llama.cpp
468ea24f - CUDA: faster non k-quant mul_mat_q kernels (#2483)

Commit

2 years ago

CUDA: faster non k-quant mul_mat_q kernels (#2483)

References

#2483 - CUDA: faster non k-quant mul_mat_q kernels

Author

JohannesGaessler

JohannesGaessler

Parents

Loading