--- license: other base_model: nvidia/mit-b5 tags: - vision - image-segmentation - generated_from_trainer model-index: - name: SegFormer_Mixed_Set2_788images_mit-b5_RGB results: [] --- # SegFormer_Mixed_Set2_788images_mit-b5_RGB This model is a fine-tuned version of [nvidia/mit-b5](https://huggingface.co/nvidia/mit-b5) on the Hasano20/Mixed_Set2_788images dataset. It achieves the following results on the evaluation set: - Loss: 0.0179 - Mean Iou: 0.9757 - Mean Accuracy: 0.9872 - Overall Accuracy: 0.9938 - Accuracy Background: 0.9959 - Accuracy Melt: 0.9697 - Accuracy Substrate: 0.9959 - Iou Background: 0.9922 - Iou Melt: 0.9437 - Iou Substrate: 0.9911 ## 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: 0.0001 - train_batch_size: 8 - eval_batch_size: 8 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: cosine - lr_scheduler_warmup_steps: 100 - num_epochs: 50 ### Training results | Training Loss | Epoch | Step | Validation Loss | Mean Iou | Mean Accuracy | Overall Accuracy | Accuracy Background | Accuracy Melt | Accuracy Substrate | Iou Background | Iou Melt | Iou Substrate | |:-------------:|:-------:|:----:|:---------------:|:--------:|:-------------:|:----------------:|:-------------------:|:-------------:|:------------------:|:--------------:|:--------:|:-------------:| | 0.1292 | 0.7042 | 50 | 0.1861 | 0.7698 | 0.8223 | 0.9387 | 0.9844 | 0.5153 | 0.9673 | 0.9318 | 0.4625 | 0.9152 | | 0.1161 | 1.4085 | 100 | 0.1307 | 0.8463 | 0.9335 | 0.9543 | 0.9851 | 0.8721 | 0.9433 | 0.9596 | 0.6514 | 0.9279 | | 0.072 | 2.1127 | 150 | 0.0675 | 0.9075 | 0.9607 | 0.9762 | 0.9887 | 0.9179 | 0.9755 | 0.9821 | 0.7779 | 0.9625 | | 0.0425 | 2.8169 | 200 | 0.0622 | 0.9078 | 0.9322 | 0.9781 | 0.9868 | 0.8138 | 0.9959 | 0.9838 | 0.7746 | 0.9652 | | 0.0214 | 3.5211 | 250 | 0.0372 | 0.9458 | 0.9688 | 0.9868 | 0.9905 | 0.9223 | 0.9935 | 0.9870 | 0.8702 | 0.9802 | | 0.0397 | 4.2254 | 300 | 0.0373 | 0.9428 | 0.9802 | 0.9858 | 0.9948 | 0.9635 | 0.9824 | 0.9892 | 0.8617 | 0.9774 | | 0.0515 | 4.9296 | 350 | 0.0411 | 0.9399 | 0.9735 | 0.9846 | 0.9902 | 0.9438 | 0.9864 | 0.9865 | 0.8583 | 0.9750 | | 0.0171 | 5.6338 | 400 | 0.0267 | 0.9587 | 0.9782 | 0.9898 | 0.9937 | 0.9477 | 0.9931 | 0.9900 | 0.9017 | 0.9843 | | 0.0274 | 6.3380 | 450 | 0.0262 | 0.9621 | 0.9780 | 0.9906 | 0.9935 | 0.9454 | 0.9951 | 0.9900 | 0.9107 | 0.9857 | | 0.0105 | 7.0423 | 500 | 0.0272 | 0.9597 | 0.9844 | 0.9900 | 0.9924 | 0.9695 | 0.9913 | 0.9898 | 0.9041 | 0.9852 | | 0.0143 | 7.7465 | 550 | 0.0250 | 0.9638 | 0.9824 | 0.9911 | 0.9946 | 0.9593 | 0.9931 | 0.9907 | 0.9142 | 0.9865 | | 0.0153 | 8.4507 | 600 | 0.0226 | 0.9670 | 0.9826 | 0.9918 | 0.9947 | 0.9585 | 0.9946 | 0.9909 | 0.9223 | 0.9878 | | 0.011 | 9.1549 | 650 | 0.0201 | 0.9711 | 0.9841 | 0.9926 | 0.9936 | 0.9622 | 0.9965 | 0.9908 | 0.9330 | 0.9893 | | 0.009 | 9.8592 | 700 | 0.0199 | 0.9707 | 0.9858 | 0.9926 | 0.9962 | 0.9676 | 0.9936 | 0.9913 | 0.9315 | 0.9891 | | 0.017 | 10.5634 | 750 | 0.0206 | 0.9692 | 0.9869 | 0.9923 | 0.9964 | 0.9723 | 0.9921 | 0.9911 | 0.9279 | 0.9886 | | 0.0095 | 11.2676 | 800 | 0.0184 | 0.9733 | 0.9870 | 0.9933 | 0.9954 | 0.9704 | 0.9950 | 0.9917 | 0.9379 | 0.9902 | | 0.0142 | 11.9718 | 850 | 0.0179 | 0.9740 | 0.9862 | 0.9935 | 0.9957 | 0.9671 | 0.9957 | 0.9919 | 0.9395 | 0.9905 | | 0.0134 | 12.6761 | 900 | 0.0180 | 0.9739 | 0.9882 | 0.9934 | 0.9948 | 0.9747 | 0.9951 | 0.9919 | 0.9394 | 0.9903 | | 0.0096 | 13.3803 | 950 | 0.0179 | 0.9744 | 0.9864 | 0.9936 | 0.9960 | 0.9675 | 0.9956 | 0.9922 | 0.9406 | 0.9905 | | 0.0089 | 14.0845 | 1000 | 0.0174 | 0.9744 | 0.9881 | 0.9936 | 0.9958 | 0.9737 | 0.9949 | 0.9922 | 0.9404 | 0.9908 | | 0.0094 | 14.7887 | 1050 | 0.0174 | 0.9754 | 0.9864 | 0.9938 | 0.9962 | 0.9671 | 0.9960 | 0.9924 | 0.9428 | 0.9911 | | 0.0089 | 15.4930 | 1100 | 0.0192 | 0.9748 | 0.9860 | 0.9935 | 0.9945 | 0.9666 | 0.9968 | 0.9918 | 0.9421 | 0.9905 | | 0.0087 | 16.1972 | 1150 | 0.0179 | 0.9757 | 0.9872 | 0.9938 | 0.9959 | 0.9697 | 0.9959 | 0.9922 | 0.9437 | 0.9911 | ### Framework versions - Transformers 4.41.2 - Pytorch 2.0.1+cu117 - Datasets 2.19.2 - Tokenizers 0.19.1