llama.cpp
f514d1b3 - CUDA: faster k-quant mul_mat_q kernels (#2525)

Commit

2 years ago

CUDA: faster k-quant mul_mat_q kernels (#2525)

References

#2525 - CUDA: faster k-quant mul_mat_q kernels

Author

JohannesGaessler

JohannesGaessler

Parents

Loading