llama.cpp
metal : dequant kv cache only for large batches
#27438

Merged

metal : dequant kv cache only for large batches #27438

ggerganov merged 1 commit into master from gg/metal-fa-q-f16-bs

metal : dequant kv cache only for large batches

3db64e3b

ggerganov requested a review 53 days ago

github-actions added ggml

github-actions added Apple Metal

forforever73 approved these changes on 2026-08-20

ggerganov merged 63b64a50 into master 53 days ago

ggerganov deleted the gg/metal-fa-q-f16-bs branch 53 days ago

Reviewers

forforever73

Assignees

No one assigned

Labels

ggml Apple Metal

Milestone

No milestone