minhah commited on
Commit
043db60
1 Parent(s): 30ad3e1

End of training

Browse files
Files changed (3) hide show
  1. all_results.json +8 -0
  2. test_results.json +8 -0
  3. trainer_state.json +359 -0
all_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 4.19,
3
+ "eval_f1": 0.9517819706498952,
4
+ "eval_loss": 0.3077031970024109,
5
+ "eval_runtime": 529.5496,
6
+ "eval_samples_per_second": 1.416,
7
+ "eval_steps_per_second": 0.178
8
+ }
test_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 4.19,
3
+ "eval_f1": 0.9517819706498952,
4
+ "eval_loss": 0.3077031970024109,
5
+ "eval_runtime": 529.5496,
6
+ "eval_samples_per_second": 1.416,
7
+ "eval_steps_per_second": 0.178
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,359 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.9642160052049447,
3
+ "best_model_checkpoint": "videomae-base-finetuned-ElderReact-Fear/checkpoint-77",
4
+ "epoch": 4.189473684210526,
5
+ "eval_steps": 500,
6
+ "global_step": 380,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.03,
13
+ "grad_norm": 6.242678642272949,
14
+ "learning_rate": 1.3157894736842106e-05,
15
+ "loss": 0.5133,
16
+ "step": 10
17
+ },
18
+ {
19
+ "epoch": 0.05,
20
+ "grad_norm": 4.866024971008301,
21
+ "learning_rate": 2.6315789473684212e-05,
22
+ "loss": 0.4234,
23
+ "step": 20
24
+ },
25
+ {
26
+ "epoch": 0.08,
27
+ "grad_norm": 6.626444339752197,
28
+ "learning_rate": 3.9473684210526316e-05,
29
+ "loss": 0.5058,
30
+ "step": 30
31
+ },
32
+ {
33
+ "epoch": 0.11,
34
+ "grad_norm": 3.6388182640075684,
35
+ "learning_rate": 4.970760233918128e-05,
36
+ "loss": 0.4791,
37
+ "step": 40
38
+ },
39
+ {
40
+ "epoch": 0.13,
41
+ "grad_norm": 2.3752734661102295,
42
+ "learning_rate": 4.824561403508772e-05,
43
+ "loss": 0.3841,
44
+ "step": 50
45
+ },
46
+ {
47
+ "epoch": 0.16,
48
+ "grad_norm": 2.3690688610076904,
49
+ "learning_rate": 4.678362573099415e-05,
50
+ "loss": 0.4903,
51
+ "step": 60
52
+ },
53
+ {
54
+ "epoch": 0.18,
55
+ "grad_norm": 2.5726335048675537,
56
+ "learning_rate": 4.5321637426900585e-05,
57
+ "loss": 0.2935,
58
+ "step": 70
59
+ },
60
+ {
61
+ "epoch": 0.2,
62
+ "eval_f1": 0.9642160052049447,
63
+ "eval_loss": 0.25238940119743347,
64
+ "eval_runtime": 1187.3545,
65
+ "eval_samples_per_second": 0.67,
66
+ "eval_steps_per_second": 0.084,
67
+ "step": 77
68
+ },
69
+ {
70
+ "epoch": 1.01,
71
+ "grad_norm": 3.456474542617798,
72
+ "learning_rate": 4.3859649122807014e-05,
73
+ "loss": 0.4967,
74
+ "step": 80
75
+ },
76
+ {
77
+ "epoch": 1.03,
78
+ "grad_norm": 1.242908239364624,
79
+ "learning_rate": 4.239766081871345e-05,
80
+ "loss": 0.4073,
81
+ "step": 90
82
+ },
83
+ {
84
+ "epoch": 1.06,
85
+ "grad_norm": 3.6170222759246826,
86
+ "learning_rate": 4.093567251461988e-05,
87
+ "loss": 0.5547,
88
+ "step": 100
89
+ },
90
+ {
91
+ "epoch": 1.09,
92
+ "grad_norm": 2.8330729007720947,
93
+ "learning_rate": 3.9473684210526316e-05,
94
+ "loss": 0.363,
95
+ "step": 110
96
+ },
97
+ {
98
+ "epoch": 1.11,
99
+ "grad_norm": 2.8961918354034424,
100
+ "learning_rate": 3.8011695906432746e-05,
101
+ "loss": 0.4056,
102
+ "step": 120
103
+ },
104
+ {
105
+ "epoch": 1.14,
106
+ "grad_norm": 5.695966720581055,
107
+ "learning_rate": 3.654970760233918e-05,
108
+ "loss": 0.3925,
109
+ "step": 130
110
+ },
111
+ {
112
+ "epoch": 1.17,
113
+ "grad_norm": 3.3888511657714844,
114
+ "learning_rate": 3.508771929824561e-05,
115
+ "loss": 0.4257,
116
+ "step": 140
117
+ },
118
+ {
119
+ "epoch": 1.19,
120
+ "grad_norm": 7.591073513031006,
121
+ "learning_rate": 3.362573099415205e-05,
122
+ "loss": 0.3966,
123
+ "step": 150
124
+ },
125
+ {
126
+ "epoch": 1.2,
127
+ "eval_f1": 0.9642160052049447,
128
+ "eval_loss": 0.25368985533714294,
129
+ "eval_runtime": 1229.9142,
130
+ "eval_samples_per_second": 0.647,
131
+ "eval_steps_per_second": 0.081,
132
+ "step": 154
133
+ },
134
+ {
135
+ "epoch": 2.02,
136
+ "grad_norm": 3.995485305786133,
137
+ "learning_rate": 3.216374269005848e-05,
138
+ "loss": 0.5168,
139
+ "step": 160
140
+ },
141
+ {
142
+ "epoch": 2.04,
143
+ "grad_norm": 2.2910969257354736,
144
+ "learning_rate": 3.0701754385964913e-05,
145
+ "loss": 0.4298,
146
+ "step": 170
147
+ },
148
+ {
149
+ "epoch": 2.07,
150
+ "grad_norm": 5.8856964111328125,
151
+ "learning_rate": 2.9239766081871346e-05,
152
+ "loss": 0.4855,
153
+ "step": 180
154
+ },
155
+ {
156
+ "epoch": 2.09,
157
+ "grad_norm": 1.2029250860214233,
158
+ "learning_rate": 2.777777777777778e-05,
159
+ "loss": 0.2937,
160
+ "step": 190
161
+ },
162
+ {
163
+ "epoch": 2.12,
164
+ "grad_norm": 1.0417933464050293,
165
+ "learning_rate": 2.6315789473684212e-05,
166
+ "loss": 0.2089,
167
+ "step": 200
168
+ },
169
+ {
170
+ "epoch": 2.15,
171
+ "grad_norm": 0.3655168116092682,
172
+ "learning_rate": 2.485380116959064e-05,
173
+ "loss": 0.5043,
174
+ "step": 210
175
+ },
176
+ {
177
+ "epoch": 2.17,
178
+ "grad_norm": 2.3576948642730713,
179
+ "learning_rate": 2.3391812865497074e-05,
180
+ "loss": 0.2995,
181
+ "step": 220
182
+ },
183
+ {
184
+ "epoch": 2.2,
185
+ "grad_norm": 2.619518280029297,
186
+ "learning_rate": 2.1929824561403507e-05,
187
+ "loss": 0.3792,
188
+ "step": 230
189
+ },
190
+ {
191
+ "epoch": 2.2,
192
+ "eval_f1": 0.9642160052049447,
193
+ "eval_loss": 0.2470169961452484,
194
+ "eval_runtime": 1353.2793,
195
+ "eval_samples_per_second": 0.588,
196
+ "eval_steps_per_second": 0.074,
197
+ "step": 231
198
+ },
199
+ {
200
+ "epoch": 3.02,
201
+ "grad_norm": 3.8343122005462646,
202
+ "learning_rate": 2.046783625730994e-05,
203
+ "loss": 0.3818,
204
+ "step": 240
205
+ },
206
+ {
207
+ "epoch": 3.05,
208
+ "grad_norm": 3.7524924278259277,
209
+ "learning_rate": 1.9005847953216373e-05,
210
+ "loss": 0.3958,
211
+ "step": 250
212
+ },
213
+ {
214
+ "epoch": 3.08,
215
+ "grad_norm": 2.071563720703125,
216
+ "learning_rate": 1.7543859649122806e-05,
217
+ "loss": 0.3903,
218
+ "step": 260
219
+ },
220
+ {
221
+ "epoch": 3.1,
222
+ "grad_norm": 2.1228742599487305,
223
+ "learning_rate": 1.608187134502924e-05,
224
+ "loss": 0.4532,
225
+ "step": 270
226
+ },
227
+ {
228
+ "epoch": 3.13,
229
+ "grad_norm": 2.22047758102417,
230
+ "learning_rate": 1.4619883040935673e-05,
231
+ "loss": 0.3909,
232
+ "step": 280
233
+ },
234
+ {
235
+ "epoch": 3.16,
236
+ "grad_norm": 1.4722857475280762,
237
+ "learning_rate": 1.3157894736842106e-05,
238
+ "loss": 0.44,
239
+ "step": 290
240
+ },
241
+ {
242
+ "epoch": 3.18,
243
+ "grad_norm": 1.8657927513122559,
244
+ "learning_rate": 1.1695906432748537e-05,
245
+ "loss": 0.3666,
246
+ "step": 300
247
+ },
248
+ {
249
+ "epoch": 3.2,
250
+ "eval_f1": 0.9642160052049447,
251
+ "eval_loss": 0.2569561004638672,
252
+ "eval_runtime": 952.0022,
253
+ "eval_samples_per_second": 0.836,
254
+ "eval_steps_per_second": 0.105,
255
+ "step": 308
256
+ },
257
+ {
258
+ "epoch": 4.01,
259
+ "grad_norm": 6.351089000701904,
260
+ "learning_rate": 1.023391812865497e-05,
261
+ "loss": 0.3108,
262
+ "step": 310
263
+ },
264
+ {
265
+ "epoch": 4.03,
266
+ "grad_norm": 2.317436933517456,
267
+ "learning_rate": 8.771929824561403e-06,
268
+ "loss": 0.2831,
269
+ "step": 320
270
+ },
271
+ {
272
+ "epoch": 4.06,
273
+ "grad_norm": 2.794163703918457,
274
+ "learning_rate": 7.3099415204678366e-06,
275
+ "loss": 0.2504,
276
+ "step": 330
277
+ },
278
+ {
279
+ "epoch": 4.08,
280
+ "grad_norm": 2.761640787124634,
281
+ "learning_rate": 5.8479532163742686e-06,
282
+ "loss": 0.4861,
283
+ "step": 340
284
+ },
285
+ {
286
+ "epoch": 4.11,
287
+ "grad_norm": 4.870150089263916,
288
+ "learning_rate": 4.3859649122807014e-06,
289
+ "loss": 0.4329,
290
+ "step": 350
291
+ },
292
+ {
293
+ "epoch": 4.14,
294
+ "grad_norm": 4.540055751800537,
295
+ "learning_rate": 2.9239766081871343e-06,
296
+ "loss": 0.3709,
297
+ "step": 360
298
+ },
299
+ {
300
+ "epoch": 4.16,
301
+ "grad_norm": 4.406496524810791,
302
+ "learning_rate": 1.4619883040935671e-06,
303
+ "loss": 0.2996,
304
+ "step": 370
305
+ },
306
+ {
307
+ "epoch": 4.19,
308
+ "grad_norm": 3.185987949371338,
309
+ "learning_rate": 0.0,
310
+ "loss": 0.2803,
311
+ "step": 380
312
+ },
313
+ {
314
+ "epoch": 4.19,
315
+ "eval_f1": 0.9642160052049447,
316
+ "eval_loss": 0.24840089678764343,
317
+ "eval_runtime": 680.4356,
318
+ "eval_samples_per_second": 1.17,
319
+ "eval_steps_per_second": 0.147,
320
+ "step": 380
321
+ },
322
+ {
323
+ "epoch": 4.19,
324
+ "step": 380,
325
+ "total_flos": 3.783052093636215e+18,
326
+ "train_loss": 0.3995331350125765,
327
+ "train_runtime": 14411.9404,
328
+ "train_samples_per_second": 0.211,
329
+ "train_steps_per_second": 0.026
330
+ },
331
+ {
332
+ "epoch": 4.19,
333
+ "eval_f1": 0.9517819706498952,
334
+ "eval_loss": 0.3077032268047333,
335
+ "eval_runtime": 515.1776,
336
+ "eval_samples_per_second": 1.456,
337
+ "eval_steps_per_second": 0.182,
338
+ "step": 380
339
+ },
340
+ {
341
+ "epoch": 4.19,
342
+ "eval_f1": 0.9517819706498952,
343
+ "eval_loss": 0.3077031970024109,
344
+ "eval_runtime": 529.5496,
345
+ "eval_samples_per_second": 1.416,
346
+ "eval_steps_per_second": 0.178,
347
+ "step": 380
348
+ }
349
+ ],
350
+ "logging_steps": 10,
351
+ "max_steps": 380,
352
+ "num_input_tokens_seen": 0,
353
+ "num_train_epochs": 9223372036854775807,
354
+ "save_steps": 500,
355
+ "total_flos": 3.783052093636215e+18,
356
+ "train_batch_size": 8,
357
+ "trial_name": null,
358
+ "trial_params": null
359
+ }