Generation / FIX: Fix multi-device generation

src/transformers/generation/utils.py

@@ -476,6 +476,7 @@ def _prepare_attention_mask_for_generation(
         )
         can_infer_attention_mask = is_pad_token_in_inputs * is_pad_token_not_equal_to_eos_token_id
         attention_mask_from_padding = inputs.ne(pad_token_id).long()
+
         attention_mask = (
             attention_mask_from_padding * can_infer_attention_mask + default_attention_mask * ~can_infer_attention_mask
         )
@@ -1340,7 +1341,10 @@ def _get_static_cache(self, max_batch_size: int, max_cache_len: int) -> StaticCa
         return self._static_cache
 
     def _prepare_special_tokens(
-        self, generation_config: GenerationConfig, kwargs_has_attention_mask: Optional[bool] = None
+        self,
+        generation_config: GenerationConfig,
+        kwargs_has_attention_mask: Optional[bool] = None,
+        device: Optional[Union[torch.device, str]] = None,
     ):
         """
         Prepares the special tokens for generation, overwriting the generation config with their processed versions
@@ -1352,15 +1356,18 @@ def _prepare_special_tokens(
         """
 
         # Convert special tokens to tensors (if they exist)
-        def _tensor_or_none(token):
+        def _tensor_or_none(token, device=None):
+            if device is None:
+                device = self.device
+
             if token is None or isinstance(token, torch.Tensor):
                 return token
-            return torch.tensor(token, device=self.device, dtype=torch.long)
+            return torch.tensor(token, device=device, dtype=torch.long)
 
-        bos_token_id = _tensor_or_none(generation_config.bos_token_id)
-        eos_token_id = _tensor_or_none(generation_config.eos_token_id)
-        pad_token_id = _tensor_or_none(generation_config.pad_token_id)
-        decoder_start_token_id = _tensor_or_none(generation_config.decoder_start_token_id)
+        bos_token_id = _tensor_or_none(generation_config.bos_token_id, device=device)
+        eos_token_id = _tensor_or_none(generation_config.eos_token_id, device=device)
+        pad_token_id = _tensor_or_none(generation_config.pad_token_id, device=device)
+        decoder_start_token_id = _tensor_or_none(generation_config.decoder_start_token_id, device=device)
         decoder_start_token_id = decoder_start_token_id if decoder_start_token_id is not None else bos_token_id
 
         # We can have more than one eos token. Always treat it as a 1D tensor (when it exists).
@@ -1511,14 +1518,16 @@ def generate(
         accepts_attention_mask = "attention_mask" in set(inspect.signature(self.forward).parameters.keys())
         requires_attention_mask = "encoder_outputs" not in model_kwargs
         kwargs_has_attention_mask = model_kwargs.get("attention_mask", None) is not None
-        self._prepare_special_tokens(generation_config, kwargs_has_attention_mask)
 
         # 3. Define model inputs
         inputs_tensor, model_input_name, model_kwargs = self._prepare_model_inputs(
             inputs, generation_config.bos_token_id, model_kwargs
         )
         batch_size = inputs_tensor.shape[0]
 
+        device = inputs_tensor.device
+        self._prepare_special_tokens(generation_config, kwargs_has_attention_mask, device=device)
+
         # decoder-only models must use left-padding for batched generation.
         if not self.config.is_encoder_decoder and not is_torchdynamo_compiling():
             # If `input_ids` was given, check if the last id in any sequence is `pad_token_id`