PyTorch implementation of IterNet, based on paper IterNet: Retinal Image Segmentation Utilizing Structural Redundancy in Vessel Networks (Li et al., 2019) and accompanying code.
python main.py --param param_valueThe following hyperparameters can also be provided. Smallest model from paper is shown for comparison.
| Argument | Default |
|---|---|
--gpus | 4,5,6 |
--size. | 592 |
--crop_size | 128 |
--image_dir | '' |
--mask_dir | '' |
--train_csv | '' |
--val_csv | '' |
--lr | 0.0001 |
--epochs | 2 |
--batch_size | 32 |
--model_dir | '' |
--seed | 2020123 |