heroza commited on
Commit
c5e62fe
·
verified ·
1 Parent(s): 4a67c80

End of training

Browse files
README.md CHANGED
@@ -2,6 +2,8 @@
2
  license: other
3
  base_model: nvidia/mit-b0
4
  tags:
 
 
5
  - generated_from_trainer
6
  model-index:
7
  - name: segformer-finetuned-biofilm_MRCNNv1_validation
@@ -13,7 +15,7 @@ should probably proofread and complete it, then remove this comment. -->
13
 
14
  # segformer-finetuned-biofilm_MRCNNv1_validation
15
 
16
- This model is a fine-tuned version of [nvidia/mit-b0](https://huggingface.co/nvidia/mit-b0) on an unknown dataset.
17
  It achieves the following results on the evaluation set:
18
  - Loss: 0.0667
19
  - Mean Iou: 0.4894
 
2
  license: other
3
  base_model: nvidia/mit-b0
4
  tags:
5
+ - image-segmentation
6
+ - vision
7
  - generated_from_trainer
8
  model-index:
9
  - name: segformer-finetuned-biofilm_MRCNNv1_validation
 
15
 
16
  # segformer-finetuned-biofilm_MRCNNv1_validation
17
 
18
+ This model is a fine-tuned version of [nvidia/mit-b0](https://huggingface.co/nvidia/mit-b0) on the heroza/biofilm_MRCNNv1_validation dataset.
19
  It achieves the following results on the evaluation set:
20
  - Loss: 0.0667
21
  - Mean Iou: 0.4894
all_results.json ADDED
@@ -0,0 +1,18 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 28.49,
3
+ "eval_accuracy_background": 0.978794326771032,
4
+ "eval_accuracy_biofilm": NaN,
5
+ "eval_iou_background": 0.978794326771032,
6
+ "eval_iou_biofilm": 0.0,
7
+ "eval_loss": 0.06670047342777252,
8
+ "eval_mean_accuracy": 0.978794326771032,
9
+ "eval_mean_iou": 0.489397163385516,
10
+ "eval_overall_accuracy": 0.978794326771032,
11
+ "eval_runtime": 165.443,
12
+ "eval_samples_per_second": 7.694,
13
+ "eval_steps_per_second": 0.967,
14
+ "train_loss": 0.037617195892333984,
15
+ "train_runtime": 12861.3706,
16
+ "train_samples_per_second": 6.22,
17
+ "train_steps_per_second": 0.778
18
+ }
eval_results.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 28.49,
3
+ "eval_accuracy_background": 0.978794326771032,
4
+ "eval_accuracy_biofilm": NaN,
5
+ "eval_iou_background": 0.978794326771032,
6
+ "eval_iou_biofilm": 0.0,
7
+ "eval_loss": 0.06670047342777252,
8
+ "eval_mean_accuracy": 0.978794326771032,
9
+ "eval_mean_iou": 0.489397163385516,
10
+ "eval_overall_accuracy": 0.978794326771032,
11
+ "eval_runtime": 165.443,
12
+ "eval_samples_per_second": 7.694,
13
+ "eval_steps_per_second": 0.967
14
+ }
runs/Jul09_14-35-54_gpu-1-130.local/events.out.tfevents.1720545193.gpu-1-130.local.109642.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7986e9c08f4d4c7a14a0c42bf91ea8e1244cb1560eaf641c2c277e6487f387f8
3
+ size 764
train_results.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 28.49,
3
+ "train_loss": 0.037617195892333984,
4
+ "train_runtime": 12861.3706,
5
+ "train_samples_per_second": 6.22,
6
+ "train_steps_per_second": 0.778
7
+ }
trainer_state.json ADDED
@@ -0,0 +1,1065 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 28.49002849002849,
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": 0.28,
13
+ "learning_rate": 5.9401e-05,
14
+ "loss": 0.3084,
15
+ "step": 100
16
+ },
17
+ {
18
+ "epoch": 0.57,
19
+ "learning_rate": 5.8802000000000004e-05,
20
+ "loss": 0.1422,
21
+ "step": 200
22
+ },
23
+ {
24
+ "epoch": 0.85,
25
+ "learning_rate": 5.8203e-05,
26
+ "loss": 0.0896,
27
+ "step": 300
28
+ },
29
+ {
30
+ "epoch": 1.0,
31
+ "eval_accuracy_background": 0.9894339098260804,
32
+ "eval_accuracy_biofilm": NaN,
33
+ "eval_iou_background": 0.9894339098260804,
34
+ "eval_iou_biofilm": 0.0,
35
+ "eval_loss": 0.04051243141293526,
36
+ "eval_mean_accuracy": 0.9894339098260804,
37
+ "eval_mean_iou": 0.4947169549130402,
38
+ "eval_overall_accuracy": 0.9894339098260804,
39
+ "eval_runtime": 172.9533,
40
+ "eval_samples_per_second": 7.36,
41
+ "eval_steps_per_second": 0.925,
42
+ "step": 351
43
+ },
44
+ {
45
+ "epoch": 1.14,
46
+ "learning_rate": 5.7604e-05,
47
+ "loss": 0.072,
48
+ "step": 400
49
+ },
50
+ {
51
+ "epoch": 1.42,
52
+ "learning_rate": 5.7005e-05,
53
+ "loss": 0.0677,
54
+ "step": 500
55
+ },
56
+ {
57
+ "epoch": 1.71,
58
+ "learning_rate": 5.6406e-05,
59
+ "loss": 0.0556,
60
+ "step": 600
61
+ },
62
+ {
63
+ "epoch": 1.99,
64
+ "learning_rate": 5.5806999999999996e-05,
65
+ "loss": 0.0556,
66
+ "step": 700
67
+ },
68
+ {
69
+ "epoch": 2.0,
70
+ "eval_accuracy_background": 0.984915374923779,
71
+ "eval_accuracy_biofilm": NaN,
72
+ "eval_iou_background": 0.984915374923779,
73
+ "eval_iou_biofilm": 0.0,
74
+ "eval_loss": 0.045945584774017334,
75
+ "eval_mean_accuracy": 0.984915374923779,
76
+ "eval_mean_iou": 0.4924576874618895,
77
+ "eval_overall_accuracy": 0.984915374923779,
78
+ "eval_runtime": 160.8979,
79
+ "eval_samples_per_second": 7.912,
80
+ "eval_steps_per_second": 0.994,
81
+ "step": 702
82
+ },
83
+ {
84
+ "epoch": 2.28,
85
+ "learning_rate": 5.5208000000000004e-05,
86
+ "loss": 0.0558,
87
+ "step": 800
88
+ },
89
+ {
90
+ "epoch": 2.56,
91
+ "learning_rate": 5.4609000000000005e-05,
92
+ "loss": 0.0527,
93
+ "step": 900
94
+ },
95
+ {
96
+ "epoch": 2.85,
97
+ "learning_rate": 5.401e-05,
98
+ "loss": 0.0532,
99
+ "step": 1000
100
+ },
101
+ {
102
+ "epoch": 3.0,
103
+ "eval_accuracy_background": 0.9862937449924256,
104
+ "eval_accuracy_biofilm": NaN,
105
+ "eval_iou_background": 0.9862937449924256,
106
+ "eval_iou_biofilm": 0.0,
107
+ "eval_loss": 0.03520960360765457,
108
+ "eval_mean_accuracy": 0.9862937449924256,
109
+ "eval_mean_iou": 0.4931468724962128,
110
+ "eval_overall_accuracy": 0.9862937449924256,
111
+ "eval_runtime": 161.2637,
112
+ "eval_samples_per_second": 7.894,
113
+ "eval_steps_per_second": 0.992,
114
+ "step": 1053
115
+ },
116
+ {
117
+ "epoch": 3.13,
118
+ "learning_rate": 5.3411e-05,
119
+ "loss": 0.0497,
120
+ "step": 1100
121
+ },
122
+ {
123
+ "epoch": 3.42,
124
+ "learning_rate": 5.2812e-05,
125
+ "loss": 0.0514,
126
+ "step": 1200
127
+ },
128
+ {
129
+ "epoch": 3.7,
130
+ "learning_rate": 5.2213e-05,
131
+ "loss": 0.0422,
132
+ "step": 1300
133
+ },
134
+ {
135
+ "epoch": 3.99,
136
+ "learning_rate": 5.1614000000000004e-05,
137
+ "loss": 0.0473,
138
+ "step": 1400
139
+ },
140
+ {
141
+ "epoch": 4.0,
142
+ "eval_accuracy_background": 0.9871781412470487,
143
+ "eval_accuracy_biofilm": NaN,
144
+ "eval_iou_background": 0.9871781412470487,
145
+ "eval_iou_biofilm": 0.0,
146
+ "eval_loss": 0.03179320693016052,
147
+ "eval_mean_accuracy": 0.9871781412470487,
148
+ "eval_mean_iou": 0.49358907062352436,
149
+ "eval_overall_accuracy": 0.9871781412470487,
150
+ "eval_runtime": 160.2722,
151
+ "eval_samples_per_second": 7.943,
152
+ "eval_steps_per_second": 0.998,
153
+ "step": 1404
154
+ },
155
+ {
156
+ "epoch": 4.27,
157
+ "learning_rate": 5.1015e-05,
158
+ "loss": 0.0451,
159
+ "step": 1500
160
+ },
161
+ {
162
+ "epoch": 4.56,
163
+ "learning_rate": 5.0416e-05,
164
+ "loss": 0.0376,
165
+ "step": 1600
166
+ },
167
+ {
168
+ "epoch": 4.84,
169
+ "learning_rate": 4.9817e-05,
170
+ "loss": 0.0387,
171
+ "step": 1700
172
+ },
173
+ {
174
+ "epoch": 5.0,
175
+ "eval_accuracy_background": 0.9856533519948681,
176
+ "eval_accuracy_biofilm": NaN,
177
+ "eval_iou_background": 0.9856533519948681,
178
+ "eval_iou_biofilm": 0.0,
179
+ "eval_loss": 0.03184612840414047,
180
+ "eval_mean_accuracy": 0.9856533519948681,
181
+ "eval_mean_iou": 0.49282667599743407,
182
+ "eval_overall_accuracy": 0.9856533519948681,
183
+ "eval_runtime": 162.9105,
184
+ "eval_samples_per_second": 7.814,
185
+ "eval_steps_per_second": 0.982,
186
+ "step": 1755
187
+ },
188
+ {
189
+ "epoch": 5.13,
190
+ "learning_rate": 4.9218e-05,
191
+ "loss": 0.0471,
192
+ "step": 1800
193
+ },
194
+ {
195
+ "epoch": 5.41,
196
+ "learning_rate": 4.8619e-05,
197
+ "loss": 0.0398,
198
+ "step": 1900
199
+ },
200
+ {
201
+ "epoch": 5.7,
202
+ "learning_rate": 4.8020000000000004e-05,
203
+ "loss": 0.0396,
204
+ "step": 2000
205
+ },
206
+ {
207
+ "epoch": 5.98,
208
+ "learning_rate": 4.7421000000000006e-05,
209
+ "loss": 0.0388,
210
+ "step": 2100
211
+ },
212
+ {
213
+ "epoch": 6.0,
214
+ "eval_accuracy_background": 0.9817061703189651,
215
+ "eval_accuracy_biofilm": NaN,
216
+ "eval_iou_background": 0.9817061703189651,
217
+ "eval_iou_biofilm": 0.0,
218
+ "eval_loss": 0.03944101929664612,
219
+ "eval_mean_accuracy": 0.9817061703189651,
220
+ "eval_mean_iou": 0.49085308515948256,
221
+ "eval_overall_accuracy": 0.9817061703189651,
222
+ "eval_runtime": 162.6848,
223
+ "eval_samples_per_second": 7.825,
224
+ "eval_steps_per_second": 0.983,
225
+ "step": 2106
226
+ },
227
+ {
228
+ "epoch": 6.27,
229
+ "learning_rate": 4.6822e-05,
230
+ "loss": 0.0361,
231
+ "step": 2200
232
+ },
233
+ {
234
+ "epoch": 6.55,
235
+ "learning_rate": 4.6223e-05,
236
+ "loss": 0.0423,
237
+ "step": 2300
238
+ },
239
+ {
240
+ "epoch": 6.84,
241
+ "learning_rate": 4.5624e-05,
242
+ "loss": 0.0344,
243
+ "step": 2400
244
+ },
245
+ {
246
+ "epoch": 7.0,
247
+ "eval_accuracy_background": 0.9811270539700527,
248
+ "eval_accuracy_biofilm": NaN,
249
+ "eval_iou_background": 0.9811270539700527,
250
+ "eval_iou_biofilm": 0.0,
251
+ "eval_loss": 0.043148938566446304,
252
+ "eval_mean_accuracy": 0.9811270539700527,
253
+ "eval_mean_iou": 0.49056352698502637,
254
+ "eval_overall_accuracy": 0.9811270539700527,
255
+ "eval_runtime": 160.6174,
256
+ "eval_samples_per_second": 7.926,
257
+ "eval_steps_per_second": 0.996,
258
+ "step": 2457
259
+ },
260
+ {
261
+ "epoch": 7.12,
262
+ "learning_rate": 4.5025000000000003e-05,
263
+ "loss": 0.0383,
264
+ "step": 2500
265
+ },
266
+ {
267
+ "epoch": 7.41,
268
+ "learning_rate": 4.4426000000000005e-05,
269
+ "loss": 0.0353,
270
+ "step": 2600
271
+ },
272
+ {
273
+ "epoch": 7.69,
274
+ "learning_rate": 4.3827e-05,
275
+ "loss": 0.041,
276
+ "step": 2700
277
+ },
278
+ {
279
+ "epoch": 7.98,
280
+ "learning_rate": 4.3228e-05,
281
+ "loss": 0.0409,
282
+ "step": 2800
283
+ },
284
+ {
285
+ "epoch": 8.0,
286
+ "eval_accuracy_background": 0.9844419901167134,
287
+ "eval_accuracy_biofilm": NaN,
288
+ "eval_iou_background": 0.9844419901167134,
289
+ "eval_iou_biofilm": 0.0,
290
+ "eval_loss": 0.034700486809015274,
291
+ "eval_mean_accuracy": 0.9844419901167134,
292
+ "eval_mean_iou": 0.4922209950583567,
293
+ "eval_overall_accuracy": 0.9844419901167134,
294
+ "eval_runtime": 160.7881,
295
+ "eval_samples_per_second": 7.917,
296
+ "eval_steps_per_second": 0.995,
297
+ "step": 2808
298
+ },
299
+ {
300
+ "epoch": 8.26,
301
+ "learning_rate": 4.2629e-05,
302
+ "loss": 0.0323,
303
+ "step": 2900
304
+ },
305
+ {
306
+ "epoch": 8.55,
307
+ "learning_rate": 4.203e-05,
308
+ "loss": 0.0406,
309
+ "step": 3000
310
+ },
311
+ {
312
+ "epoch": 8.83,
313
+ "learning_rate": 4.1431e-05,
314
+ "loss": 0.0322,
315
+ "step": 3100
316
+ },
317
+ {
318
+ "epoch": 9.0,
319
+ "eval_accuracy_background": 0.9819413501892381,
320
+ "eval_accuracy_biofilm": NaN,
321
+ "eval_iou_background": 0.9819413501892381,
322
+ "eval_iou_biofilm": 0.0,
323
+ "eval_loss": 0.04153432324528694,
324
+ "eval_mean_accuracy": 0.9819413501892381,
325
+ "eval_mean_iou": 0.49097067509461906,
326
+ "eval_overall_accuracy": 0.9819413501892381,
327
+ "eval_runtime": 160.6617,
328
+ "eval_samples_per_second": 7.923,
329
+ "eval_steps_per_second": 0.996,
330
+ "step": 3159
331
+ },
332
+ {
333
+ "epoch": 9.12,
334
+ "learning_rate": 4.0832e-05,
335
+ "loss": 0.0417,
336
+ "step": 3200
337
+ },
338
+ {
339
+ "epoch": 9.4,
340
+ "learning_rate": 4.0233e-05,
341
+ "loss": 0.0335,
342
+ "step": 3300
343
+ },
344
+ {
345
+ "epoch": 9.69,
346
+ "learning_rate": 3.9634e-05,
347
+ "loss": 0.0308,
348
+ "step": 3400
349
+ },
350
+ {
351
+ "epoch": 9.97,
352
+ "learning_rate": 3.9035e-05,
353
+ "loss": 0.0331,
354
+ "step": 3500
355
+ },
356
+ {
357
+ "epoch": 10.0,
358
+ "eval_accuracy_background": 0.9767117788684989,
359
+ "eval_accuracy_biofilm": NaN,
360
+ "eval_iou_background": 0.9767117788684989,
361
+ "eval_iou_biofilm": 0.0,
362
+ "eval_loss": 0.05577180162072182,
363
+ "eval_mean_accuracy": 0.9767117788684989,
364
+ "eval_mean_iou": 0.48835588943424946,
365
+ "eval_overall_accuracy": 0.9767117788684989,
366
+ "eval_runtime": 150.1821,
367
+ "eval_samples_per_second": 8.476,
368
+ "eval_steps_per_second": 1.065,
369
+ "step": 3510
370
+ },
371
+ {
372
+ "epoch": 10.26,
373
+ "learning_rate": 3.8436e-05,
374
+ "loss": 0.0365,
375
+ "step": 3600
376
+ },
377
+ {
378
+ "epoch": 10.54,
379
+ "learning_rate": 3.7837000000000004e-05,
380
+ "loss": 0.0342,
381
+ "step": 3700
382
+ },
383
+ {
384
+ "epoch": 10.83,
385
+ "learning_rate": 3.7238000000000005e-05,
386
+ "loss": 0.0337,
387
+ "step": 3800
388
+ },
389
+ {
390
+ "epoch": 11.0,
391
+ "eval_accuracy_background": 0.9846524558893199,
392
+ "eval_accuracy_biofilm": NaN,
393
+ "eval_iou_background": 0.9846524558893199,
394
+ "eval_iou_biofilm": 0.0,
395
+ "eval_loss": 0.04216180369257927,
396
+ "eval_mean_accuracy": 0.9846524558893199,
397
+ "eval_mean_iou": 0.49232622794465997,
398
+ "eval_overall_accuracy": 0.9846524558893199,
399
+ "eval_runtime": 162.4381,
400
+ "eval_samples_per_second": 7.837,
401
+ "eval_steps_per_second": 0.985,
402
+ "step": 3861
403
+ },
404
+ {
405
+ "epoch": 11.11,
406
+ "learning_rate": 3.6639e-05,
407
+ "loss": 0.0322,
408
+ "step": 3900
409
+ },
410
+ {
411
+ "epoch": 11.4,
412
+ "learning_rate": 3.604e-05,
413
+ "loss": 0.0291,
414
+ "step": 4000
415
+ },
416
+ {
417
+ "epoch": 11.68,
418
+ "learning_rate": 3.544100000000001e-05,
419
+ "loss": 0.0319,
420
+ "step": 4100
421
+ },
422
+ {
423
+ "epoch": 11.97,
424
+ "learning_rate": 3.4842e-05,
425
+ "loss": 0.0357,
426
+ "step": 4200
427
+ },
428
+ {
429
+ "epoch": 12.0,
430
+ "eval_accuracy_background": 0.9816231085527228,
431
+ "eval_accuracy_biofilm": NaN,
432
+ "eval_iou_background": 0.9816231085527228,
433
+ "eval_iou_biofilm": 0.0,
434
+ "eval_loss": 0.04213234409689903,
435
+ "eval_mean_accuracy": 0.9816231085527228,
436
+ "eval_mean_iou": 0.4908115542763614,
437
+ "eval_overall_accuracy": 0.9816231085527228,
438
+ "eval_runtime": 162.7573,
439
+ "eval_samples_per_second": 7.821,
440
+ "eval_steps_per_second": 0.983,
441
+ "step": 4212
442
+ },
443
+ {
444
+ "epoch": 12.25,
445
+ "learning_rate": 3.4243000000000004e-05,
446
+ "loss": 0.0296,
447
+ "step": 4300
448
+ },
449
+ {
450
+ "epoch": 12.54,
451
+ "learning_rate": 3.3644000000000005e-05,
452
+ "loss": 0.0316,
453
+ "step": 4400
454
+ },
455
+ {
456
+ "epoch": 12.82,
457
+ "learning_rate": 3.3045000000000006e-05,
458
+ "loss": 0.0306,
459
+ "step": 4500
460
+ },
461
+ {
462
+ "epoch": 13.0,
463
+ "eval_accuracy_background": 0.9826610410148492,
464
+ "eval_accuracy_biofilm": NaN,
465
+ "eval_iou_background": 0.9826610410148492,
466
+ "eval_iou_biofilm": 0.0,
467
+ "eval_loss": 0.03981821611523628,
468
+ "eval_mean_accuracy": 0.9826610410148492,
469
+ "eval_mean_iou": 0.4913305205074246,
470
+ "eval_overall_accuracy": 0.9826610410148492,
471
+ "eval_runtime": 165.8395,
472
+ "eval_samples_per_second": 7.676,
473
+ "eval_steps_per_second": 0.965,
474
+ "step": 4563
475
+ },
476
+ {
477
+ "epoch": 13.11,
478
+ "learning_rate": 3.2446e-05,
479
+ "loss": 0.0389,
480
+ "step": 4600
481
+ },
482
+ {
483
+ "epoch": 13.39,
484
+ "learning_rate": 3.1847e-05,
485
+ "loss": 0.0285,
486
+ "step": 4700
487
+ },
488
+ {
489
+ "epoch": 13.68,
490
+ "learning_rate": 3.1248e-05,
491
+ "loss": 0.0307,
492
+ "step": 4800
493
+ },
494
+ {
495
+ "epoch": 13.96,
496
+ "learning_rate": 3.0649000000000004e-05,
497
+ "loss": 0.0324,
498
+ "step": 4900
499
+ },
500
+ {
501
+ "epoch": 14.0,
502
+ "eval_accuracy_background": 0.9810405695381865,
503
+ "eval_accuracy_biofilm": NaN,
504
+ "eval_iou_background": 0.9810405695381865,
505
+ "eval_iou_biofilm": 0.0,
506
+ "eval_loss": 0.04880182817578316,
507
+ "eval_mean_accuracy": 0.9810405695381865,
508
+ "eval_mean_iou": 0.49052028476909326,
509
+ "eval_overall_accuracy": 0.9810405695381865,
510
+ "eval_runtime": 166.4755,
511
+ "eval_samples_per_second": 7.647,
512
+ "eval_steps_per_second": 0.961,
513
+ "step": 4914
514
+ },
515
+ {
516
+ "epoch": 14.25,
517
+ "learning_rate": 3.0050000000000002e-05,
518
+ "loss": 0.0315,
519
+ "step": 5000
520
+ },
521
+ {
522
+ "epoch": 14.53,
523
+ "learning_rate": 2.9451e-05,
524
+ "loss": 0.0331,
525
+ "step": 5100
526
+ },
527
+ {
528
+ "epoch": 14.81,
529
+ "learning_rate": 2.8851999999999998e-05,
530
+ "loss": 0.0293,
531
+ "step": 5200
532
+ },
533
+ {
534
+ "epoch": 15.0,
535
+ "eval_accuracy_background": 0.9835488963816809,
536
+ "eval_accuracy_biofilm": NaN,
537
+ "eval_iou_background": 0.9835488963816809,
538
+ "eval_iou_biofilm": 0.0,
539
+ "eval_loss": 0.040080878883600235,
540
+ "eval_mean_accuracy": 0.9835488963816809,
541
+ "eval_mean_iou": 0.49177444819084043,
542
+ "eval_overall_accuracy": 0.9835488963816809,
543
+ "eval_runtime": 168.172,
544
+ "eval_samples_per_second": 7.57,
545
+ "eval_steps_per_second": 0.951,
546
+ "step": 5265
547
+ },
548
+ {
549
+ "epoch": 15.1,
550
+ "learning_rate": 2.8253e-05,
551
+ "loss": 0.0283,
552
+ "step": 5300
553
+ },
554
+ {
555
+ "epoch": 15.38,
556
+ "learning_rate": 2.7653999999999996e-05,
557
+ "loss": 0.0358,
558
+ "step": 5400
559
+ },
560
+ {
561
+ "epoch": 15.67,
562
+ "learning_rate": 2.7054999999999998e-05,
563
+ "loss": 0.0276,
564
+ "step": 5500
565
+ },
566
+ {
567
+ "epoch": 15.95,
568
+ "learning_rate": 2.6455999999999995e-05,
569
+ "loss": 0.0243,
570
+ "step": 5600
571
+ },
572
+ {
573
+ "epoch": 16.0,
574
+ "eval_accuracy_background": 0.9788283678810232,
575
+ "eval_accuracy_biofilm": NaN,
576
+ "eval_iou_background": 0.9788283678810232,
577
+ "eval_iou_biofilm": 0.0,
578
+ "eval_loss": 0.049858421087265015,
579
+ "eval_mean_accuracy": 0.9788283678810232,
580
+ "eval_mean_iou": 0.4894141839405116,
581
+ "eval_overall_accuracy": 0.9788283678810232,
582
+ "eval_runtime": 161.6437,
583
+ "eval_samples_per_second": 7.875,
584
+ "eval_steps_per_second": 0.99,
585
+ "step": 5616
586
+ },
587
+ {
588
+ "epoch": 16.24,
589
+ "learning_rate": 2.5857e-05,
590
+ "loss": 0.0291,
591
+ "step": 5700
592
+ },
593
+ {
594
+ "epoch": 16.52,
595
+ "learning_rate": 2.5258e-05,
596
+ "loss": 0.0282,
597
+ "step": 5800
598
+ },
599
+ {
600
+ "epoch": 16.81,
601
+ "learning_rate": 2.4659e-05,
602
+ "loss": 0.0306,
603
+ "step": 5900
604
+ },
605
+ {
606
+ "epoch": 17.0,
607
+ "eval_accuracy_background": 0.9804685668999177,
608
+ "eval_accuracy_biofilm": NaN,
609
+ "eval_iou_background": 0.9804685668999177,
610
+ "eval_iou_biofilm": 0.0,
611
+ "eval_loss": 0.04945218190550804,
612
+ "eval_mean_accuracy": 0.9804685668999177,
613
+ "eval_mean_iou": 0.49023428344995884,
614
+ "eval_overall_accuracy": 0.9804685668999177,
615
+ "eval_runtime": 165.5559,
616
+ "eval_samples_per_second": 7.689,
617
+ "eval_steps_per_second": 0.966,
618
+ "step": 5967
619
+ },
620
+ {
621
+ "epoch": 17.09,
622
+ "learning_rate": 2.406e-05,
623
+ "loss": 0.0323,
624
+ "step": 6000
625
+ },
626
+ {
627
+ "epoch": 17.38,
628
+ "learning_rate": 2.3460999999999998e-05,
629
+ "loss": 0.027,
630
+ "step": 6100
631
+ },
632
+ {
633
+ "epoch": 17.66,
634
+ "learning_rate": 2.2862e-05,
635
+ "loss": 0.0275,
636
+ "step": 6200
637
+ },
638
+ {
639
+ "epoch": 17.95,
640
+ "learning_rate": 2.2263e-05,
641
+ "loss": 0.0267,
642
+ "step": 6300
643
+ },
644
+ {
645
+ "epoch": 18.0,
646
+ "eval_accuracy_background": 0.9813212612526138,
647
+ "eval_accuracy_biofilm": NaN,
648
+ "eval_iou_background": 0.9813212612526138,
649
+ "eval_iou_biofilm": 0.0,
650
+ "eval_loss": 0.04980457201600075,
651
+ "eval_mean_accuracy": 0.9813212612526138,
652
+ "eval_mean_iou": 0.4906606306263069,
653
+ "eval_overall_accuracy": 0.9813212612526138,
654
+ "eval_runtime": 159.3977,
655
+ "eval_samples_per_second": 7.986,
656
+ "eval_steps_per_second": 1.004,
657
+ "step": 6318
658
+ },
659
+ {
660
+ "epoch": 18.23,
661
+ "learning_rate": 2.1663999999999998e-05,
662
+ "loss": 0.0296,
663
+ "step": 6400
664
+ },
665
+ {
666
+ "epoch": 18.52,
667
+ "learning_rate": 2.1065e-05,
668
+ "loss": 0.0307,
669
+ "step": 6500
670
+ },
671
+ {
672
+ "epoch": 18.8,
673
+ "learning_rate": 2.0465999999999997e-05,
674
+ "loss": 0.0295,
675
+ "step": 6600
676
+ },
677
+ {
678
+ "epoch": 19.0,
679
+ "eval_accuracy_background": 0.9805983299943016,
680
+ "eval_accuracy_biofilm": NaN,
681
+ "eval_iou_background": 0.9805983299943016,
682
+ "eval_iou_biofilm": 0.0,
683
+ "eval_loss": 0.05660165846347809,
684
+ "eval_mean_accuracy": 0.9805983299943016,
685
+ "eval_mean_iou": 0.4902991649971508,
686
+ "eval_overall_accuracy": 0.9805983299943016,
687
+ "eval_runtime": 161.2791,
688
+ "eval_samples_per_second": 7.893,
689
+ "eval_steps_per_second": 0.992,
690
+ "step": 6669
691
+ },
692
+ {
693
+ "epoch": 19.09,
694
+ "learning_rate": 1.9866999999999998e-05,
695
+ "loss": 0.0276,
696
+ "step": 6700
697
+ },
698
+ {
699
+ "epoch": 19.37,
700
+ "learning_rate": 1.9267999999999996e-05,
701
+ "loss": 0.0299,
702
+ "step": 6800
703
+ },
704
+ {
705
+ "epoch": 19.66,
706
+ "learning_rate": 1.8669e-05,
707
+ "loss": 0.0289,
708
+ "step": 6900
709
+ },
710
+ {
711
+ "epoch": 19.94,
712
+ "learning_rate": 1.807e-05,
713
+ "loss": 0.0263,
714
+ "step": 7000
715
+ },
716
+ {
717
+ "epoch": 20.0,
718
+ "eval_accuracy_background": 0.9785523280761178,
719
+ "eval_accuracy_biofilm": NaN,
720
+ "eval_iou_background": 0.9785523280761178,
721
+ "eval_iou_biofilm": 0.0,
722
+ "eval_loss": 0.06579634547233582,
723
+ "eval_mean_accuracy": 0.9785523280761178,
724
+ "eval_mean_iou": 0.4892761640380589,
725
+ "eval_overall_accuracy": 0.9785523280761178,
726
+ "eval_runtime": 169.7419,
727
+ "eval_samples_per_second": 7.5,
728
+ "eval_steps_per_second": 0.943,
729
+ "step": 7020
730
+ },
731
+ {
732
+ "epoch": 20.23,
733
+ "learning_rate": 1.7471e-05,
734
+ "loss": 0.0259,
735
+ "step": 7100
736
+ },
737
+ {
738
+ "epoch": 20.51,
739
+ "learning_rate": 1.6872e-05,
740
+ "loss": 0.0282,
741
+ "step": 7200
742
+ },
743
+ {
744
+ "epoch": 20.8,
745
+ "learning_rate": 1.6272999999999998e-05,
746
+ "loss": 0.0319,
747
+ "step": 7300
748
+ },
749
+ {
750
+ "epoch": 21.0,
751
+ "eval_accuracy_background": 0.9769664311549116,
752
+ "eval_accuracy_biofilm": NaN,
753
+ "eval_iou_background": 0.9769664311549116,
754
+ "eval_iou_biofilm": 0.0,
755
+ "eval_loss": 0.0646413117647171,
756
+ "eval_mean_accuracy": 0.9769664311549116,
757
+ "eval_mean_iou": 0.4884832155774558,
758
+ "eval_overall_accuracy": 0.9769664311549116,
759
+ "eval_runtime": 170.9793,
760
+ "eval_samples_per_second": 7.445,
761
+ "eval_steps_per_second": 0.936,
762
+ "step": 7371
763
+ },
764
+ {
765
+ "epoch": 21.08,
766
+ "learning_rate": 1.5674e-05,
767
+ "loss": 0.0288,
768
+ "step": 7400
769
+ },
770
+ {
771
+ "epoch": 21.37,
772
+ "learning_rate": 1.5075000000000002e-05,
773
+ "loss": 0.031,
774
+ "step": 7500
775
+ },
776
+ {
777
+ "epoch": 21.65,
778
+ "learning_rate": 1.4476e-05,
779
+ "loss": 0.0238,
780
+ "step": 7600
781
+ },
782
+ {
783
+ "epoch": 21.94,
784
+ "learning_rate": 1.3877e-05,
785
+ "loss": 0.0236,
786
+ "step": 7700
787
+ },
788
+ {
789
+ "epoch": 22.0,
790
+ "eval_accuracy_background": 0.9793223251284174,
791
+ "eval_accuracy_biofilm": NaN,
792
+ "eval_iou_background": 0.9793223251284174,
793
+ "eval_iou_biofilm": 0.0,
794
+ "eval_loss": 0.06082683056592941,
795
+ "eval_mean_accuracy": 0.9793223251284174,
796
+ "eval_mean_iou": 0.4896611625642087,
797
+ "eval_overall_accuracy": 0.9793223251284174,
798
+ "eval_runtime": 168.1724,
799
+ "eval_samples_per_second": 7.57,
800
+ "eval_steps_per_second": 0.951,
801
+ "step": 7722
802
+ },
803
+ {
804
+ "epoch": 22.22,
805
+ "learning_rate": 1.3277999999999999e-05,
806
+ "loss": 0.025,
807
+ "step": 7800
808
+ },
809
+ {
810
+ "epoch": 22.51,
811
+ "learning_rate": 1.2678999999999998e-05,
812
+ "loss": 0.0262,
813
+ "step": 7900
814
+ },
815
+ {
816
+ "epoch": 22.79,
817
+ "learning_rate": 1.2079999999999998e-05,
818
+ "loss": 0.0249,
819
+ "step": 8000
820
+ },
821
+ {
822
+ "epoch": 23.0,
823
+ "eval_accuracy_background": 0.9794974907272019,
824
+ "eval_accuracy_biofilm": NaN,
825
+ "eval_iou_background": 0.9794974907272019,
826
+ "eval_iou_biofilm": 0.0,
827
+ "eval_loss": 0.05781866982579231,
828
+ "eval_mean_accuracy": 0.9794974907272019,
829
+ "eval_mean_iou": 0.48974874536360097,
830
+ "eval_overall_accuracy": 0.9794974907272019,
831
+ "eval_runtime": 159.7253,
832
+ "eval_samples_per_second": 7.97,
833
+ "eval_steps_per_second": 1.002,
834
+ "step": 8073
835
+ },
836
+ {
837
+ "epoch": 23.08,
838
+ "learning_rate": 1.1480999999999997e-05,
839
+ "loss": 0.0243,
840
+ "step": 8100
841
+ },
842
+ {
843
+ "epoch": 23.36,
844
+ "learning_rate": 1.0882000000000004e-05,
845
+ "loss": 0.0247,
846
+ "step": 8200
847
+ },
848
+ {
849
+ "epoch": 23.65,
850
+ "learning_rate": 1.0283000000000003e-05,
851
+ "loss": 0.0249,
852
+ "step": 8300
853
+ },
854
+ {
855
+ "epoch": 23.93,
856
+ "learning_rate": 9.684000000000002e-06,
857
+ "loss": 0.0242,
858
+ "step": 8400
859
+ },
860
+ {
861
+ "epoch": 24.0,
862
+ "eval_accuracy_background": 0.9804170976958486,
863
+ "eval_accuracy_biofilm": NaN,
864
+ "eval_iou_background": 0.9804170976958486,
865
+ "eval_iou_biofilm": 0.0,
866
+ "eval_loss": 0.05583757162094116,
867
+ "eval_mean_accuracy": 0.9804170976958486,
868
+ "eval_mean_iou": 0.4902085488479243,
869
+ "eval_overall_accuracy": 0.9804170976958486,
870
+ "eval_runtime": 174.0592,
871
+ "eval_samples_per_second": 7.314,
872
+ "eval_steps_per_second": 0.919,
873
+ "step": 8424
874
+ },
875
+ {
876
+ "epoch": 24.22,
877
+ "learning_rate": 9.085000000000002e-06,
878
+ "loss": 0.0227,
879
+ "step": 8500
880
+ },
881
+ {
882
+ "epoch": 24.5,
883
+ "learning_rate": 8.486000000000001e-06,
884
+ "loss": 0.0278,
885
+ "step": 8600
886
+ },
887
+ {
888
+ "epoch": 24.79,
889
+ "learning_rate": 7.887000000000001e-06,
890
+ "loss": 0.0264,
891
+ "step": 8700
892
+ },
893
+ {
894
+ "epoch": 25.0,
895
+ "eval_accuracy_background": 0.979809171978494,
896
+ "eval_accuracy_biofilm": NaN,
897
+ "eval_iou_background": 0.979809171978494,
898
+ "eval_iou_biofilm": 0.0,
899
+ "eval_loss": 0.05787277594208717,
900
+ "eval_mean_accuracy": 0.979809171978494,
901
+ "eval_mean_iou": 0.489904585989247,
902
+ "eval_overall_accuracy": 0.979809171978494,
903
+ "eval_runtime": 161.6401,
904
+ "eval_samples_per_second": 7.876,
905
+ "eval_steps_per_second": 0.99,
906
+ "step": 8775
907
+ },
908
+ {
909
+ "epoch": 25.07,
910
+ "learning_rate": 7.2879999999999995e-06,
911
+ "loss": 0.0247,
912
+ "step": 8800
913
+ },
914
+ {
915
+ "epoch": 25.36,
916
+ "learning_rate": 6.688999999999999e-06,
917
+ "loss": 0.0251,
918
+ "step": 8900
919
+ },
920
+ {
921
+ "epoch": 25.64,
922
+ "learning_rate": 6.0899999999999984e-06,
923
+ "loss": 0.025,
924
+ "step": 9000
925
+ },
926
+ {
927
+ "epoch": 25.93,
928
+ "learning_rate": 5.490999999999998e-06,
929
+ "loss": 0.0235,
930
+ "step": 9100
931
+ },
932
+ {
933
+ "epoch": 26.0,
934
+ "eval_accuracy_background": 0.9800799999218055,
935
+ "eval_accuracy_biofilm": NaN,
936
+ "eval_iou_background": 0.9800799999218055,
937
+ "eval_iou_biofilm": 0.0,
938
+ "eval_loss": 0.05818142741918564,
939
+ "eval_mean_accuracy": 0.9800799999218055,
940
+ "eval_mean_iou": 0.49003999996090275,
941
+ "eval_overall_accuracy": 0.9800799999218055,
942
+ "eval_runtime": 168.4683,
943
+ "eval_samples_per_second": 7.556,
944
+ "eval_steps_per_second": 0.95,
945
+ "step": 9126
946
+ },
947
+ {
948
+ "epoch": 26.21,
949
+ "learning_rate": 4.891999999999997e-06,
950
+ "loss": 0.0255,
951
+ "step": 9200
952
+ },
953
+ {
954
+ "epoch": 26.5,
955
+ "learning_rate": 4.292999999999997e-06,
956
+ "loss": 0.0225,
957
+ "step": 9300
958
+ },
959
+ {
960
+ "epoch": 26.78,
961
+ "learning_rate": 3.694000000000003e-06,
962
+ "loss": 0.0235,
963
+ "step": 9400
964
+ },
965
+ {
966
+ "epoch": 27.0,
967
+ "eval_accuracy_background": 0.9793572707763714,
968
+ "eval_accuracy_biofilm": NaN,
969
+ "eval_iou_background": 0.9793572707763714,
970
+ "eval_iou_biofilm": 0.0,
971
+ "eval_loss": 0.06092679873108864,
972
+ "eval_mean_accuracy": 0.9793572707763714,
973
+ "eval_mean_iou": 0.4896786353881857,
974
+ "eval_overall_accuracy": 0.9793572707763714,
975
+ "eval_runtime": 160.4565,
976
+ "eval_samples_per_second": 7.934,
977
+ "eval_steps_per_second": 0.997,
978
+ "step": 9477
979
+ },
980
+ {
981
+ "epoch": 27.07,
982
+ "learning_rate": 3.0950000000000026e-06,
983
+ "loss": 0.022,
984
+ "step": 9500
985
+ },
986
+ {
987
+ "epoch": 27.35,
988
+ "learning_rate": 2.496000000000002e-06,
989
+ "loss": 0.0252,
990
+ "step": 9600
991
+ },
992
+ {
993
+ "epoch": 27.64,
994
+ "learning_rate": 1.8970000000000013e-06,
995
+ "loss": 0.0245,
996
+ "step": 9700
997
+ },
998
+ {
999
+ "epoch": 27.92,
1000
+ "learning_rate": 1.298000000000001e-06,
1001
+ "loss": 0.0204,
1002
+ "step": 9800
1003
+ },
1004
+ {
1005
+ "epoch": 28.0,
1006
+ "eval_accuracy_background": 0.979131934797259,
1007
+ "eval_accuracy_biofilm": NaN,
1008
+ "eval_iou_background": 0.979131934797259,
1009
+ "eval_iou_biofilm": 0.0,
1010
+ "eval_loss": 0.06482358276844025,
1011
+ "eval_mean_accuracy": 0.979131934797259,
1012
+ "eval_mean_iou": 0.4895659673986295,
1013
+ "eval_overall_accuracy": 0.979131934797259,
1014
+ "eval_runtime": 170.4459,
1015
+ "eval_samples_per_second": 7.469,
1016
+ "eval_steps_per_second": 0.939,
1017
+ "step": 9828
1018
+ },
1019
+ {
1020
+ "epoch": 28.21,
1021
+ "learning_rate": 6.990000000000005e-07,
1022
+ "loss": 0.0203,
1023
+ "step": 9900
1024
+ },
1025
+ {
1026
+ "epoch": 28.49,
1027
+ "learning_rate": 1e-07,
1028
+ "loss": 0.023,
1029
+ "step": 10000
1030
+ },
1031
+ {
1032
+ "epoch": 28.49,
1033
+ "eval_accuracy_background": 0.978794326771032,
1034
+ "eval_accuracy_biofilm": NaN,
1035
+ "eval_iou_background": 0.978794326771032,
1036
+ "eval_iou_biofilm": 0.0,
1037
+ "eval_loss": 0.06670047342777252,
1038
+ "eval_mean_accuracy": 0.978794326771032,
1039
+ "eval_mean_iou": 0.489397163385516,
1040
+ "eval_overall_accuracy": 0.978794326771032,
1041
+ "eval_runtime": 163.9845,
1042
+ "eval_samples_per_second": 7.763,
1043
+ "eval_steps_per_second": 0.976,
1044
+ "step": 10000
1045
+ },
1046
+ {
1047
+ "epoch": 28.49,
1048
+ "step": 10000,
1049
+ "total_flos": 1.398800962440659e+18,
1050
+ "train_loss": 0.037617195892333984,
1051
+ "train_runtime": 12861.3706,
1052
+ "train_samples_per_second": 6.22,
1053
+ "train_steps_per_second": 0.778
1054
+ }
1055
+ ],
1056
+ "logging_steps": 100,
1057
+ "max_steps": 10000,
1058
+ "num_input_tokens_seen": 0,
1059
+ "num_train_epochs": 29,
1060
+ "save_steps": 500,
1061
+ "total_flos": 1.398800962440659e+18,
1062
+ "train_batch_size": 8,
1063
+ "trial_name": null,
1064
+ "trial_params": null
1065
+ }