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

Arbitrary token len (>32) work; Fix bug in mulmat

This commit is contained in:
Yu, Zijun 2025-04-17 17:42:44 +08:00 committed by Mustafa Cavus
parent 8d263bd6a5
commit 8b408869ae
1 changed files with 2 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
ggml/src/ggml-openvino/ggml-decoder.cpp
View File

@ -56,13 +56,8 @@ void GgmlOvDecoder::set_input_output(ggml_tensor* node, std::map<std::string, gg
break;
}
case GGML_OP_MUL_MAT: {
if (!ggml_is_contiguous(node->src[1]) ||
node->src[1]->ne[0] * node->src[1]->nb[0] != node->src[1]->nb[1]) {
m_continuous = false;
} else {
m_continuous = true;
}
break;
m_continuous = node->src[0]->view_src == nullptr;
break;
}
default:
break;