Commit
•
8ebab4f
1
Parent(s):
40cf913
End of training
Browse files- README.md +2 -2
- all_results.json +5 -5
- runs/Apr07_14-34-51_2f21d1345b5e/events.out.tfevents.1712501216.2f21d1345b5e.11416.1 +3 -0
- test_results.json +5 -5
- trainer_state.json +165 -543
README.md
CHANGED
@@ -17,8 +17,8 @@ should probably proofread and complete it, then remove this comment. -->
|
|
17 |
|
18 |
This model is a fine-tuned version of [MCG-NJU/videomae-base-finetuned-kinetics](https://huggingface.co/MCG-NJU/videomae-base-finetuned-kinetics) on an unknown dataset.
|
19 |
It achieves the following results on the evaluation set:
|
20 |
-
- Loss: 0.
|
21 |
-
- Accuracy:
|
22 |
|
23 |
## Model description
|
24 |
|
|
|
17 |
|
18 |
This model is a fine-tuned version of [MCG-NJU/videomae-base-finetuned-kinetics](https://huggingface.co/MCG-NJU/videomae-base-finetuned-kinetics) on an unknown dataset.
|
19 |
It achieves the following results on the evaluation set:
|
20 |
+
- Loss: 0.0953
|
21 |
+
- Accuracy: 0.9762
|
22 |
|
23 |
## Model description
|
24 |
|
all_results.json
CHANGED
@@ -1,8 +1,8 @@
|
|
1 |
{
|
2 |
"epoch": 3.25,
|
3 |
-
"eval_accuracy": 0.
|
4 |
-
"eval_loss": 0.
|
5 |
-
"eval_runtime":
|
6 |
-
"eval_samples_per_second": 1.
|
7 |
-
"eval_steps_per_second": 0.
|
8 |
}
|
|
|
1 |
{
|
2 |
"epoch": 3.25,
|
3 |
+
"eval_accuracy": 0.9761904761904762,
|
4 |
+
"eval_loss": 0.09526590257883072,
|
5 |
+
"eval_runtime": 122.5253,
|
6 |
+
"eval_samples_per_second": 1.714,
|
7 |
+
"eval_steps_per_second": 0.433
|
8 |
}
|
runs/Apr07_14-34-51_2f21d1345b5e/events.out.tfevents.1712501216.2f21d1345b5e.11416.1
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7cfa6f367726c37195a1938e893b4a7267c3028bc5881c55498da2df2d4792b9
|
3 |
+
size 411
|
test_results.json
CHANGED
@@ -1,8 +1,8 @@
|
|
1 |
{
|
2 |
"epoch": 3.25,
|
3 |
-
"eval_accuracy": 0.
|
4 |
-
"eval_loss": 0.
|
5 |
-
"eval_runtime":
|
6 |
-
"eval_samples_per_second": 1.
|
7 |
-
"eval_steps_per_second": 0.
|
8 |
}
|
|
|
1 |
{
|
2 |
"epoch": 3.25,
|
3 |
+
"eval_accuracy": 0.9761904761904762,
|
4 |
+
"eval_loss": 0.09526590257883072,
|
5 |
+
"eval_runtime": 122.5253,
|
6 |
+
"eval_samples_per_second": 1.714,
|
7 |
+
"eval_steps_per_second": 0.433
|
8 |
}
|
trainer_state.json
CHANGED
@@ -1,663 +1,285 @@
|
|
1 |
{
|
2 |
-
"best_metric": 0
|
3 |
-
"best_model_checkpoint": "videomae-base-finetuned-bekhoaxe/checkpoint-
|
4 |
"epoch": 3.25,
|
5 |
"eval_steps": 500,
|
6 |
-
"global_step":
|
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.
|
13 |
-
"grad_norm":
|
14 |
-
"learning_rate":
|
15 |
-
"loss": 0.
|
16 |
"step": 10
|
17 |
},
|
18 |
{
|
19 |
-
"epoch": 0.
|
20 |
-
"grad_norm": 6.
|
21 |
-
"learning_rate":
|
22 |
-
"loss": 0.
|
23 |
"step": 20
|
24 |
},
|
25 |
{
|
26 |
-
"epoch": 0.
|
27 |
-
"grad_norm":
|
28 |
-
"learning_rate":
|
29 |
-
"loss": 0.
|
30 |
"step": 30
|
31 |
},
|
32 |
{
|
33 |
-
"epoch": 0.
|
34 |
-
"grad_norm":
|
35 |
-
"learning_rate":
|
36 |
-
"loss": 0.
|
37 |
"step": 40
|
38 |
},
|
39 |
{
|
40 |
-
"epoch": 0.
|
41 |
-
"grad_norm":
|
42 |
-
"learning_rate":
|
43 |
-
"loss": 0.
|
44 |
"step": 50
|
45 |
},
|
46 |
{
|
47 |
-
"epoch": 0.
|
48 |
-
"grad_norm":
|
49 |
-
"learning_rate":
|
50 |
-
"loss": 0.
|
51 |
"step": 60
|
52 |
},
|
53 |
{
|
54 |
-
"epoch": 0.
|
55 |
-
"grad_norm": 0.
|
56 |
-
"learning_rate": 4.
|
57 |
-
"loss": 0.
|
58 |
"step": 70
|
59 |
},
|
60 |
-
{
|
61 |
-
"epoch": 0.09,
|
62 |
-
"grad_norm": 106.07608795166016,
|
63 |
-
"learning_rate": 4.705882352941177e-05,
|
64 |
-
"loss": 0.3572,
|
65 |
-
"step": 80
|
66 |
-
},
|
67 |
-
{
|
68 |
-
"epoch": 0.11,
|
69 |
-
"grad_norm": 0.02139919251203537,
|
70 |
-
"learning_rate": 4.9672346002621236e-05,
|
71 |
-
"loss": 0.2095,
|
72 |
-
"step": 90
|
73 |
-
},
|
74 |
-
{
|
75 |
-
"epoch": 0.12,
|
76 |
-
"grad_norm": 87.85816955566406,
|
77 |
-
"learning_rate": 4.90170380078637e-05,
|
78 |
-
"loss": 0.389,
|
79 |
-
"step": 100
|
80 |
-
},
|
81 |
-
{
|
82 |
-
"epoch": 0.13,
|
83 |
-
"grad_norm": 0.4829784035682678,
|
84 |
-
"learning_rate": 4.836173001310616e-05,
|
85 |
-
"loss": 0.1773,
|
86 |
-
"step": 110
|
87 |
-
},
|
88 |
-
{
|
89 |
-
"epoch": 0.14,
|
90 |
-
"grad_norm": 0.020957980304956436,
|
91 |
-
"learning_rate": 4.7706422018348626e-05,
|
92 |
-
"loss": 0.0059,
|
93 |
-
"step": 120
|
94 |
-
},
|
95 |
-
{
|
96 |
-
"epoch": 0.15,
|
97 |
-
"grad_norm": 0.018635602667927742,
|
98 |
-
"learning_rate": 4.705111402359109e-05,
|
99 |
-
"loss": 1.1257,
|
100 |
-
"step": 130
|
101 |
-
},
|
102 |
-
{
|
103 |
-
"epoch": 0.17,
|
104 |
-
"grad_norm": 0.2680661976337433,
|
105 |
-
"learning_rate": 4.6395806028833554e-05,
|
106 |
-
"loss": 0.4074,
|
107 |
-
"step": 140
|
108 |
-
},
|
109 |
-
{
|
110 |
-
"epoch": 0.18,
|
111 |
-
"grad_norm": 0.034663423895835876,
|
112 |
-
"learning_rate": 4.5740498034076015e-05,
|
113 |
-
"loss": 0.0125,
|
114 |
-
"step": 150
|
115 |
-
},
|
116 |
-
{
|
117 |
-
"epoch": 0.19,
|
118 |
-
"grad_norm": 0.027404017746448517,
|
119 |
-
"learning_rate": 4.508519003931848e-05,
|
120 |
-
"loss": 0.1217,
|
121 |
-
"step": 160
|
122 |
-
},
|
123 |
-
{
|
124 |
-
"epoch": 0.2,
|
125 |
-
"grad_norm": 25.80691146850586,
|
126 |
-
"learning_rate": 4.4429882044560943e-05,
|
127 |
-
"loss": 0.3167,
|
128 |
-
"step": 170
|
129 |
-
},
|
130 |
-
{
|
131 |
-
"epoch": 0.21,
|
132 |
-
"grad_norm": 0.13836145401000977,
|
133 |
-
"learning_rate": 4.3774574049803404e-05,
|
134 |
-
"loss": 0.5605,
|
135 |
-
"step": 180
|
136 |
-
},
|
137 |
-
{
|
138 |
-
"epoch": 0.22,
|
139 |
-
"grad_norm": 0.010564002208411694,
|
140 |
-
"learning_rate": 4.311926605504588e-05,
|
141 |
-
"loss": 0.002,
|
142 |
-
"step": 190
|
143 |
-
},
|
144 |
-
{
|
145 |
-
"epoch": 0.24,
|
146 |
-
"grad_norm": 0.02207314595580101,
|
147 |
-
"learning_rate": 4.246395806028834e-05,
|
148 |
-
"loss": 0.1598,
|
149 |
-
"step": 200
|
150 |
-
},
|
151 |
-
{
|
152 |
-
"epoch": 0.25,
|
153 |
-
"grad_norm": 0.039745017886161804,
|
154 |
-
"learning_rate": 4.18086500655308e-05,
|
155 |
-
"loss": 0.337,
|
156 |
-
"step": 210
|
157 |
-
},
|
158 |
{
|
159 |
"epoch": 0.25,
|
160 |
-
"eval_accuracy": 0.
|
161 |
-
"eval_loss": 0.
|
162 |
-
"eval_runtime":
|
163 |
-
"eval_samples_per_second":
|
164 |
-
"eval_steps_per_second": 0.
|
165 |
-
"step":
|
166 |
},
|
167 |
{
|
168 |
"epoch": 1.01,
|
169 |
-
"grad_norm": 1.
|
170 |
-
"learning_rate": 4.
|
171 |
-
"loss": 0.
|
172 |
-
"step":
|
173 |
-
},
|
174 |
-
{
|
175 |
-
"epoch": 1.02,
|
176 |
-
"grad_norm": 0.03908234462141991,
|
177 |
-
"learning_rate": 4.049803407601573e-05,
|
178 |
-
"loss": 0.0017,
|
179 |
-
"step": 230
|
180 |
-
},
|
181 |
-
{
|
182 |
-
"epoch": 1.03,
|
183 |
-
"grad_norm": 0.07498504221439362,
|
184 |
-
"learning_rate": 3.984272608125819e-05,
|
185 |
-
"loss": 0.0012,
|
186 |
-
"step": 240
|
187 |
},
|
188 |
{
|
189 |
"epoch": 1.04,
|
190 |
-
"grad_norm": 0.
|
191 |
-
"learning_rate": 3.
|
192 |
-
"loss": 0.
|
193 |
-
"step":
|
194 |
-
},
|
195 |
-
{
|
196 |
-
"epoch": 1.06,
|
197 |
-
"grad_norm": 0.13020159304141998,
|
198 |
-
"learning_rate": 3.8532110091743125e-05,
|
199 |
-
"loss": 0.0009,
|
200 |
-
"step": 260
|
201 |
},
|
202 |
{
|
203 |
"epoch": 1.07,
|
204 |
-
"grad_norm": 0.
|
205 |
-
"learning_rate": 3.
|
206 |
-
"loss": 0.
|
207 |
-
"step":
|
208 |
-
},
|
209 |
-
{
|
210 |
-
"epoch": 1.08,
|
211 |
-
"grad_norm": 0.013056197203695774,
|
212 |
-
"learning_rate": 3.722149410222805e-05,
|
213 |
-
"loss": 0.2843,
|
214 |
-
"step": 280
|
215 |
-
},
|
216 |
-
{
|
217 |
-
"epoch": 1.09,
|
218 |
-
"grad_norm": 0.00816601887345314,
|
219 |
-
"learning_rate": 3.6566186107470514e-05,
|
220 |
-
"loss": 0.2704,
|
221 |
-
"step": 290
|
222 |
-
},
|
223 |
-
{
|
224 |
-
"epoch": 1.1,
|
225 |
-
"grad_norm": 0.4981946647167206,
|
226 |
-
"learning_rate": 3.5910878112712975e-05,
|
227 |
-
"loss": 0.0053,
|
228 |
-
"step": 300
|
229 |
-
},
|
230 |
-
{
|
231 |
-
"epoch": 1.12,
|
232 |
-
"grad_norm": 0.014170478098094463,
|
233 |
-
"learning_rate": 3.5255570117955436e-05,
|
234 |
-
"loss": 0.2984,
|
235 |
-
"step": 310
|
236 |
},
|
237 |
{
|
238 |
-
"epoch": 1.
|
239 |
-
"grad_norm": 0.
|
240 |
-
"learning_rate": 3.
|
241 |
-
"loss": 0.
|
242 |
-
"step":
|
243 |
},
|
244 |
{
|
245 |
"epoch": 1.14,
|
246 |
-
"grad_norm": 0.
|
247 |
-
"learning_rate": 3.
|
248 |
-
"loss": 0.
|
249 |
-
"step":
|
250 |
-
},
|
251 |
-
{
|
252 |
-
"epoch": 1.15,
|
253 |
-
"grad_norm": 2.8213207721710205,
|
254 |
-
"learning_rate": 3.328964613368283e-05,
|
255 |
-
"loss": 0.4437,
|
256 |
-
"step": 340
|
257 |
-
},
|
258 |
-
{
|
259 |
-
"epoch": 1.16,
|
260 |
-
"grad_norm": 0.026097455993294716,
|
261 |
-
"learning_rate": 3.26343381389253e-05,
|
262 |
-
"loss": 0.0015,
|
263 |
-
"step": 350
|
264 |
},
|
265 |
{
|
266 |
"epoch": 1.17,
|
267 |
-
"grad_norm": 0.
|
268 |
-
"learning_rate": 3.
|
269 |
-
"loss": 0.
|
270 |
-
"step":
|
271 |
-
},
|
272 |
-
{
|
273 |
-
"epoch": 1.19,
|
274 |
-
"grad_norm": 0.3815198540687561,
|
275 |
-
"learning_rate": 3.132372214941022e-05,
|
276 |
-
"loss": 0.1395,
|
277 |
-
"step": 370
|
278 |
},
|
279 |
{
|
280 |
"epoch": 1.2,
|
281 |
-
"grad_norm": 0.
|
282 |
-
"learning_rate": 3.
|
283 |
-
"loss": 0.
|
284 |
-
"step":
|
285 |
-
},
|
286 |
-
{
|
287 |
-
"epoch": 1.21,
|
288 |
-
"grad_norm": 7.192620754241943,
|
289 |
-
"learning_rate": 3.0013106159895154e-05,
|
290 |
-
"loss": 0.0059,
|
291 |
-
"step": 390
|
292 |
-
},
|
293 |
-
{
|
294 |
-
"epoch": 1.22,
|
295 |
-
"grad_norm": 0.009683230891823769,
|
296 |
-
"learning_rate": 2.9357798165137618e-05,
|
297 |
-
"loss": 0.0054,
|
298 |
-
"step": 400
|
299 |
-
},
|
300 |
-
{
|
301 |
-
"epoch": 1.23,
|
302 |
-
"grad_norm": 0.0035556950606405735,
|
303 |
-
"learning_rate": 2.8702490170380082e-05,
|
304 |
-
"loss": 0.0122,
|
305 |
-
"step": 410
|
306 |
},
|
307 |
{
|
308 |
-
"epoch": 1.
|
309 |
-
"grad_norm": 0.
|
310 |
-
"learning_rate": 2.
|
311 |
-
"loss": 0.
|
312 |
-
"step":
|
313 |
},
|
314 |
{
|
315 |
"epoch": 1.25,
|
316 |
-
"eval_accuracy": 0.
|
317 |
-
"eval_loss": 0.
|
318 |
-
"eval_runtime":
|
319 |
-
"eval_samples_per_second": 3.
|
320 |
-
"eval_steps_per_second":
|
321 |
-
"step":
|
322 |
-
},
|
323 |
-
{
|
324 |
-
"epoch": 2.01,
|
325 |
-
"grad_norm": 0.017384065315127373,
|
326 |
-
"learning_rate": 2.7391874180865007e-05,
|
327 |
-
"loss": 0.0032,
|
328 |
-
"step": 430
|
329 |
},
|
330 |
{
|
331 |
"epoch": 2.02,
|
332 |
-
"grad_norm": 0.
|
333 |
-
"learning_rate": 2.
|
334 |
-
"loss": 0.
|
335 |
-
"step":
|
336 |
-
},
|
337 |
-
{
|
338 |
-
"epoch": 2.03,
|
339 |
-
"grad_norm": 0.014596754685044289,
|
340 |
-
"learning_rate": 2.6081258191349932e-05,
|
341 |
-
"loss": 0.002,
|
342 |
-
"step": 450
|
343 |
-
},
|
344 |
-
{
|
345 |
-
"epoch": 2.04,
|
346 |
-
"grad_norm": 0.004826271440833807,
|
347 |
-
"learning_rate": 2.5425950196592403e-05,
|
348 |
-
"loss": 0.1837,
|
349 |
-
"step": 460
|
350 |
},
|
351 |
{
|
352 |
"epoch": 2.05,
|
353 |
-
"grad_norm": 0.
|
354 |
-
"learning_rate": 2.
|
355 |
-
"loss": 0.
|
356 |
-
"step":
|
357 |
-
},
|
358 |
-
{
|
359 |
-
"epoch": 2.07,
|
360 |
-
"grad_norm": 0.012596765533089638,
|
361 |
-
"learning_rate": 2.411533420707733e-05,
|
362 |
-
"loss": 0.001,
|
363 |
-
"step": 480
|
364 |
},
|
365 |
{
|
366 |
"epoch": 2.08,
|
367 |
-
"grad_norm": 0.
|
368 |
-
"learning_rate": 2.
|
369 |
-
"loss": 0.
|
370 |
-
"step":
|
371 |
-
},
|
372 |
-
{
|
373 |
-
"epoch": 2.09,
|
374 |
-
"grad_norm": 0.0014000836526975036,
|
375 |
-
"learning_rate": 2.2804718217562254e-05,
|
376 |
-
"loss": 0.3089,
|
377 |
-
"step": 500
|
378 |
-
},
|
379 |
-
{
|
380 |
-
"epoch": 2.1,
|
381 |
-
"grad_norm": 0.0013152866158634424,
|
382 |
-
"learning_rate": 2.214941022280472e-05,
|
383 |
-
"loss": 0.0019,
|
384 |
-
"step": 510
|
385 |
-
},
|
386 |
-
{
|
387 |
-
"epoch": 2.11,
|
388 |
-
"grad_norm": 0.0013162486720830202,
|
389 |
-
"learning_rate": 2.1494102228047182e-05,
|
390 |
-
"loss": 0.0012,
|
391 |
-
"step": 520
|
392 |
},
|
393 |
{
|
394 |
"epoch": 2.12,
|
395 |
-
"grad_norm": 0.
|
396 |
-
"learning_rate": 2.
|
397 |
-
"loss": 0.
|
398 |
-
"step":
|
399 |
-
},
|
400 |
-
{
|
401 |
-
"epoch": 2.14,
|
402 |
-
"grad_norm": 0.0011886212741956115,
|
403 |
-
"learning_rate": 2.018348623853211e-05,
|
404 |
-
"loss": 0.0002,
|
405 |
-
"step": 540
|
406 |
},
|
407 |
{
|
408 |
"epoch": 2.15,
|
409 |
-
"grad_norm": 0.
|
410 |
-
"learning_rate": 1.
|
411 |
-
"loss": 0.
|
412 |
-
"step":
|
413 |
-
},
|
414 |
-
{
|
415 |
-
"epoch": 2.16,
|
416 |
-
"grad_norm": 141.24769592285156,
|
417 |
-
"learning_rate": 1.887287024901704e-05,
|
418 |
-
"loss": 0.0858,
|
419 |
-
"step": 560
|
420 |
-
},
|
421 |
-
{
|
422 |
-
"epoch": 2.17,
|
423 |
-
"grad_norm": 89.11790466308594,
|
424 |
-
"learning_rate": 1.82175622542595e-05,
|
425 |
-
"loss": 0.6766,
|
426 |
-
"step": 570
|
427 |
},
|
428 |
{
|
429 |
"epoch": 2.18,
|
430 |
-
"grad_norm": 0.
|
431 |
-
"learning_rate": 1.
|
432 |
-
"loss": 0.
|
433 |
-
"step":
|
434 |
-
},
|
435 |
-
{
|
436 |
-
"epoch": 2.2,
|
437 |
-
"grad_norm": 0.03515917435288429,
|
438 |
-
"learning_rate": 1.6906946264744432e-05,
|
439 |
-
"loss": 0.0003,
|
440 |
-
"step": 590
|
441 |
},
|
442 |
{
|
443 |
"epoch": 2.21,
|
444 |
-
"grad_norm": 0.
|
445 |
-
"learning_rate": 1.
|
446 |
-
"loss": 0.
|
447 |
-
"step":
|
448 |
-
},
|
449 |
-
{
|
450 |
-
"epoch": 2.22,
|
451 |
-
"grad_norm": 0.011713879182934761,
|
452 |
-
"learning_rate": 1.559633027522936e-05,
|
453 |
-
"loss": 0.5946,
|
454 |
-
"step": 610
|
455 |
-
},
|
456 |
-
{
|
457 |
-
"epoch": 2.23,
|
458 |
-
"grad_norm": 0.06233768165111542,
|
459 |
-
"learning_rate": 1.4941022280471823e-05,
|
460 |
-
"loss": 0.1258,
|
461 |
-
"step": 620
|
462 |
-
},
|
463 |
-
{
|
464 |
-
"epoch": 2.24,
|
465 |
-
"grad_norm": 0.0129940714687109,
|
466 |
-
"learning_rate": 1.4285714285714285e-05,
|
467 |
-
"loss": 0.1747,
|
468 |
-
"step": 630
|
469 |
},
|
470 |
{
|
471 |
"epoch": 2.25,
|
472 |
-
"
|
473 |
-
"
|
474 |
-
"
|
475 |
-
"
|
476 |
-
"eval_steps_per_second": 1.882,
|
477 |
-
"step": 636
|
478 |
-
},
|
479 |
-
{
|
480 |
-
"epoch": 3.0,
|
481 |
-
"grad_norm": 0.0608808733522892,
|
482 |
-
"learning_rate": 1.3630406290956751e-05,
|
483 |
-
"loss": 0.0226,
|
484 |
-
"step": 640
|
485 |
},
|
486 |
{
|
487 |
-
"epoch":
|
488 |
-
"
|
489 |
-
"
|
490 |
-
"
|
491 |
-
"
|
|
|
|
|
492 |
},
|
493 |
{
|
494 |
"epoch": 3.03,
|
495 |
-
"grad_norm": 0.
|
496 |
-
"learning_rate": 1.
|
497 |
-
"loss": 0.
|
498 |
-
"step":
|
499 |
-
},
|
500 |
-
{
|
501 |
-
"epoch": 3.04,
|
502 |
-
"grad_norm": 0.0013453299179673195,
|
503 |
-
"learning_rate": 1.1664482306684142e-05,
|
504 |
-
"loss": 0.1834,
|
505 |
-
"step": 670
|
506 |
-
},
|
507 |
-
{
|
508 |
-
"epoch": 3.05,
|
509 |
-
"grad_norm": 0.0041399141773581505,
|
510 |
-
"learning_rate": 1.1009174311926607e-05,
|
511 |
-
"loss": 0.2115,
|
512 |
-
"step": 680
|
513 |
},
|
514 |
{
|
515 |
"epoch": 3.06,
|
516 |
-
"grad_norm": 0.
|
517 |
-
"learning_rate": 1.
|
518 |
-
"loss": 0.
|
519 |
-
"step":
|
520 |
-
},
|
521 |
-
{
|
522 |
-
"epoch": 3.08,
|
523 |
-
"grad_norm": 0.03554755076766014,
|
524 |
-
"learning_rate": 9.698558322411533e-06,
|
525 |
-
"loss": 0.0316,
|
526 |
-
"step": 700
|
527 |
},
|
528 |
{
|
529 |
"epoch": 3.09,
|
530 |
-
"grad_norm": 0.
|
531 |
-
"learning_rate":
|
532 |
-
"loss": 0.
|
533 |
-
"step":
|
534 |
-
},
|
535 |
-
{
|
536 |
-
"epoch": 3.1,
|
537 |
-
"grad_norm": 0.016770780086517334,
|
538 |
-
"learning_rate": 8.387942332896462e-06,
|
539 |
-
"loss": 0.0004,
|
540 |
-
"step": 720
|
541 |
-
},
|
542 |
-
{
|
543 |
-
"epoch": 3.11,
|
544 |
-
"grad_norm": 0.6642473936080933,
|
545 |
-
"learning_rate": 7.732634338138926e-06,
|
546 |
-
"loss": 0.0012,
|
547 |
-
"step": 730
|
548 |
-
},
|
549 |
-
{
|
550 |
-
"epoch": 3.12,
|
551 |
-
"grad_norm": 0.0030567694921046495,
|
552 |
-
"learning_rate": 7.07732634338139e-06,
|
553 |
-
"loss": 0.2103,
|
554 |
-
"step": 740
|
555 |
},
|
556 |
{
|
557 |
"epoch": 3.13,
|
558 |
-
"grad_norm": 0.
|
559 |
-
"learning_rate": 6.
|
560 |
-
"loss": 0.
|
561 |
-
"step":
|
562 |
-
},
|
563 |
-
{
|
564 |
-
"epoch": 3.15,
|
565 |
-
"grad_norm": 0.02660321444272995,
|
566 |
-
"learning_rate": 5.766710353866317e-06,
|
567 |
-
"loss": 0.0006,
|
568 |
-
"step": 760
|
569 |
},
|
570 |
{
|
571 |
"epoch": 3.16,
|
572 |
-
"grad_norm": 0.
|
573 |
-
"learning_rate": 5.
|
574 |
-
"loss": 0.
|
575 |
-
"step":
|
576 |
-
},
|
577 |
-
{
|
578 |
-
"epoch": 3.17,
|
579 |
-
"grad_norm": 0.008924220688641071,
|
580 |
-
"learning_rate": 4.456094364351245e-06,
|
581 |
-
"loss": 0.0021,
|
582 |
-
"step": 780
|
583 |
-
},
|
584 |
-
{
|
585 |
-
"epoch": 3.18,
|
586 |
-
"grad_norm": 0.02253740094602108,
|
587 |
-
"learning_rate": 3.800786369593709e-06,
|
588 |
-
"loss": 0.0005,
|
589 |
-
"step": 790
|
590 |
},
|
591 |
{
|
592 |
"epoch": 3.19,
|
593 |
-
"grad_norm": 0.
|
594 |
-
"learning_rate": 3.
|
595 |
-
"loss": 0.
|
596 |
-
"step":
|
597 |
-
},
|
598 |
-
{
|
599 |
-
"epoch": 3.21,
|
600 |
-
"grad_norm": 0.01928607001900673,
|
601 |
-
"learning_rate": 2.490170380078637e-06,
|
602 |
-
"loss": 0.0002,
|
603 |
-
"step": 810
|
604 |
},
|
605 |
{
|
606 |
"epoch": 3.22,
|
607 |
-
"grad_norm": 0.
|
608 |
-
"learning_rate": 1.
|
609 |
-
"loss": 0.
|
610 |
-
"step":
|
611 |
-
},
|
612 |
-
{
|
613 |
-
"epoch": 3.23,
|
614 |
-
"grad_norm": 0.012002573348581791,
|
615 |
-
"learning_rate": 1.179554390563565e-06,
|
616 |
-
"loss": 0.0003,
|
617 |
-
"step": 830
|
618 |
-
},
|
619 |
-
{
|
620 |
-
"epoch": 3.24,
|
621 |
-
"grad_norm": 0.009785789065063,
|
622 |
-
"learning_rate": 5.242463958060289e-07,
|
623 |
-
"loss": 0.001,
|
624 |
-
"step": 840
|
625 |
},
|
626 |
{
|
627 |
"epoch": 3.25,
|
628 |
-
"eval_accuracy": 0
|
629 |
-
"eval_loss": 0.
|
630 |
-
"eval_runtime":
|
631 |
-
"eval_samples_per_second": 3.
|
632 |
-
"eval_steps_per_second":
|
633 |
-
"step":
|
634 |
},
|
635 |
{
|
636 |
"epoch": 3.25,
|
637 |
-
"step":
|
638 |
-
"total_flos":
|
639 |
-
"train_loss": 0.
|
640 |
-
"train_runtime":
|
641 |
-
"train_samples_per_second":
|
642 |
-
"train_steps_per_second": 0.
|
643 |
},
|
644 |
{
|
645 |
"epoch": 3.25,
|
646 |
-
"eval_accuracy": 0.
|
647 |
-
"eval_loss": 0.
|
648 |
-
"eval_runtime":
|
649 |
-
"eval_samples_per_second": 1.
|
650 |
-
"eval_steps_per_second": 0.
|
651 |
-
"step":
|
652 |
}
|
653 |
],
|
654 |
"logging_steps": 10,
|
655 |
-
"max_steps":
|
656 |
"num_input_tokens_seen": 0,
|
657 |
"num_train_epochs": 9223372036854775807,
|
658 |
"save_steps": 500,
|
659 |
-
"total_flos":
|
660 |
-
"train_batch_size":
|
661 |
"trial_name": null,
|
662 |
"trial_params": null
|
663 |
}
|
|
|
1 |
{
|
2 |
+
"best_metric": 1.0,
|
3 |
+
"best_model_checkpoint": "videomae-base-finetuned-bekhoaxe/checkpoint-231",
|
4 |
"epoch": 3.25,
|
5 |
"eval_steps": 500,
|
6 |
+
"global_step": 308,
|
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": 4.9310173988342285,
|
14 |
+
"learning_rate": 1.6129032258064517e-05,
|
15 |
+
"loss": 0.7032,
|
16 |
"step": 10
|
17 |
},
|
18 |
{
|
19 |
+
"epoch": 0.06,
|
20 |
+
"grad_norm": 6.229051113128662,
|
21 |
+
"learning_rate": 3.2258064516129034e-05,
|
22 |
+
"loss": 0.5738,
|
23 |
"step": 20
|
24 |
},
|
25 |
{
|
26 |
+
"epoch": 0.1,
|
27 |
+
"grad_norm": 9.958233833312988,
|
28 |
+
"learning_rate": 4.8387096774193554e-05,
|
29 |
+
"loss": 0.3415,
|
30 |
"step": 30
|
31 |
},
|
32 |
{
|
33 |
+
"epoch": 0.13,
|
34 |
+
"grad_norm": 7.560524940490723,
|
35 |
+
"learning_rate": 4.837545126353791e-05,
|
36 |
+
"loss": 0.4689,
|
37 |
"step": 40
|
38 |
},
|
39 |
{
|
40 |
+
"epoch": 0.16,
|
41 |
+
"grad_norm": 5.026527404785156,
|
42 |
+
"learning_rate": 4.657039711191336e-05,
|
43 |
+
"loss": 0.2069,
|
44 |
"step": 50
|
45 |
},
|
46 |
{
|
47 |
+
"epoch": 0.19,
|
48 |
+
"grad_norm": 0.358181893825531,
|
49 |
+
"learning_rate": 4.4765342960288806e-05,
|
50 |
+
"loss": 0.0961,
|
51 |
"step": 60
|
52 |
},
|
53 |
{
|
54 |
+
"epoch": 0.23,
|
55 |
+
"grad_norm": 0.05433480069041252,
|
56 |
+
"learning_rate": 4.296028880866426e-05,
|
57 |
+
"loss": 0.2172,
|
58 |
"step": 70
|
59 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
60 |
{
|
61 |
"epoch": 0.25,
|
62 |
+
"eval_accuracy": 0.9915254237288136,
|
63 |
+
"eval_loss": 0.013951542787253857,
|
64 |
+
"eval_runtime": 38.5113,
|
65 |
+
"eval_samples_per_second": 3.064,
|
66 |
+
"eval_steps_per_second": 0.779,
|
67 |
+
"step": 77
|
68 |
},
|
69 |
{
|
70 |
"epoch": 1.01,
|
71 |
+
"grad_norm": 1.3378506898880005,
|
72 |
+
"learning_rate": 4.115523465703972e-05,
|
73 |
+
"loss": 0.3363,
|
74 |
+
"step": 80
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
75 |
},
|
76 |
{
|
77 |
"epoch": 1.04,
|
78 |
+
"grad_norm": 0.04731794446706772,
|
79 |
+
"learning_rate": 3.935018050541516e-05,
|
80 |
+
"loss": 0.195,
|
81 |
+
"step": 90
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
82 |
},
|
83 |
{
|
84 |
"epoch": 1.07,
|
85 |
+
"grad_norm": 0.04140738397836685,
|
86 |
+
"learning_rate": 3.754512635379062e-05,
|
87 |
+
"loss": 0.0049,
|
88 |
+
"step": 100
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
89 |
},
|
90 |
{
|
91 |
+
"epoch": 1.11,
|
92 |
+
"grad_norm": 0.08140890300273895,
|
93 |
+
"learning_rate": 3.574007220216607e-05,
|
94 |
+
"loss": 0.2396,
|
95 |
+
"step": 110
|
96 |
},
|
97 |
{
|
98 |
"epoch": 1.14,
|
99 |
+
"grad_norm": 0.09178142994642258,
|
100 |
+
"learning_rate": 3.3935018050541516e-05,
|
101 |
+
"loss": 0.0746,
|
102 |
+
"step": 120
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
103 |
},
|
104 |
{
|
105 |
"epoch": 1.17,
|
106 |
+
"grad_norm": 0.031925346702337265,
|
107 |
+
"learning_rate": 3.212996389891697e-05,
|
108 |
+
"loss": 0.3279,
|
109 |
+
"step": 130
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
110 |
},
|
111 |
{
|
112 |
"epoch": 1.2,
|
113 |
+
"grad_norm": 0.11948370933532715,
|
114 |
+
"learning_rate": 3.032490974729242e-05,
|
115 |
+
"loss": 0.0021,
|
116 |
+
"step": 140
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
117 |
},
|
118 |
{
|
119 |
+
"epoch": 1.24,
|
120 |
+
"grad_norm": 0.05676674470305443,
|
121 |
+
"learning_rate": 2.851985559566787e-05,
|
122 |
+
"loss": 0.1826,
|
123 |
+
"step": 150
|
124 |
},
|
125 |
{
|
126 |
"epoch": 1.25,
|
127 |
+
"eval_accuracy": 0.9830508474576272,
|
128 |
+
"eval_loss": 0.042561955749988556,
|
129 |
+
"eval_runtime": 38.1931,
|
130 |
+
"eval_samples_per_second": 3.09,
|
131 |
+
"eval_steps_per_second": 0.785,
|
132 |
+
"step": 154
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
133 |
},
|
134 |
{
|
135 |
"epoch": 2.02,
|
136 |
+
"grad_norm": 0.11792542040348053,
|
137 |
+
"learning_rate": 2.6714801444043324e-05,
|
138 |
+
"loss": 0.1442,
|
139 |
+
"step": 160
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
140 |
},
|
141 |
{
|
142 |
"epoch": 2.05,
|
143 |
+
"grad_norm": 0.10978111624717712,
|
144 |
+
"learning_rate": 2.4909747292418774e-05,
|
145 |
+
"loss": 0.0912,
|
146 |
+
"step": 170
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
147 |
},
|
148 |
{
|
149 |
"epoch": 2.08,
|
150 |
+
"grad_norm": 0.016077380627393723,
|
151 |
+
"learning_rate": 2.3104693140794227e-05,
|
152 |
+
"loss": 0.1236,
|
153 |
+
"step": 180
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
154 |
},
|
155 |
{
|
156 |
"epoch": 2.12,
|
157 |
+
"grad_norm": 0.01301741786301136,
|
158 |
+
"learning_rate": 2.1299638989169676e-05,
|
159 |
+
"loss": 0.1501,
|
160 |
+
"step": 190
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
161 |
},
|
162 |
{
|
163 |
"epoch": 2.15,
|
164 |
+
"grad_norm": 0.026157772168517113,
|
165 |
+
"learning_rate": 1.9494584837545125e-05,
|
166 |
+
"loss": 0.18,
|
167 |
+
"step": 200
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
168 |
},
|
169 |
{
|
170 |
"epoch": 2.18,
|
171 |
+
"grad_norm": 0.04890529438853264,
|
172 |
+
"learning_rate": 1.768953068592058e-05,
|
173 |
+
"loss": 0.0038,
|
174 |
+
"step": 210
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
175 |
},
|
176 |
{
|
177 |
"epoch": 2.21,
|
178 |
+
"grad_norm": 0.038881637156009674,
|
179 |
+
"learning_rate": 1.588447653429603e-05,
|
180 |
+
"loss": 0.0019,
|
181 |
+
"step": 220
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
182 |
},
|
183 |
{
|
184 |
"epoch": 2.25,
|
185 |
+
"grad_norm": 0.01524051371961832,
|
186 |
+
"learning_rate": 1.407942238267148e-05,
|
187 |
+
"loss": 0.1427,
|
188 |
+
"step": 230
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
189 |
},
|
190 |
{
|
191 |
+
"epoch": 2.25,
|
192 |
+
"eval_accuracy": 1.0,
|
193 |
+
"eval_loss": 0.0019937974866479635,
|
194 |
+
"eval_runtime": 38.3352,
|
195 |
+
"eval_samples_per_second": 3.078,
|
196 |
+
"eval_steps_per_second": 0.783,
|
197 |
+
"step": 231
|
198 |
},
|
199 |
{
|
200 |
"epoch": 3.03,
|
201 |
+
"grad_norm": 0.01709669642150402,
|
202 |
+
"learning_rate": 1.2274368231046932e-05,
|
203 |
+
"loss": 0.0012,
|
204 |
+
"step": 240
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
205 |
},
|
206 |
{
|
207 |
"epoch": 3.06,
|
208 |
+
"grad_norm": 0.1900138109922409,
|
209 |
+
"learning_rate": 1.0469314079422383e-05,
|
210 |
+
"loss": 0.1073,
|
211 |
+
"step": 250
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
212 |
},
|
213 |
{
|
214 |
"epoch": 3.09,
|
215 |
+
"grad_norm": 0.06225905194878578,
|
216 |
+
"learning_rate": 8.664259927797834e-06,
|
217 |
+
"loss": 0.0029,
|
218 |
+
"step": 260
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
219 |
},
|
220 |
{
|
221 |
"epoch": 3.13,
|
222 |
+
"grad_norm": 0.015592777170240879,
|
223 |
+
"learning_rate": 6.859205776173286e-06,
|
224 |
+
"loss": 0.0009,
|
225 |
+
"step": 270
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
226 |
},
|
227 |
{
|
228 |
"epoch": 3.16,
|
229 |
+
"grad_norm": 0.016208168119192123,
|
230 |
+
"learning_rate": 5.054151624548736e-06,
|
231 |
+
"loss": 0.0036,
|
232 |
+
"step": 280
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
233 |
},
|
234 |
{
|
235 |
"epoch": 3.19,
|
236 |
+
"grad_norm": 0.007616001646965742,
|
237 |
+
"learning_rate": 3.2490974729241876e-06,
|
238 |
+
"loss": 0.0008,
|
239 |
+
"step": 290
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
240 |
},
|
241 |
{
|
242 |
"epoch": 3.22,
|
243 |
+
"grad_norm": 0.008425813168287277,
|
244 |
+
"learning_rate": 1.4440433212996392e-06,
|
245 |
+
"loss": 0.2853,
|
246 |
+
"step": 300
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
247 |
},
|
248 |
{
|
249 |
"epoch": 3.25,
|
250 |
+
"eval_accuracy": 1.0,
|
251 |
+
"eval_loss": 0.0012017178814858198,
|
252 |
+
"eval_runtime": 38.094,
|
253 |
+
"eval_samples_per_second": 3.098,
|
254 |
+
"eval_steps_per_second": 0.788,
|
255 |
+
"step": 308
|
256 |
},
|
257 |
{
|
258 |
"epoch": 3.25,
|
259 |
+
"step": 308,
|
260 |
+
"total_flos": 1.5351515742291886e+18,
|
261 |
+
"train_loss": 0.16999648642641577,
|
262 |
+
"train_runtime": 584.5872,
|
263 |
+
"train_samples_per_second": 2.107,
|
264 |
+
"train_steps_per_second": 0.527
|
265 |
},
|
266 |
{
|
267 |
"epoch": 3.25,
|
268 |
+
"eval_accuracy": 0.9761904761904762,
|
269 |
+
"eval_loss": 0.09526590257883072,
|
270 |
+
"eval_runtime": 122.5253,
|
271 |
+
"eval_samples_per_second": 1.714,
|
272 |
+
"eval_steps_per_second": 0.433,
|
273 |
+
"step": 308
|
274 |
}
|
275 |
],
|
276 |
"logging_steps": 10,
|
277 |
+
"max_steps": 308,
|
278 |
"num_input_tokens_seen": 0,
|
279 |
"num_train_epochs": 9223372036854775807,
|
280 |
"save_steps": 500,
|
281 |
+
"total_flos": 1.5351515742291886e+18,
|
282 |
+
"train_batch_size": 4,
|
283 |
"trial_name": null,
|
284 |
"trial_params": null
|
285 |
}
|