--- library_name: transformers license: other base_model: nvidia/mit-b0 tags: - image-segmentation - vision - generated_from_trainer datasets: - generator model-index: - name: autocrop-av-abm results: [] --- # autocrop-av-abm This model is a fine-tuned version of [nvidia/mit-b0](https://huggingface.co/nvidia/mit-b0) on the /mnt/disk1/autocrop-data/datasets/av-abm dataset. It achieves the following results on the evaluation set: - Loss: 0.0146 - Mean Iou: 0.4940 - Mean Accuracy: 0.9880 - Overall Accuracy: 0.9880 - Accuracy Background: nan - Accuracy Crop: 0.9880 - Iou Background: 0.0 - Iou Crop: 0.9880 ## 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: 8 - eval_batch_size: 8 - seed: 42 - optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments - lr_scheduler_type: cosine - lr_scheduler_warmup_steps: 0.1 - num_epochs: 50.0 - mixed_precision_training: Native AMP ### Training results | Training Loss | Epoch | Step | Validation Loss | Mean Iou | Mean Accuracy | Overall Accuracy | Accuracy Background | Accuracy Crop | Iou Background | Iou Crop | |:-------------:|:-----:|:-----:|:---------------:|:--------:|:-------------:|:----------------:|:-------------------:|:-------------:|:--------------:|:--------:| | 0.1919 | 1.0 | 225 | 0.1796 | 0.4930 | 0.9859 | 0.9859 | nan | 0.9859 | 0.0 | 0.9859 | | 0.1073 | 2.0 | 450 | 0.1057 | 0.4895 | 0.9790 | 0.9790 | nan | 0.9790 | 0.0 | 0.9790 | | 0.0572 | 3.0 | 675 | 0.0490 | 0.4853 | 0.9706 | 0.9706 | nan | 0.9706 | 0.0 | 0.9706 | | 0.0375 | 4.0 | 900 | 0.0371 | 0.4902 | 0.9804 | 0.9804 | nan | 0.9804 | 0.0 | 0.9804 | | 0.0440 | 5.0 | 1125 | 0.0316 | 0.4919 | 0.9837 | 0.9837 | nan | 0.9837 | 0.0 | 0.9837 | | 0.0228 | 6.0 | 1350 | 0.0250 | 0.4902 | 0.9803 | 0.9803 | nan | 0.9803 | 0.0 | 0.9803 | | 0.0294 | 7.0 | 1575 | 0.0233 | 0.4873 | 0.9746 | 0.9746 | nan | 0.9746 | 0.0 | 0.9746 | | 0.0213 | 8.0 | 1800 | 0.0209 | 0.4931 | 0.9863 | 0.9863 | nan | 0.9863 | 0.0 | 0.9863 | | 0.0215 | 9.0 | 2025 | 0.0220 | 0.4940 | 0.9881 | 0.9881 | nan | 0.9881 | 0.0 | 0.9881 | | 0.0186 | 10.0 | 2250 | 0.0180 | 0.4924 | 0.9849 | 0.9849 | nan | 0.9849 | 0.0 | 0.9849 | | 0.0195 | 11.0 | 2475 | 0.0186 | 0.4927 | 0.9855 | 0.9855 | nan | 0.9855 | 0.0 | 0.9855 | | 0.0136 | 12.0 | 2700 | 0.0174 | 0.4917 | 0.9835 | 0.9835 | nan | 0.9835 | 0.0 | 0.9835 | | 0.0197 | 13.0 | 2925 | 0.0182 | 0.4949 | 0.9898 | 0.9898 | nan | 0.9898 | 0.0 | 0.9898 | | 0.0141 | 14.0 | 3150 | 0.0164 | 0.4931 | 0.9861 | 0.9861 | nan | 0.9861 | 0.0 | 0.9861 | | 0.0138 | 15.0 | 3375 | 0.0169 | 0.4917 | 0.9835 | 0.9835 | nan | 0.9835 | 0.0 | 0.9835 | | 0.0123 | 16.0 | 3600 | 0.0175 | 0.4926 | 0.9852 | 0.9852 | nan | 0.9852 | 0.0 | 0.9852 | | 0.0171 | 17.0 | 3825 | 0.0166 | 0.4920 | 0.9839 | 0.9839 | nan | 0.9839 | 0.0 | 0.9839 | | 0.0135 | 18.0 | 4050 | 0.0159 | 0.4930 | 0.9860 | 0.9860 | nan | 0.9860 | 0.0 | 0.9860 | | 0.0155 | 19.0 | 4275 | 0.0151 | 0.4931 | 0.9862 | 0.9862 | nan | 0.9862 | 0.0 | 0.9862 | | 0.0141 | 20.0 | 4500 | 0.0151 | 0.4936 | 0.9871 | 0.9871 | nan | 0.9871 | 0.0 | 0.9871 | | 0.0174 | 21.0 | 4725 | 0.0151 | 0.4935 | 0.9870 | 0.9870 | nan | 0.9870 | 0.0 | 0.9870 | | 0.0125 | 22.0 | 4950 | 0.0155 | 0.4936 | 0.9871 | 0.9871 | nan | 0.9871 | 0.0 | 0.9871 | | 0.0115 | 23.0 | 5175 | 0.0157 | 0.4926 | 0.9852 | 0.9852 | nan | 0.9852 | 0.0 | 0.9852 | | 0.0124 | 24.0 | 5400 | 0.0156 | 0.4933 | 0.9866 | 0.9866 | nan | 0.9866 | 0.0 | 0.9866 | | 0.0122 | 25.0 | 5625 | 0.0149 | 0.4934 | 0.9867 | 0.9867 | nan | 0.9867 | 0.0 | 0.9867 | | 0.0116 | 26.0 | 5850 | 0.0164 | 0.4922 | 0.9844 | 0.9844 | nan | 0.9844 | 0.0 | 0.9844 | | 0.0122 | 27.0 | 6075 | 0.0146 | 0.4940 | 0.9880 | 0.9880 | nan | 0.9880 | 0.0 | 0.9880 | | 0.0085 | 28.0 | 6300 | 0.0161 | 0.4932 | 0.9864 | 0.9864 | nan | 0.9864 | 0.0 | 0.9864 | | 0.0122 | 29.0 | 6525 | 0.0151 | 0.4925 | 0.9851 | 0.9851 | nan | 0.9851 | 0.0 | 0.9851 | | 0.0115 | 30.0 | 6750 | 0.0165 | 0.4930 | 0.9861 | 0.9861 | nan | 0.9861 | 0.0 | 0.9861 | | 0.0117 | 31.0 | 6975 | 0.0160 | 0.4924 | 0.9848 | 0.9848 | nan | 0.9848 | 0.0 | 0.9848 | | 0.0105 | 32.0 | 7200 | 0.0161 | 0.4941 | 0.9883 | 0.9883 | nan | 0.9883 | 0.0 | 0.9883 | | 0.0113 | 33.0 | 7425 | 0.0155 | 0.4930 | 0.9859 | 0.9859 | nan | 0.9859 | 0.0 | 0.9859 | | 0.0089 | 34.0 | 7650 | 0.0161 | 0.4937 | 0.9873 | 0.9873 | nan | 0.9873 | 0.0 | 0.9873 | | 0.0114 | 35.0 | 7875 | 0.0157 | 0.4939 | 0.9877 | 0.9877 | nan | 0.9877 | 0.0 | 0.9877 | | 0.0094 | 36.0 | 8100 | 0.0161 | 0.4943 | 0.9887 | 0.9887 | nan | 0.9887 | 0.0 | 0.9887 | | 0.0082 | 37.0 | 8325 | 0.0162 | 0.4936 | 0.9871 | 0.9871 | nan | 0.9871 | 0.0 | 0.9871 | | 0.0081 | 38.0 | 8550 | 0.0161 | 0.4927 | 0.9853 | 0.9853 | nan | 0.9853 | 0.0 | 0.9853 | | 0.0090 | 39.0 | 8775 | 0.0169 | 0.4931 | 0.9862 | 0.9862 | nan | 0.9862 | 0.0 | 0.9862 | | 0.0084 | 40.0 | 9000 | 0.0161 | 0.4934 | 0.9868 | 0.9868 | nan | 0.9868 | 0.0 | 0.9868 | | 0.0115 | 41.0 | 9225 | 0.0158 | 0.4935 | 0.9870 | 0.9870 | nan | 0.9870 | 0.0 | 0.9870 | | 0.0080 | 42.0 | 9450 | 0.0162 | 0.4935 | 0.9870 | 0.9870 | nan | 0.9870 | 0.0 | 0.9870 | | 0.0097 | 43.0 | 9675 | 0.0167 | 0.4937 | 0.9874 | 0.9874 | nan | 0.9874 | 0.0 | 0.9874 | | 0.0082 | 44.0 | 9900 | 0.0166 | 0.4934 | 0.9867 | 0.9867 | nan | 0.9867 | 0.0 | 0.9867 | | 0.0097 | 45.0 | 10125 | 0.0168 | 0.4936 | 0.9872 | 0.9872 | nan | 0.9872 | 0.0 | 0.9872 | | 0.0076 | 46.0 | 10350 | 0.0167 | 0.4933 | 0.9866 | 0.9866 | nan | 0.9866 | 0.0 | 0.9866 | | 0.0085 | 47.0 | 10575 | 0.0164 | 0.4936 | 0.9873 | 0.9873 | nan | 0.9873 | 0.0 | 0.9873 | | 0.0090 | 48.0 | 10800 | 0.0164 | 0.4935 | 0.9870 | 0.9870 | nan | 0.9870 | 0.0 | 0.9870 | | 0.0081 | 49.0 | 11025 | 0.0164 | 0.4935 | 0.9869 | 0.9869 | nan | 0.9869 | 0.0 | 0.9869 | | 0.0099 | 50.0 | 11250 | 0.0165 | 0.4935 | 0.9869 | 0.9869 | nan | 0.9869 | 0.0 | 0.9869 | ### Framework versions - Transformers 5.8.0 - Pytorch 2.11.0+cu130 - Datasets 4.8.5 - Tokenizers 0.22.2