llama.cpp
feat: MTP support for dense Qwen 3.5 with FastMTP vocabulary trimming
#20700

Open

feat: MTP support for dense Qwen 3.5 with FastMTP vocabulary trimming #20700

itigges22 wants to merge 1 commit into ggml-org:master from itigges22:feat/qwen35-dense-mtp

github-actions added model

github-actions added examples

github-actions added python

github-actions added server

itigges22 force pushed from fef9ada3 to affba2a6 14 days ago

itigges22 changed the title ~~feat: MTP support for dense Qwen 3.5 (0.8B-27B)~~ feat: MTP support for dense Qwen 3.5 with FastMTP vocabulary trimming 14 days ago

feat: MTP support for dense Qwen 3.5 with FastMTP vocabulary trimming

19fdba56

itigges22 force pushed from affba2a6 to 19fdba56 14 days ago

itigges22 marked this pull request as ready for review 14 days ago

itigges22 requested a review 14 days ago

itigges22 requested a review from

CISC 14 days ago

itigges22 requested a review from

ggerganov 14 days ago

itigges22 requested a review 14 days ago

Reviewers

CISC

ggerganov

Assignees

No one assigned

Labels

model examples python server

Milestone

No milestone

llama.cpp feat: MTP support for dense Qwen 3.5 with FastMTP vocabulary trimming #20700 Open

feat: MTP support for dense Qwen 3.5 with FastMTP vocabulary trimming #20700

llama.cpp
feat: MTP support for dense Qwen 3.5 with FastMTP vocabulary trimming
#20700

Open