rishitunu commited on
Commit
8db463e
1 Parent(s): cbda0d1

End of training

Browse files
Files changed (5) hide show
  1. README.md +3 -1
  2. all_results.json +18 -0
  3. eval_results.json +14 -0
  4. train_results.json +7 -0
  5. trainer_state.json +913 -0
README.md CHANGED
@@ -2,6 +2,8 @@
2
  license: other
3
  base_model: nvidia/mit-b5
4
  tags:
 
 
5
  - generated_from_trainer
6
  model-index:
7
  - name: FINAL_ecc_segformer
@@ -13,7 +15,7 @@ should probably proofread and complete it, then remove this comment. -->
13
 
14
  # FINAL_ecc_segformer
15
 
16
- This model is a fine-tuned version of [nvidia/mit-b5](https://huggingface.co/nvidia/mit-b5) on an unknown dataset.
17
  It achieves the following results on the evaluation set:
18
  - Loss: 0.0749
19
  - Mean Iou: 0.1968
 
2
  license: other
3
  base_model: nvidia/mit-b5
4
  tags:
5
+ - image-segmentation
6
+ - vision
7
  - generated_from_trainer
8
  model-index:
9
  - name: FINAL_ecc_segformer
 
15
 
16
  # FINAL_ecc_segformer
17
 
18
+ This model is a fine-tuned version of [nvidia/mit-b5](https://huggingface.co/nvidia/mit-b5) on the rishitunu/ecc_crackdetector_dataset_exhaustive dataset.
19
  It achieves the following results on the evaluation set:
20
  - Loss: 0.0749
21
  - Mean Iou: 0.1968
all_results.json ADDED
@@ -0,0 +1,18 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 18.25,
3
+ "eval_accuracy_background": NaN,
4
+ "eval_accuracy_crack": 0.39388432615865027,
5
+ "eval_iou_background": 0.0,
6
+ "eval_iou_crack": 0.3935762911195175,
7
+ "eval_loss": 0.07488969713449478,
8
+ "eval_mean_accuracy": 0.39388432615865027,
9
+ "eval_mean_iou": 0.19678814555975874,
10
+ "eval_overall_accuracy": 0.39388432615865027,
11
+ "eval_runtime": 392.3604,
12
+ "eval_samples_per_second": 0.494,
13
+ "eval_steps_per_second": 0.247,
14
+ "train_loss": 0.039843981289863585,
15
+ "train_runtime": 169492.4367,
16
+ "train_samples_per_second": 0.118,
17
+ "train_steps_per_second": 0.059
18
+ }
eval_results.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 18.25,
3
+ "eval_accuracy_background": NaN,
4
+ "eval_accuracy_crack": 0.39388432615865027,
5
+ "eval_iou_background": 0.0,
6
+ "eval_iou_crack": 0.3935762911195175,
7
+ "eval_loss": 0.07488969713449478,
8
+ "eval_mean_accuracy": 0.39388432615865027,
9
+ "eval_mean_iou": 0.19678814555975874,
10
+ "eval_overall_accuracy": 0.39388432615865027,
11
+ "eval_runtime": 392.3604,
12
+ "eval_samples_per_second": 0.494,
13
+ "eval_steps_per_second": 0.247
14
+ }
train_results.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 18.25,
3
+ "train_loss": 0.039843981289863585,
4
+ "train_runtime": 169492.4367,
5
+ "train_samples_per_second": 0.118,
6
+ "train_steps_per_second": 0.059
7
+ }
trainer_state.json ADDED
@@ -0,0 +1,913 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 18.248175182481752,
5
+ "eval_steps": 500,
6
+ "global_step": 10000,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.18,
13
+ "learning_rate": 5.9401e-05,
14
+ "loss": 0.1021,
15
+ "step": 100
16
+ },
17
+ {
18
+ "epoch": 0.36,
19
+ "learning_rate": 5.8802000000000004e-05,
20
+ "loss": 0.0544,
21
+ "step": 200
22
+ },
23
+ {
24
+ "epoch": 0.55,
25
+ "learning_rate": 5.8203e-05,
26
+ "loss": 0.0664,
27
+ "step": 300
28
+ },
29
+ {
30
+ "epoch": 0.73,
31
+ "learning_rate": 5.7604e-05,
32
+ "loss": 0.0563,
33
+ "step": 400
34
+ },
35
+ {
36
+ "epoch": 0.91,
37
+ "learning_rate": 5.7005e-05,
38
+ "loss": 0.0534,
39
+ "step": 500
40
+ },
41
+ {
42
+ "epoch": 1.0,
43
+ "eval_accuracy_background": NaN,
44
+ "eval_accuracy_crack": 0.2749723182564086,
45
+ "eval_iou_background": 0.0,
46
+ "eval_iou_crack": 0.27359924384134937,
47
+ "eval_loss": 0.061391595751047134,
48
+ "eval_mean_accuracy": 0.2749723182564086,
49
+ "eval_mean_iou": 0.13679962192067469,
50
+ "eval_overall_accuracy": 0.2749723182564086,
51
+ "eval_runtime": 371.8778,
52
+ "eval_samples_per_second": 0.522,
53
+ "eval_steps_per_second": 0.261,
54
+ "step": 548
55
+ },
56
+ {
57
+ "epoch": 1.09,
58
+ "learning_rate": 5.6406e-05,
59
+ "loss": 0.0555,
60
+ "step": 600
61
+ },
62
+ {
63
+ "epoch": 1.28,
64
+ "learning_rate": 5.5806999999999996e-05,
65
+ "loss": 0.0437,
66
+ "step": 700
67
+ },
68
+ {
69
+ "epoch": 1.46,
70
+ "learning_rate": 5.5208000000000004e-05,
71
+ "loss": 0.0619,
72
+ "step": 800
73
+ },
74
+ {
75
+ "epoch": 1.64,
76
+ "learning_rate": 5.4609000000000005e-05,
77
+ "loss": 0.0452,
78
+ "step": 900
79
+ },
80
+ {
81
+ "epoch": 1.82,
82
+ "learning_rate": 5.401e-05,
83
+ "loss": 0.058,
84
+ "step": 1000
85
+ },
86
+ {
87
+ "epoch": 2.0,
88
+ "eval_accuracy_background": NaN,
89
+ "eval_accuracy_crack": 0.42382043908423217,
90
+ "eval_iou_background": 0.0,
91
+ "eval_iou_crack": 0.4185528527078574,
92
+ "eval_loss": 0.10176104307174683,
93
+ "eval_mean_accuracy": 0.42382043908423217,
94
+ "eval_mean_iou": 0.2092764263539287,
95
+ "eval_overall_accuracy": 0.42382043908423217,
96
+ "eval_runtime": 378.5472,
97
+ "eval_samples_per_second": 0.512,
98
+ "eval_steps_per_second": 0.256,
99
+ "step": 1096
100
+ },
101
+ {
102
+ "epoch": 2.01,
103
+ "learning_rate": 5.3411e-05,
104
+ "loss": 0.0574,
105
+ "step": 1100
106
+ },
107
+ {
108
+ "epoch": 2.19,
109
+ "learning_rate": 5.2812e-05,
110
+ "loss": 0.0416,
111
+ "step": 1200
112
+ },
113
+ {
114
+ "epoch": 2.37,
115
+ "learning_rate": 5.2213e-05,
116
+ "loss": 0.0675,
117
+ "step": 1300
118
+ },
119
+ {
120
+ "epoch": 2.55,
121
+ "learning_rate": 5.1614000000000004e-05,
122
+ "loss": 0.0482,
123
+ "step": 1400
124
+ },
125
+ {
126
+ "epoch": 2.74,
127
+ "learning_rate": 5.1015e-05,
128
+ "loss": 0.0502,
129
+ "step": 1500
130
+ },
131
+ {
132
+ "epoch": 2.92,
133
+ "learning_rate": 5.0416e-05,
134
+ "loss": 0.0482,
135
+ "step": 1600
136
+ },
137
+ {
138
+ "epoch": 3.0,
139
+ "eval_accuracy_background": NaN,
140
+ "eval_accuracy_crack": 0.43149166759602975,
141
+ "eval_iou_background": 0.0,
142
+ "eval_iou_crack": 0.35824652777777777,
143
+ "eval_loss": 0.05083267390727997,
144
+ "eval_mean_accuracy": 0.43149166759602975,
145
+ "eval_mean_iou": 0.17912326388888888,
146
+ "eval_overall_accuracy": 0.43149166759602975,
147
+ "eval_runtime": 380.6012,
148
+ "eval_samples_per_second": 0.51,
149
+ "eval_steps_per_second": 0.255,
150
+ "step": 1644
151
+ },
152
+ {
153
+ "epoch": 3.1,
154
+ "learning_rate": 4.9817e-05,
155
+ "loss": 0.0507,
156
+ "step": 1700
157
+ },
158
+ {
159
+ "epoch": 3.28,
160
+ "learning_rate": 4.9218e-05,
161
+ "loss": 0.0492,
162
+ "step": 1800
163
+ },
164
+ {
165
+ "epoch": 3.47,
166
+ "learning_rate": 4.8619e-05,
167
+ "loss": 0.0458,
168
+ "step": 1900
169
+ },
170
+ {
171
+ "epoch": 3.65,
172
+ "learning_rate": 4.8020000000000004e-05,
173
+ "loss": 0.0469,
174
+ "step": 2000
175
+ },
176
+ {
177
+ "epoch": 3.83,
178
+ "learning_rate": 4.7421000000000006e-05,
179
+ "loss": 0.0338,
180
+ "step": 2100
181
+ },
182
+ {
183
+ "epoch": 4.0,
184
+ "eval_accuracy_background": NaN,
185
+ "eval_accuracy_crack": 0.3716453311453471,
186
+ "eval_iou_background": 0.0,
187
+ "eval_iou_crack": 0.3698312917782239,
188
+ "eval_loss": 0.05686204507946968,
189
+ "eval_mean_accuracy": 0.3716453311453471,
190
+ "eval_mean_iou": 0.18491564588911194,
191
+ "eval_overall_accuracy": 0.3716453311453471,
192
+ "eval_runtime": 385.9938,
193
+ "eval_samples_per_second": 0.503,
194
+ "eval_steps_per_second": 0.251,
195
+ "step": 2192
196
+ },
197
+ {
198
+ "epoch": 4.01,
199
+ "learning_rate": 4.6822e-05,
200
+ "loss": 0.051,
201
+ "step": 2200
202
+ },
203
+ {
204
+ "epoch": 4.2,
205
+ "learning_rate": 4.6223e-05,
206
+ "loss": 0.0353,
207
+ "step": 2300
208
+ },
209
+ {
210
+ "epoch": 4.38,
211
+ "learning_rate": 4.5624e-05,
212
+ "loss": 0.0503,
213
+ "step": 2400
214
+ },
215
+ {
216
+ "epoch": 4.56,
217
+ "learning_rate": 4.5025000000000003e-05,
218
+ "loss": 0.0425,
219
+ "step": 2500
220
+ },
221
+ {
222
+ "epoch": 4.74,
223
+ "learning_rate": 4.4426000000000005e-05,
224
+ "loss": 0.0528,
225
+ "step": 2600
226
+ },
227
+ {
228
+ "epoch": 4.93,
229
+ "learning_rate": 4.3827e-05,
230
+ "loss": 0.0395,
231
+ "step": 2700
232
+ },
233
+ {
234
+ "epoch": 5.0,
235
+ "eval_accuracy_background": NaN,
236
+ "eval_accuracy_crack": 0.3505922698233148,
237
+ "eval_iou_background": 0.0,
238
+ "eval_iou_crack": 0.34898750111508686,
239
+ "eval_loss": 0.0596516877412796,
240
+ "eval_mean_accuracy": 0.3505922698233148,
241
+ "eval_mean_iou": 0.17449375055754343,
242
+ "eval_overall_accuracy": 0.3505922698233148,
243
+ "eval_runtime": 382.2611,
244
+ "eval_samples_per_second": 0.508,
245
+ "eval_steps_per_second": 0.254,
246
+ "step": 2740
247
+ },
248
+ {
249
+ "epoch": 5.11,
250
+ "learning_rate": 4.3228e-05,
251
+ "loss": 0.0445,
252
+ "step": 2800
253
+ },
254
+ {
255
+ "epoch": 5.29,
256
+ "learning_rate": 4.2629e-05,
257
+ "loss": 0.0495,
258
+ "step": 2900
259
+ },
260
+ {
261
+ "epoch": 5.47,
262
+ "learning_rate": 4.203e-05,
263
+ "loss": 0.0468,
264
+ "step": 3000
265
+ },
266
+ {
267
+ "epoch": 5.66,
268
+ "learning_rate": 4.1431e-05,
269
+ "loss": 0.0488,
270
+ "step": 3100
271
+ },
272
+ {
273
+ "epoch": 5.84,
274
+ "learning_rate": 4.0832e-05,
275
+ "loss": 0.0372,
276
+ "step": 3200
277
+ },
278
+ {
279
+ "epoch": 6.0,
280
+ "eval_accuracy_background": NaN,
281
+ "eval_accuracy_crack": 0.46350689852948207,
282
+ "eval_iou_background": 0.0,
283
+ "eval_iou_crack": 0.459698930762766,
284
+ "eval_loss": 0.050851527601480484,
285
+ "eval_mean_accuracy": 0.46350689852948207,
286
+ "eval_mean_iou": 0.229849465381383,
287
+ "eval_overall_accuracy": 0.46350689852948207,
288
+ "eval_runtime": 381.5402,
289
+ "eval_samples_per_second": 0.508,
290
+ "eval_steps_per_second": 0.254,
291
+ "step": 3288
292
+ },
293
+ {
294
+ "epoch": 6.02,
295
+ "learning_rate": 4.0233e-05,
296
+ "loss": 0.0452,
297
+ "step": 3300
298
+ },
299
+ {
300
+ "epoch": 6.2,
301
+ "learning_rate": 3.9634e-05,
302
+ "loss": 0.0588,
303
+ "step": 3400
304
+ },
305
+ {
306
+ "epoch": 6.39,
307
+ "learning_rate": 3.9035e-05,
308
+ "loss": 0.0408,
309
+ "step": 3500
310
+ },
311
+ {
312
+ "epoch": 6.57,
313
+ "learning_rate": 3.8436e-05,
314
+ "loss": 0.0441,
315
+ "step": 3600
316
+ },
317
+ {
318
+ "epoch": 6.75,
319
+ "learning_rate": 3.7837000000000004e-05,
320
+ "loss": 0.0319,
321
+ "step": 3700
322
+ },
323
+ {
324
+ "epoch": 6.93,
325
+ "learning_rate": 3.7238000000000005e-05,
326
+ "loss": 0.0402,
327
+ "step": 3800
328
+ },
329
+ {
330
+ "epoch": 7.0,
331
+ "eval_accuracy_background": NaN,
332
+ "eval_accuracy_crack": 0.35071275511016936,
333
+ "eval_iou_background": 0.0,
334
+ "eval_iou_crack": 0.3502962805306196,
335
+ "eval_loss": 0.061992477625608444,
336
+ "eval_mean_accuracy": 0.35071275511016936,
337
+ "eval_mean_iou": 0.1751481402653098,
338
+ "eval_overall_accuracy": 0.35071275511016936,
339
+ "eval_runtime": 381.7043,
340
+ "eval_samples_per_second": 0.508,
341
+ "eval_steps_per_second": 0.254,
342
+ "step": 3836
343
+ },
344
+ {
345
+ "epoch": 7.12,
346
+ "learning_rate": 3.6639e-05,
347
+ "loss": 0.0347,
348
+ "step": 3900
349
+ },
350
+ {
351
+ "epoch": 7.3,
352
+ "learning_rate": 3.604e-05,
353
+ "loss": 0.0356,
354
+ "step": 4000
355
+ },
356
+ {
357
+ "epoch": 7.48,
358
+ "learning_rate": 3.544100000000001e-05,
359
+ "loss": 0.046,
360
+ "step": 4100
361
+ },
362
+ {
363
+ "epoch": 7.66,
364
+ "learning_rate": 3.4842e-05,
365
+ "loss": 0.0417,
366
+ "step": 4200
367
+ },
368
+ {
369
+ "epoch": 7.85,
370
+ "learning_rate": 3.4243000000000004e-05,
371
+ "loss": 0.038,
372
+ "step": 4300
373
+ },
374
+ {
375
+ "epoch": 8.0,
376
+ "eval_accuracy_background": NaN,
377
+ "eval_accuracy_crack": 0.38145343094301143,
378
+ "eval_iou_background": 0.0,
379
+ "eval_iou_crack": 0.381008030267165,
380
+ "eval_loss": 0.06807420402765274,
381
+ "eval_mean_accuracy": 0.38145343094301143,
382
+ "eval_mean_iou": 0.1905040151335825,
383
+ "eval_overall_accuracy": 0.38145343094301143,
384
+ "eval_runtime": 380.8562,
385
+ "eval_samples_per_second": 0.509,
386
+ "eval_steps_per_second": 0.255,
387
+ "step": 4384
388
+ },
389
+ {
390
+ "epoch": 8.03,
391
+ "learning_rate": 3.3644000000000005e-05,
392
+ "loss": 0.0471,
393
+ "step": 4400
394
+ },
395
+ {
396
+ "epoch": 8.21,
397
+ "learning_rate": 3.3045000000000006e-05,
398
+ "loss": 0.0504,
399
+ "step": 4500
400
+ },
401
+ {
402
+ "epoch": 8.39,
403
+ "learning_rate": 3.2446e-05,
404
+ "loss": 0.0406,
405
+ "step": 4600
406
+ },
407
+ {
408
+ "epoch": 8.58,
409
+ "learning_rate": 3.1847e-05,
410
+ "loss": 0.0336,
411
+ "step": 4700
412
+ },
413
+ {
414
+ "epoch": 8.76,
415
+ "learning_rate": 3.1248e-05,
416
+ "loss": 0.0322,
417
+ "step": 4800
418
+ },
419
+ {
420
+ "epoch": 8.94,
421
+ "learning_rate": 3.0649000000000004e-05,
422
+ "loss": 0.0393,
423
+ "step": 4900
424
+ },
425
+ {
426
+ "epoch": 9.0,
427
+ "eval_accuracy_background": NaN,
428
+ "eval_accuracy_crack": 0.4433271066006022,
429
+ "eval_iou_background": 0.0,
430
+ "eval_iou_crack": 0.4425224978083646,
431
+ "eval_loss": 0.06852483004331589,
432
+ "eval_mean_accuracy": 0.4433271066006022,
433
+ "eval_mean_iou": 0.2212612489041823,
434
+ "eval_overall_accuracy": 0.4433271066006022,
435
+ "eval_runtime": 380.8095,
436
+ "eval_samples_per_second": 0.509,
437
+ "eval_steps_per_second": 0.255,
438
+ "step": 4932
439
+ },
440
+ {
441
+ "epoch": 9.12,
442
+ "learning_rate": 3.0050000000000002e-05,
443
+ "loss": 0.0475,
444
+ "step": 5000
445
+ },
446
+ {
447
+ "epoch": 9.31,
448
+ "learning_rate": 2.9451e-05,
449
+ "loss": 0.0335,
450
+ "step": 5100
451
+ },
452
+ {
453
+ "epoch": 9.49,
454
+ "learning_rate": 2.8851999999999998e-05,
455
+ "loss": 0.0402,
456
+ "step": 5200
457
+ },
458
+ {
459
+ "epoch": 9.67,
460
+ "learning_rate": 2.8253e-05,
461
+ "loss": 0.0405,
462
+ "step": 5300
463
+ },
464
+ {
465
+ "epoch": 9.85,
466
+ "learning_rate": 2.7653999999999996e-05,
467
+ "loss": 0.0376,
468
+ "step": 5400
469
+ },
470
+ {
471
+ "epoch": 10.0,
472
+ "eval_accuracy_background": NaN,
473
+ "eval_accuracy_crack": 0.39291446938677965,
474
+ "eval_iou_background": 0.0,
475
+ "eval_iou_crack": 0.39239480430627627,
476
+ "eval_loss": 0.05900894105434418,
477
+ "eval_mean_accuracy": 0.39291446938677965,
478
+ "eval_mean_iou": 0.19619740215313813,
479
+ "eval_overall_accuracy": 0.39291446938677965,
480
+ "eval_runtime": 380.8979,
481
+ "eval_samples_per_second": 0.509,
482
+ "eval_steps_per_second": 0.255,
483
+ "step": 5480
484
+ },
485
+ {
486
+ "epoch": 10.04,
487
+ "learning_rate": 2.7054999999999998e-05,
488
+ "loss": 0.0282,
489
+ "step": 5500
490
+ },
491
+ {
492
+ "epoch": 10.22,
493
+ "learning_rate": 2.6455999999999995e-05,
494
+ "loss": 0.0246,
495
+ "step": 5600
496
+ },
497
+ {
498
+ "epoch": 10.4,
499
+ "learning_rate": 2.5857e-05,
500
+ "loss": 0.0362,
501
+ "step": 5700
502
+ },
503
+ {
504
+ "epoch": 10.58,
505
+ "learning_rate": 2.5258e-05,
506
+ "loss": 0.0333,
507
+ "step": 5800
508
+ },
509
+ {
510
+ "epoch": 10.77,
511
+ "learning_rate": 2.4659e-05,
512
+ "loss": 0.0437,
513
+ "step": 5900
514
+ },
515
+ {
516
+ "epoch": 10.95,
517
+ "learning_rate": 2.406e-05,
518
+ "loss": 0.0381,
519
+ "step": 6000
520
+ },
521
+ {
522
+ "epoch": 11.0,
523
+ "eval_accuracy_background": NaN,
524
+ "eval_accuracy_crack": 0.3801490034572307,
525
+ "eval_iou_background": 0.0,
526
+ "eval_iou_crack": 0.37829945440961094,
527
+ "eval_loss": 0.06262247264385223,
528
+ "eval_mean_accuracy": 0.3801490034572307,
529
+ "eval_mean_iou": 0.18914972720480547,
530
+ "eval_overall_accuracy": 0.3801490034572307,
531
+ "eval_runtime": 379.9216,
532
+ "eval_samples_per_second": 0.511,
533
+ "eval_steps_per_second": 0.255,
534
+ "step": 6028
535
+ },
536
+ {
537
+ "epoch": 11.13,
538
+ "learning_rate": 2.3460999999999998e-05,
539
+ "loss": 0.0338,
540
+ "step": 6100
541
+ },
542
+ {
543
+ "epoch": 11.31,
544
+ "learning_rate": 2.2862e-05,
545
+ "loss": 0.0453,
546
+ "step": 6200
547
+ },
548
+ {
549
+ "epoch": 11.5,
550
+ "learning_rate": 2.2263e-05,
551
+ "loss": 0.0305,
552
+ "step": 6300
553
+ },
554
+ {
555
+ "epoch": 11.68,
556
+ "learning_rate": 2.1663999999999998e-05,
557
+ "loss": 0.0358,
558
+ "step": 6400
559
+ },
560
+ {
561
+ "epoch": 11.86,
562
+ "learning_rate": 2.1065e-05,
563
+ "loss": 0.034,
564
+ "step": 6500
565
+ },
566
+ {
567
+ "epoch": 12.0,
568
+ "eval_accuracy_background": NaN,
569
+ "eval_accuracy_crack": 0.4161940191501904,
570
+ "eval_iou_background": 0.0,
571
+ "eval_iou_crack": 0.41221175296407997,
572
+ "eval_loss": 0.0623166486620903,
573
+ "eval_mean_accuracy": 0.4161940191501904,
574
+ "eval_mean_iou": 0.20610587648203998,
575
+ "eval_overall_accuracy": 0.4161940191501904,
576
+ "eval_runtime": 381.2562,
577
+ "eval_samples_per_second": 0.509,
578
+ "eval_steps_per_second": 0.254,
579
+ "step": 6576
580
+ },
581
+ {
582
+ "epoch": 12.04,
583
+ "learning_rate": 2.0465999999999997e-05,
584
+ "loss": 0.0294,
585
+ "step": 6600
586
+ },
587
+ {
588
+ "epoch": 12.23,
589
+ "learning_rate": 1.9866999999999998e-05,
590
+ "loss": 0.0294,
591
+ "step": 6700
592
+ },
593
+ {
594
+ "epoch": 12.41,
595
+ "learning_rate": 1.9267999999999996e-05,
596
+ "loss": 0.0452,
597
+ "step": 6800
598
+ },
599
+ {
600
+ "epoch": 12.59,
601
+ "learning_rate": 1.8669e-05,
602
+ "loss": 0.0347,
603
+ "step": 6900
604
+ },
605
+ {
606
+ "epoch": 12.77,
607
+ "learning_rate": 1.807e-05,
608
+ "loss": 0.0279,
609
+ "step": 7000
610
+ },
611
+ {
612
+ "epoch": 12.96,
613
+ "learning_rate": 1.7471e-05,
614
+ "loss": 0.0301,
615
+ "step": 7100
616
+ },
617
+ {
618
+ "epoch": 13.0,
619
+ "eval_accuracy_background": NaN,
620
+ "eval_accuracy_crack": 0.3669474007041917,
621
+ "eval_iou_background": 0.0,
622
+ "eval_iou_crack": 0.3664464100154627,
623
+ "eval_loss": 0.08311110734939575,
624
+ "eval_mean_accuracy": 0.3669474007041917,
625
+ "eval_mean_iou": 0.18322320500773134,
626
+ "eval_overall_accuracy": 0.3669474007041917,
627
+ "eval_runtime": 380.1764,
628
+ "eval_samples_per_second": 0.51,
629
+ "eval_steps_per_second": 0.255,
630
+ "step": 7124
631
+ },
632
+ {
633
+ "epoch": 13.14,
634
+ "learning_rate": 1.6872e-05,
635
+ "loss": 0.0321,
636
+ "step": 7200
637
+ },
638
+ {
639
+ "epoch": 13.32,
640
+ "learning_rate": 1.6272999999999998e-05,
641
+ "loss": 0.0293,
642
+ "step": 7300
643
+ },
644
+ {
645
+ "epoch": 13.5,
646
+ "learning_rate": 1.5674e-05,
647
+ "loss": 0.0342,
648
+ "step": 7400
649
+ },
650
+ {
651
+ "epoch": 13.69,
652
+ "learning_rate": 1.5075000000000002e-05,
653
+ "loss": 0.0305,
654
+ "step": 7500
655
+ },
656
+ {
657
+ "epoch": 13.87,
658
+ "learning_rate": 1.4476e-05,
659
+ "loss": 0.034,
660
+ "step": 7600
661
+ },
662
+ {
663
+ "epoch": 14.0,
664
+ "eval_accuracy_background": NaN,
665
+ "eval_accuracy_crack": 0.4119352127710421,
666
+ "eval_iou_background": 0.0,
667
+ "eval_iou_crack": 0.4118392523290055,
668
+ "eval_loss": 0.06355985999107361,
669
+ "eval_mean_accuracy": 0.4119352127710421,
670
+ "eval_mean_iou": 0.20591962616450274,
671
+ "eval_overall_accuracy": 0.4119352127710421,
672
+ "eval_runtime": 381.445,
673
+ "eval_samples_per_second": 0.509,
674
+ "eval_steps_per_second": 0.254,
675
+ "step": 7672
676
+ },
677
+ {
678
+ "epoch": 14.05,
679
+ "learning_rate": 1.3877e-05,
680
+ "loss": 0.0331,
681
+ "step": 7700
682
+ },
683
+ {
684
+ "epoch": 14.23,
685
+ "learning_rate": 1.3277999999999999e-05,
686
+ "loss": 0.0312,
687
+ "step": 7800
688
+ },
689
+ {
690
+ "epoch": 14.42,
691
+ "learning_rate": 1.2678999999999998e-05,
692
+ "loss": 0.0295,
693
+ "step": 7900
694
+ },
695
+ {
696
+ "epoch": 14.6,
697
+ "learning_rate": 1.2079999999999998e-05,
698
+ "loss": 0.0346,
699
+ "step": 8000
700
+ },
701
+ {
702
+ "epoch": 14.78,
703
+ "learning_rate": 1.1480999999999997e-05,
704
+ "loss": 0.0302,
705
+ "step": 8100
706
+ },
707
+ {
708
+ "epoch": 14.96,
709
+ "learning_rate": 1.0882000000000004e-05,
710
+ "loss": 0.0303,
711
+ "step": 8200
712
+ },
713
+ {
714
+ "epoch": 15.0,
715
+ "eval_accuracy_background": NaN,
716
+ "eval_accuracy_crack": 0.3864062724042889,
717
+ "eval_iou_background": 0.0,
718
+ "eval_iou_crack": 0.38623359121586986,
719
+ "eval_loss": 0.07051627337932587,
720
+ "eval_mean_accuracy": 0.3864062724042889,
721
+ "eval_mean_iou": 0.19311679560793493,
722
+ "eval_overall_accuracy": 0.3864062724042889,
723
+ "eval_runtime": 379.3942,
724
+ "eval_samples_per_second": 0.511,
725
+ "eval_steps_per_second": 0.256,
726
+ "step": 8220
727
+ },
728
+ {
729
+ "epoch": 15.15,
730
+ "learning_rate": 1.0283000000000003e-05,
731
+ "loss": 0.0254,
732
+ "step": 8300
733
+ },
734
+ {
735
+ "epoch": 15.33,
736
+ "learning_rate": 9.684000000000002e-06,
737
+ "loss": 0.0325,
738
+ "step": 8400
739
+ },
740
+ {
741
+ "epoch": 15.51,
742
+ "learning_rate": 9.085000000000002e-06,
743
+ "loss": 0.0377,
744
+ "step": 8500
745
+ },
746
+ {
747
+ "epoch": 15.69,
748
+ "learning_rate": 8.486000000000001e-06,
749
+ "loss": 0.0246,
750
+ "step": 8600
751
+ },
752
+ {
753
+ "epoch": 15.88,
754
+ "learning_rate": 7.887000000000001e-06,
755
+ "loss": 0.0338,
756
+ "step": 8700
757
+ },
758
+ {
759
+ "epoch": 16.0,
760
+ "eval_accuracy_background": NaN,
761
+ "eval_accuracy_crack": 0.4206479917791196,
762
+ "eval_iou_background": 0.0,
763
+ "eval_iou_crack": 0.42018106442487874,
764
+ "eval_loss": 0.06845217943191528,
765
+ "eval_mean_accuracy": 0.4206479917791196,
766
+ "eval_mean_iou": 0.21009053221243937,
767
+ "eval_overall_accuracy": 0.4206479917791196,
768
+ "eval_runtime": 380.9614,
769
+ "eval_samples_per_second": 0.509,
770
+ "eval_steps_per_second": 0.255,
771
+ "step": 8768
772
+ },
773
+ {
774
+ "epoch": 16.06,
775
+ "learning_rate": 7.2879999999999995e-06,
776
+ "loss": 0.0206,
777
+ "step": 8800
778
+ },
779
+ {
780
+ "epoch": 16.24,
781
+ "learning_rate": 6.688999999999999e-06,
782
+ "loss": 0.0259,
783
+ "step": 8900
784
+ },
785
+ {
786
+ "epoch": 16.42,
787
+ "learning_rate": 6.0899999999999984e-06,
788
+ "loss": 0.0333,
789
+ "step": 9000
790
+ },
791
+ {
792
+ "epoch": 16.61,
793
+ "learning_rate": 5.490999999999998e-06,
794
+ "loss": 0.0326,
795
+ "step": 9100
796
+ },
797
+ {
798
+ "epoch": 16.79,
799
+ "learning_rate": 4.891999999999997e-06,
800
+ "loss": 0.0325,
801
+ "step": 9200
802
+ },
803
+ {
804
+ "epoch": 16.97,
805
+ "learning_rate": 4.292999999999997e-06,
806
+ "loss": 0.0229,
807
+ "step": 9300
808
+ },
809
+ {
810
+ "epoch": 17.0,
811
+ "eval_accuracy_background": NaN,
812
+ "eval_accuracy_crack": 0.4204358978444087,
813
+ "eval_iou_background": 0.0,
814
+ "eval_iou_crack": 0.4197487662985776,
815
+ "eval_loss": 0.07058978080749512,
816
+ "eval_mean_accuracy": 0.4204358978444087,
817
+ "eval_mean_iou": 0.2098743831492888,
818
+ "eval_overall_accuracy": 0.4204358978444087,
819
+ "eval_runtime": 383.066,
820
+ "eval_samples_per_second": 0.506,
821
+ "eval_steps_per_second": 0.253,
822
+ "step": 9316
823
+ },
824
+ {
825
+ "epoch": 17.15,
826
+ "learning_rate": 3.694000000000003e-06,
827
+ "loss": 0.0273,
828
+ "step": 9400
829
+ },
830
+ {
831
+ "epoch": 17.34,
832
+ "learning_rate": 3.0950000000000026e-06,
833
+ "loss": 0.0381,
834
+ "step": 9500
835
+ },
836
+ {
837
+ "epoch": 17.52,
838
+ "learning_rate": 2.496000000000002e-06,
839
+ "loss": 0.0262,
840
+ "step": 9600
841
+ },
842
+ {
843
+ "epoch": 17.7,
844
+ "learning_rate": 1.8970000000000013e-06,
845
+ "loss": 0.0242,
846
+ "step": 9700
847
+ },
848
+ {
849
+ "epoch": 17.88,
850
+ "learning_rate": 1.298000000000001e-06,
851
+ "loss": 0.0337,
852
+ "step": 9800
853
+ },
854
+ {
855
+ "epoch": 18.0,
856
+ "eval_accuracy_background": NaN,
857
+ "eval_accuracy_crack": 0.39679987095129604,
858
+ "eval_iou_background": 0.0,
859
+ "eval_iou_crack": 0.39646904318930265,
860
+ "eval_loss": 0.07418648153543472,
861
+ "eval_mean_accuracy": 0.39679987095129604,
862
+ "eval_mean_iou": 0.19823452159465133,
863
+ "eval_overall_accuracy": 0.39679987095129604,
864
+ "eval_runtime": 380.0658,
865
+ "eval_samples_per_second": 0.51,
866
+ "eval_steps_per_second": 0.255,
867
+ "step": 9864
868
+ },
869
+ {
870
+ "epoch": 18.07,
871
+ "learning_rate": 6.990000000000005e-07,
872
+ "loss": 0.0243,
873
+ "step": 9900
874
+ },
875
+ {
876
+ "epoch": 18.25,
877
+ "learning_rate": 1e-07,
878
+ "loss": 0.0257,
879
+ "step": 10000
880
+ },
881
+ {
882
+ "epoch": 18.25,
883
+ "eval_accuracy_background": NaN,
884
+ "eval_accuracy_crack": 0.39388432615865027,
885
+ "eval_iou_background": 0.0,
886
+ "eval_iou_crack": 0.3935762911195175,
887
+ "eval_loss": 0.07488969713449478,
888
+ "eval_mean_accuracy": 0.39388432615865027,
889
+ "eval_mean_iou": 0.19678814555975874,
890
+ "eval_overall_accuracy": 0.39388432615865027,
891
+ "eval_runtime": 375.6599,
892
+ "eval_samples_per_second": 0.516,
893
+ "eval_steps_per_second": 0.258,
894
+ "step": 10000
895
+ },
896
+ {
897
+ "epoch": 18.25,
898
+ "step": 10000,
899
+ "total_flos": 7.976189631720849e+18,
900
+ "train_loss": 0.039843981289863585,
901
+ "train_runtime": 169492.4367,
902
+ "train_samples_per_second": 0.118,
903
+ "train_steps_per_second": 0.059
904
+ }
905
+ ],
906
+ "logging_steps": 100,
907
+ "max_steps": 10000,
908
+ "num_train_epochs": 19,
909
+ "save_steps": 500,
910
+ "total_flos": 7.976189631720849e+18,
911
+ "trial_name": null,
912
+ "trial_params": null
913
+ }