Big file_utils cleanup (#16396)

* Big file_utils cleanup

* This one still needs to be treated separately

src/transformers/models/blenderbot/modeling_flax_blenderbot.py

@@ -124,7 +124,7 @@ BLENDERBOT_INPUTS_DOCSTRING = r"""
             Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
             more detail.
         return_dict (`bool`, *optional*):
-            Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple.
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
 """
 
 
@@ -155,7 +155,7 @@ BLENDERBOT_ENCODE_INPUTS_DOCSTRING = r"""
             Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
             more detail.
         return_dict (`bool`, *optional*):
-            Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple.
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
 """
 
 BLENDERBOT_DECODE_INPUTS_DOCSTRING = r"""
@@ -201,7 +201,7 @@ BLENDERBOT_DECODE_INPUTS_DOCSTRING = r"""
             Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
             more detail.
         return_dict (`bool`, *optional*):
-            Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple.
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
 """
 
 

src/transformers/models/blenderbot/modeling_tf_blenderbot.py

@@ -608,8 +608,8 @@ BLENDERBOT_INPUTS_DOCSTRING = r"""
             more detail. This argument can be used only in eager mode, in graph mode the value in the config will be
             used instead.
         return_dict (`bool`, *optional*):
-            Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple. This argument can be used
-            in eager mode, in graph mode the value will always be set to True.
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple. This argument can be used in
+            eager mode, in graph mode the value will always be set to True.
         training (`bool`, *optional*, defaults to `False`):
             Whether or not to use the model in training mode (some modules like dropout modules have different
             behaviors between training and evaluation).
@@ -700,8 +700,8 @@ class TFBlenderbotEncoder(tf.keras.layers.Layer):
                 for more detail. This argument can be used only in eager mode, in graph mode the value in the config
                 will be used instead.
             return_dict (`bool`, *optional*):
-                Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple. This argument can be
-                used in eager mode, in graph mode the value will always be set to True.
+                Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple. This argument can be used
+                in eager mode, in graph mode the value will always be set to True.
             training (`bool`, *optional*, defaults to `False`):
                 Whether or not to use the model in training mode (some modules like dropout modules have different
                 behaviors between training and evaluation).
@@ -885,8 +885,8 @@ class TFBlenderbotDecoder(tf.keras.layers.Layer):
                 for more detail. This argument can be used only in eager mode, in graph mode the value in the config
                 will be used instead.
             return_dict (`bool`, *optional*):
-                Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple. This argument can be
-                used in eager mode, in graph mode the value will always be set to True.
+                Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple. This argument can be used
+                in eager mode, in graph mode the value will always be set to True.
             training (`bool`, *optional*, defaults to `False`):
                 Whether or not to use the model in training mode (some modules like dropout modules have different
                 behaviors between training and evaluation).

src/transformers/models/blenderbot_small/modeling_blenderbot_small.py

@@ -623,7 +623,7 @@ BLENDERBOT_SMALL_INPUTS_DOCSTRING = r"""
             Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
             more detail.
         return_dict (`bool`, *optional*):
-            Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple.
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
 """
 
 
@@ -708,7 +708,7 @@ class BlenderbotSmallEncoder(BlenderbotSmallPreTrainedModel):
                 Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors
                 for more detail.
             return_dict (`bool`, *optional*):
-                Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple.
+                Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
         """
         output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
         output_hidden_states = (
@@ -930,7 +930,7 @@ class BlenderbotSmallDecoder(BlenderbotSmallPreTrainedModel):
                 Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors
                 for more detail.
             return_dict (`bool`, *optional*):
-                Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple.
+                Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
         """
         output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
         output_hidden_states = (
@@ -1489,7 +1489,7 @@ class BlenderbotSmallForCausalLM(BlenderbotSmallPreTrainedModel):
                 Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors
                 for more detail.
             return_dict (`bool`, *optional*):
-                Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple.
+                Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
 
         Returns:
 

src/transformers/models/blenderbot_small/modeling_flax_blenderbot_small.py

@@ -136,7 +136,7 @@ BLENDERBOT_SMALL_INPUTS_DOCSTRING = r"""
             Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
             more detail.
         return_dict (`bool`, *optional*):
-            Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple.
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
 """
 
 
@@ -167,7 +167,7 @@ BLENDERBOT_SMALL_ENCODE_INPUTS_DOCSTRING = r"""
             Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
             more detail.
         return_dict (`bool`, *optional*):
-            Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple.
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
 """
 
 BLENDERBOT_SMALL_DECODE_INPUTS_DOCSTRING = r"""
@@ -213,7 +213,7 @@ BLENDERBOT_SMALL_DECODE_INPUTS_DOCSTRING = r"""
             Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
             more detail.
         return_dict (`bool`, *optional*):
-            Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple.
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
 """
 
 

src/transformers/models/blenderbot_small/modeling_tf_blenderbot_small.py

@@ -613,8 +613,8 @@ BLENDERBOT_SMALL_INPUTS_DOCSTRING = r"""
             more detail. This argument can be used only in eager mode, in graph mode the value in the config will be
             used instead.
         return_dict (`bool`, *optional*):
-            Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple. This argument can be used
-            in eager mode, in graph mode the value will always be set to True.
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple. This argument can be used in
+            eager mode, in graph mode the value will always be set to True.
         training (`bool`, *optional*, defaults to `False`):
             Whether or not to use the model in training mode (some modules like dropout modules have different
             behaviors between training and evaluation).
@@ -705,8 +705,8 @@ class TFBlenderbotSmallEncoder(tf.keras.layers.Layer):
                 for more detail. This argument can be used only in eager mode, in graph mode the value in the config
                 will be used instead.
             return_dict (`bool`, *optional*):
-                Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple. This argument can be
-                used in eager mode, in graph mode the value will always be set to True.
+                Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple. This argument can be used
+                in eager mode, in graph mode the value will always be set to True.
             training (`bool`, *optional*, defaults to `False`):
                 Whether or not to use the model in training mode (some modules like dropout modules have different
                 behaviors between training and evaluation).
@@ -889,8 +889,8 @@ class TFBlenderbotSmallDecoder(tf.keras.layers.Layer):
                 for more detail. This argument can be used only in eager mode, in graph mode the value in the config
                 will be used instead.
             return_dict (`bool`, *optional*):
-                Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple. This argument can be
-                used in eager mode, in graph mode the value will always be set to True.
+                Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple. This argument can be used
+                in eager mode, in graph mode the value will always be set to True.
             training (`bool`, *optional*, defaults to `False`):
                 Whether or not to use the model in training mode (some modules like dropout modules have different
                 behaviors between training and evaluation).

src/transformers/models/canine/modeling_canine.py

@@ -975,7 +975,7 @@ CANINE_INPUTS_DOCSTRING = r"""
             Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
             more detail.
         return_dict (`bool`, *optional*):
-            Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple.
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
 """
 
 

src/transformers/models/clip/feature_extraction_clip.py

@@ -104,7 +104,7 @@ class CLIPFeatureExtractor(FeatureExtractionMixin, ImageFeatureExtractionMixin):
                 tensor. In case of a NumPy array/PyTorch tensor, each image should be of shape (C, H, W), where C is a
                 number of channels, H and W are image height and width.
 
-            return_tensors (`str` or [`~file_utils.TensorType`], *optional*, defaults to `'np'`):
+            return_tensors (`str` or [`~utils.TensorType`], *optional*, defaults to `'np'`):
                 If set, will return tensors of a particular framework. Acceptable values are:
 
                 - `'tf'`: Return TensorFlow `tf.constant` objects.

src/transformers/models/clip/modeling_clip.py

@@ -433,7 +433,7 @@ CLIP_TEXT_INPUTS_DOCSTRING = r"""
             Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
             more detail.
         return_dict (`bool`, *optional*):
-            Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple.
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
 """
 
 CLIP_VISION_INPUTS_DOCSTRING = r"""
@@ -448,7 +448,7 @@ CLIP_VISION_INPUTS_DOCSTRING = r"""
             Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
             more detail.
         return_dict (`bool`, *optional*):
-            Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple.
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
 """
 
 CLIP_INPUTS_DOCSTRING = r"""
@@ -485,7 +485,7 @@ CLIP_INPUTS_DOCSTRING = r"""
             Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
             more detail.
         return_dict (`bool`, *optional*):
-            Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple.
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
 """
 
 
@@ -540,7 +540,7 @@ class CLIPEncoder(nn.Module):
                 Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors
                 for more detail.
             return_dict (`bool`, *optional*):
-                Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple.
+                Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
         """
         output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
         output_hidden_states = (

src/transformers/models/clip/modeling_flax_clip.py

@@ -100,7 +100,7 @@ CLIP_TEXT_INPUTS_DOCSTRING = r"""
             Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
             more detail.
         return_dict (`bool`, *optional*):
-            Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple.
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
 """
 
 CLIP_VISION_INPUTS_DOCSTRING = r"""
@@ -115,7 +115,7 @@ CLIP_VISION_INPUTS_DOCSTRING = r"""
             Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
             more detail.
         return_dict (`bool`, *optional*):
-            Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple.
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
 """
 
 CLIP_INPUTS_DOCSTRING = r"""
@@ -150,7 +150,7 @@ CLIP_INPUTS_DOCSTRING = r"""
             Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
             more detail.
         return_dict (`bool`, *optional*):
-            Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple.
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
 """
 
 

src/transformers/models/clip/modeling_tf_clip.py

@@ -968,8 +968,8 @@ CLIP_TEXT_INPUTS_DOCSTRING = r"""
             more detail. This argument can be used only in eager mode, in graph mode the value in the config will be
             used instead.
         return_dict (`bool`, *optional*):
-            Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple. This argument can be used
-            in eager mode, in graph mode the value will always be set to True.
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple. This argument can be used in
+            eager mode, in graph mode the value will always be set to True.
         training (`bool`, *optional*, defaults to `False``):
             Whether or not to use the model in training mode (some modules like dropout modules have different
             behaviors between training and evaluation).
@@ -988,8 +988,8 @@ CLIP_VISION_INPUTS_DOCSTRING = r"""
             more detail. This argument can be used only in eager mode, in graph mode the value in the config will be
             used instead.
         return_dict (`bool`, *optional*):
-            Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple. This argument can be used
-            in eager mode, in graph mode the value will always be set to True.
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple. This argument can be used in
+            eager mode, in graph mode the value will always be set to True.
         training (`bool`, *optional*, defaults to `False``):
             Whether or not to use the model in training mode (some modules like dropout modules have different
             behaviors between training and evaluation).
@@ -1030,8 +1030,8 @@ CLIP_INPUTS_DOCSTRING = r"""
             more detail. This argument can be used only in eager mode, in graph mode the value in the config will be
             used instead.
         return_dict (`bool`, *optional*):
-            Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple. This argument can be used
-            in eager mode, in graph mode the value will always be set to True.
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple. This argument can be used in
+            eager mode, in graph mode the value will always be set to True.
         training (`bool`, *optional*, defaults to `False``):
             Whether or not to use the model in training mode (some modules like dropout modules have different
             behaviors between training and evaluation).

src/transformers/models/clip/processing_clip.py

@@ -57,7 +57,7 @@ class CLIPProcessor(ProcessorMixin):
                 tensor. In case of a NumPy array/PyTorch tensor, each image should be of shape (C, H, W), where C is a
                 number of channels, H and W are image height and width.
 
-            return_tensors (`str` or [`~file_utils.TensorType`], *optional*):
+            return_tensors (`str` or [`~utils.TensorType`], *optional*):
                 If set, will return tensors of a particular framework. Acceptable values are:
 
                 - `'tf'`: Return TensorFlow `tf.constant` objects.

src/transformers/models/convbert/modeling_convbert.py

@@ -755,7 +755,7 @@ CONVBERT_INPUTS_DOCSTRING = r"""
             Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
             more detail.
         return_dict (`bool`, *optional*):
-            Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple.
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
 """
 
 

src/transformers/models/convbert/modeling_tf_convbert.py

@@ -736,8 +736,8 @@ CONVBERT_INPUTS_DOCSTRING = r"""
             more detail. This argument can be used only in eager mode, in graph mode the value in the config will be
             used instead.
         return_dict (`bool`, *optional*):
-            Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple. This argument can be used
-            in eager mode, in graph mode the value will always be set to True.
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple. This argument can be used in
+            eager mode, in graph mode the value will always be set to True.
         training (`bool`, *optional*, defaults to `False`):
             Whether or not to use the model in training mode (some modules like dropout modules have different
             behaviors between training and evaluation).

src/transformers/models/convnext/feature_extraction_convnext.py

@@ -103,7 +103,7 @@ class ConvNextFeatureExtractor(FeatureExtractionMixin, ImageFeatureExtractionMix
                 tensor. In case of a NumPy array/PyTorch tensor, each image should be of shape (C, H, W), where C is a
                 number of channels, H and W are image height and width.
 
-            return_tensors (`str` or [`~file_utils.TensorType`], *optional*, defaults to `'np'`):
+            return_tensors (`str` or [`~utils.TensorType`], *optional*, defaults to `'np'`):
                 If set, will return tensors of a particular framework. Acceptable values are:
 
                 - `'tf'`: Return TensorFlow `tf.constant` objects.

src/transformers/models/convnext/modeling_convnext.py

@@ -357,7 +357,7 @@ CONVNEXT_INPUTS_DOCSTRING = r"""
             Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
             more detail.
         return_dict (`bool`, *optional*):
-            Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple.
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
 """
 
 

src/transformers/models/convnext/modeling_tf_convnext.py

@@ -416,8 +416,8 @@ CONVNEXT_INPUTS_DOCSTRING = r"""
             more detail. This argument can be used only in eager mode, in graph mode the value in the config will be
             used instead.
         return_dict (`bool`, *optional*):
-            Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple. This argument can be used
-            in eager mode, in graph mode the value will always be set to True.
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple. This argument can be used in
+            eager mode, in graph mode the value will always be set to True.
 """
 
 

src/transformers/models/ctrl/modeling_ctrl.py

@@ -309,7 +309,7 @@ CTRL_INPUTS_DOCSTRING = r"""
             Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
             more detail.
         return_dict (`bool`, *optional*):
-            Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple.
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
 """
 
 

src/transformers/models/ctrl/modeling_tf_ctrl.py

@@ -502,8 +502,8 @@ CTRL_INPUTS_DOCSTRING = r"""
             more detail. This argument can be used only in eager mode, in graph mode the value in the config will be
             used instead.
         return_dict (`bool`, *optional*):
-            Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple. This argument can be used
-            in eager mode, in graph mode the value will always be set to True.
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple. This argument can be used in
+            eager mode, in graph mode the value will always be set to True.
         training (`bool`, *optional*, defaults to `False`):
             Whether or not to use the model in training mode (some modules like dropout modules have different
             behaviors between training and evaluation).

src/transformers/models/data2vec/modeling_data2vec_audio.py

@@ -887,7 +887,7 @@ DATA2VEC_AUDIO_INPUTS_DOCSTRING = r"""
             Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
             more detail.
         return_dict (`bool`, *optional*):
-            Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple.
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
 """
 
 

src/transformers/models/data2vec/modeling_data2vec_text.py

@@ -689,7 +689,7 @@ DATA2VECTEXT_INPUTS_DOCSTRING = r"""
             Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
             more detail.
         return_dict (`bool`, *optional*):
-            Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple.
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
 """