xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2025-12-21 15:15:45 +08:00
Code Issues Packages Projects Releases Wiki Activity

[Core] Drop overly aggressive whisper assertion (#25408)

Browse Source 
Signed-off-by: Russell Bryant <rbryant@redhat.com>
Signed-off-by: yewentao256 <zhyanwentao@126.com>
This commit is contained in:
Russell Bryant 2025-09-22 20:09:52 -04:00 committed by yewentao256
parent d96a3fc653
commit e97cf2e32b
1 changed files with 0 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
vllm/v1/core/sched/scheduler.py
View File

@ -463,10 +463,6 @@ class Scheduler(SchedulerInterface):
# always padded to the maximum length. If we support other # always padded to the maximum length. If we support other
# encoder-decoder models, this will need to be updated if we # encoder-decoder models, this will need to be updated if we
# want to only allocate what is needed. # want to only allocate what is needed.
assert ("whisper"
in self.vllm_config.model_config.model.lower()), (
"Whisper is the only supported "
"encoder-decoder model.")
num_encoder_tokens =\ num_encoder_tokens =\
self.scheduler_config.max_num_encoder_input_tokens self.scheduler_config.max_num_encoder_input_tokens
else: else: