happyz
/
llama.cpp
mirror of https://github.com/ggerganov/llama.cpp.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

force patch_merger tensors to f16/f32

This commit is contained in:
Sigbjørn Skjæret 2025-12-17 05:36:01 +01:00 committed by GitHub
parent 58062860af
commit a7ab470832
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
convert_hf_to_gguf.py
View File

@ -1838,7 +1838,7 @@ class MmprojModel(ModelBase):
def tensor_force_quant(self, name, new_name, bid, n_dims): def tensor_force_quant(self, name, new_name, bid, n_dims):
del bid, name, n_dims # unused del bid, name, n_dims # unused
if ".patch_embd.weight" in new_name: if ".patch_embd.weight" in new_name or ".patch_merger.weight" in new_name:
return gguf.GGMLQuantizationType.F16 if self.ftype == gguf.LlamaFileType.MOSTLY_F16 else gguf.GGMLQuantizationType.F32 return gguf.GGMLQuantizationType.F16 if self.ftype == gguf.LlamaFileType.MOSTLY_F16 else gguf.GGMLQuantizationType.F32
return False return False