[DeepSeek V3.2] [Bugfix] slice indexer and padding fa3 when can not run cuda graph (#17076)
This commit is contained in:
@@ -189,6 +189,9 @@ class NSAIndexerMetadata(BaseIndexerMetadata):
|
||||
def get_indexer_seq_len_cpu(self) -> torch.Tensor:
|
||||
return self.attn_metadata.indexer_seq_lens_cpu
|
||||
|
||||
def get_nsa_extend_len_cpu(self) -> List[int]:
|
||||
return self.attn_metadata.nsa_extend_seq_lens_list
|
||||
|
||||
def get_token_to_batch_idx(self) -> torch.Tensor:
|
||||
return self.attn_metadata.token_to_batch_idx
|
||||
|
||||
|
||||
Reference in New Issue
Block a user