llama.cpp
8afe2280 - CUDA: mul_mat_q=true llama_context_params default (#2912)

Commit

2 years ago

CUDA: mul_mat_q=true llama_context_params default (#2912)

References

#2912 - CUDA: mul_mat_q=true as default for llama_context_params

Author

JohannesGaessler

JohannesGaessler

Parents

Loading