public_git/whisper.cpp
1
0
Fork 0
mirror of https://github.com/ggml-org/whisper.cpp.git synced 2026-10-21 09:41:29 +02:00
Code Issues Packages Projects Releases Wiki Activity

ggml-webgpu: fix flash_attn supports_op check for overlapping KV (llama/28205)

Browse Source
This commit is contained in:
Masashi Yoshimura
2026-10-19 09:47:10 +02:00
committed by Daniel Bevenius
parent b35b457a14
commit 7f56abab92
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
ggml/src/ggml-webgpu/ggml-webgpu.cpp
+1 -1
View File
@@ -4544,7 +4544,7 @@ static bool ggml_backend_webgpu_device_supports_op(ggml_backend_dev_t dev, const
break; break;
} }
if (ggml_webgpu_tensor_binding_overlap(ctx->webgpu_global_ctx, src1, src2) && if (ggml_webgpu_tensor_binding_overlap(ctx->webgpu_global_ctx, src1, src2) &&
src1->type != src2->type && !ggml_is_quantized(src1->type) && !ggml_is_quantized(src2->type)) { src1->type != src2->type && !(ggml_is_quantized(src1->type) && ggml_is_quantized(src2->type))) {
supports_op = false; supports_op = false;
break; break;
} }