mibaumgartner / hackathon_health

AI-HERO Health Challenge on energy efficient AI - Predicting Covid-19 from chest x-ray images

Prototyping repository for our solution to the AI-HERO Health Challenge Hackathon.

Setup Env

Run following commands from the root dir of this repository (tested with python=3.8):

pip install -r requirements.txt
pip install .

Running Inference

Inference can be run with the following command:

sbatch eval_conda.sh --save_dir [dir to save results] --csv_filepath [Filepath of csv] --data_dir [directory containing imgs of csv file] --ckpt_path [path to checkpoint]

Details:

1. save_dir: Directory where predictions will be saved to as 'predictions.csv'
2. data_dir: Directory to the folder containing the images of the csv file.
3. csv_filepath: Path to the csv file containing the "images" in the first column and the first row is the header
4. ckpt_path: Path to the checkpoint file loaded from Zenodo.