Lumpiasty/llama.cpp
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

model: MTP support for Qwen3-Next (#25589)

Browse Source 
* mtp for qwen3nex

* fix for python type-check

* Fix to compute num_mtp from directly mtp layer

* define opt_num_mtp_layers in _QwenMtpMixin and fix some comments

* Fix for python type check

* Update gguf-py/gguf/constants.py

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@huggingface.co>

* rebase and add load_mtp flags

* Update src/models/qwen3next.cpp

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@huggingface.co>

* Update src/models/qwen3next.cpp

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@huggingface.co>

---------

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@huggingface.co>
This commit is contained in:
Masashi Yoshimura
2026-08-03 17:15:01 +09:00
committed by GitHub
co-authored by Sigbjørn Skjæret
parent 2b63e0610b
commit 4ed2b13f75
5 changed files with 389 additions and 152 deletions
Download Patch File Download Diff File Expand all files Collapse all files
gguf-py/gguf/constants.py
+7 -1
View File
@@ -2329,7 +2329,13 @@ MODEL_TENSORS: dict[MODEL_ARCH, list[MODEL_TENSOR]] = {
MODEL_TENSOR.SSM_NORM,
MODEL_TENSOR.SSM_IN,
MODEL_TENSOR.SSM_BETA_ALPHA,
MODEL_TENSOR.SSM_OUT
MODEL_TENSOR.SSM_OUT,
MODEL_TENSOR.NEXTN_EH_PROJ,
MODEL_TENSOR.NEXTN_EMBED_TOKENS,
MODEL_TENSOR.NEXTN_ENORM,
MODEL_TENSOR.NEXTN_HNORM,
MODEL_TENSOR.NEXTN_SHARED_HEAD_HEAD,
MODEL_TENSOR.NEXTN_SHARED_HEAD_NORM,
],
MODEL_ARCH.QWEN3VL: [
MODEL_TENSOR.TOKEN_EMBD,