Support chat-stream

modeling_qwen.py

@@ -1026,7 +1026,7 @@ class QWenLMHeadModel(QWenPreTrainedModel):
                     seed=-1,
                     **kwargs):
                 outputs.append(token.item())
-                yield tokenizer.decode(outputs, skip_special_tokens=True, errors='ignore', keep_image_special=True))
+                yield tokenizer.decode(outputs, skip_special_tokens=True, errors='ignore', keep_image_special=True)
 
         return stream_generator()