- opencv
- numpy
- torch
- torchvision
- matplotlib
python main.py
- Code by default trains for 50 epochs, and runs inference on the first image of train dataset
- To visualize training data, uncomment line 119. Press 'q' to exit.
- To test forward pass on dummy data, to check output dimensions, uncomment line 120.
- Add and test different loss functions (Focal Loss, Dice Loss etc)
- Test different model architectures