A tool for reproducibly benchmarking machine learning models.
git clone https://github.com/gautomdas/modelbench
cd modelbench
npm install
npm run electron-dev
Make sure you have your python environment setup from the shell you are running Electron from. In my case, I have to run conda activate before running the app so that the models find their dependencies and run properly.
- Test + build on Windows
- Distribute binaries
- Create CI/Release system
- Update website for releases
- Fix warnings
- Convert from mostly React local state to redux
- Build backend and analysis pages