Skip to content

A pytorch style, machine learning framework from scratch

Notifications You must be signed in to change notification settings

adithyabsk/simple_ml

Repository files navigation

simple_ml

Implementation of OpenCL for the simple_ml framework

Installation

Python: 3.6.13

macOS

brew install gcc@7 opencl-headers
sudo curl -o /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk/System/Library/Frameworks/OpenCL.framework/Headers/cl.hpp https://raw.githubusercontent.com/KhronosGroup/OpenCL-CLHPP/main/include/CL/opencl.hpp
pip install -r requirements.txt

Set up pre-commit

pre-commit runs some tooling before files are committed.

Developers

OpenCL

OpenCL C++ headers do not ship with macOS by default. This project is set up to run using a legacy version of the OpenCL API (1.1). It is running using compatibility mode to for example make kernels (cl::compatibility::make_kernel).

TODOs

  • Check if the project still passes all tests on Google Colab
    • Because of the CMakeLists.txt changes
  • Update Makefile to only add brew specific paths if on macOS
  • Find a way to only pass the DCMAKE flags in the Makefile when on MacOS
  • Two main bugs right now
    • cannot find apps folder
    • some Nonetype array error (I've seen this before, I just need to break out the debugger)
  • Add fixture to tests that downloads PTB and CIFAR to a data folder to run in unit tests

Acknowledgements

Based on the following course.

About

A pytorch style, machine learning framework from scratch

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published