heroza commited on
Commit
43ef69b
·
verified ·
1 Parent(s): 09b1791

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_concat
@@ -13,7 +15,7 @@ should probably proofread and complete it, then remove this comment. -->
13
 
14
  # segformer-finetuned-biofilm_MRCNNv1_concat
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: 0.0530
19
  - Mean Iou: 0.8677
 
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_concat
 
15
 
16
  # segformer-finetuned-biofilm_MRCNNv1_concat
17
 
18
+ This model is a fine-tuned version of [nvidia/mit-b0](https://huggingface.co/nvidia/mit-b0) on the heroza/biofilm_MRCNNv1_concat dataset.
19
  It achieves the following results on the evaluation set:
20
  - Loss: 0.0530
21
  - Mean Iou: 0.8677
all_results.json ADDED
@@ -0,0 +1,18 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 23.09,
3
+ "eval_accuracy_background": 0.9820080627772998,
4
+ "eval_accuracy_biofilm": 0.973973103250941,
5
+ "eval_iou_background": 0.9804257765494778,
6
+ "eval_iou_biofilm": 0.7549272284294235,
7
+ "eval_loss": 0.05298486724495888,
8
+ "eval_mean_accuracy": 0.9779905830141205,
9
+ "eval_mean_iou": 0.8676765024894506,
10
+ "eval_overall_accuracy": 0.9815389213004198,
11
+ "eval_runtime": 46.249,
12
+ "eval_samples_per_second": 13.233,
13
+ "eval_steps_per_second": 1.665,
14
+ "train_loss": 0.03242814379930496,
15
+ "train_runtime": 14425.5946,
16
+ "train_samples_per_second": 5.546,
17
+ "train_steps_per_second": 0.693
18
+ }
eval_results.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 23.09,
3
+ "eval_accuracy_background": 0.9820080627772998,
4
+ "eval_accuracy_biofilm": 0.973973103250941,
5
+ "eval_iou_background": 0.9804257765494778,
6
+ "eval_iou_biofilm": 0.7549272284294235,
7
+ "eval_loss": 0.05298486724495888,
8
+ "eval_mean_accuracy": 0.9779905830141205,
9
+ "eval_mean_iou": 0.8676765024894506,
10
+ "eval_overall_accuracy": 0.9815389213004198,
11
+ "eval_runtime": 46.249,
12
+ "eval_samples_per_second": 13.233,
13
+ "eval_steps_per_second": 1.665
14
+ }
runs/Jul10_10-42-15_gpu-1-130.local/events.out.tfevents.1720619145.gpu-1-130.local.221659.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3c05c0f880f45b17f04462cb0d845bba8e6a22d83f6688166b658ea83c699847
3
+ size 764
train_results.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 23.09,
3
+ "train_loss": 0.03242814379930496,
4
+ "train_runtime": 14425.5946,
5
+ "train_samples_per_second": 5.546,
6
+ "train_steps_per_second": 0.693
7
+ }
trainer_state.json ADDED
@@ -0,0 +1,990 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 23.094688221709006,
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": 0.23,
13
+ "learning_rate": 5.9401e-05,
14
+ "loss": 0.3181,
15
+ "step": 100
16
+ },
17
+ {
18
+ "epoch": 0.46,
19
+ "learning_rate": 5.8802000000000004e-05,
20
+ "loss": 0.149,
21
+ "step": 200
22
+ },
23
+ {
24
+ "epoch": 0.69,
25
+ "learning_rate": 5.8203e-05,
26
+ "loss": 0.0917,
27
+ "step": 300
28
+ },
29
+ {
30
+ "epoch": 0.92,
31
+ "learning_rate": 5.7604e-05,
32
+ "loss": 0.0759,
33
+ "step": 400
34
+ },
35
+ {
36
+ "epoch": 1.0,
37
+ "eval_accuracy_background": 0.9828615415666171,
38
+ "eval_accuracy_biofilm": 0.9650712037505778,
39
+ "eval_iou_background": 0.980737395917761,
40
+ "eval_iou_biofilm": 0.7560937377942886,
41
+ "eval_loss": 0.06491132825613022,
42
+ "eval_mean_accuracy": 0.9739663726585974,
43
+ "eval_mean_iou": 0.8684155668560247,
44
+ "eval_overall_accuracy": 0.981822807602889,
45
+ "eval_runtime": 59.9845,
46
+ "eval_samples_per_second": 10.203,
47
+ "eval_steps_per_second": 1.284,
48
+ "step": 433
49
+ },
50
+ {
51
+ "epoch": 1.15,
52
+ "learning_rate": 5.7005e-05,
53
+ "loss": 0.0566,
54
+ "step": 500
55
+ },
56
+ {
57
+ "epoch": 1.39,
58
+ "learning_rate": 5.6406e-05,
59
+ "loss": 0.0504,
60
+ "step": 600
61
+ },
62
+ {
63
+ "epoch": 1.62,
64
+ "learning_rate": 5.5806999999999996e-05,
65
+ "loss": 0.0468,
66
+ "step": 700
67
+ },
68
+ {
69
+ "epoch": 1.85,
70
+ "learning_rate": 5.5208000000000004e-05,
71
+ "loss": 0.0435,
72
+ "step": 800
73
+ },
74
+ {
75
+ "epoch": 2.0,
76
+ "eval_accuracy_background": 0.9886900685144177,
77
+ "eval_accuracy_biofilm": 0.9344582131930536,
78
+ "eval_iou_background": 0.9846881804227252,
79
+ "eval_iou_biofilm": 0.7903099298609719,
80
+ "eval_loss": 0.03914250433444977,
81
+ "eval_mean_accuracy": 0.9615741408537356,
82
+ "eval_mean_iou": 0.8874990551418486,
83
+ "eval_overall_accuracy": 0.9855236041886248,
84
+ "eval_runtime": 56.557,
85
+ "eval_samples_per_second": 10.821,
86
+ "eval_steps_per_second": 1.361,
87
+ "step": 866
88
+ },
89
+ {
90
+ "epoch": 2.08,
91
+ "learning_rate": 5.4609000000000005e-05,
92
+ "loss": 0.0447,
93
+ "step": 900
94
+ },
95
+ {
96
+ "epoch": 2.31,
97
+ "learning_rate": 5.401e-05,
98
+ "loss": 0.0458,
99
+ "step": 1000
100
+ },
101
+ {
102
+ "epoch": 2.54,
103
+ "learning_rate": 5.3411e-05,
104
+ "loss": 0.0348,
105
+ "step": 1100
106
+ },
107
+ {
108
+ "epoch": 2.77,
109
+ "learning_rate": 5.2812e-05,
110
+ "loss": 0.0407,
111
+ "step": 1200
112
+ },
113
+ {
114
+ "epoch": 3.0,
115
+ "eval_accuracy_background": 0.9887603228091695,
116
+ "eval_accuracy_biofilm": 0.9503892545066361,
117
+ "eval_iou_background": 0.9857279627494347,
118
+ "eval_iou_biofilm": 0.8045544051549277,
119
+ "eval_loss": 0.035304486751556396,
120
+ "eval_mean_accuracy": 0.9695747886579028,
121
+ "eval_mean_iou": 0.8951411839521812,
122
+ "eval_overall_accuracy": 0.9865199307332844,
123
+ "eval_runtime": 54.7086,
124
+ "eval_samples_per_second": 11.187,
125
+ "eval_steps_per_second": 1.407,
126
+ "step": 1299
127
+ },
128
+ {
129
+ "epoch": 3.0,
130
+ "learning_rate": 5.2213e-05,
131
+ "loss": 0.0336,
132
+ "step": 1300
133
+ },
134
+ {
135
+ "epoch": 3.23,
136
+ "learning_rate": 5.1614000000000004e-05,
137
+ "loss": 0.0313,
138
+ "step": 1400
139
+ },
140
+ {
141
+ "epoch": 3.46,
142
+ "learning_rate": 5.1015e-05,
143
+ "loss": 0.0338,
144
+ "step": 1500
145
+ },
146
+ {
147
+ "epoch": 3.7,
148
+ "learning_rate": 5.0416e-05,
149
+ "loss": 0.0375,
150
+ "step": 1600
151
+ },
152
+ {
153
+ "epoch": 3.93,
154
+ "learning_rate": 4.9817e-05,
155
+ "loss": 0.0372,
156
+ "step": 1700
157
+ },
158
+ {
159
+ "epoch": 4.0,
160
+ "eval_accuracy_background": 0.9832945045568867,
161
+ "eval_accuracy_biofilm": 0.9788046133867453,
162
+ "eval_iou_background": 0.9820038724017046,
163
+ "eval_iou_biofilm": 0.7710713706092136,
164
+ "eval_loss": 0.048879224807024,
165
+ "eval_mean_accuracy": 0.981049558971816,
166
+ "eval_mean_iou": 0.8765376215054591,
167
+ "eval_overall_accuracy": 0.9830323508811052,
168
+ "eval_runtime": 53.641,
169
+ "eval_samples_per_second": 11.409,
170
+ "eval_steps_per_second": 1.435,
171
+ "step": 1732
172
+ },
173
+ {
174
+ "epoch": 4.16,
175
+ "learning_rate": 4.9218e-05,
176
+ "loss": 0.0362,
177
+ "step": 1800
178
+ },
179
+ {
180
+ "epoch": 4.39,
181
+ "learning_rate": 4.8619e-05,
182
+ "loss": 0.0329,
183
+ "step": 1900
184
+ },
185
+ {
186
+ "epoch": 4.62,
187
+ "learning_rate": 4.8020000000000004e-05,
188
+ "loss": 0.0321,
189
+ "step": 2000
190
+ },
191
+ {
192
+ "epoch": 4.85,
193
+ "learning_rate": 4.7421000000000006e-05,
194
+ "loss": 0.0378,
195
+ "step": 2100
196
+ },
197
+ {
198
+ "epoch": 5.0,
199
+ "eval_accuracy_background": 0.9919392085185956,
200
+ "eval_accuracy_biofilm": 0.9229355314419034,
201
+ "eval_iou_background": 0.9872216553975541,
202
+ "eval_iou_biofilm": 0.8167600506475522,
203
+ "eval_loss": 0.03106585703790188,
204
+ "eval_mean_accuracy": 0.9574373699802495,
205
+ "eval_mean_iou": 0.9019908530225531,
206
+ "eval_overall_accuracy": 0.9879102539398317,
207
+ "eval_runtime": 53.6924,
208
+ "eval_samples_per_second": 11.398,
209
+ "eval_steps_per_second": 1.434,
210
+ "step": 2165
211
+ },
212
+ {
213
+ "epoch": 5.08,
214
+ "learning_rate": 4.6822e-05,
215
+ "loss": 0.0306,
216
+ "step": 2200
217
+ },
218
+ {
219
+ "epoch": 5.31,
220
+ "learning_rate": 4.6223e-05,
221
+ "loss": 0.028,
222
+ "step": 2300
223
+ },
224
+ {
225
+ "epoch": 5.54,
226
+ "learning_rate": 4.5624e-05,
227
+ "loss": 0.0315,
228
+ "step": 2400
229
+ },
230
+ {
231
+ "epoch": 5.77,
232
+ "learning_rate": 4.5025000000000003e-05,
233
+ "loss": 0.0325,
234
+ "step": 2500
235
+ },
236
+ {
237
+ "epoch": 6.0,
238
+ "eval_accuracy_background": 0.9823462728892177,
239
+ "eval_accuracy_biofilm": 0.9665873924240089,
240
+ "eval_iou_background": 0.9803152068358649,
241
+ "eval_iou_biofilm": 0.7523833558585936,
242
+ "eval_loss": 0.051038164645433426,
243
+ "eval_mean_accuracy": 0.9744668326566133,
244
+ "eval_mean_iou": 0.8663492813472293,
245
+ "eval_overall_accuracy": 0.9814261507117296,
246
+ "eval_runtime": 53.252,
247
+ "eval_samples_per_second": 11.493,
248
+ "eval_steps_per_second": 1.446,
249
+ "step": 2598
250
+ },
251
+ {
252
+ "epoch": 6.0,
253
+ "learning_rate": 4.4426000000000005e-05,
254
+ "loss": 0.0389,
255
+ "step": 2600
256
+ },
257
+ {
258
+ "epoch": 6.24,
259
+ "learning_rate": 4.3827e-05,
260
+ "loss": 0.0322,
261
+ "step": 2700
262
+ },
263
+ {
264
+ "epoch": 6.47,
265
+ "learning_rate": 4.3228e-05,
266
+ "loss": 0.0268,
267
+ "step": 2800
268
+ },
269
+ {
270
+ "epoch": 6.7,
271
+ "learning_rate": 4.2629e-05,
272
+ "loss": 0.028,
273
+ "step": 2900
274
+ },
275
+ {
276
+ "epoch": 6.93,
277
+ "learning_rate": 4.203e-05,
278
+ "loss": 0.0306,
279
+ "step": 3000
280
+ },
281
+ {
282
+ "epoch": 7.0,
283
+ "eval_accuracy_background": 0.9862434447863753,
284
+ "eval_accuracy_biofilm": 0.9657386701296414,
285
+ "eval_iou_background": 0.9841526345692085,
286
+ "eval_iou_biofilm": 0.7903897627132261,
287
+ "eval_loss": 0.04278487339615822,
288
+ "eval_mean_accuracy": 0.9759910574580084,
289
+ "eval_mean_iou": 0.8872711986412174,
290
+ "eval_overall_accuracy": 0.9850462215478519,
291
+ "eval_runtime": 53.1423,
292
+ "eval_samples_per_second": 11.516,
293
+ "eval_steps_per_second": 1.449,
294
+ "step": 3031
295
+ },
296
+ {
297
+ "epoch": 7.16,
298
+ "learning_rate": 4.1431e-05,
299
+ "loss": 0.0292,
300
+ "step": 3100
301
+ },
302
+ {
303
+ "epoch": 7.39,
304
+ "learning_rate": 4.0832e-05,
305
+ "loss": 0.0308,
306
+ "step": 3200
307
+ },
308
+ {
309
+ "epoch": 7.62,
310
+ "learning_rate": 4.0233e-05,
311
+ "loss": 0.0268,
312
+ "step": 3300
313
+ },
314
+ {
315
+ "epoch": 7.85,
316
+ "learning_rate": 3.9634e-05,
317
+ "loss": 0.0318,
318
+ "step": 3400
319
+ },
320
+ {
321
+ "epoch": 8.0,
322
+ "eval_accuracy_background": 0.9859234209489122,
323
+ "eval_accuracy_biofilm": 0.9617940703893646,
324
+ "eval_iou_background": 0.983593215391998,
325
+ "eval_iou_biofilm": 0.7838504582636022,
326
+ "eval_loss": 0.03987625986337662,
327
+ "eval_mean_accuracy": 0.9738587456691383,
328
+ "eval_mean_iou": 0.8837218368278001,
329
+ "eval_overall_accuracy": 0.9845145676596673,
330
+ "eval_runtime": 52.5703,
331
+ "eval_samples_per_second": 11.642,
332
+ "eval_steps_per_second": 1.465,
333
+ "step": 3464
334
+ },
335
+ {
336
+ "epoch": 8.08,
337
+ "learning_rate": 3.9035e-05,
338
+ "loss": 0.0269,
339
+ "step": 3500
340
+ },
341
+ {
342
+ "epoch": 8.31,
343
+ "learning_rate": 3.8436e-05,
344
+ "loss": 0.0321,
345
+ "step": 3600
346
+ },
347
+ {
348
+ "epoch": 8.55,
349
+ "learning_rate": 3.7837000000000004e-05,
350
+ "loss": 0.0305,
351
+ "step": 3700
352
+ },
353
+ {
354
+ "epoch": 8.78,
355
+ "learning_rate": 3.7238000000000005e-05,
356
+ "loss": 0.0302,
357
+ "step": 3800
358
+ },
359
+ {
360
+ "epoch": 9.0,
361
+ "eval_accuracy_background": 0.9859386046114496,
362
+ "eval_accuracy_biofilm": 0.9519521052979112,
363
+ "eval_iou_background": 0.9830098689255221,
364
+ "eval_iou_biofilm": 0.7759842341679627,
365
+ "eval_loss": 0.043639492243528366,
366
+ "eval_mean_accuracy": 0.9689453549546804,
367
+ "eval_mean_iou": 0.8794970515467424,
368
+ "eval_overall_accuracy": 0.9839542167088564,
369
+ "eval_runtime": 53.0115,
370
+ "eval_samples_per_second": 11.545,
371
+ "eval_steps_per_second": 1.453,
372
+ "step": 3897
373
+ },
374
+ {
375
+ "epoch": 9.01,
376
+ "learning_rate": 3.6639e-05,
377
+ "loss": 0.0243,
378
+ "step": 3900
379
+ },
380
+ {
381
+ "epoch": 9.24,
382
+ "learning_rate": 3.604e-05,
383
+ "loss": 0.029,
384
+ "step": 4000
385
+ },
386
+ {
387
+ "epoch": 9.47,
388
+ "learning_rate": 3.544100000000001e-05,
389
+ "loss": 0.0279,
390
+ "step": 4100
391
+ },
392
+ {
393
+ "epoch": 9.7,
394
+ "learning_rate": 3.4842e-05,
395
+ "loss": 0.0312,
396
+ "step": 4200
397
+ },
398
+ {
399
+ "epoch": 9.93,
400
+ "learning_rate": 3.4243000000000004e-05,
401
+ "loss": 0.0236,
402
+ "step": 4300
403
+ },
404
+ {
405
+ "epoch": 10.0,
406
+ "eval_accuracy_background": 0.9866936591469345,
407
+ "eval_accuracy_biofilm": 0.9559968525080889,
408
+ "eval_iou_background": 0.9840087435512237,
409
+ "eval_iou_biofilm": 0.7870938963524574,
410
+ "eval_loss": 0.03914466127753258,
411
+ "eval_mean_accuracy": 0.9713452558275117,
412
+ "eval_mean_iou": 0.8855513199518406,
413
+ "eval_overall_accuracy": 0.9849013482890138,
414
+ "eval_runtime": 52.0282,
415
+ "eval_samples_per_second": 11.763,
416
+ "eval_steps_per_second": 1.48,
417
+ "step": 4330
418
+ },
419
+ {
420
+ "epoch": 10.16,
421
+ "learning_rate": 3.3644000000000005e-05,
422
+ "loss": 0.0246,
423
+ "step": 4400
424
+ },
425
+ {
426
+ "epoch": 10.39,
427
+ "learning_rate": 3.3045000000000006e-05,
428
+ "loss": 0.0235,
429
+ "step": 4500
430
+ },
431
+ {
432
+ "epoch": 10.62,
433
+ "learning_rate": 3.2446e-05,
434
+ "loss": 0.0268,
435
+ "step": 4600
436
+ },
437
+ {
438
+ "epoch": 10.85,
439
+ "learning_rate": 3.1847e-05,
440
+ "loss": 0.0247,
441
+ "step": 4700
442
+ },
443
+ {
444
+ "epoch": 11.0,
445
+ "eval_accuracy_background": 0.9834379952859539,
446
+ "eval_accuracy_biofilm": 0.9627888319063236,
447
+ "eval_iou_background": 0.9811740428860415,
448
+ "eval_iou_biofilm": 0.7598397926761448,
449
+ "eval_loss": 0.0451015830039978,
450
+ "eval_mean_accuracy": 0.9731134135961388,
451
+ "eval_mean_iou": 0.8705069177810931,
452
+ "eval_overall_accuracy": 0.9822323415458852,
453
+ "eval_runtime": 52.2459,
454
+ "eval_samples_per_second": 11.714,
455
+ "eval_steps_per_second": 1.474,
456
+ "step": 4763
457
+ },
458
+ {
459
+ "epoch": 11.09,
460
+ "learning_rate": 3.1248e-05,
461
+ "loss": 0.023,
462
+ "step": 4800
463
+ },
464
+ {
465
+ "epoch": 11.32,
466
+ "learning_rate": 3.0649000000000004e-05,
467
+ "loss": 0.0242,
468
+ "step": 4900
469
+ },
470
+ {
471
+ "epoch": 11.55,
472
+ "learning_rate": 3.0050000000000002e-05,
473
+ "loss": 0.0263,
474
+ "step": 5000
475
+ },
476
+ {
477
+ "epoch": 11.78,
478
+ "learning_rate": 2.9451e-05,
479
+ "loss": 0.0213,
480
+ "step": 5100
481
+ },
482
+ {
483
+ "epoch": 12.0,
484
+ "eval_accuracy_background": 0.9823563316391414,
485
+ "eval_accuracy_biofilm": 0.9646785376268351,
486
+ "eval_iou_background": 0.9802094628384811,
487
+ "eval_iou_biofilm": 0.7509923460988596,
488
+ "eval_loss": 0.04865105450153351,
489
+ "eval_mean_accuracy": 0.9735174346329882,
490
+ "eval_mean_iou": 0.8656009044686703,
491
+ "eval_overall_accuracy": 0.9813241688306411,
492
+ "eval_runtime": 52.4181,
493
+ "eval_samples_per_second": 11.675,
494
+ "eval_steps_per_second": 1.469,
495
+ "step": 5196
496
+ },
497
+ {
498
+ "epoch": 12.01,
499
+ "learning_rate": 2.8851999999999998e-05,
500
+ "loss": 0.0309,
501
+ "step": 5200
502
+ },
503
+ {
504
+ "epoch": 12.24,
505
+ "learning_rate": 2.8253e-05,
506
+ "loss": 0.0252,
507
+ "step": 5300
508
+ },
509
+ {
510
+ "epoch": 12.47,
511
+ "learning_rate": 2.7653999999999996e-05,
512
+ "loss": 0.0234,
513
+ "step": 5400
514
+ },
515
+ {
516
+ "epoch": 12.7,
517
+ "learning_rate": 2.7054999999999998e-05,
518
+ "loss": 0.0235,
519
+ "step": 5500
520
+ },
521
+ {
522
+ "epoch": 12.93,
523
+ "learning_rate": 2.6455999999999995e-05,
524
+ "loss": 0.0256,
525
+ "step": 5600
526
+ },
527
+ {
528
+ "epoch": 13.0,
529
+ "eval_accuracy_background": 0.9863652279961024,
530
+ "eval_accuracy_biofilm": 0.9473298043272511,
531
+ "eval_iou_background": 0.9831542706255657,
532
+ "eval_iou_biofilm": 0.7765716428707975,
533
+ "eval_loss": 0.0444067120552063,
534
+ "eval_mean_accuracy": 0.9668475161616767,
535
+ "eval_mean_iou": 0.8798629567481816,
536
+ "eval_overall_accuracy": 0.9840860458483189,
537
+ "eval_runtime": 52.5359,
538
+ "eval_samples_per_second": 11.649,
539
+ "eval_steps_per_second": 1.466,
540
+ "step": 5629
541
+ },
542
+ {
543
+ "epoch": 13.16,
544
+ "learning_rate": 2.5857e-05,
545
+ "loss": 0.0228,
546
+ "step": 5700
547
+ },
548
+ {
549
+ "epoch": 13.39,
550
+ "learning_rate": 2.5258e-05,
551
+ "loss": 0.0255,
552
+ "step": 5800
553
+ },
554
+ {
555
+ "epoch": 13.63,
556
+ "learning_rate": 2.4659e-05,
557
+ "loss": 0.0212,
558
+ "step": 5900
559
+ },
560
+ {
561
+ "epoch": 13.86,
562
+ "learning_rate": 2.406e-05,
563
+ "loss": 0.0218,
564
+ "step": 6000
565
+ },
566
+ {
567
+ "epoch": 14.0,
568
+ "eval_accuracy_background": 0.9821700673384857,
569
+ "eval_accuracy_biofilm": 0.9724741927673717,
570
+ "eval_iou_background": 0.9804965351674374,
571
+ "eval_iou_biofilm": 0.7552949378009881,
572
+ "eval_loss": 0.049166884273290634,
573
+ "eval_mean_accuracy": 0.9773221300529287,
574
+ "eval_mean_iou": 0.8678957364842128,
575
+ "eval_overall_accuracy": 0.9816039491270605,
576
+ "eval_runtime": 52.7983,
577
+ "eval_samples_per_second": 11.591,
578
+ "eval_steps_per_second": 1.458,
579
+ "step": 6062
580
+ },
581
+ {
582
+ "epoch": 14.09,
583
+ "learning_rate": 2.3460999999999998e-05,
584
+ "loss": 0.0297,
585
+ "step": 6100
586
+ },
587
+ {
588
+ "epoch": 14.32,
589
+ "learning_rate": 2.2862e-05,
590
+ "loss": 0.0286,
591
+ "step": 6200
592
+ },
593
+ {
594
+ "epoch": 14.55,
595
+ "learning_rate": 2.2263e-05,
596
+ "loss": 0.0219,
597
+ "step": 6300
598
+ },
599
+ {
600
+ "epoch": 14.78,
601
+ "learning_rate": 2.1663999999999998e-05,
602
+ "loss": 0.0216,
603
+ "step": 6400
604
+ },
605
+ {
606
+ "epoch": 15.0,
607
+ "eval_accuracy_background": 0.983388424892843,
608
+ "eval_accuracy_biofilm": 0.9662580062949838,
609
+ "eval_iou_background": 0.981335202197703,
610
+ "eval_iou_biofilm": 0.7620968774425835,
611
+ "eval_loss": 0.05022722855210304,
612
+ "eval_mean_accuracy": 0.9748232155939134,
613
+ "eval_mean_iou": 0.8717160398201432,
614
+ "eval_overall_accuracy": 0.9823882219852604,
615
+ "eval_runtime": 52.3676,
616
+ "eval_samples_per_second": 11.687,
617
+ "eval_steps_per_second": 1.47,
618
+ "step": 6495
619
+ },
620
+ {
621
+ "epoch": 15.01,
622
+ "learning_rate": 2.1065e-05,
623
+ "loss": 0.022,
624
+ "step": 6500
625
+ },
626
+ {
627
+ "epoch": 15.24,
628
+ "learning_rate": 2.0465999999999997e-05,
629
+ "loss": 0.0211,
630
+ "step": 6600
631
+ },
632
+ {
633
+ "epoch": 15.47,
634
+ "learning_rate": 1.9866999999999998e-05,
635
+ "loss": 0.0236,
636
+ "step": 6700
637
+ },
638
+ {
639
+ "epoch": 15.7,
640
+ "learning_rate": 1.9267999999999996e-05,
641
+ "loss": 0.0225,
642
+ "step": 6800
643
+ },
644
+ {
645
+ "epoch": 15.94,
646
+ "learning_rate": 1.8669e-05,
647
+ "loss": 0.0206,
648
+ "step": 6900
649
+ },
650
+ {
651
+ "epoch": 16.0,
652
+ "eval_accuracy_background": 0.9811236366128779,
653
+ "eval_accuracy_biofilm": 0.9721258776660137,
654
+ "eval_iou_background": 0.9794307694409038,
655
+ "eval_iou_biofilm": 0.7452563861896737,
656
+ "eval_loss": 0.056503016501665115,
657
+ "eval_mean_accuracy": 0.9766247571394457,
658
+ "eval_mean_iou": 0.8623435778152888,
659
+ "eval_overall_accuracy": 0.9805982796515773,
660
+ "eval_runtime": 53.0323,
661
+ "eval_samples_per_second": 11.54,
662
+ "eval_steps_per_second": 1.452,
663
+ "step": 6928
664
+ },
665
+ {
666
+ "epoch": 16.17,
667
+ "learning_rate": 1.807e-05,
668
+ "loss": 0.022,
669
+ "step": 7000
670
+ },
671
+ {
672
+ "epoch": 16.4,
673
+ "learning_rate": 1.7471e-05,
674
+ "loss": 0.0208,
675
+ "step": 7100
676
+ },
677
+ {
678
+ "epoch": 16.63,
679
+ "learning_rate": 1.6872e-05,
680
+ "loss": 0.0234,
681
+ "step": 7200
682
+ },
683
+ {
684
+ "epoch": 16.86,
685
+ "learning_rate": 1.6272999999999998e-05,
686
+ "loss": 0.0223,
687
+ "step": 7300
688
+ },
689
+ {
690
+ "epoch": 17.0,
691
+ "eval_accuracy_background": 0.9826374580040367,
692
+ "eval_accuracy_biofilm": 0.963460810424141,
693
+ "eval_iou_background": 0.9804161065014417,
694
+ "eval_iou_biofilm": 0.7527009753594974,
695
+ "eval_loss": 0.0508936271071434,
696
+ "eval_mean_accuracy": 0.9730491342140888,
697
+ "eval_mean_iou": 0.8665585409304696,
698
+ "eval_overall_accuracy": 0.981517780831087,
699
+ "eval_runtime": 53.2214,
700
+ "eval_samples_per_second": 11.499,
701
+ "eval_steps_per_second": 1.447,
702
+ "step": 7361
703
+ },
704
+ {
705
+ "epoch": 17.09,
706
+ "learning_rate": 1.5674e-05,
707
+ "loss": 0.0247,
708
+ "step": 7400
709
+ },
710
+ {
711
+ "epoch": 17.32,
712
+ "learning_rate": 1.5075000000000002e-05,
713
+ "loss": 0.022,
714
+ "step": 7500
715
+ },
716
+ {
717
+ "epoch": 17.55,
718
+ "learning_rate": 1.4476e-05,
719
+ "loss": 0.0265,
720
+ "step": 7600
721
+ },
722
+ {
723
+ "epoch": 17.78,
724
+ "learning_rate": 1.3877e-05,
725
+ "loss": 0.0226,
726
+ "step": 7700
727
+ },
728
+ {
729
+ "epoch": 18.0,
730
+ "eval_accuracy_background": 0.9841817422579221,
731
+ "eval_accuracy_biofilm": 0.9742679330002422,
732
+ "eval_iou_background": 0.9826138886496493,
733
+ "eval_iou_biofilm": 0.7762471663828775,
734
+ "eval_loss": 0.04640285670757294,
735
+ "eval_mean_accuracy": 0.9792248376290822,
736
+ "eval_mean_iou": 0.8794305275162634,
737
+ "eval_overall_accuracy": 0.9836028993774575,
738
+ "eval_runtime": 53.7901,
739
+ "eval_samples_per_second": 11.378,
740
+ "eval_steps_per_second": 1.431,
741
+ "step": 7794
742
+ },
743
+ {
744
+ "epoch": 18.01,
745
+ "learning_rate": 1.3277999999999999e-05,
746
+ "loss": 0.0189,
747
+ "step": 7800
748
+ },
749
+ {
750
+ "epoch": 18.24,
751
+ "learning_rate": 1.2678999999999998e-05,
752
+ "loss": 0.0196,
753
+ "step": 7900
754
+ },
755
+ {
756
+ "epoch": 18.48,
757
+ "learning_rate": 1.2079999999999998e-05,
758
+ "loss": 0.0239,
759
+ "step": 8000
760
+ },
761
+ {
762
+ "epoch": 18.71,
763
+ "learning_rate": 1.1480999999999997e-05,
764
+ "loss": 0.0224,
765
+ "step": 8100
766
+ },
767
+ {
768
+ "epoch": 18.94,
769
+ "learning_rate": 1.0882000000000004e-05,
770
+ "loss": 0.0243,
771
+ "step": 8200
772
+ },
773
+ {
774
+ "epoch": 19.0,
775
+ "eval_accuracy_background": 0.9806471221015685,
776
+ "eval_accuracy_biofilm": 0.9842518653841921,
777
+ "eval_iou_background": 0.9796904437901697,
778
+ "eval_iou_biofilm": 0.7501332139181935,
779
+ "eval_loss": 0.05462173372507095,
780
+ "eval_mean_accuracy": 0.9824494937428803,
781
+ "eval_mean_iou": 0.8649118288541816,
782
+ "eval_overall_accuracy": 0.9808575941744937,
783
+ "eval_runtime": 53.6405,
784
+ "eval_samples_per_second": 11.409,
785
+ "eval_steps_per_second": 1.435,
786
+ "step": 8227
787
+ },
788
+ {
789
+ "epoch": 19.17,
790
+ "learning_rate": 1.0283000000000003e-05,
791
+ "loss": 0.0201,
792
+ "step": 8300
793
+ },
794
+ {
795
+ "epoch": 19.4,
796
+ "learning_rate": 9.684000000000002e-06,
797
+ "loss": 0.0203,
798
+ "step": 8400
799
+ },
800
+ {
801
+ "epoch": 19.63,
802
+ "learning_rate": 9.085000000000002e-06,
803
+ "loss": 0.017,
804
+ "step": 8500
805
+ },
806
+ {
807
+ "epoch": 19.86,
808
+ "learning_rate": 8.486000000000001e-06,
809
+ "loss": 0.02,
810
+ "step": 8600
811
+ },
812
+ {
813
+ "epoch": 20.0,
814
+ "eval_accuracy_background": 0.9816362984772036,
815
+ "eval_accuracy_biofilm": 0.9715923447714216,
816
+ "eval_iou_background": 0.9799101839896647,
817
+ "eval_iou_biofilm": 0.7495984750713218,
818
+ "eval_loss": 0.05672093480825424,
819
+ "eval_mean_accuracy": 0.9766143216243126,
820
+ "eval_mean_iou": 0.8647543295304932,
821
+ "eval_overall_accuracy": 0.98104985678328,
822
+ "eval_runtime": 52.5058,
823
+ "eval_samples_per_second": 11.656,
824
+ "eval_steps_per_second": 1.467,
825
+ "step": 8660
826
+ },
827
+ {
828
+ "epoch": 20.09,
829
+ "learning_rate": 7.887000000000001e-06,
830
+ "loss": 0.0213,
831
+ "step": 8700
832
+ },
833
+ {
834
+ "epoch": 20.32,
835
+ "learning_rate": 7.2879999999999995e-06,
836
+ "loss": 0.0218,
837
+ "step": 8800
838
+ },
839
+ {
840
+ "epoch": 20.55,
841
+ "learning_rate": 6.688999999999999e-06,
842
+ "loss": 0.0233,
843
+ "step": 8900
844
+ },
845
+ {
846
+ "epoch": 20.79,
847
+ "learning_rate": 6.0899999999999984e-06,
848
+ "loss": 0.0196,
849
+ "step": 9000
850
+ },
851
+ {
852
+ "epoch": 21.0,
853
+ "eval_accuracy_background": 0.9813258625255226,
854
+ "eval_accuracy_biofilm": 0.9755059098012457,
855
+ "eval_iou_background": 0.9798376554115104,
856
+ "eval_iou_biofilm": 0.7497220481513676,
857
+ "eval_loss": 0.05593876540660858,
858
+ "eval_mean_accuracy": 0.9784158861633842,
859
+ "eval_mean_iou": 0.8647798517814389,
860
+ "eval_overall_accuracy": 0.9809860498348075,
861
+ "eval_runtime": 53.7718,
862
+ "eval_samples_per_second": 11.381,
863
+ "eval_steps_per_second": 1.432,
864
+ "step": 9093
865
+ },
866
+ {
867
+ "epoch": 21.02,
868
+ "learning_rate": 5.490999999999998e-06,
869
+ "loss": 0.0201,
870
+ "step": 9100
871
+ },
872
+ {
873
+ "epoch": 21.25,
874
+ "learning_rate": 4.891999999999997e-06,
875
+ "loss": 0.0167,
876
+ "step": 9200
877
+ },
878
+ {
879
+ "epoch": 21.48,
880
+ "learning_rate": 4.292999999999997e-06,
881
+ "loss": 0.0199,
882
+ "step": 9300
883
+ },
884
+ {
885
+ "epoch": 21.71,
886
+ "learning_rate": 3.694000000000003e-06,
887
+ "loss": 0.0225,
888
+ "step": 9400
889
+ },
890
+ {
891
+ "epoch": 21.94,
892
+ "learning_rate": 3.0950000000000026e-06,
893
+ "loss": 0.0206,
894
+ "step": 9500
895
+ },
896
+ {
897
+ "epoch": 22.0,
898
+ "eval_accuracy_background": 0.9815057872205616,
899
+ "eval_accuracy_biofilm": 0.9742263332819756,
900
+ "eval_iou_background": 0.9799396728065195,
901
+ "eval_iou_biofilm": 0.7504120847099512,
902
+ "eval_loss": 0.05521002784371376,
903
+ "eval_mean_accuracy": 0.9778660602512685,
904
+ "eval_mean_iou": 0.8651758787582353,
905
+ "eval_overall_accuracy": 0.9810807578522835,
906
+ "eval_runtime": 44.0076,
907
+ "eval_samples_per_second": 13.907,
908
+ "eval_steps_per_second": 1.75,
909
+ "step": 9526
910
+ },
911
+ {
912
+ "epoch": 22.17,
913
+ "learning_rate": 2.496000000000002e-06,
914
+ "loss": 0.0191,
915
+ "step": 9600
916
+ },
917
+ {
918
+ "epoch": 22.4,
919
+ "learning_rate": 1.8970000000000013e-06,
920
+ "loss": 0.0162,
921
+ "step": 9700
922
+ },
923
+ {
924
+ "epoch": 22.63,
925
+ "learning_rate": 1.298000000000001e-06,
926
+ "loss": 0.0213,
927
+ "step": 9800
928
+ },
929
+ {
930
+ "epoch": 22.86,
931
+ "learning_rate": 6.990000000000005e-07,
932
+ "loss": 0.0189,
933
+ "step": 9900
934
+ },
935
+ {
936
+ "epoch": 23.0,
937
+ "eval_accuracy_background": 0.9815976262005296,
938
+ "eval_accuracy_biofilm": 0.9753409416063213,
939
+ "eval_iou_background": 0.9800989963754594,
940
+ "eval_iou_biofilm": 0.7521286753495444,
941
+ "eval_loss": 0.05435158312320709,
942
+ "eval_mean_accuracy": 0.9784692839034255,
943
+ "eval_mean_iou": 0.8661138358625019,
944
+ "eval_overall_accuracy": 0.9812323138126737,
945
+ "eval_runtime": 44.1254,
946
+ "eval_samples_per_second": 13.87,
947
+ "eval_steps_per_second": 1.745,
948
+ "step": 9959
949
+ },
950
+ {
951
+ "epoch": 23.09,
952
+ "learning_rate": 1e-07,
953
+ "loss": 0.0208,
954
+ "step": 10000
955
+ },
956
+ {
957
+ "epoch": 23.09,
958
+ "eval_accuracy_background": 0.9820080627772998,
959
+ "eval_accuracy_biofilm": 0.973973103250941,
960
+ "eval_iou_background": 0.9804257765494778,
961
+ "eval_iou_biofilm": 0.7549272284294235,
962
+ "eval_loss": 0.05298486724495888,
963
+ "eval_mean_accuracy": 0.9779905830141205,
964
+ "eval_mean_iou": 0.8676765024894506,
965
+ "eval_overall_accuracy": 0.9815389213004198,
966
+ "eval_runtime": 44.7948,
967
+ "eval_samples_per_second": 13.662,
968
+ "eval_steps_per_second": 1.719,
969
+ "step": 10000
970
+ },
971
+ {
972
+ "epoch": 23.09,
973
+ "step": 10000,
974
+ "total_flos": 1.4014301557688893e+18,
975
+ "train_loss": 0.03242814379930496,
976
+ "train_runtime": 14425.5946,
977
+ "train_samples_per_second": 5.546,
978
+ "train_steps_per_second": 0.693
979
+ }
980
+ ],
981
+ "logging_steps": 100,
982
+ "max_steps": 10000,
983
+ "num_input_tokens_seen": 0,
984
+ "num_train_epochs": 24,
985
+ "save_steps": 500,
986
+ "total_flos": 1.4014301557688893e+18,
987
+ "train_batch_size": 8,
988
+ "trial_name": null,
989
+ "trial_params": null
990
+ }