ggml : group all experts in a single ggml_mul_mat_id (#6505)

Browse Source

* ggml : group all experts in a single ggml_mul_mat_id
cuda : improve mmid row copy

* cuda : fix bin bcast with non-cont src0

* test-backend-ops : only run all mul mat tests for base types

* llama : disable moe offloading with SYCL

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

...

This commit is contained in:

slaren

2024-04-18 15:18:48 +02:00

committed by GitHub

co-authored by Georgi Gerganov

parent 03c0946d73

commit 0d56246f4b

12 changed files with 971 additions and 821 deletions

Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL

Download Patch File Download Diff File Expand all files Collapse all files

ggml-metal.metal

+400 -478

File diff suppressed because it is too large Load Diff