Hugging Face's logo

Ventalik
/
segformer-roof

Transformers
PyTorch
segformer
Generated from Trainer
Inference Endpoints
Model card Files Community
Ventalik commited on
Commit
5be0603
•
1 Parent(s): 40b3fb3

End of training

Browse files
Files changed (1) hide show
  1. README.md +19 -19
README.md CHANGED
@@ -15,12 +15,12 @@ should probably proofread and complete it, then remove this comment. -->
15
 
16
  This model is a fine-tuned version of [nvidia/mit-b3](https://huggingface.co/nvidia/mit-b3) on the None dataset.
17
  It achieves the following results on the evaluation set:
18
- - Loss: 0.1489
19
- - Mean Iou: 0.5505
20
- - Mean Accuracy: 0.6156
21
- - Overall Accuracy: 0.9466
22
- - Per Category Iou: [0.9462854167649989, 0.4158813402104387, 0.5888025421987559, 0.2511892496218515]
23
- - Per Category Accuracy: [0.9833455188401141, 0.5148994995987893, 0.676924986222792, 0.2870358541186019]
24
 
25
  ## Model description
26
 
@@ -39,7 +39,7 @@ More information needed
39
  ### Training hyperparameters
40
 
41
  The following hyperparameters were used during training:
42
- - learning_rate: 3e-05
43
  - train_batch_size: 8
44
  - eval_batch_size: 8
45
  - seed: 42
@@ -49,18 +49,18 @@ The following hyperparameters were used during training:
49
 
50
  ### Training results
51
 
52
- | Training Loss | Epoch | Step | Validation Loss | Mean Iou | Mean Accuracy | Overall Accuracy | Per Category Iou | Per Category Accuracy |
53
- |:-------------:|:-----:|:----:|:---------------:|:--------:|:-------------:|:----------------:|:----------------------------------------------------------------------------------:|:-----------------------------------------------------------------------------------:|
54
- | 0.2661 | 1.0 | 930 | 0.1927 | 0.4344 | 0.4872 | 0.9342 | [0.9344825418450485, 0.26395183336511713, 0.5002992114037405, 0.03903434157031228] | [0.9830912179324156, 0.30179383206489235, 0.6231286015017626, 0.040693785837209825] |
55
- | 0.1861 | 2.0 | 1860 | 0.1771 | 0.4734 | 0.5249 | 0.9386 | [0.9386647658288735, 0.3531834631146987, 0.5069223720303438, 0.0946611168820477] | [0.9840564875171536, 0.44512255737917406, 0.5725907870810009, 0.09793061640627153] |
56
- | 0.1687 | 3.0 | 2790 | 0.1698 | 0.4767 | 0.5130 | 0.9406 | [0.9402758185411285, 0.29429541402749515, 0.5250142510383131, 0.14727973718586107] | [0.99111024766829, 0.3301444083968103, 0.5794189494390111, 0.15138489132041252] |
57
- | 0.1613 | 4.0 | 3720 | 0.1647 | 0.5172 | 0.5770 | 0.9419 | [0.9422327506285707, 0.34453359555127, 0.5465939758230205, 0.23538538010815044] | [0.9851811183921607, 0.4048120021595627, 0.644126865023268, 0.2740198546596827] |
58
- | 0.1525 | 5.0 | 4650 | 0.1590 | 0.5292 | 0.5928 | 0.9432 | [0.9431927537543411, 0.38474057129503686, 0.5631121296564262, 0.22593269100097432] | [0.9819667553267772, 0.474550498451453, 0.6688971173848546, 0.24571080733765696] |
59
- | 0.1457 | 6.0 | 5580 | 0.1589 | 0.5390 | 0.6114 | 0.9433 | [0.94302864691369, 0.4180438013881352, 0.5733125443615176, 0.22156761999995392] | [0.9783425645654029, 0.5634737323691095, 0.6628879372323019, 0.24089368513605927] |
60
- | 0.1433 | 7.0 | 6510 | 0.1540 | 0.5457 | 0.6187 | 0.9444 | [0.9443213253119109, 0.42535580593183475, 0.5784090659962025, 0.23479628926382032] | [0.9791198670671988, 0.5719683051771836, 0.6633544698685145, 0.26019974899862475] |
61
- | 0.1392 | 8.0 | 7440 | 0.1537 | 0.5498 | 0.6256 | 0.9451 | [0.9449256983807501, 0.42292996141110306, 0.5822954997871862, 0.24924828811831987] | [0.9800876046026772, 0.5468931021964158, 0.6790202671476468, 0.29652230330084345] |
62
- | 0.135 | 9.0 | 8370 | 0.1523 | 0.5500 | 0.6217 | 0.9456 | [0.9454458420307357, 0.4169163774539352, 0.5848382526881205, 0.25277234460532344] | [0.9813638665751383, 0.5271156382106387, 0.6824384642647543, 0.29588853735336335] |
63
- | 0.135 | 10.0 | 9300 | 0.1489 | 0.5505 | 0.6156 | 0.9466 | [0.9462854167649989, 0.4158813402104387, 0.5888025421987559, 0.2511892496218515] | [0.9833455188401141, 0.5148994995987893, 0.676924986222792, 0.2870358541186019] |
64
 
65
 
66
  ### Framework versions
 
15
 
16
  This model is a fine-tuned version of [nvidia/mit-b3](https://huggingface.co/nvidia/mit-b3) on the None dataset.
17
  It achieves the following results on the evaluation set:
18
+ - Loss: 0.1331
19
+ - Mean Iou: 0.6705
20
+ - Mean Accuracy: 0.7459
21
+ - Overall Accuracy: 0.9509
22
+ - Per Category Iou: [0.9501174871140823, 0.44914356298751956, 0.612314004780354]
23
+ - Per Category Accuracy: [0.9835400780911927, 0.5547023488814158, 0.6996065789512536]
24
 
25
  ## Model description
26
 
 
39
  ### Training hyperparameters
40
 
41
  The following hyperparameters were used during training:
42
+ - learning_rate: 6e-05
43
  - train_batch_size: 8
44
  - eval_batch_size: 8
45
  - seed: 42
 
49
 
50
  ### Training results
51
 
52
+ | Training Loss | Epoch | Step | Validation Loss | Mean Iou | Mean Accuracy | Overall Accuracy | Per Category Iou | Per Category Accuracy |
53
+ |:-------------:|:-----:|:----:|:---------------:|:--------:|:-------------:|:----------------:|:-------------------------------------------------------------:|:-------------------------------------------------------------:|
54
+ | 0.2357 | 1.0 | 930 | 0.1790 | 0.5725 | 0.6427 | 0.9370 | [0.9372296551214157, 0.265369517396097, 0.5150004888948623] | [0.9831748407739025, 0.30066901406516805, 0.6442311056295564] |
55
+ | 0.1732 | 2.0 | 1860 | 0.1700 | 0.5944 | 0.6623 | 0.9398 | [0.9401268036898566, 0.32807672100401025, 0.5150202791005837] | [0.9839554901228429, 0.3979733503813561, 0.6048359409441622] |
56
+ | 0.1559 | 3.0 | 2790 | 0.1600 | 0.6127 | 0.6753 | 0.9429 | [0.9427325757540327, 0.35965960312996503, 0.5357965624353673] | [0.9857619070576414, 0.448355090982288, 0.5917792010373596] |
57
+ | 0.1482 | 4.0 | 3720 | 0.1550 | 0.6070 | 0.6703 | 0.9437 | [0.944172393520638, 0.32282759103373876, 0.553957021364391] | [0.9867471837507854, 0.3693595842127839, 0.6546589242888398] |
58
+ | 0.1388 | 5.0 | 4650 | 0.1459 | 0.6224 | 0.6804 | 0.9463 | [0.9459092954263936, 0.3388316205746287, 0.582315028654454] | [0.9880828845164424, 0.3845685735297591, 0.6686608983285444] |
59
+ | 0.1311 | 6.0 | 5580 | 0.1462 | 0.6577 | 0.7466 | 0.9468 | [0.9461722360255241, 0.43412167635821636, 0.5928671376046599] | [0.9789455448939886, 0.5812880374397429, 0.679673538106026] |
60
+ | 0.1279 | 7.0 | 6510 | 0.1423 | 0.6611 | 0.7569 | 0.9469 | [0.9465891089044499, 0.4381184809600582, 0.5986930952368954] | [0.9773687408863051, 0.6022321705637107, 0.6909579192751797] |
61
+ | 0.1232 | 8.0 | 7440 | 0.1388 | 0.6682 | 0.7548 | 0.9491 | [0.9484426711464405, 0.44975791193706466, 0.6064948465370358] | [0.9802070573066378, 0.5898759789294347, 0.6944388397098907] |
62
+ | 0.1175 | 9.0 | 8370 | 0.1353 | 0.6665 | 0.7392 | 0.9505 | [0.9497153698000098, 0.44248964964215, 0.6074386624389524] | [0.9841990715683068, 0.5444332218950657, 0.6888275633995394] |
63
+ | 0.1174 | 10.0 | 9300 | 0.1331 | 0.6705 | 0.7459 | 0.9509 | [0.9501174871140823, 0.44914356298751956, 0.612314004780354] | [0.9835400780911927, 0.5547023488814158, 0.6996065789512536] |
64
 
65
 
66
  ### Framework versions