is a toolkit for delightful deep learning with PyTorch in public clouds.
Kaen simplifies
- setup of PyTorch distributed training environments in AWS, Azure, and GCP
- containerization of PyTorch models using Docker
- integration of Optuna hyperparameter tuning with PyTorch
- experiment tracking with MLFlow
- Kaen (็ซ็) means ๐ฅflame๐ฅ in Japanese.
- Kaen is all in on ๐open-source๐ with the strongest copyleft license
- Kaen can neither confirm nor deny a relationship to a ๐ถPolish rapper๐ถ