Helpful Commands

1. Create conda environment (if needed)
   • conda env create -f environment.yml
2. Activate conda environment
   • conda activate pact
3. Set up SSH for GitHub push/pull access
   • source ssh.sh
   • Note: bash ssh.sh command will run in sub-shell and not give permissions in outer shell
4. Test model configuration before training
   • python train.py --size_check [other_args]
5. Train/test model(s)
   • Modify train_test.sh as needed
   • bash train_test.sh
6. Launch Tensorboard
   • Open new terminal
   • conda activate aht
   • tensorboard --logdir /scratch/zb7df/checkpoints/PACT
7. Run benchmarking
   • source benchmark_setup.sh
   • bash benchmark.sh test (Use 'test', 'validation', or 'test2')

Important Notes:

• train_test.sh script contains 2-stage training process for I/Q and Amp loss
• The NGA dataset used here includes 9,000 training patches, 1,000 validation patches, 1,000 256x256 test patches, and 2 1024x1024 test images.

Overleaf Projects:

• TBD