This project came with actual global situation and reciently release X-ray dataset.
The baseline models uses only covid-chestxray-dataset which is rather small (79 images at the time of writing).
They are also accessible with following python packages:
Download the project:
git clone https://github.com/PyTorchLightning/lightning-Covid19.git
Setup your development environment:
From Python Virtual Environments
python3.6 -m venv venv # from repo root, use python3.6+ source venv/bin/activate pip install -r requirements
To install all the needed dependencies, or update the conda environment:
Then to activate the conda environment:
- You are good to go to run your experiments -> See section Experiments.
See the Experiments on Wiki.
Anyone is welcome to contribute and use this project!
The project's features, bugs, and questions are discussed and planned in the project's Github issue's.
For informal, longer discussion we use PyTorch-Lightning (PL) Slack Workspace. See how to join it - See point 4 in Section "Asking for Help" in PL README.
For easier combination and coordination we are using separate channel in PL Slack. Join our open-source community!
Notes on contributing code and results:
- when mering PR, do "Squash & merge" so in master is only one commit and does not overlap with other commits from different PRs in parallel
- for writing particular result or how to launch anything, lets ise Wiki for now as it much more flexible