llama.cpp
server : encode images on the GPU by borrowing the idle compute buffers
#28320

Open

server : encode images on the GPU by borrowing the idle compute buffers #28320

efschu wants to merge 1 commit into ggml-org:master from efschu:mmproj-vram-swap

server : encode images on the GPU by borrowing the idle compute buffers

1710a36f

efschu requested a review 7 hours ago

efschu requested a review from

ggerganov 7 hours ago

efschu requested a review 7 hours ago

github-actions added server

github-actions added mtmd

ggml-gh-bot added draft

github-actions marked this pull request as draft 7 hours ago

github-actions removed draft

Reviewers

ggerganov

Assignees

No one assigned

Labels

server mtmd

Milestone

No milestone