search

ggerganov / llama.cpp

LLM inference in C/C++
MIT License
60.95k stars 8.7k forks source link

ggml : add GGML_CUDA_USE_GRAPHS option, restore GGML_CUDA_FORCE_CUBLAS #8140

Closed slaren closed 2 days ago