Fork of PyTorch SimCLR adapted for pre-training VGG19 weights for usage in Medical Out-of-Distribution Analysis Challenge (MOOD).
See original code and documentation in https://github.com/sthalles/SimCLR
See a description of our solution to the MOOD Challenge in https://github.com/ninatu/mood_challenge
- Install the anomaly detection framework
pip install git+https://github.com/ninatu/mood_challenge.git
- Perform "Data Preparation" step. See git@github.com:ninatu/mood_challenge.git
- Put correct paths in
config/exp_1_mood.yaml
and run
python run.py configs/exp_1_mood_tmp.yaml