llama.cpp
Make the drafter probabilistic and the target verify by rejection sampling for simple draft and MTP
#27694
Merged
Go
Login via GitHub
Home
Pricing
FAQ
Install
Login
via GitHub
Overview
Commits
9
Changes
View On
GitHub
Make the drafter probabilistic and the target verify by rejection sampling for simple draft and MTP
#27694
ggerganov
merged 9 commits into
ggml-org:master
from
praneshgo:pgonegandla/spec-draft-rejection-sampling
github-actions
added
server
praneshgo
force pushed
42 days ago
praneshgo
force pushed
to
88e706e9
39 days ago
praneshgo
marked this pull request as ready for review
39 days ago
praneshgo
requested a review
39 days ago
praneshgo
requested a review
39 days ago
praneshgo
force pushed
from
88e706e9
to
b11dae78
35 days ago
gaugarg-nv
commented on 2026-09-08
ggerganov
assigned
ggerganov
34 days ago
danbev
commented on 2026-09-09
danbev
approved these changes on 2026-09-15
praneshgo
requested a review
from
gaugarg-nv
25 days ago
ggerganov
commented on 2026-09-21
praneshgo
requested a review
from
ggerganov
12 days ago
ggerganov
approved these changes on 2026-10-02
Make the drafter probabilistic and the target verify by rejection sam…
269bb4fa
Drop stale spec_draft_q before drafting
2f82a679
Fallback to argmax sampling for grammar-constrained requests and addi…
94d05d5f
Support grammar-constrained requests in rejection sampling
b1b5b01c
Fix - renormalize distribution after masking
bba1f2e1
copy rng on sampler copy and re-accept drafted tokens on replay
d9eb9f0d
Fix draft sampler sharing the target's rng stream
b5c26731
Simplify the rejection sampler's inputs and move replay to the server
61de64e1
Truncate the draft candidates along with the draft
c19c18a7
praneshgo
force pushed
from
75471a12
to
c19c18a7
10 days ago
ggerganov
added
merge ready
ServeurpersoCom
approved these changes on 2026-10-02
ggerganov
merged
1fb7ef3e
into master
10 days ago
Login to write a write a comment.
Login via GitHub
Reviewers
ServeurpersoCom
ggerganov
danbev
gaugarg-nv
Assignees
ggerganov
Labels
server
merge ready
Milestone
No milestone
Login to write a write a comment.
Login via GitHub