Cleanup bart caching logic (#5640)

src/transformers/modeling_bart.py

@@ -628,8 +628,8 @@ class SelfAttention(nn.Module):
         self.out_proj = nn.Linear(embed_dim, embed_dim, bias=bias)
         self.cache_key = "encoder_decoder" if self.encoder_decoder_attention else "self"
 
-    def _shape(self, tensor, dim_0, bsz):
-        return tensor.contiguous().view(dim_0, bsz * self.num_heads, self.head_dim).transpose(0, 1)
+    def _shape(self, tensor, seq_len, bsz):
+        return tensor.contiguous().view(seq_len, bsz * self.num_heads, self.head_dim).transpose(0, 1)
 
     def forward(
         self,
@@ -648,9 +648,8 @@ class SelfAttention(nn.Module):
         # get here for encoder decoder cause of static_kv
         if layer_state is not None:  # reuse k,v and encoder_padding_mask
             saved_state = layer_state.get(self.cache_key, {})
-            if "prev_key" in saved_state:
+            if "prev_key" in saved_state and static_kv:
                 # previous time steps are cached - no need to recompute key and value if they are static
-                if static_kv:
                 key = None
         else:
             saved_state = None
@@ -738,37 +737,14 @@ class SelfAttention(nn.Module):
                 v = torch.cat([prev_value, v], dim=1)
         assert k is not None and v is not None
         prev_key_padding_mask: Optional[Tensor] = saved_state.get("prev_key_padding_mask", None)
-        key_padding_mask = self._cat_prev_key_padding_mask(
-            key_padding_mask, prev_key_padding_mask, bsz, k.size(1), static_kv
-        )
-        return k, v, key_padding_mask
-
-    @staticmethod
-    def _cat_prev_key_padding_mask(
-        key_padding_mask: Optional[Tensor],
-        prev_key_padding_mask: Optional[Tensor],
-        batch_size: int,
-        src_len: int,
-        static_kv: bool,
-    ) -> Optional[Tensor]:
-        # saved key padding masks have shape (bsz, seq_len)
         if prev_key_padding_mask is not None:
             if static_kv:
                 new_key_padding_mask = prev_key_padding_mask
             else:
                 new_key_padding_mask = torch.cat([prev_key_padding_mask, key_padding_mask], dim=1)
-
-        elif key_padding_mask is not None:
-            filler = torch.zeros(
-                batch_size,
-                src_len - key_padding_mask.size(1),
-                dtype=key_padding_mask.dtype,
-                device=key_padding_mask.device,
-            )
-            new_key_padding_mask = torch.cat([filler, key_padding_mask], dim=1)
         else:
-            new_key_padding_mask = prev_key_padding_mask
-        return new_key_padding_mask
+            new_key_padding_mask = key_padding_mask
+        return k, v, new_key_padding_mask
 
 
 class BartClassificationHead(nn.Module):