Fix broken streaming response with --incremental-streaming-output (#22549)
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
This commit is contained in:
co-authored by
Claude Opus 4.6
parent
701a0e0c25
commit
1f8df97054
@@ -738,7 +738,11 @@ class OpenAIServingChat(OpenAIServingBase):
|
||||
stream_started = True
|
||||
|
||||
stream_buffer = stream_buffers.get(index, "")
|
||||
delta = content["text"][len(stream_buffer) :]
|
||||
if self.tokenizer_manager.server_args.incremental_streaming_output:
|
||||
# content["text"] is already the incremental delta
|
||||
delta = content["text"]
|
||||
else:
|
||||
delta = content["text"][len(stream_buffer) :]
|
||||
stream_buffers[index] = stream_buffer + delta
|
||||
|
||||
# Handle reasoning content
|
||||
|
||||
Reference in New Issue
Block a user