Lumpiasty/llama.cpp
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

ggml-webgpu: add mulmat with overlapping src0/src1 (e.g., for minimax-01) (#27321)

Browse Source
This commit is contained in:
Masashi Yoshimura
2026-08-19 16:29:33 +09:00
committed by GitHub
parent 0adcc3bb57
commit 5112b9738b
10 changed files with 189 additions and 92 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tests/test-llama-archs.cpp
+1 -1
View File
@@ -456,7 +456,7 @@ static bool arch_supported(const llm_arch arch) {
// FIXME: these hit scheduler/view-backed-output issues with WebGPU on CI.
#ifdef GGML_USE_WEBGPU
if (arch == LLM_ARCH_DEEPSEEK32 || arch == LLM_ARCH_GLM_DSA || arch == LLM_ARCH_MINIMAX_01) {
if (arch == LLM_ARCH_DEEPSEEK32 || arch == LLM_ARCH_GLM_DSA) {
return false;
}
#endif // GGML_USE_WEBGPU