Fix BertGeneration (#35043)

fix

Co-authored-by: ydshieh <[email protected]>

src/transformers/models/bert_generation/modeling_bert_generation.py

@@ -785,9 +785,7 @@ def forward(
 
         # We can provide a self-attention mask of dimensions [batch_size, from_seq_length, to_seq_length]
         # ourselves in which case we just need to make it broadcastable to all heads.
-        extended_attention_mask = None
-        if not use_cache:
-            extended_attention_mask: torch.Tensor = self.get_extended_attention_mask(attention_mask, input_shape)
+        extended_attention_mask: torch.Tensor = self.get_extended_attention_mask(attention_mask, input_shape)
 
         # If a 2D or 3D attention mask is provided for the cross-attention
         # we need to make broadcastable to [batch_size, num_heads, seq_length, seq_length]