thek0tyara/llama-cpp-turboquant
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 4

mtmd : Add Nemotron Nano 12B v2 VL support (#19547)

Browse source 
* nemotron nano v2 vlm support added

* simplified code; addressed reviews

* pre-downsample position embeddings during GGUF conversion for fixed input size
This commit is contained in:
Anav Prasad 2026-02-14 05:07:00 -08:00 • committed by GitHub
parent 1725e316c1
commit 01d8eaa28d
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
9 changed files with 167 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

2
tools/mtmd/clip-impl.h
View file

@ -236,6 +236,7 @@ enum projector_type {
PROJECTOR_TYPE_GLM4V,
PROJECTOR_TYPE_YOUTUVL,
PROJECTOR_TYPE_KIMIK25,
PROJECTOR_TYPE_NEMOTRON_V2_VL,
PROJECTOR_TYPE_UNKNOWN,
};
@ -270,6 +271,7 @@ static std::map<projector_type, std::string> PROJECTOR_TYPE_NAMES = {
{ PROJECTOR_TYPE_GLM4V, "glm4v"},
{ PROJECTOR_TYPE_YOUTUVL, "youtuvl"},
{ PROJECTOR_TYPE_KIMIK25, "kimik25"},
{ PROJECTOR_TYPE_NEMOTRON_V2_VL, "nemotron_v2_vl"},
};
static projector_type clip_projector_type_from_string(const std::string & str) {