Public/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2026-09-20 01:31:31 +02:00
Code Issues Packages Projects Releases Wiki Activity

force f16 or f32 on step3-vl conv weights

Browse Source
This commit is contained in:
Sigbjørn Skjæret
2026-04-08 21:35:44 +02:00
committed by GitHub
parent d12cc3d1ca
commit f5059efad9
1 changed files with 2 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
convert_hf_to_gguf.py
+2
View File
@@ -4984,6 +4984,8 @@ class Step3VLVisionModel(MmprojModel):
def tensor_force_quant(self, name, new_name, bid, n_dims):
if ".position_embd." in new_name:
return gguf.GGMLQuantizationType.F32
if ("mm.0." in new_name or "mm.1." in new_name) and new_name.endswith(".weight"):
return gguf.GGMLQuantizationType.F16 if self.ftype == gguf.LlamaFileType.MOSTLY_F16 else gguf.GGMLQuantizationType.F32
return super().tensor_force_quant(name, new_name, bid, n_dims)
def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]: