Skip to content

[bugfix] refactor FunASR's _get_data_parser #34397

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
DarkLight1337 merged 2 commits into from
Feb 12, 2026
Merged

[bugfix] refactor FunASR's _get_data_parser #34397

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
b24268b
update.
Feb 12, 2026
f87da21
update.
Feb 12, 2026
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
18 changes: 7 additions & 11 deletions vllm/model_executor/models/funasr.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -714,10 +714,6 @@ class FunASRProcessingInfo(BaseProcessingInfo):
def get_hf_config(self) -> Qwen3Config:
return self.ctx.get_hf_config(Qwen3Config)

@property
def skip_prompt_length_check(self) -> bool:
return True # Because the encoder prompt is padded

def get_supported_mm_limits(self) -> Mapping[str, int | None]:
return {"audio": 1}

Expand All @@ -727,6 +723,13 @@ def get_feature_extractor(self, **kwargs: object) -> FunASRFeatureExtractor:
assert isinstance(feature_extractor, FunASRFeatureExtractor)
return feature_extractor

def get_data_parser(self) -> MultiModalDataParser:
feature_extractor = self.get_feature_extractor()
return MultiModalDataParser(
target_sr=feature_extractor.sampling_rate,
target_channels=self.get_target_channels(),
)

def get_target_channels(self) -> int:
return 1

Expand Down Expand Up @@ -765,13 +768,6 @@ def get_dummy_mm_data(


class FunASRMultiModalProcessor(BaseMultiModalProcessor[FunASRProcessingInfo]):
def _get_data_parser(self) -> MultiModalDataParser:
feature_extractor = self.info.get_feature_extractor()
return MultiModalDataParser(
target_sr=feature_extractor.sampling_rate,
target_channels=self.info.get_target_channels(),
)

def _call_hf_processor(
self,
prompt: str,
Expand Down