CUDA: refactor mmq, dmmv, mmvq (#7716)
* CUDA: refactor mmq, dmmv, mmvq * fix out-of-bounds write * struct for qk, qr, qi * fix cmake build * mmq_type_traits
J
Johannes Gäßler committed
7d1a378b8fb266782d9248538a661405aad80768
Parent: 2b33896
Committed by GitHub <noreply@github.com>
on 6/5/2024, 2:53:00 PM