Hugging Face's logo

nateraw
/
mit-b0-finetuned-sidewalks-v2

Transformers
TensorFlow
segformer
generated_from_keras_callback
Inference Endpoints
Model card Files Community
nateraw commited on
Commit
bcea238
•
1 Parent(s): ad3fd58

Training in progress epoch 8

Browse files
Files changed (2) hide show
  1. README.md +29 -24
  2. tf_model.h5 +1 -1
README.md CHANGED
@@ -14,30 +14,24 @@ probably proofread and complete it, then remove this comment. -->
14
 
15
  This model is a fine-tuned version of [nvidia/mit-b0](https://huggingface.co/nvidia/mit-b0) on an unknown dataset.
16
  It achieves the following results on the evaluation set:
17
- - Train Loss: 0.4495
18
- - Validation Loss: 0.4797
19
- - Validation Mean Iou: 0.3035
20
- - Validation Mean Accuracy: 0.3702
21
- - Validation Overall Accuracy: 0.8468
22
- - Validation Per Category Iou: [0.00000000e+00 7.52163526e-01 8.46563375e-01 7.16396797e-01
23
- 7.38850637e-01 3.93073019e-01 nan 3.31795957e-01
24
- 4.92991567e-01 0.00000000e+00 8.11302090e-01 0.00000000e+00
25
- 0.00000000e+00 0.00000000e+00 0.00000000e+00 5.16059849e-01
26
- 0.00000000e+00 0.00000000e+00 6.56058294e-01 1.25948501e-02
27
- 2.66942435e-01 5.34406894e-01 0.00000000e+00 nan
28
- 0.00000000e+00 2.27750085e-01 4.86381323e-04 0.00000000e+00
29
- 8.48618960e-01 7.25828093e-01 9.17747637e-01 8.28380212e-03
30
- 6.74590297e-02 1.51281596e-01 0.00000000e+00]
31
- - Validation Per Category Accuracy: [0.00000000e+00 8.75360044e-01 9.43650850e-01 8.78658645e-01
32
- 7.76578096e-01 4.85757596e-01 nan 4.30901582e-01
33
- 7.54126335e-01 0.00000000e+00 9.30112537e-01 0.00000000e+00
34
- 0.00000000e+00 0.00000000e+00 0.00000000e+00 6.42914247e-01
35
- 0.00000000e+00 0.00000000e+00 7.57605356e-01 1.27102686e-02
36
- 6.50888458e-01 6.94757080e-01 0.00000000e+00 nan
37
- 0.00000000e+00 2.91727649e-01 4.86381323e-04 0.00000000e+00
38
- 9.42251577e-01 8.60753175e-01 9.56778008e-01 8.51551074e-03
39
- 1.38756779e-01 1.83583708e-01 0.00000000e+00]
40
- - Epoch: 7
41
 
42
  ## Model description
43
 
@@ -169,6 +163,17 @@ The following hyperparameters were used during training:
169
  0.00000000e+00 2.91727649e-01 4.86381323e-04 0.00000000e+00
170
  9.42251577e-01 8.60753175e-01 9.56778008e-01 8.51551074e-03
171
  1.38756779e-01 1.83583708e-01 0.00000000e+00] | 7 |
 
 
 
 
 
 
 
 
 
 
 
172
 
173
 
174
  ### Framework versions
 
14
 
15
  This model is a fine-tuned version of [nvidia/mit-b0](https://huggingface.co/nvidia/mit-b0) on an unknown dataset.
16
  It achieves the following results on the evaluation set:
17
+ - Train Loss: 0.4193
18
+ - Validation Loss: 0.4487
19
+ - Validation Mean Iou: 0.3073
20
+ - Validation Mean Accuracy: 0.3633
21
+ - Validation Overall Accuracy: 0.8594
22
+ - Validation Per Category Iou: [0. 0.77081114 0.86089485 0.64464211 0.82962632 0.36186873
23
+ nan 0.39092332 0.5399988 0. 0.81734925 0.
24
+ 0. 0. 0. 0.50271555 0. 0.
25
+ 0.70239658 0. 0.30875695 0.52195319 0. nan
26
+ 0. 0.20124517 0.00696273 0. 0.84526591 0.72563399
27
+ 0.91703372 0. 0.03526147 0.15693635 0. ]
28
+ - Validation Per Category Accuracy: [0. 0.8654775 0.95711297 0.70665759 0.93130714 0.42436958
29
+ nan 0.52892143 0.69243377 0. 0.91682626 0.
30
+ 0. 0. 0. 0.62315913 0. 0.
31
+ 0.86251114 0. 0.5607807 0.70416055 0. nan
32
+ 0. 0.24483525 0.00698305 0. 0.921099 0.81848055
33
+ 0.96789871 0. 0.06891948 0.18778302 0. ]
34
+ - Epoch: 8
 
 
 
 
 
 
35
 
36
  ## Model description
37
 
 
163
  0.00000000e+00 2.91727649e-01 4.86381323e-04 0.00000000e+00
164
  9.42251577e-01 8.60753175e-01 9.56778008e-01 8.51551074e-03
165
  1.38756779e-01 1.83583708e-01 0.00000000e+00] | 7 |
166
+ | 0.4193 | 0.4487 | 0.3073 | 0.3633 | 0.8594 | [0. 0.77081114 0.86089485 0.64464211 0.82962632 0.36186873
167
+ nan 0.39092332 0.5399988 0. 0.81734925 0.
168
+ 0. 0. 0. 0.50271555 0. 0.
169
+ 0.70239658 0. 0.30875695 0.52195319 0. nan
170
+ 0. 0.20124517 0.00696273 0. 0.84526591 0.72563399
171
+ 0.91703372 0. 0.03526147 0.15693635 0. ] | [0. 0.8654775 0.95711297 0.70665759 0.93130714 0.42436958
172
+ nan 0.52892143 0.69243377 0. 0.91682626 0.
173
+ 0. 0. 0. 0.62315913 0. 0.
174
+ 0.86251114 0. 0.5607807 0.70416055 0. nan
175
+ 0. 0.24483525 0.00698305 0. 0.921099 0.81848055
176
+ 0.96789871 0. 0.06891948 0.18778302 0. ] | 8 |
177
 
178
 
179
  ### Framework versions
tf_model.h5 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:94b84b3db99e12356df44a67a3034eecaa80a42d80f80e90a1d76c90dbcb34fc
3
  size 15167588
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e63d034662ba8c744c7eb1e9a8e53c427ff6af286ff57083994304ee36f4bc93
3
  size 15167588