Skip to content

Latest commit

 

History

History
22 lines (18 loc) · 2.28 KB

CONTRIBUTING.md

File metadata and controls

22 lines (18 loc) · 2.28 KB
Raw

How to contribute

  • Tensorflex is a library still in it's nascent stages, so we really appreciate third-party pull requests for improving it.
  • There are only very minimal guidelines that need to be followed by developers before submitting code.

Guidelines

  • Please open an issue if extant functionality is not working as expected with proper usage (or failure) examples, and mention the current Tensorflow C API version you are running along with your OS.
  • If you would like to submit a pull request as a fix for opened issues, refer to the issue in your PR and ensure that adding your code does not break any other tests.
  • If you are submitting a PR with new functionality, please write documentation adhering to the existing documentation style (read the docs here), and also write tests describing this functionality. You can also open an issue for new functionality to discuss your ideas with the contributors/maintainers.
  • If your PR contains C code please format it using clang-format: clang-format -i <your-C-code-file.c>

Key contribution areas

Tensorflex basically involves four main functionalities that can be improved upon:

  • Improving code for matrices, linear algebra, and numeric computation through NIFs [Intermediate]
  • Writing/using Python examples for pre-existing graphs (or creating your own graph models) and running them in Tensorflex, replete with the entire prediction pipeline [Beginner]
  • Porting over the Tensorflow C API's functionality in the form of NIFs that adds value to Tensorflex [Intermediate]
  • Adding low-level C code that will help improve upon the functionality missing in the Tensorflow C API and then porting those to Tensorflex using NIFs [Advanced]

Getting started

  • Pick any one of the four contribution areas, depending on your comfort level with NIFs, Elixir, or Tensorflow, and think of a contribution. Feel free to discuss your ideas with any of the contributors beforehand by opening an issue for new functionality.
  • Fork the repository, and then clone it.
  • Add the functionality locally, make sure tests pass (mix test), and write (then view) documentation (mix docs). If all is fine, push those commits to your branch and submit a PR to master.