Hugging Face's logo

luca-g97
/
mit-b0-finetuned-sidewalks

Transformers
TensorFlow
segformer
generated_from_keras_callback
Inference Endpoints
Model card Files Community
luca-g97 commited on
Commit
c1a70e3
·
1 Parent(s): a8a2fcf

Training in progress epoch 30

Browse files
Files changed (2) hide show
  1. README.md +57 -56
  2. tf_model.h5 +1 -1
README.md CHANGED
@@ -15,82 +15,82 @@ probably proofread and complete it, then remove this comment. -->
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
- - Train Loss: 0.1959
19
- - Validation Loss: 0.5240
20
- - Validation Mean Iou: 0.3521
21
- - Validation Mean Accuracy: 0.4343
22
- - Validation Overall Accuracy: 0.8771
23
  - Validation Accuracy Unlabeled: 0.0
24
- - Validation Accuracy Flat-road: 0.8744
25
- - Validation Accuracy Flat-sidewalk: 0.9749
26
- - Validation Accuracy Flat-crosswalk: 0.4993
27
- - Validation Accuracy Flat-cyclinglane: 0.8143
28
- - Validation Accuracy Flat-parkingdriveway: 0.3366
29
  - Validation Accuracy Flat-railtrack: nan
30
- - Validation Accuracy Flat-curb: 0.6376
31
- - Validation Accuracy Human-person: 0.7596
32
- - Validation Accuracy Human-rider: 0.1736
33
- - Validation Accuracy Vehicle-car: 0.9184
34
- - Validation Accuracy Vehicle-truck: 0.0424
35
  - Validation Accuracy Vehicle-bus: nan
36
  - Validation Accuracy Vehicle-tramtrain: nan
37
- - Validation Accuracy Vehicle-motorcycle: 0.0002
38
- - Validation Accuracy Vehicle-bicycle: 0.7469
39
  - Validation Accuracy Vehicle-caravan: 0.0
40
  - Validation Accuracy Vehicle-cartrailer: 0.0
41
- - Validation Accuracy Construction-building: 0.8957
42
- - Validation Accuracy Construction-door: 0.1516
43
- - Validation Accuracy Construction-wall: 0.5852
44
- - Validation Accuracy Construction-fenceguardrail: 0.4130
45
  - Validation Accuracy Construction-bridge: 0.0
46
  - Validation Accuracy Construction-tunnel: nan
47
- - Validation Accuracy Construction-stairs: 0.2630
48
- - Validation Accuracy Object-pole: 0.4594
49
- - Validation Accuracy Object-trafficsign: 0.3941
50
  - Validation Accuracy Object-trafficlight: 0.0
51
- - Validation Accuracy Nature-vegetation: 0.9456
52
- - Validation Accuracy Nature-terrain: 0.9363
53
- - Validation Accuracy Sky: 0.9789
54
- - Validation Accuracy Void-ground: 0.1007
55
- - Validation Accuracy Void-dynamic: 0.1004
56
- - Validation Accuracy Void-static: 0.4621
57
  - Validation Accuracy Void-unclear: 0.0
58
  - Validation Iou Unlabeled: 0.0
59
- - Validation Iou Flat-road: 0.8044
60
- - Validation Iou Flat-sidewalk: 0.8775
61
- - Validation Iou Flat-crosswalk: 0.4029
62
- - Validation Iou Flat-cyclinglane: 0.7728
63
- - Validation Iou Flat-parkingdriveway: 0.2994
64
  - Validation Iou Flat-railtrack: nan
65
- - Validation Iou Flat-curb: 0.5577
66
- - Validation Iou Human-person: 0.4748
67
- - Validation Iou Human-rider: 0.1513
68
- - Validation Iou Vehicle-car: 0.7884
69
- - Validation Iou Vehicle-truck: 0.0378
70
  - Validation Iou Vehicle-bus: nan
71
  - Validation Iou Vehicle-tramtrain: 0.0
72
- - Validation Iou Vehicle-motorcycle: 0.0002
73
- - Validation Iou Vehicle-bicycle: 0.5007
74
  - Validation Iou Vehicle-caravan: 0.0
75
  - Validation Iou Vehicle-cartrailer: 0.0
76
- - Validation Iou Construction-building: 0.7343
77
- - Validation Iou Construction-door: 0.1427
78
- - Validation Iou Construction-wall: 0.4677
79
- - Validation Iou Construction-fenceguardrail: 0.3440
80
  - Validation Iou Construction-bridge: 0.0
81
  - Validation Iou Construction-tunnel: nan
82
- - Validation Iou Construction-stairs: 0.1765
83
- - Validation Iou Object-pole: 0.3631
84
- - Validation Iou Object-trafficsign: 0.2372
85
  - Validation Iou Object-trafficlight: 0.0
86
- - Validation Iou Nature-vegetation: 0.8800
87
- - Validation Iou Nature-terrain: 0.8324
88
- - Validation Iou Sky: 0.9400
89
- - Validation Iou Void-ground: 0.0701
90
- - Validation Iou Void-dynamic: 0.0558
91
- - Validation Iou Void-static: 0.3558
92
  - Validation Iou Void-unclear: 0.0
93
- - Epoch: 29
94
 
95
  ## Model description
96
 
@@ -146,6 +146,7 @@ The following hyperparameters were used during training:
146
  | 0.2208 | 0.5502 | 0.3493 | 0.4366 | 0.8689 | 0.0 | 0.8383 | 0.9673 | 0.5042 | 0.7920 | 0.3196 | nan | 0.6306 | 0.7779 | 0.1291 | 0.9021 | 0.1445 | nan | nan | 0.0273 | 0.7410 | 0.0 | 0.0 | 0.8969 | 0.1208 | 0.5939 | 0.4402 | 0.0 | nan | 0.2468 | 0.4770 | 0.3748 | 0.0 | 0.9438 | 0.9466 | 0.9747 | 0.1148 | 0.1817 | 0.4475 | 0.0 | 0.0 | 0.7682 | 0.8633 | 0.4140 | 0.7395 | 0.2747 | nan | 0.5426 | 0.4367 | 0.1124 | 0.7849 | 0.1196 | nan | 0.0 | 0.0272 | 0.5120 | 0.0 | 0.0 | 0.7431 | 0.1154 | 0.4755 | 0.3410 | 0.0 | nan | 0.1788 | 0.3618 | 0.2128 | 0.0 | 0.8779 | 0.8179 | 0.9403 | 0.0889 | 0.0963 | 0.3342 | 0.0 | 27 |
147
  | 0.2110 | 0.5231 | 0.3414 | 0.4290 | 0.8722 | 0.0 | 0.8684 | 0.9551 | 0.5060 | 0.8743 | 0.3936 | nan | 0.6295 | 0.7991 | 0.1233 | 0.9409 | 0.0481 | nan | nan | 0.0 | 0.7064 | 0.0 | 0.0 | 0.9129 | 0.1463 | 0.4067 | 0.4126 | 0.0 | nan | 0.2611 | 0.4535 | 0.3979 | 0.0 | 0.9440 | 0.9430 | 0.9801 | 0.1153 | 0.0838 | 0.3970 | 0.0 | 0.0 | 0.7971 | 0.8795 | 0.3698 | 0.7345 | 0.3286 | nan | 0.5408 | 0.4354 | 0.1101 | 0.7568 | 0.0463 | nan | 0.0 | 0.0 | 0.4965 | 0.0 | 0.0 | 0.7422 | 0.1314 | 0.3671 | 0.3160 | 0.0 | nan | 0.1858 | 0.3615 | 0.2297 | 0.0 | 0.8787 | 0.8337 | 0.9380 | 0.0751 | 0.0511 | 0.3184 | 0.0 | 28 |
148
  | 0.1959 | 0.5240 | 0.3521 | 0.4343 | 0.8771 | 0.0 | 0.8744 | 0.9749 | 0.4993 | 0.8143 | 0.3366 | nan | 0.6376 | 0.7596 | 0.1736 | 0.9184 | 0.0424 | nan | nan | 0.0002 | 0.7469 | 0.0 | 0.0 | 0.8957 | 0.1516 | 0.5852 | 0.4130 | 0.0 | nan | 0.2630 | 0.4594 | 0.3941 | 0.0 | 0.9456 | 0.9363 | 0.9789 | 0.1007 | 0.1004 | 0.4621 | 0.0 | 0.0 | 0.8044 | 0.8775 | 0.4029 | 0.7728 | 0.2994 | nan | 0.5577 | 0.4748 | 0.1513 | 0.7884 | 0.0378 | nan | 0.0 | 0.0002 | 0.5007 | 0.0 | 0.0 | 0.7343 | 0.1427 | 0.4677 | 0.3440 | 0.0 | nan | 0.1765 | 0.3631 | 0.2372 | 0.0 | 0.8800 | 0.8324 | 0.9400 | 0.0701 | 0.0558 | 0.3558 | 0.0 | 29 |
 
149
 
150
 
151
  ### Framework versions
 
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
+ - Train Loss: 0.1941
19
+ - Validation Loss: 0.5077
20
+ - Validation Mean Iou: 0.3541
21
+ - Validation Mean Accuracy: 0.4368
22
+ - Validation Overall Accuracy: 0.8760
23
  - Validation Accuracy Unlabeled: 0.0
24
+ - Validation Accuracy Flat-road: 0.8810
25
+ - Validation Accuracy Flat-sidewalk: 0.9600
26
+ - Validation Accuracy Flat-crosswalk: 0.5080
27
+ - Validation Accuracy Flat-cyclinglane: 0.7956
28
+ - Validation Accuracy Flat-parkingdriveway: 0.4939
29
  - Validation Accuracy Flat-railtrack: nan
30
+ - Validation Accuracy Flat-curb: 0.5978
31
+ - Validation Accuracy Human-person: 0.7554
32
+ - Validation Accuracy Human-rider: 0.1367
33
+ - Validation Accuracy Vehicle-car: 0.9179
34
+ - Validation Accuracy Vehicle-truck: 0.0599
35
  - Validation Accuracy Vehicle-bus: nan
36
  - Validation Accuracy Vehicle-tramtrain: nan
37
+ - Validation Accuracy Vehicle-motorcycle: 0.0
38
+ - Validation Accuracy Vehicle-bicycle: 0.7221
39
  - Validation Accuracy Vehicle-caravan: 0.0
40
  - Validation Accuracy Vehicle-cartrailer: 0.0
41
+ - Validation Accuracy Construction-building: 0.8944
42
+ - Validation Accuracy Construction-door: 0.1175
43
+ - Validation Accuracy Construction-wall: 0.6129
44
+ - Validation Accuracy Construction-fenceguardrail: 0.4032
45
  - Validation Accuracy Construction-bridge: 0.0
46
  - Validation Accuracy Construction-tunnel: nan
47
+ - Validation Accuracy Construction-stairs: 0.2636
48
+ - Validation Accuracy Object-pole: 0.4817
49
+ - Validation Accuracy Object-trafficsign: 0.3849
50
  - Validation Accuracy Object-trafficlight: 0.0
51
+ - Validation Accuracy Nature-vegetation: 0.9453
52
+ - Validation Accuracy Nature-terrain: 0.9461
53
+ - Validation Accuracy Sky: 0.9760
54
+ - Validation Accuracy Void-ground: 0.0649
55
+ - Validation Accuracy Void-dynamic: 0.1096
56
+ - Validation Accuracy Void-static: 0.5128
57
  - Validation Accuracy Void-unclear: 0.0
58
  - Validation Iou Unlabeled: 0.0
59
+ - Validation Iou Flat-road: 0.7899
60
+ - Validation Iou Flat-sidewalk: 0.8798
61
+ - Validation Iou Flat-crosswalk: 0.3854
62
+ - Validation Iou Flat-cyclinglane: 0.7444
63
+ - Validation Iou Flat-parkingdriveway: 0.3950
64
  - Validation Iou Flat-railtrack: nan
65
+ - Validation Iou Flat-curb: 0.5344
66
+ - Validation Iou Human-person: 0.4893
67
+ - Validation Iou Human-rider: 0.1164
68
+ - Validation Iou Vehicle-car: 0.7795
69
+ - Validation Iou Vehicle-truck: 0.0515
70
  - Validation Iou Vehicle-bus: nan
71
  - Validation Iou Vehicle-tramtrain: 0.0
72
+ - Validation Iou Vehicle-motorcycle: 0.0
73
+ - Validation Iou Vehicle-bicycle: 0.5232
74
  - Validation Iou Vehicle-caravan: 0.0
75
  - Validation Iou Vehicle-cartrailer: 0.0
76
+ - Validation Iou Construction-building: 0.7498
77
+ - Validation Iou Construction-door: 0.1094
78
+ - Validation Iou Construction-wall: 0.4920
79
+ - Validation Iou Construction-fenceguardrail: 0.3325
80
  - Validation Iou Construction-bridge: 0.0
81
  - Validation Iou Construction-tunnel: nan
82
+ - Validation Iou Construction-stairs: 0.2083
83
+ - Validation Iou Object-pole: 0.3678
84
+ - Validation Iou Object-trafficsign: 0.2519
85
  - Validation Iou Object-trafficlight: 0.0
86
+ - Validation Iou Nature-vegetation: 0.8789
87
+ - Validation Iou Nature-terrain: 0.8304
88
+ - Validation Iou Sky: 0.9409
89
+ - Validation Iou Void-ground: 0.0447
90
+ - Validation Iou Void-dynamic: 0.0765
91
+ - Validation Iou Void-static: 0.3601
92
  - Validation Iou Void-unclear: 0.0
93
+ - Epoch: 30
94
 
95
  ## Model description
96
 
 
146
  | 0.2208 | 0.5502 | 0.3493 | 0.4366 | 0.8689 | 0.0 | 0.8383 | 0.9673 | 0.5042 | 0.7920 | 0.3196 | nan | 0.6306 | 0.7779 | 0.1291 | 0.9021 | 0.1445 | nan | nan | 0.0273 | 0.7410 | 0.0 | 0.0 | 0.8969 | 0.1208 | 0.5939 | 0.4402 | 0.0 | nan | 0.2468 | 0.4770 | 0.3748 | 0.0 | 0.9438 | 0.9466 | 0.9747 | 0.1148 | 0.1817 | 0.4475 | 0.0 | 0.0 | 0.7682 | 0.8633 | 0.4140 | 0.7395 | 0.2747 | nan | 0.5426 | 0.4367 | 0.1124 | 0.7849 | 0.1196 | nan | 0.0 | 0.0272 | 0.5120 | 0.0 | 0.0 | 0.7431 | 0.1154 | 0.4755 | 0.3410 | 0.0 | nan | 0.1788 | 0.3618 | 0.2128 | 0.0 | 0.8779 | 0.8179 | 0.9403 | 0.0889 | 0.0963 | 0.3342 | 0.0 | 27 |
147
  | 0.2110 | 0.5231 | 0.3414 | 0.4290 | 0.8722 | 0.0 | 0.8684 | 0.9551 | 0.5060 | 0.8743 | 0.3936 | nan | 0.6295 | 0.7991 | 0.1233 | 0.9409 | 0.0481 | nan | nan | 0.0 | 0.7064 | 0.0 | 0.0 | 0.9129 | 0.1463 | 0.4067 | 0.4126 | 0.0 | nan | 0.2611 | 0.4535 | 0.3979 | 0.0 | 0.9440 | 0.9430 | 0.9801 | 0.1153 | 0.0838 | 0.3970 | 0.0 | 0.0 | 0.7971 | 0.8795 | 0.3698 | 0.7345 | 0.3286 | nan | 0.5408 | 0.4354 | 0.1101 | 0.7568 | 0.0463 | nan | 0.0 | 0.0 | 0.4965 | 0.0 | 0.0 | 0.7422 | 0.1314 | 0.3671 | 0.3160 | 0.0 | nan | 0.1858 | 0.3615 | 0.2297 | 0.0 | 0.8787 | 0.8337 | 0.9380 | 0.0751 | 0.0511 | 0.3184 | 0.0 | 28 |
148
  | 0.1959 | 0.5240 | 0.3521 | 0.4343 | 0.8771 | 0.0 | 0.8744 | 0.9749 | 0.4993 | 0.8143 | 0.3366 | nan | 0.6376 | 0.7596 | 0.1736 | 0.9184 | 0.0424 | nan | nan | 0.0002 | 0.7469 | 0.0 | 0.0 | 0.8957 | 0.1516 | 0.5852 | 0.4130 | 0.0 | nan | 0.2630 | 0.4594 | 0.3941 | 0.0 | 0.9456 | 0.9363 | 0.9789 | 0.1007 | 0.1004 | 0.4621 | 0.0 | 0.0 | 0.8044 | 0.8775 | 0.4029 | 0.7728 | 0.2994 | nan | 0.5577 | 0.4748 | 0.1513 | 0.7884 | 0.0378 | nan | 0.0 | 0.0002 | 0.5007 | 0.0 | 0.0 | 0.7343 | 0.1427 | 0.4677 | 0.3440 | 0.0 | nan | 0.1765 | 0.3631 | 0.2372 | 0.0 | 0.8800 | 0.8324 | 0.9400 | 0.0701 | 0.0558 | 0.3558 | 0.0 | 29 |
149
+ | 0.1941 | 0.5077 | 0.3541 | 0.4368 | 0.8760 | 0.0 | 0.8810 | 0.9600 | 0.5080 | 0.7956 | 0.4939 | nan | 0.5978 | 0.7554 | 0.1367 | 0.9179 | 0.0599 | nan | nan | 0.0 | 0.7221 | 0.0 | 0.0 | 0.8944 | 0.1175 | 0.6129 | 0.4032 | 0.0 | nan | 0.2636 | 0.4817 | 0.3849 | 0.0 | 0.9453 | 0.9461 | 0.9760 | 0.0649 | 0.1096 | 0.5128 | 0.0 | 0.0 | 0.7899 | 0.8798 | 0.3854 | 0.7444 | 0.3950 | nan | 0.5344 | 0.4893 | 0.1164 | 0.7795 | 0.0515 | nan | 0.0 | 0.0 | 0.5232 | 0.0 | 0.0 | 0.7498 | 0.1094 | 0.4920 | 0.3325 | 0.0 | nan | 0.2083 | 0.3678 | 0.2519 | 0.0 | 0.8789 | 0.8304 | 0.9409 | 0.0447 | 0.0765 | 0.3601 | 0.0 | 30 |
150
 
151
 
152
  ### Framework versions
tf_model.h5 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:914ed17528f45c4d5ce98b876f3907b453beafbe899fbc136f98f2f26335a023
3
  size 15167476
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a7f9ccf7e433a70910fd658c98f0ab7a0364fd6f7a8820ccee01666b6d084d35
3
  size 15167476