Fix GLM45v launch server cuda torch compile bug (#9554)

2025-08-25 13:46:28 -07:00
parent 3affa9dcc3
commit 24a8cee66d
1 changed files with 1 additions and 0 deletions
--- a/python/sglang/srt/models/qwen2_5_vl.py
+++ b/python/sglang/srt/models/qwen2_5_vl.py
@@ -526,6 +526,7 @@ class Qwen2_5_VLForConditionalGeneration(nn.Module):
    def get_input_embeddings(self):
        return self.model.embed_tokens

+    @torch.no_grad()
    def forward(
        self,
        input_ids: torch.Tensor,