SAIL - Simple AI Library
SAIL is a python package designed for speed and simplicity when developing and running deep learning models. Built on top of a c++ library with python bindings, SAIL is currently in development, changes are being released daily with new features and bug fixes.
SAIL's api is based on influences from PyTorch, Tensorflow, and NumPy. Implementation of different features and modules are influenced by PyTorch, NumPy, and Chainer
SAIL only works on linux for now.
Click here to view the documentation.
NOT IN ORDER
- Matrix ops
- Documentation
- More dtype support
- GPU support
- Tighter integration with python
Just run pip install sail-ml
Note: SAIL only works on linux
Right now, YOU MUST HAVE NVIDIA TOOLKIT AND AVX2 COMPATIBLE CPU TO RUN. This is being built to a specific system for now, and in the future that will no longer be and issue.
SAIL is built using GCC and G++ 8.0. To install, you need Python >= 3.7 and NumPy. Then just run
python setup.py install
That is a super simple explanation but the project is new so ¯_(ツ)_/¯
To learn more about making a contribution to SAIL, please see our Contribution page. No concrete policies are set yet, as the base of the project is still being setup.
SAIL is currently maintained solely by Tucker Siegel (Github Profile)
SAIL has a BSD License found in the LICENSE file.