- conda create -n question1 python==3.7 -y
- conda activate question1
- bash setupfor1.sh
- `download train_best_loss.pth.tar file from here and place under ckpt1 directory.`
- `python testQ1.py path2imagefolder`
- `python evalQ1.py path2predicted_csv path2GroundTruth csv`
- The code assumes that GroundTruth.csv file will be identical in structure to the file provided for training.
- conda create -n question2 python==3.8 -y
- conda activate question2
- bash setupfor2.sh
`download model_best.pth from here and keep in ckpt2 dir in root`
- python convertDataForQ2.py path2test_projections.csv path2test_report.csv -1 path2ImageDir
- `please include the -1 Parameter as above`
- python testQ2.py path2image_dir
- python evalQ2.py path2image_dir path2GTreports.csv path2Predicted.csv
- conda activate question2
`download model_best.pth from here and keep in ckpt2 dir in root`
- python convertDataForQ2.py path2test_projections.csv path2test_report.csv 1 path2ImageDir
- `please include the 1 Parameter as above`
- python testQ3.py path2image_dir
- python evalQ3.py path2image_dir path2GTreports.csv path2Predicted.csv