Update vision docstring bool masked pos (#22237)

* Add bool_masked_pos to forward docstrings

* Add note about mask ratio - videomae

* Fix up

* Fix indenting

src/transformers/models/beit/modeling_beit.py

@@ -659,6 +659,10 @@ class BeitModel(BeitPreTrainedModel):
         output_hidden_states: Optional[bool] = None,
         return_dict: Optional[bool] = None,
     ) -> Union[tuple, BeitModelOutputWithPooling]:
+        r"""
+        bool_masked_pos (`torch.BoolTensor` of shape `(batch_size, num_patches)`, *optional*):
+            Boolean masked positions. Indicates which patches are masked (1) and which aren't (0).
+        """
         output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
         output_hidden_states = (
             output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states

src/transformers/models/beit/modeling_flax_beit.py

@@ -831,7 +831,7 @@ class FlaxBeitForMaskedImageModeling(FlaxBeitPreTrainedModel):
 
 FLAX_BEIT_MLM_DOCSTRING = """
     bool_masked_pos (`numpy.ndarray` of shape `(batch_size, num_patches)`):
-            Boolean masked positions. Indicates which patches are masked (1) and which aren't (0).
+        Boolean masked positions. Indicates which patches are masked (1) and which aren't (0).
 
     Returns:
 

src/transformers/models/data2vec/modeling_data2vec_vision.py

@@ -673,6 +673,10 @@ class Data2VecVisionModel(Data2VecVisionPreTrainedModel):
         output_hidden_states: Optional[bool] = None,
         return_dict: Optional[bool] = None,
     ) -> Union[tuple, Data2VecVisionModelOutputWithPooling]:
+        r"""
+        bool_masked_pos (`torch.BoolTensor` of shape `(batch_size, num_patches)`, *optional*):
+            Boolean masked positions. Indicates which patches are masked (1) and which aren't (0).
+        """
         output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
         output_hidden_states = (
             output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states

src/transformers/models/data2vec/modeling_tf_data2vec_vision.py

@@ -907,6 +907,10 @@ class TFData2VecVisionModel(TFData2VecVisionPreTrainedModel):
         return_dict: Optional[bool] = None,
         training: bool = False,
     ) -> Union[tuple, TFData2VecVisionModelOutputWithPooling]:
+        r"""
+        bool_masked_pos (`tf.Tensor` of shape `(batch_size, num_patches)`, *optional*):
+            Boolean masked positions. Indicates which patches are masked (1) and which aren't (0).
+        """
         outputs = self.data2vec_vision(
             pixel_values=pixel_values,
             bool_masked_pos=bool_masked_pos,

src/transformers/models/deit/modeling_deit.py

@@ -495,6 +495,10 @@ class DeiTModel(DeiTPreTrainedModel):
         output_hidden_states: Optional[bool] = None,
         return_dict: Optional[bool] = None,
     ) -> Union[Tuple, BaseModelOutputWithPooling]:
+        r"""
+        bool_masked_pos (`torch.BoolTensor` of shape `(batch_size, num_patches)`, *optional*):
+            Boolean masked positions. Indicates which patches are masked (1) and which aren't (0).
+        """
         output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
         output_hidden_states = (
             output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states

src/transformers/models/deit/modeling_tf_deit.py

@@ -666,6 +666,10 @@ class TFDeiTModel(TFDeiTPreTrainedModel):
         return_dict: Optional[bool] = None,
         training: bool = False,
     ):
+        r"""
+        bool_masked_pos (`tf.Tensor` of shape `(batch_size, num_patches)`, *optional*):
+            Boolean masked positions. Indicates which patches are masked (1) and which aren't (0).
+        """
         outputs = self.deit(
             pixel_values=pixel_values,
             bool_masked_pos=bool_masked_pos,

src/transformers/models/donut/modeling_donut_swin.py

@@ -911,6 +911,10 @@ class DonutSwinModel(DonutSwinPreTrainedModel):
         output_hidden_states: Optional[bool] = None,
         return_dict: Optional[bool] = None,
     ) -> Union[Tuple, DonutSwinModelOutput]:
+        r"""
+        bool_masked_pos (`torch.BoolTensor` of shape `(batch_size, num_patches)`):
+            Boolean masked positions. Indicates which patches are masked (1) and which aren't (0).
+        """
         output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
         output_hidden_states = (
             output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states

src/transformers/models/swin/modeling_swin.py

@@ -976,6 +976,10 @@ class SwinModel(SwinPreTrainedModel):
         output_hidden_states: Optional[bool] = None,
         return_dict: Optional[bool] = None,
     ) -> Union[Tuple, SwinModelOutput]:
+        r"""
+        bool_masked_pos (`torch.BoolTensor` of shape `(batch_size, num_patches)`, *optional*):
+            Boolean masked positions. Indicates which patches are masked (1) and which aren't (0).
+        """
         output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
         output_hidden_states = (
             output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states

src/transformers/models/swin/modeling_tf_swin.py

@@ -1208,6 +1208,10 @@ class TFSwinModel(TFSwinPreTrainedModel):
         return_dict: Optional[bool] = None,
         training: bool = False,
     ) -> Union[TFSwinModelOutput, Tuple[tf.Tensor, ...]]:
+        r"""
+        bool_masked_pos (`tf.Tensor` of shape `(batch_size, num_patches)`, *optional*):
+            Boolean masked positions. Indicates which patches are masked (1) and which aren't (0).
+        """
         output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
         output_hidden_states = (
             output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states

src/transformers/models/swinv2/modeling_swinv2.py

@@ -1055,6 +1055,10 @@ class Swinv2Model(Swinv2PreTrainedModel):
         output_hidden_states: Optional[bool] = None,
         return_dict: Optional[bool] = None,
     ) -> Union[Tuple, Swinv2ModelOutput]:
+        r"""
+        bool_masked_pos (`torch.BoolTensor` of shape `(batch_size, num_patches)`, *optional*):
+            Boolean masked positions. Indicates which patches are masked (1) and which aren't (0).
+        """
         output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
         output_hidden_states = (
             output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states

src/transformers/models/videomae/modeling_videomae.py

@@ -571,6 +571,11 @@ class VideoMAEModel(VideoMAEPreTrainedModel):
         return_dict: Optional[bool] = None,
     ) -> Union[Tuple, BaseModelOutput]:
         r"""
+        bool_masked_pos (`torch.BoolTensor` of shape `(batch_size, sequence_length)`, *optional*):
+            Boolean masked positions. Indicates which patches are masked (1) and which aren't (0). Each video in the
+            batch must have the same number of masked patches. If `None`, then all patches are considered. Sequence
+            length is `(num_frames // tubelet_size) * (image_size // patch_size) ** 2`.
+
         Returns:
 
         Examples:
@@ -781,6 +786,11 @@ class VideoMAEForPreTraining(VideoMAEPreTrainedModel):
         return_dict: Optional[bool] = None,
     ) -> Union[tuple, VideoMAEForPreTrainingOutput]:
         r"""
+        bool_masked_pos (`torch.BoolTensor` of shape `(batch_size, sequence_length)`):
+            Boolean masked positions. Indicates which patches are masked (1) and which aren't (0). Each video in the
+            batch must have the same number of masked patches. Sequence length is `(num_frames // tubelet_size) *
+            (image_size // patch_size) ** 2`.
+
         Returns:
 
         Examples:

src/transformers/models/vit/modeling_vit.py

@@ -549,6 +549,10 @@ class ViTModel(ViTPreTrainedModel):
         interpolate_pos_encoding: Optional[bool] = None,
         return_dict: Optional[bool] = None,
     ) -> Union[Tuple, BaseModelOutputWithPooling]:
+        r"""
+        bool_masked_pos (`torch.BoolTensor` of shape `(batch_size, num_patches)`, *optional*):
+            Boolean masked positions. Indicates which patches are masked (1) and which aren't (0).
+        """
         output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
         output_hidden_states = (
             output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states

src/transformers/models/vit_hybrid/modeling_vit_hybrid.py

@@ -573,6 +573,10 @@ class ViTHybridModel(ViTHybridPreTrainedModel):
         interpolate_pos_encoding: Optional[bool] = None,
         return_dict: Optional[bool] = None,
     ) -> Union[Tuple, BaseModelOutputWithPooling]:
+        r"""
+        bool_masked_pos (`torch.BoolTensor` of shape `(batch_size, num_patches)`, *optional*):
+            Boolean masked positions. Indicates which patches are masked (1) and which aren't (0).
+        """
         output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
         output_hidden_states = (
             output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states

src/transformers/models/vit_msn/modeling_vit_msn.py

@@ -525,6 +525,9 @@ class ViTMSNModel(ViTMSNPreTrainedModel):
         return_dict: Optional[bool] = None,
     ) -> Union[tuple, BaseModelOutput]:
         r"""
+        bool_masked_pos (`torch.BoolTensor` of shape `(batch_size, num_patches)`, *optional*):
+            Boolean masked positions. Indicates which patches are masked (1) and which aren't (0).
+
         Returns:
 
         Examples: