search

harrisonvanderbyl / rwkv-cpp-accelerated

A torchless, c++ rwkv implementation using 8bit quantization, written in cuda/hip/vulkan for maximum compatibility and minimum dependencies
MIT License
306 stars 19 forks source link

Radeon Open Compute support #31

Open erkinalp opened 1 year ago

erkinalp commented 1 year ago
harrisonvanderbyl commented 1 year ago

https://github.com/harrisonvanderbyl/rwkv-cpp-cuda/blob/main/include/rwkv/cuda/rwkv.cu#L1-L12

Look, you can just compile the source code with a hip compiler and it works