A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 5, 06:00 UTC
[Bugfix] stream failure when model name not in audio endpoints (#36679)
Signed-off-by: Ekagra Ranjan <3116519+ekagra-ranjan@users.noreply.github.com>
Ekagra Ranjan committed
Apr 13, 2026 at 10:20 UTC
1ce0318c6811e8291d6d56777b1f33efb4a0e0f9
1 file changed
+3
vllm/entrypoints/openai/speech_to_text/speech_to_text.py
+3
@@ -389,6 +389,9 @@ class OpenAISpeechToText(OpenAIServing):
389
if error_check_ret is not None:
390
return error_check_ret
391
392
+ if not request.model:
393
+ request.model = self.models.model_name()
394
+
395
# If the engine is dead, raise the engine's DEAD_ERROR.
396
# This is required for the streaming case, where we return a
397
# success status before we actually start generating text :).