diff options
| author | Iwan Kawrakow <iwan.kawrakow@gmail.com> | 2024-07-31 08:01:45 +0200 |
|---|---|---|
| committer | Kawrakow <48489457+ikawrakow@users.noreply.github.com> | 2024-08-01 09:38:06 +0200 |
| commit | 57df5ccdd7495e67c4d3707cd0a0318f6d04f190 (patch) | |
| tree | af93b81b7a7d359fab93e7a2ad904c230210dc00 /ggml/src/ggml-cann.cpp | |
| parent | 30d2d1b1ebfe0da401c3859adbb9e8512a36bd9d (diff) | |
iq2_k: Metal dot product finally works
It is slow: 45.4 t/s for 7B model vs 50 t/s for iq2_xs,
or 63.3 t/s for q2_K_S.
Diffstat (limited to 'ggml/src/ggml-cann.cpp')
0 files changed, 0 insertions, 0 deletions