cuda : allocate all temporary ggml_tensor_extra_gpu from a fixed-size buffer (#2220)
B
Bach Le committed
7cdd30bf1f84339c55a5e3de29384f6bbdebb61c
Parent: e8035f1
Committed by GitHub <noreply@github.com>
on 7/14/2023, 7:00:58 PM