0.49141716566866267,
776
+ "eval_loss": 0.9953572154045105,
777
+ "eval_runtime": 61.4077,
778
+ "eval_samples_per_second": 81.586,
779
+ "eval_steps_per_second": 10.21,
780
+ "step": 51000
781
+ },
782
+ {
783
+ "epoch": 2.12,
784
+ "learning_rate": 1.4438055555555558e-06,
785
+ "loss": 1.0233,
786
+ "step": 52000
787
+ },
788
+ {
789
+ "epoch": 2.12,
790
+ "eval_accuracy": 0.4728542914171657,
791
+ "eval_loss": 1.017760992050171,
792
+ "eval_runtime": 61.6076,
793
+ "eval_samples_per_second": 81.321,
794
+ "eval_steps_per_second": 10.177,
795
+ "step": 52000
796
+ },
797
+ {
798
+ "epoch": 2.16,
799
+ "learning_rate": 1.4715833333333333e-06,
800
+ "loss": 1.0275,
801
+ "step": 53000
802
+ },
803
+ {
804
+ "epoch": 2.16,
805
+ "eval_accuracy": 0.49860279441117766,
806
+ "eval_loss": 0.9917733073234558,
807
+ "eval_runtime": 61.3839,
808
+ "eval_samples_per_second": 81.617,
809
+ "eval_steps_per_second": 10.214,
810
+ "step": 53000
811
+ },
812
+ {
813
+ "epoch": 2.2,
814
+ "learning_rate": 1.4993333333333334e-06,
815
+ "loss": 1.0216,
816
+ "step": 54000
817
+ },
818
+ {
819
+ "epoch": 2.2,
820
+ "eval_accuracy": 0.4940119760479042,
821
+ "eval_loss": 0.9957506656646729,
822
+ "eval_runtime": 61.555,
823
+ "eval_samples_per_second": 81.391,
824
+ "eval_steps_per_second": 10.186,
825
+ "step": 54000
826
+ },
827
+ {
828
+ "epoch": 2.24,
829
+ "learning_rate": 1.5271111111111112e-06,
830
+ "loss": 1.0218,
831
+ "step": 55000
832
+ },
833
+ {
834
+ "epoch": 2.24,
835
+ "eval_accuracy": 0.5,
836
+ "eval_loss": 0.9929438233375549,
837
+ "eval_runtime": 60.9099,
838
+ "eval_samples_per_second": 82.253,
839
+ "eval_steps_per_second": 10.294,
840
+ "step": 55000
841
+ },
842
+ {
843
+ "epoch": 2.28,
844
+ "learning_rate": 1.554861111111111e-06,
845
+ "loss": 1.0168,
846
+ "step": 56000
847
+ },
848
+ {
849
+ "epoch": 2.28,
850
+ "eval_accuracy": 0.4624750499001996,
851
+ "eval_loss": 1.0249923467636108,
852
+ "eval_runtime": 61.6569,
853
+ "eval_samples_per_second": 81.256,
854
+ "eval_steps_per_second": 10.169,
855
+ "step": 56000
856
+ },
857
+ {
858
+ "epoch": 2.32,
859
+ "learning_rate": 1.582638888888889e-06,
860
+ "loss": 1.0155,
861
+ "step": 57000
862
+ },
863
+ {
864
+ "epoch": 2.32,
865
+ "eval_accuracy": 0.4932135728542914,
866
+ "eval_loss": 0.9934276938438416,
867
+ "eval_runtime": 61.608,
868
+ "eval_samples_per_second": 81.321,
869
+ "eval_steps_per_second": 10.177,
870
+ "step": 57000
871
+ },
872
+ {
873
+ "epoch": 2.36,
874
+ "learning_rate": 1.610388888888889e-06,
875
+ "loss": 1.0168,
876
+ "step": 58000
877
+ },
878
+ {
879
+ "epoch": 2.36,
880
+ "eval_accuracy": 0.48582834331337327,
881
+ "eval_loss": 1.0029484033584595,
882
+ "eval_runtime": 61.694,
883
+ "eval_samples_per_second": 81.207,
884
+ "eval_steps_per_second": 10.163,
885
+ "step": 58000
886
+ },
887
+ {
888
+ "epoch": 2.4,
889
+ "learning_rate": 1.6381666666666666e-06,
890
+ "loss": 1.0154,
891
+ "step": 59000
892
+ },
893
+ {
894
+ "epoch": 2.4,
895
+ "eval_accuracy": 0.49880239520958086,
896
+ "eval_loss": 0.9892131686210632,
897
+ "eval_runtime": 61.2271,
898
+ "eval_samples_per_second": 81.827,
899
+ "eval_steps_per_second": 10.241,
900
+ "step": 59000
901
+ },
902
+ {
903
+ "epoch": 2.44,
904
+ "learning_rate": 1.6659166666666667e-06,
905
+ "loss": 1.0195,
906
+ "step": 60000
907
+ },
908
+ {
909
+ "epoch": 2.44,
910
+ "eval_accuracy": 0.5007984031936128,
911
+ "eval_loss": 0.988898754119873,
912
+ "eval_runtime": 61.1883,
913
+ "eval_samples_per_second": 81.878,
914
+ "eval_steps_per_second": 10.247,
915
+ "step": 60000
916
+ },
917
+ {
918
+ "epoch": 2.49,
919
+ "learning_rate": 1.6936944444444447e-06,
920
+ "loss": 1.0181,
921
+ "step": 61000
922
+ },
923
+ {
924
+ "epoch": 2.49,
925
+ "eval_accuracy": 0.49880239520958086,
926
+ "eval_loss": 0.9839370250701904,
927
+ "eval_runtime": 61.1847,
928
+ "eval_samples_per_second": 81.883,
929
+ "eval_steps_per_second": 10.248,
930
+ "step": 61000
931
+ },
932
+ {
933
+ "epoch": 2.53,
934
+ "learning_rate": 1.7214444444444446e-06,
935
+ "loss": 1.019,
936
+ "step": 62000
937
+ },
938
+ {
939
+ "epoch": 2.53,
940
+ "eval_accuracy": 0.4848303393213573,
941
+ "eval_loss": 1.0039244890213013,
942
+ "eval_runtime": 61.204,
943
+ "eval_samples_per_second": 81.857,
944
+ "eval_steps_per_second": 10.244,
945
+ "step": 62000
946
+ },
947
+ {
948
+ "epoch": 2.57,
949
+ "learning_rate": 1.7492222222222223e-06,
950
+ "loss": 1.017,
951
+ "step": 63000
952
+ },
953
+ {
954
+ "epoch": 2.57,
955
+ "eval_accuracy": 0.49700598802395207,
956
+ "eval_loss": 0.9895579218864441,
957
+ "eval_runtime": 61.1812,
958
+ "eval_samples_per_second": 81.888,
959
+ "eval_steps_per_second": 10.248,
960
+ "step": 63000
961
+ },
962
+ {
963
+ "epoch": 2.61,
964
+ "learning_rate": 1.7769722222222222e-06,
965
+ "loss": 1.0139,
966
+ "step": 64000
967
+ },
968
+ {
969
+ "epoch": 2.61,
970
+ "eval_accuracy": 0.5017964071856288,
971
+ "eval_loss": 0.9869150519371033,
972
+ "eval_runtime": 61.3945,
973
+ "eval_samples_per_second": 81.603,
974
+ "eval_steps_per_second": 10.213,
975
+ "step": 64000
976
+ },
977
+ {
978
+ "epoch": 2.65,
979
+ "learning_rate": 1.8047500000000002e-06,
980
+ "loss": 1.0132,
981
+ "step": 65000
982
+ },
983
+ {
984
+ "epoch": 2.65,
985
+ "eval_accuracy": 0.5059880239520959,
986
+ "eval_loss": 0.9817209243774414,
987
+ "eval_runtime": 61.4781,
988
+ "eval_samples_per_second": 81.492,
989
+ "eval_steps_per_second": 10.199,
990
+ "step": 65000
991
+ },
992
+ {
993
+ "epoch": 2.69,
994
+ "learning_rate": 1.8325000000000002e-06,
995
+ "loss": 1.0166,
996
+ "step": 66000
997
+ },
998
+ {
999
+ "epoch": 2.69,
1000
+ "eval_accuracy": 0.5049900199600799,
1001
+ "eval_loss": 0.9807708859443665,
1002
+ "eval_runtime": 61.3424,
1003
+ "eval_samples_per_second": 81.673,
1004
+ "eval_steps_per_second": 10.221,
1005
+ "step": 66000
1006
+ },
1007
+ {
1008
+ "epoch": 2.73,
1009
+ "learning_rate": 1.8602777777777778e-06,
1010
+ "loss": 1.0133,
1011
+ "step": 67000
1012
+ },
1013
+ {
1014
+ "epoch": 2.73,
1015
+ "eval_accuracy": 0.49221556886227547,
1016
+ "eval_loss": 0.9878746271133423,
1017
+ "eval_runtime": 61.1619,
1018
+ "eval_samples_per_second": 81.914,
1019
+ "eval_steps_per_second": 10.251,
1020
+ "step": 67000
1021
+ },
1022
+ {
1023
+ "epoch": 2.77,
1024
+ "learning_rate": 1.8880277777777779e-06,
1025
+ "loss": 1.0145,
1026
+ "step": 68000
1027
+ },
1028
+ {
1029
+ "epoch": 2.77,
1030
+ "eval_accuracy": 0.5071856287425149,
1031
+ "eval_loss": 0.9776698350906372,
1032
+ "eval_runtime": 60.8697,
1033
+ "eval_samples_per_second": 82.307,
1034
+ "eval_steps_per_second": 10.301,
1035
+ "step": 68000
1036
+ }
1037
+ ],
1038
+ "logging_steps": 1000,
1039
+ "max_steps": 10000000,
1040
+ "num_train_epochs": 408,
1041
+ "save_steps": 1000,
1042
+ "total_flos": 1.421436262762414e+17,
1043
+ "trial_name": null,
1044
+ "trial_params": null
1045
+ }
checkpoint-68000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e31537d78ddfb65ec62a3bb3689996ba9eab16c9072c351cad313c78d3b7a608
3
+ size 4091
checkpoint-68000/vocab.json ADDED
The diff for this file is too large to render. See raw diff