finish (#9721)

src/transformers/models/t5/modeling_t5.py

@@ -934,9 +934,9 @@ class T5Stack(T5PreTrainedModel):
                     encoder_extended_attention_mask = encoder_extended_attention_mask.to(hidden_states.device)
                 if encoder_decoder_position_bias is not None:
                     encoder_decoder_position_bias = encoder_decoder_position_bias.to(hidden_states.device)
-                if head_mask is not None:
+                if not (isinstance(head_mask, list) and head_mask[0] is None):
                     head_mask = head_mask.to(hidden_states.device)
-                if encoder_head_mask is not None:
+                if not (isinstance(encoder_head_mask, list) and encoder_head_mask[0] is None):
                     encoder_head_mask = encoder_head_mask.to(hidden_states.device)
             if output_hidden_states:
                 all_hidden_states = all_hidden_states + (hidden_states,)