[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
+19 -11
View File
@@ -158,27 +158,35 @@ class DeepseekOCR2DummyInputsBuilder(
class DeepseekOCR2MultiModalProcessor(
BaseMultiModalProcessor[DeepseekOCR2ProcessingInfo]
):
def _call_hf_processor(
def _apply_hf_processor_main(
self,
prompt: str,
mm_data: Mapping[str, object],
mm_kwargs: Mapping[str, object],
mm_items: MultiModalDataItems,
hf_processor_mm_kwargs: Mapping[str, object],
) -> BatchFeature:
valid_mm_items = mm_items.select(
{k for k, c in mm_items.get_all_counts().items() if c > 0}
)
mm_data, passthrough_data = self._get_hf_mm_data(valid_mm_items)
prompt_text = self.dummy_inputs.get_dummy_text(mm_items.get_all_counts())
if mm_data:
processed_outputs = self.info.ctx.call_hf_processor(
self.info.get_hf_processor(**mm_kwargs),
dict(prompt=prompt, **mm_data),
mm_kwargs,
processed_data = self.info.ctx.call_hf_processor(
self.info.get_hf_processor(**hf_processor_mm_kwargs),
dict(prompt=prompt_text, **mm_data),
hf_processor_mm_kwargs,
)
else:
tokenizer = self.info.get_tokenizer()
assert isinstance(tokenizer, HfTokenizer)
processed_outputs = tokenizer(
prompt, add_special_tokens=True, return_tensors="pt"
processed_data = tokenizer(
prompt_text, add_special_tokens=True, return_tensors="pt"
)
return processed_outputs
processed_data.update(passthrough_data)
return processed_data
def _get_mm_fields_config(
self,