Tiny TF Bart fixes (#8023)

src/transformers/modeling_tf_bart.py

@@ -822,7 +822,7 @@ class TFBartModel(TFPretrainedBartModel):
         if decoder_attn_mask is None:
             decoder_padding_mask = make_padding_mask(decoder_input_ids, pad_token_id)
         else:
-            decoder_padding_mask = invert_mask(tf.Tensor)
+            decoder_padding_mask = invert_mask(decoder_attn_mask)
 
         causal_lm_mask = causal_attention_mask(tgt_len, tgt_len, mask_dtype)
         return decoder_input_ids, decoder_padding_mask, causal_lm_mask
@@ -903,6 +903,7 @@ class TFBartModel(TFPretrainedBartModel):
                 output_attentions=output_attentions,
                 output_hidden_states=output_hidden_states,
                 return_dict=True,
+                training=training,
             )
         decoder_outputs = self.decoder(
             decoder_input_ids,
@@ -915,6 +916,7 @@ class TFBartModel(TFPretrainedBartModel):
             output_attentions=output_attentions,
             output_hidden_states=output_hidden_states,
             return_dict=return_dict,
+            training=training,
         )
         if not return_dict:
             # Attention and hidden_states will be [] or None if they aren't needed