llama.cpp
c63bb1d1 - CUDA: use mul_mat_q kernels by default (#2683)

Commit

2 years ago

CUDA: use mul_mat_q kernels by default (#2683)

References

#2683 - CUDA: use mul_mat_q kernels by default

Author

JohannesGaessler

JohannesGaessler

Parents

Loading