imatrix : offload to GPU support #4957
ggerganov
merged
10 commits into
master
from
gg/imatrix-gpu-4931
backend : add eval callback
65648b34
backend : group nodes in a single compute when user don't need them
01b6f68a
backend : clean-up the implementation
83f3d7a8
simple : do not perform tensor data copy if not needed
e1b1db9f
simple : fix
e0493800
imatrix : offload to GPU support
0b2fca9a
ikawrakow
approved these changes
on 2024-01-15
imatrix : fix ggml_mul_mat_id hanlding
a722d05a
ci : add imatrix test
10b25e03
ci : rearrange output
4fb52843
Base automatically changed from
gg/sched-eval-callback-4931
to
master
1 year ago
Merge branch 'master' into gg/imatrix-gpu-4931
2917e6b5
Login to write a write a comment.
Login via GitHub