add padding_mask

modeling_decilm.py

@@ -62,6 +62,7 @@ class DeciLMAttention(LlamaAttention):
             past_key_value: Optional[Tuple[torch.Tensor]] = None,
             output_attentions: bool = False,
             use_cache: bool = False,
+            padding_mask: Optional[torch.LongTensor] = None,
     ) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]:
         bsz, q_len, _ = hidden_states.size()
         if past_key_value is None: