qwen3vl: enable flash attention by default (#12862)

75e75d9a · Michael Yang · GitHub · ed78e127 · 75e75d9a
Unverified Commit 75e75d9a authored Oct 30, 2025 by Michael Yang Committed by GitHub Oct 30, 2025
Hide whitespace changes
Inline Side-by-side

Showing with 2 additions and 2 deletions

fs/ggml/ggml.go fs/ggml/ggml.go +2 -2

No files found.
--- a/fs/ggml/ggml.go
+++ b/fs/ggml/ggml.go
@@ -895,8 +895,8 @@ func (f GGML) FlashAttention() bool {
 	return slices.Contains([]string{
 		"gemma3",
 		"gptoss", "gpt-oss",
-		"qwen3",
-		"qwen3moe",
+		"qwen3", "qwen3moe",
+		"qwen3vl", "qwen3vlmoe",
 	}, f.KV().String("general.architecture"))
 }