infinitejoy commited on
Commit
4adcb46
1 Parent(s): 003cecd

End of training

Browse files
all_results.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 200.0,
3
+ "eval_loss": 0.9669056534767151,
4
+ "eval_runtime": 11.048,
5
+ "eval_samples": 212,
6
+ "eval_samples_per_second": 19.189,
7
+ "eval_steps_per_second": 0.634,
8
+ "eval_wer": 0.6942355889724311,
9
+ "train_loss": 0.9929841698540581,
10
+ "train_runtime": 8863.3906,
11
+ "train_samples": 554,
12
+ "train_samples_per_second": 12.501,
13
+ "train_steps_per_second": 0.406
14
+ }
eval_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 200.0,
3
+ "eval_loss": 0.9669056534767151,
4
+ "eval_runtime": 11.048,
5
+ "eval_samples": 212,
6
+ "eval_samples_per_second": 19.189,
7
+ "eval_steps_per_second": 0.634,
8
+ "eval_wer": 0.6942355889724311
9
+ }
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:82116bae17d8d846b4f1d5d1ee4779d86795903a135aac0ce8a4ff84c687ff1a
3
  size 1262128689
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:968322e4a8e10bd8bc3d58fe86b1074b5647a5ed1afbb43978844ef7b37b72f1
3
  size 1262128689
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 200.0,
3
+ "train_loss": 0.9929841698540581,
4
+ "train_runtime": 8863.3906,
5
+ "train_samples": 554,
6
+ "train_samples_per_second": 12.501,
7
+ "train_steps_per_second": 0.406
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,304 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 200.0,
5
+ "global_step": 3600,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 5.56,
12
+ "learning_rate": 5.82e-05,
13
+ "loss": 9.8118,
14
+ "step": 100
15
+ },
16
+ {
17
+ "epoch": 11.11,
18
+ "learning_rate": 0.0001182,
19
+ "loss": 3.4789,
20
+ "step": 200
21
+ },
22
+ {
23
+ "epoch": 16.67,
24
+ "learning_rate": 0.00017819999999999997,
25
+ "loss": 3.114,
26
+ "step": 300
27
+ },
28
+ {
29
+ "epoch": 22.22,
30
+ "learning_rate": 0.0002382,
31
+ "loss": 2.721,
32
+ "step": 400
33
+ },
34
+ {
35
+ "epoch": 27.78,
36
+ "learning_rate": 0.0002982,
37
+ "loss": 1.7294,
38
+ "step": 500
39
+ },
40
+ {
41
+ "epoch": 27.78,
42
+ "eval_loss": 0.8540233373641968,
43
+ "eval_runtime": 10.9769,
44
+ "eval_samples_per_second": 19.313,
45
+ "eval_steps_per_second": 0.638,
46
+ "eval_wer": 0.9943609022556391,
47
+ "step": 500
48
+ },
49
+ {
50
+ "epoch": 33.33,
51
+ "learning_rate": 0.0002906129032258064,
52
+ "loss": 1.351,
53
+ "step": 600
54
+ },
55
+ {
56
+ "epoch": 38.89,
57
+ "learning_rate": 0.0002809354838709677,
58
+ "loss": 1.1775,
59
+ "step": 700
60
+ },
61
+ {
62
+ "epoch": 44.44,
63
+ "learning_rate": 0.000271258064516129,
64
+ "loss": 1.0539,
65
+ "step": 800
66
+ },
67
+ {
68
+ "epoch": 50.0,
69
+ "learning_rate": 0.0002615806451612903,
70
+ "loss": 0.9587,
71
+ "step": 900
72
+ },
73
+ {
74
+ "epoch": 55.56,
75
+ "learning_rate": 0.0002519032258064516,
76
+ "loss": 0.8863,
77
+ "step": 1000
78
+ },
79
+ {
80
+ "epoch": 55.56,
81
+ "eval_loss": 0.7282317280769348,
82
+ "eval_runtime": 11.4744,
83
+ "eval_samples_per_second": 18.476,
84
+ "eval_steps_per_second": 0.61,
85
+ "eval_wer": 0.731203007518797,
86
+ "step": 1000
87
+ },
88
+ {
89
+ "epoch": 61.11,
90
+ "learning_rate": 0.00024222580645161287,
91
+ "loss": 0.7985,
92
+ "step": 1100
93
+ },
94
+ {
95
+ "epoch": 66.67,
96
+ "learning_rate": 0.00023254838709677416,
97
+ "loss": 0.7323,
98
+ "step": 1200
99
+ },
100
+ {
101
+ "epoch": 72.22,
102
+ "learning_rate": 0.00022287096774193548,
103
+ "loss": 0.6712,
104
+ "step": 1300
105
+ },
106
+ {
107
+ "epoch": 77.78,
108
+ "learning_rate": 0.00021319354838709678,
109
+ "loss": 0.6214,
110
+ "step": 1400
111
+ },
112
+ {
113
+ "epoch": 83.33,
114
+ "learning_rate": 0.00020351612903225804,
115
+ "loss": 0.5789,
116
+ "step": 1500
117
+ },
118
+ {
119
+ "epoch": 83.33,
120
+ "eval_loss": 0.817844033241272,
121
+ "eval_runtime": 11.023,
122
+ "eval_samples_per_second": 19.232,
123
+ "eval_steps_per_second": 0.635,
124
+ "eval_wer": 0.8101503759398496,
125
+ "step": 1500
126
+ },
127
+ {
128
+ "epoch": 88.89,
129
+ "learning_rate": 0.00019383870967741934,
130
+ "loss": 0.5519,
131
+ "step": 1600
132
+ },
133
+ {
134
+ "epoch": 94.44,
135
+ "learning_rate": 0.00018416129032258063,
136
+ "loss": 0.4957,
137
+ "step": 1700
138
+ },
139
+ {
140
+ "epoch": 100.0,
141
+ "learning_rate": 0.00017448387096774193,
142
+ "loss": 0.4599,
143
+ "step": 1800
144
+ },
145
+ {
146
+ "epoch": 105.56,
147
+ "learning_rate": 0.0001648064516129032,
148
+ "loss": 0.4272,
149
+ "step": 1900
150
+ },
151
+ {
152
+ "epoch": 111.11,
153
+ "learning_rate": 0.0001551290322580645,
154
+ "loss": 0.3899,
155
+ "step": 2000
156
+ },
157
+ {
158
+ "epoch": 111.11,
159
+ "eval_loss": 0.803404688835144,
160
+ "eval_runtime": 10.8752,
161
+ "eval_samples_per_second": 19.494,
162
+ "eval_steps_per_second": 0.644,
163
+ "eval_wer": 0.7700501253132832,
164
+ "step": 2000
165
+ },
166
+ {
167
+ "epoch": 116.67,
168
+ "learning_rate": 0.0001454516129032258,
169
+ "loss": 0.3725,
170
+ "step": 2100
171
+ },
172
+ {
173
+ "epoch": 122.22,
174
+ "learning_rate": 0.00013577419354838708,
175
+ "loss": 0.3373,
176
+ "step": 2200
177
+ },
178
+ {
179
+ "epoch": 127.78,
180
+ "learning_rate": 0.00012609677419354837,
181
+ "loss": 0.3134,
182
+ "step": 2300
183
+ },
184
+ {
185
+ "epoch": 133.33,
186
+ "learning_rate": 0.00011641935483870967,
187
+ "loss": 0.3057,
188
+ "step": 2400
189
+ },
190
+ {
191
+ "epoch": 138.89,
192
+ "learning_rate": 0.00010674193548387096,
193
+ "loss": 0.2869,
194
+ "step": 2500
195
+ },
196
+ {
197
+ "epoch": 138.89,
198
+ "eval_loss": 0.9060825705528259,
199
+ "eval_runtime": 10.9301,
200
+ "eval_samples_per_second": 19.396,
201
+ "eval_steps_per_second": 0.64,
202
+ "eval_wer": 0.699874686716792,
203
+ "step": 2500
204
+ },
205
+ {
206
+ "epoch": 144.44,
207
+ "learning_rate": 9.706451612903224e-05,
208
+ "loss": 0.2579,
209
+ "step": 2600
210
+ },
211
+ {
212
+ "epoch": 150.0,
213
+ "learning_rate": 8.738709677419355e-05,
214
+ "loss": 0.2383,
215
+ "step": 2700
216
+ },
217
+ {
218
+ "epoch": 155.56,
219
+ "learning_rate": 7.770967741935483e-05,
220
+ "loss": 0.2253,
221
+ "step": 2800
222
+ },
223
+ {
224
+ "epoch": 161.11,
225
+ "learning_rate": 6.803225806451613e-05,
226
+ "loss": 0.2144,
227
+ "step": 2900
228
+ },
229
+ {
230
+ "epoch": 166.67,
231
+ "learning_rate": 5.835483870967742e-05,
232
+ "loss": 0.1934,
233
+ "step": 3000
234
+ },
235
+ {
236
+ "epoch": 166.67,
237
+ "eval_loss": 0.9400397539138794,
238
+ "eval_runtime": 11.0289,
239
+ "eval_samples_per_second": 19.222,
240
+ "eval_steps_per_second": 0.635,
241
+ "eval_wer": 0.7105263157894737,
242
+ "step": 3000
243
+ },
244
+ {
245
+ "epoch": 172.22,
246
+ "learning_rate": 4.8677419354838705e-05,
247
+ "loss": 0.1874,
248
+ "step": 3100
249
+ },
250
+ {
251
+ "epoch": 177.78,
252
+ "learning_rate": 3.9e-05,
253
+ "loss": 0.1778,
254
+ "step": 3200
255
+ },
256
+ {
257
+ "epoch": 183.33,
258
+ "learning_rate": 2.932258064516129e-05,
259
+ "loss": 0.171,
260
+ "step": 3300
261
+ },
262
+ {
263
+ "epoch": 188.89,
264
+ "learning_rate": 1.9645161290322582e-05,
265
+ "loss": 0.1568,
266
+ "step": 3400
267
+ },
268
+ {
269
+ "epoch": 194.44,
270
+ "learning_rate": 9.96774193548387e-06,
271
+ "loss": 0.1551,
272
+ "step": 3500
273
+ },
274
+ {
275
+ "epoch": 194.44,
276
+ "eval_loss": 0.966656506061554,
277
+ "eval_runtime": 11.4149,
278
+ "eval_samples_per_second": 18.572,
279
+ "eval_steps_per_second": 0.613,
280
+ "eval_wer": 0.6954887218045113,
281
+ "step": 3500
282
+ },
283
+ {
284
+ "epoch": 200.0,
285
+ "learning_rate": 2.903225806451613e-07,
286
+ "loss": 0.145,
287
+ "step": 3600
288
+ },
289
+ {
290
+ "epoch": 200.0,
291
+ "step": 3600,
292
+ "total_flos": 2.405940879390409e+19,
293
+ "train_loss": 0.9929841698540581,
294
+ "train_runtime": 8863.3906,
295
+ "train_samples_per_second": 12.501,
296
+ "train_steps_per_second": 0.406
297
+ }
298
+ ],
299
+ "max_steps": 3600,
300
+ "num_train_epochs": 200,
301
+ "total_flos": 2.405940879390409e+19,
302
+ "trial_name": null,
303
+ "trial_params": null
304
+ }