ggerganov / whisper.cpp

Port of OpenAI's Whisper model in C/C++
MIT License
35.38k stars 3.61k forks source link

Added CMake changes for building HIP/ROCm CUDA for AMD GPUs. #2102

Closed aldorof closed 5 months ago

aldorof commented 6 months ago

The same way it's realised on llama.cpp. So that one can build CUDA BLAS for HIP with something like this:

CXX=hipcc CC=hipcc cmake -DWHISPER_HIPBLAS=ON -DAMDGPU_TARGETS="gfx1100;gfx1101;gfx1102" -DBUILD_SHARED_LIBS=ON ../whisper.cpp