segformer-b0-scene-parse-150-50
This model is a fine-tuned version of klentree/segformer-b0-scene-parse-150-20 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.9992
- Mean Iou: 0.1414
- Mean Accuracy: 0.5623
- Overall Accuracy: 0.2681
- Per Category Iou: [0.03415709043952773, 0.2486167920519975]
- Per Category Accuracy: [0.8749923648579864, 0.24956422933712585]
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 6e-05
- train_batch_size: 1
- eval_batch_size: 1
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 20
Training results
Training Loss | Epoch | Step | Validation Loss | Mean Iou | Mean Accuracy | Overall Accuracy | Per Category Iou | Per Category Accuracy |
---|---|---|---|---|---|---|---|---|
No log | 1.0 | 133 | 1.3809 | 0.1831 | 0.6100 | 0.3452 | [0.038710505981588275, 0.3274145847863719] | [0.8914028300926398, 0.32849851170741207] |
No log | 2.0 | 266 | 3.1399 | 0.0278 | 0.5071 | 0.0541 | [0.02998980464207664, 0.02561876180112447] | [0.9885405001866368, 0.025627711476269972] |
No log | 3.0 | 399 | 1.4873 | 0.1612 | 0.5930 | 0.3045 | [0.036857115032605504, 0.28551314243518766] | [0.8996386032780209, 0.28638666931766954] |
0.131 | 4.0 | 532 | 1.8598 | 0.0884 | 0.5443 | 0.1683 | [0.03248204249077747, 0.14439283966619093] | [0.9438817061997353, 0.14463986050373331] |
0.131 | 5.0 | 665 | 1.1300 | 0.2009 | 0.6049 | 0.3790 | [0.03869589759308316, 0.3630560542281187] | [0.8450235841053311, 0.3647712858322449] |
0.131 | 6.0 | 798 | 3.4677 | 0.0281 | 0.5067 | 0.0547 | [0.029968966006210712, 0.026287373046977498] | [0.9871950184940107, 0.02629763451787782] |
0.131 | 7.0 | 931 | 0.8035 | 0.3046 | 0.6395 | 0.5715 | [0.0468379925244376, 0.5622884059713553] | [0.7117632087956836, 0.567229146162601] |
0.1093 | 8.0 | 1064 | 3.0148 | 0.0372 | 0.5090 | 0.0717 | [0.03010087836424667, 0.04420013882895897] | [0.9738284298754624, 0.0442354032603611] |
0.1093 | 9.0 | 1197 | 1.3918 | 0.1977 | 0.6147 | 0.3726 | [0.03949337936321034, 0.35595848950467285] | [0.8720570769282975, 0.35734683993080635] |
0.1093 | 10.0 | 1330 | 0.8904 | 0.3048 | 0.6555 | 0.5705 | [0.04885666809307599, 0.5607852599315016] | [0.7458023685907225, 0.5651308780936614] |
0.1093 | 11.0 | 1463 | 2.1101 | 0.1091 | 0.5544 | 0.2072 | [0.03330984189332642, 0.18492787307291492] | [0.9234432793783298, 0.18535946088367358] |
0.0978 | 12.0 | 1596 | 0.3487 | 0.4522 | 0.6510 | 0.8320 | [0.07471728817756591, 0.8296494449285194] | [0.4586853982150735, 0.8433402139408958] |
0.0978 | 13.0 | 1729 | 1.0341 | 0.2902 | 0.6500 | 0.5437 | [0.0471396698931736, 0.5331783341429923] | [0.7630459126539754, 0.5370297478519739] |
0.0978 | 14.0 | 1862 | 8.8510 | 0.0191 | 0.5026 | 0.0378 | [0.029729543963973613, 0.008555530995168384] | [0.9966099969459432, 0.008556415154495053] |
0.0978 | 15.0 | 1995 | 1.5468 | 0.2099 | 0.6104 | 0.3958 | [0.039432935991880434, 0.38042719301704037] | [0.8384437883877973, 0.38230080082492807] |
0.0938 | 16.0 | 2128 | 0.3316 | 0.4622 | 0.6499 | 0.8478 | [0.07871638020749482, 0.8457624711227079] | [0.43967728799755673, 0.8602092195727595] |
0.0938 | 17.0 | 2261 | 6.8985 | 0.0200 | 0.5030 | 0.0394 | [0.02975190949797357, 0.010246433961607446] | [0.9957090501883336, 0.01024777428517577] |
0.0938 | 18.0 | 2394 | 3.6067 | 0.0460 | 0.5068 | 0.0885 | [0.02995386117285437, 0.06208246891194747] | [0.9514455868879161, 0.062174361643823145] |
0.092 | 19.0 | 2527 | 0.2255 | 0.4956 | 0.6091 | 0.9069 | [0.08506947955817622, 0.9060937507594746] | [0.29259221554854253, 0.9256338355411345] |
0.092 | 20.0 | 2660 | 1.9992 | 0.1414 | 0.5623 | 0.2681 | [0.03415709043952773, 0.2486167920519975] | [0.8749923648579864, 0.24956422933712585] |
Framework versions
- Transformers 4.37.0
- Pytorch 2.1.0+cu121
- Datasets 2.16.1
- Tokenizers 0.15.0
- Downloads last month
- 1
Unable to determine this model’s pipeline type. Check the
docs
.