Musicgen special tokens in tensors (#31420)

fix

src/transformers/models/musicgen/modeling_musicgen.py

@@ -1666,6 +1666,8 @@ class MusicgenForCausalLM(MusicgenPreTrainedModel):
             inputs, generation_config.bos_token_id, model_kwargs
         )
         batch_size = input_ids.shape[0] // self.num_codebooks
+        kwargs_has_attention_mask = model_kwargs.get("attention_mask", None) is not None
+        self._prepare_special_tokens(generation_config, kwargs_has_attention_mask, device=input_ids.device)
 
         # 4. Define other model kwargs
         model_kwargs["use_cache"] = generation_config.use_cache
@@ -2738,6 +2740,8 @@ class MusicgenForConditionalGeneration(PreTrainedModel):
             inputs, generation_config.bos_token_id, model_kwargs
         )
         batch_size = inputs_tensor.shape[0]
+        kwargs_has_attention_mask = model_kwargs.get("attention_mask", None) is not None
+        self._prepare_special_tokens(generation_config, kwargs_has_attention_mask, device=inputs_tensor.device)
 
         # 4. Define other model kwargs
         model_kwargs["use_cache"] = generation_config.use_cache

src/transformers/models/musicgen_melody/modeling_musicgen_melody.py

@@ -1587,6 +1587,8 @@ class MusicgenMelodyForCausalLM(MusicgenMelodyPreTrainedModel):
             inputs, generation_config.bos_token_id, model_kwargs
         )
         batch_size = input_ids.shape[0] // self.num_codebooks
+        kwargs_has_attention_mask = model_kwargs.get("attention_mask", None) is not None
+        self._prepare_special_tokens(generation_config, kwargs_has_attention_mask, device=input_ids.device)
 
         # 4. Define other model kwargs
         model_kwargs["use_cache"] = generation_config.use_cache
@@ -2588,6 +2590,8 @@ class MusicgenMelodyForConditionalGeneration(PreTrainedModel):
             inputs, generation_config.bos_token_id, model_kwargs
         )
         batch_size = inputs_tensor.shape[0]
+        kwargs_has_attention_mask = model_kwargs.get("attention_mask", None) is not None
+        self._prepare_special_tokens(generation_config, kwargs_has_attention_mask, device=inputs_tensor.device)
 
         # 4. Define other model kwargs
         model_kwargs["use_cache"] = generation_config.use_cache