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

don't quantize conv1d of Kimi Linear

This commit is contained in:
Yee Man Chan 2026-01-03 08:27:29 +08:00
parent 8bd617eb1c
commit a4020d867f
1 changed files with 2 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
src/llama-quant.cpp
View File

@ -869,9 +869,9 @@ static void llama_model_quantize_impl(const std::string & fname_inp, const std::
quantize &= name != LLM_TN(model.arch)(LLM_TENSOR_POS_EMBD, "weight");
quantize &= name != LLM_TN(model.arch)(LLM_TENSOR_TOKEN_TYPES, "weight");
// do not quantize Mamba's small yet 2D weights
// do not quantize Mamba /Kimi's small conv1d weights
// NOTE: can't use LLM_TN here because the layer number is not known
quantize &= name.find("ssm_conv1d.weight") == std::string::npos;
quantize &= name.find("ssm_conv1d") == std::string::npos;
quantize &= name.find("shortconv.conv.weight") == std::string::npos;
// do not quantize RWKV's small yet 2D weights