Skip to content
Merged
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
19 changes: 13 additions & 6 deletions python/sglang/srt/managers/tokenizer_manager.py
Original file line number Diff line number Diff line change
Expand Up @@ -958,7 +958,11 @@ async def _wait_one_response(
finish_reason.get("type") == "abort"
and finish_reason.get("status_code") == HTTPStatus.BAD_REQUEST
):
raise ValueError(finish_reason["message"])
if not obj.stream:
raise ValueError(finish_reason["message"])
else:
yield out
break

if finish_reason.get("type") == "abort" and finish_reason.get(
"status_code"
Expand All @@ -975,11 +979,14 @@ async def _wait_one_response(
# Mark ongoing LoRA request as finished.
if self.server_args.enable_lora and state.obj.lora_path:
await self.lora_registry.release(state.obj.lora_id)

raise fastapi.HTTPException(
status_code=finish_reason["status_code"],
detail=finish_reason["message"],
)
if not obj.stream:
raise fastapi.HTTPException(
status_code=finish_reason["status_code"],
detail=finish_reason["message"],
)
else:
yield out
break
yield out
break

Expand Down
Loading