segformer-b0-finetuned-segments-greenhouse-oct-23
This model is a fine-tuned version of nvidia/mit-b0 on the MexicanVanGogh/greenhouse dataset. It achieves the following results on the evaluation set:
- Loss: 0.7058
- Mean Iou: 0.2227
- Mean Accuracy: 0.2804
- Overall Accuracy: 0.9101
- Accuracy Unlabeled: nan
- Accuracy Object: nan
- Accuracy Road: 0.9378
- Accuracy Plant: 0.9667
- Accuracy Iron: 0.0
- Accuracy Wood: 0.0
- Accuracy Wall: 0.1932
- Accuracy Raw Road: nan
- Accuracy Bottom Wall: 0.0
- Accuracy Roof: 0.1457
- Accuracy Grass: 0.0
- Iou Unlabeled: nan
- Iou Object: nan
- Iou Road: 0.9039
- Iou Plant: 0.8421
- Iou Iron: 0.0
- Iou Wood: 0.0
- Iou Wall: 0.1521
- Iou Raw Road: 0.0
- Iou Bottom Wall: 0.0
- Iou Roof: 0.1061
- Iou Grass: 0.0
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: 2
- eval_batch_size: 2
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 30
Training results
Training Loss | Epoch | Step | Validation Loss | Mean Iou | Mean Accuracy | Overall Accuracy | Accuracy Unlabeled | Accuracy Object | Accuracy Road | Accuracy Plant | Accuracy Iron | Accuracy Wood | Accuracy Wall | Accuracy Raw Road | Accuracy Bottom Wall | Accuracy Roof | Accuracy Grass | Iou Unlabeled | Iou Object | Iou Road | Iou Plant | Iou Iron | Iou Wood | Iou Wall | Iou Raw Road | Iou Bottom Wall | Iou Roof | Iou Grass |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
1.8756 | 2.86 | 20 | 2.0063 | 0.1415 | 0.2269 | 0.8216 | nan | nan | 0.7882 | 0.9674 | 0.0 | 0.0594 | 0.0 | nan | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.7760 | 0.7552 | 0.0 | 0.0256 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
1.3624 | 5.71 | 40 | 1.0910 | 0.1715 | 0.2380 | 0.8991 | nan | nan | 0.9206 | 0.9757 | 0.0 | 0.0077 | 0.0 | nan | 0.0 | 0.0 | 0.0 | 0.0 | nan | 0.8888 | 0.8220 | 0.0 | 0.0045 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
1.4095 | 8.57 | 60 | 0.9033 | 0.1734 | 0.2392 | 0.9068 | nan | nan | 0.9264 | 0.9873 | 0.0 | 0.0 | 0.0 | nan | 0.0 | 0.0 | 0.0 | 0.0 | nan | 0.9000 | 0.8338 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
0.8802 | 11.43 | 80 | 0.7784 | 0.1764 | 0.2414 | 0.9165 | nan | nan | 0.9470 | 0.9823 | 0.0 | 0.0022 | 0.0 | nan | 0.0 | 0.0 | 0.0 | 0.0 | nan | 0.9155 | 0.8463 | 0.0 | 0.0021 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
1.0936 | 14.29 | 100 | 0.8060 | 0.1946 | 0.2405 | 0.9132 | nan | nan | 0.9400 | 0.9839 | 0.0 | 0.0 | 0.0 | nan | 0.0 | 0.0 | 0.0 | nan | nan | 0.9100 | 0.8418 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
0.8086 | 17.14 | 120 | 0.7786 | 0.1940 | 0.2402 | 0.9115 | nan | nan | 0.9361 | 0.9852 | 0.0 | 0.0 | 0.0 | nan | 0.0 | 0.0006 | 0.0 | nan | nan | 0.9071 | 0.8380 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0006 | 0.0 |
1.0669 | 20.0 | 140 | 0.7462 | 0.2072 | 0.2562 | 0.9088 | nan | nan | 0.9282 | 0.9853 | 0.0 | 0.0 | 0.0113 | nan | 0.0 | 0.1246 | 0.0 | nan | nan | 0.9010 | 0.8385 | 0.0 | 0.0 | 0.0102 | 0.0 | 0.0 | 0.1155 | 0.0 |
0.7399 | 22.86 | 160 | 0.7328 | 0.2137 | 0.2662 | 0.9080 | nan | nan | 0.9290 | 0.9788 | 0.0 | 0.0 | 0.0814 | nan | 0.0 | 0.1405 | 0.0 | nan | nan | 0.8997 | 0.8389 | 0.0 | 0.0 | 0.0663 | 0.0 | 0.0 | 0.1181 | 0.0 |
0.808 | 25.71 | 180 | 0.7296 | 0.2218 | 0.2797 | 0.9072 | nan | nan | 0.9277 | 0.9742 | 0.0 | 0.0 | 0.1840 | nan | 0.0 | 0.1515 | 0.0 | nan | nan | 0.8981 | 0.8404 | 0.0 | 0.0 | 0.1423 | 0.0 | 0.0 | 0.1155 | 0.0 |
0.8494 | 28.57 | 200 | 0.7058 | 0.2227 | 0.2804 | 0.9101 | nan | nan | 0.9378 | 0.9667 | 0.0 | 0.0 | 0.1932 | nan | 0.0 | 0.1457 | 0.0 | nan | nan | 0.9039 | 0.8421 | 0.0 | 0.0 | 0.1521 | 0.0 | 0.0 | 0.1061 | 0.0 |
Framework versions
- Transformers 4.34.1
- Pytorch 2.1.0+cu118
- Datasets 2.14.5
- Tokenizers 0.14.1
- Downloads last month
- 8
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.
Model tree for MexicanVanGogh/segformer-b0-finetuned-segments-greenhouse-oct-23
Base model
nvidia/mit-b0