llama.cpp
Move to backend sampling for MTP draft path
#23287
Merged

Move to backend sampling for MTP draft path #23287

gaugarg-nv
gaugarg-nv gaugarg-nv requested a review from ggerganov ggerganov 123 days ago
gaugarg-nv gaugarg-nv requested a review 123 days ago
gaugarg-nv gaugarg-nv requested a review from am17an am17an 123 days ago
am17an
gaugarg-nv
ggerganov
ORippler
gaugarg-nv gaugarg-nv force pushed 122 days ago
gaugarg-nv
ggerganov
ggerganov
ggerganov commented on 2026-05-19
ggerganov
ggerganov ggerganov assigned ggerganov ggerganov 121 days ago
am17an
gaugarg-nv
am17an
gaugarg-nv
gaugarg-nv
ggerganov
gaugarg-nv
gaugarg-nv Move to backend sampling for MTP draft path
e20085ba
gaugarg-nv Allow sampler chains to be partially offloaded to backend
e961ca57
gaugarg-nv Add --spec-draft-backend-sampling argument. Enabled by default.
b0061f7f
gaugarg-nv gaugarg-nv force pushed to b0061f7f 121 days ago
ggerganov
ggerganov approved these changes on 2026-05-20
am17an
am17an approved these changes on 2026-05-20
gaugarg-nv gaugarg-nv merged ad277572 into master 121 days ago
homemdesgraca
ggerganov
homemdesgraca
gaugarg-nv gaugarg-nv deleted the mtp_backend_sampling branch 111 days ago

Login to write a write a comment.

Login via GitHub

Reviewers
Assignees
Labels
Milestone