[MM] Simplify _apply_hf_processor_main (#53275)

Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
This commit is contained in:
Cyrus Leung
2026-08-22 01:49:37 -07:00
committed by GitHub
parent 7ca49fbe4b
commit 0b19ebcacd
90 changed files with 1848 additions and 1706 deletions
@@ -461,24 +461,29 @@ class MossTranscribeDiarizeDummyInputsBuilder(
class MossTranscribeDiarizeMultiModalProcessor(
BaseMultiModalProcessor[MossTranscribeDiarizeProcessingInfo]
):
def _call_hf_processor(
self,
prompt: str,
mm_data: Mapping[str, object],
mm_kwargs: Mapping[str, object],
) -> BatchFeature:
tokenizer = self.info.get_tokenizer()
audios = _get_audios_from_mm_data(mm_data)
if not audios:
input_ids = tokenizer.encode(prompt, add_special_tokens=False)
return BatchFeature({"input_ids": [input_ids]}, tensor_type="pt")
def _get_hf_processor_text(self, mm_counts: Mapping[str, int]) -> str:
return self.dummy_inputs.get_dummy_text(mm_counts)
processed = self.info.ctx.call_hf_processor(
self.info.get_hf_processor(**mm_kwargs),
dict(text=prompt, audio=audios),
mm_kwargs,
)
return _add_vllm_audio_metadata(processed, len(audios))
def _preprocess_hf_mm_data(
self,
mm_data: Mapping[str, object],
hf_processor_mm_kwargs: Mapping[str, object],
) -> tuple[Mapping[str, object], Mapping[str, object]]:
audios = _get_audios_from_mm_data(mm_data)
return dict(audio=audios), hf_processor_mm_kwargs
def _postprocess_hf_mm_data(
self,
mm_data: Mapping[str, object],
hf_processor_mm_kwargs: Mapping[str, object],
processed_data: BatchFeature,
) -> BatchFeature:
audios = _get_audios_from_mm_data(mm_data)
if audios:
processed_data = _add_vllm_audio_metadata(processed_data, len(audios))
return processed_data
def _get_mm_fields_config(
self,