[VLM] Merged multi-modal processor for InternVL-based models (#12553)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk> Signed-off-by: Isotr0py <2037008807@qq.com> Co-authored-by: Isotr0py <2037008807@qq.com>
This commit is contained in:
@@ -92,7 +92,11 @@ class UltravoxProcessingInfo(BaseProcessingInfo):
|
||||
def get_supported_mm_limits(self) -> Mapping[str, Optional[int]]:
|
||||
return {"audio": None}
|
||||
|
||||
def get_mm_max_tokens_per_item(self, seq_len: int) -> Mapping[str, int]:
|
||||
def get_mm_max_tokens_per_item(
|
||||
self,
|
||||
seq_len: int,
|
||||
mm_counts: Mapping[str, int],
|
||||
) -> Mapping[str, int]:
|
||||
feature_extractor = self.get_feature_extractor()
|
||||
max_audio_tokens = math.ceil(feature_extractor.chunk_length *
|
||||
_AUDIO_TOKENS_PER_SECOND)
|
||||
|
||||
Reference in New Issue
Block a user