Vit deit fixes (#11309)

* Improve docs of DeiT and ViT, add community notebook

* Add gitignore for test_samples

* Add notebook with Trainer

Co-authored-by: Lysandre Debut <lysandre@huggingface.co>

docs/source/community.md

@@ -52,6 +52,8 @@ This page regroups resources around 🤗 Transformers developed by the community
 |[Fine-tune BART for summarization in two languages with Trainer class](https://github.com/elsanns/xai-nlp-notebooks/blob/master/fine_tune_bart_summarization_two_langs.ipynb) | How to fine-tune BART for summarization in two languages with Trainer class | [Eliza Szczechla](https://github.com/elsanns) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/elsanns/xai-nlp-notebooks/blob/master/fine_tune_bart_summarization_two_langs.ipynb)|
 |[Evaluate Big Bird on Trivia QA](https://github.com/patrickvonplaten/notebooks/blob/master/Evaluating_Big_Bird_on_TriviaQA.ipynb) | How to evaluate BigBird on long document question answering on Trivia QA | [Patrick von Platen](https://github.com/patrickvonplaten) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/patrickvonplaten/notebooks/blob/master/Evaluating_Big_Bird_on_TriviaQA.ipynb)|
 | [Create video captions using Wav2Vec2](https://github.com/Muennighoff/ytclipcc/blob/main/wav2vec_youtube_captions.ipynb) | How to create YouTube captions from any video by transcribing the audio with Wav2Vec | [Niklas Muennighoff](https://github.com/Muennighoff) |[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/Muennighoff/ytclipcc/blob/main/wav2vec_youtube_captions.ipynb) |
+| [Fine-tune the Vision Transformer on CIFAR-10 using PyTorch Lightning](https://github.com/NielsRogge/Transformers-Tutorials/blob/master/VisionTransformer/Fine_tuning_the_Vision_Transformer_on_CIFAR_10_with_PyTorch_Lightning.ipynb) | How to fine-tune the Vision Transformer (ViT) on CIFAR-10 using HuggingFace Transformers, Datasets and PyTorch Lightning | [Niels Rogge](https://github.com/nielsrogge) |[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/NielsRogge/Transformers-Tutorials/blob/master/VisionTransformer/Fine_tuning_the_Vision_Transformer_on_CIFAR_10_with_PyTorch_Lightning.ipynb) |
+| [Fine-tune the Vision Transformer on CIFAR-10 using the 🤗 Trainer](https://github.com/NielsRogge/Transformers-Tutorials/blob/master/VisionTransformer/Fine_tuning_the_Vision_Transformer_on_CIFAR_10_with_the_%F0%9F%A4%97_Trainer.ipynb) | How to fine-tune the Vision Transformer (ViT) on CIFAR-10 using HuggingFace Transformers, Datasets and the 🤗 Trainer | [Niels Rogge](https://github.com/nielsrogge) |[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/NielsRogge/Transformers-Tutorials/blob/master/VisionTransformer/Fine_tuning_the_Vision_Transformer_on_CIFAR_10_with_the_%F0%9F%A4%97_Trainer.ipynb) |
 | [Evaluate LUKE on Open Entity, an entity typing dataset](https://github.com/studio-ousia/luke/blob/master/notebooks/huggingface_open_entity.ipynb) | How to evaluate *LukeForEntityClassification* on the Open Entity dataset | [Ikuya Yamada](https://github.com/ikuyamada) |[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/studio-ousia/luke/blob/master/notebooks/huggingface_open_entity.ipynb) |
 | [Evaluate LUKE on TACRED, a relation extraction dataset](https://github.com/studio-ousia/luke/blob/master/notebooks/huggingface_tacred.ipynb) | How to evaluate *LukeForEntityPairClassification* on the TACRED dataset | [Ikuya Yamada](https://github.com/ikuyamada) |[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/studio-ousia/luke/blob/master/notebooks/huggingface_tacred.ipynb) |
 | [Evaluate LUKE on CoNLL-2003, an important NER benchmark](https://github.com/studio-ousia/luke/blob/master/notebooks/huggingface_conll_2003.ipynb) | How to evaluate *LukeForEntitySpanClassification* on the CoNLL-2003 dataset | [Ikuya Yamada](https://github.com/ikuyamada) |[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/studio-ousia/luke/blob/master/notebooks/huggingface_conll_2003.ipynb) |

src/transformers/models/deit/feature_extraction_deit.py

@@ -38,8 +38,10 @@ class DeiTFeatureExtractor(FeatureExtractionMixin, ImageFeatureExtractionMixin):
     Args:
         do_resize (:obj:`bool`, `optional`, defaults to :obj:`True`):
             Whether to resize the input to a certain :obj:`size`.
-        size (:obj:`int`, `optional`, defaults to 256):
+        size (:obj:`int` or :obj:`Tuple(int)`, `optional`, defaults to 256):
-            Resize the input to the given size. Only has an effect if :obj:`do_resize` is set to :obj:`True`.
+            Resize the input to the given size. If a tuple is provided, it should be (width, height). If only an
+            integer is provided, then the input will be resized to (size, size). Only has an effect if :obj:`do_resize`
+            is set to :obj:`True`.
         resample (:obj:`int`, `optional`, defaults to :obj:`PIL.Image.BICUBIC`):
             An optional resampling filter. This can be one of :obj:`PIL.Image.NEAREST`, :obj:`PIL.Image.BOX`,
             :obj:`PIL.Image.BILINEAR`, :obj:`PIL.Image.HAMMING`, :obj:`PIL.Image.BICUBIC` or :obj:`PIL.Image.LANCZOS`.
@@ -115,7 +117,8 @@ class DeiTFeatureExtractor(FeatureExtractionMixin, ImageFeatureExtractionMixin):
         Returns:
             :class:`~transformers.BatchFeature`: A :class:`~transformers.BatchFeature` with the following fields:
 
-            - **pixel_values** -- Pixel values to be fed to a model.
+            - **pixel_values** -- Pixel values to be fed to a model, of shape (batch_size, num_channels, height,
+              width).
         """
         # Input type checking for clearer error
         valid_images = False

src/transformers/models/deit/modeling_deit.py

@@ -417,9 +417,8 @@ DEIT_START_DOCSTRING = r"""
 DEIT_INPUTS_DOCSTRING = r"""
     Args:
         pixel_values (:obj:`torch.FloatTensor` of shape :obj:`(batch_size, num_channels, height, width)`):
-            Pixel values. Padding will be ignored by default should you provide it. Pixel values can be obtained using
+            Pixel values. Pixel values can be obtained using :class:`~transformers.DeiTFeatureExtractor`. See
-            :class:`~transformers.DeiTFeatureExtractor`. See :meth:`transformers.DeiTFeatureExtractor.__call__` for
+            :meth:`transformers.DeiTFeatureExtractor.__call__` for details.
-            details.
 
         head_mask (:obj:`torch.FloatTensor` of shape :obj:`(num_heads,)` or :obj:`(num_layers, num_heads)`, `optional`):
             Mask to nullify selected heads of the self-attention modules. Mask values selected in ``[0, 1]``:

src/transformers/models/vit/feature_extraction_vit.py

@@ -38,8 +38,10 @@ class ViTFeatureExtractor(FeatureExtractionMixin, ImageFeatureExtractionMixin):
     Args:
         do_resize (:obj:`bool`, `optional`, defaults to :obj:`True`):
             Whether to resize the input to a certain :obj:`size`.
-        size (:obj:`int`, `optional`, defaults to 224):
+        size (:obj:`int` or :obj:`Tuple(int)`, `optional`, defaults to 224):
-            Resize the input to the given size. Only has an effect if :obj:`do_resize` is set to :obj:`True`.
+            Resize the input to the given size. If a tuple is provided, it should be (width, height). If only an
+            integer is provided, then the input will be resized to (size, size). Only has an effect if :obj:`do_resize`
+            is set to :obj:`True`.
         resample (:obj:`int`, `optional`, defaults to :obj:`PIL.Image.BILINEAR`):
             An optional resampling filter. This can be one of :obj:`PIL.Image.NEAREST`, :obj:`PIL.Image.BOX`,
             :obj:`PIL.Image.BILINEAR`, :obj:`PIL.Image.HAMMING`, :obj:`PIL.Image.BICUBIC` or :obj:`PIL.Image.LANCZOS`.
@@ -105,7 +107,8 @@ class ViTFeatureExtractor(FeatureExtractionMixin, ImageFeatureExtractionMixin):
         Returns:
             :class:`~transformers.BatchFeature`: A :class:`~transformers.BatchFeature` with the following fields:
 
-            - **pixel_values** -- Pixel values to be fed to a model.
+            - **pixel_values** -- Pixel values to be fed to a model, of shape (batch_size, num_channels, height,
+              width).
         """
         # Input type checking for clearer error
         valid_images = False

src/transformers/models/vit/modeling_vit.py

@@ -403,9 +403,8 @@ VIT_START_DOCSTRING = r"""
 VIT_INPUTS_DOCSTRING = r"""
     Args:
         pixel_values (:obj:`torch.FloatTensor` of shape :obj:`(batch_size, num_channels, height, width)`):
-            Pixel values. Padding will be ignored by default should you provide it. Pixel values can be obtained using
+            Pixel values. Pixel values can be obtained using :class:`~transformers.ViTFeatureExtractor`. See
-            :class:`~transformers.ViTFeatureExtractor`. See :meth:`transformers.ViTFeatureExtractor.__call__` for
+            :meth:`transformers.ViTFeatureExtractor.__call__` for details.
-            details.
 
         head_mask (:obj:`torch.FloatTensor` of shape :obj:`(num_heads,)` or :obj:`(num_layers, num_heads)`, `optional`):
             Mask to nullify selected heads of the self-attention modules. Mask values selected in ``[0, 1]``: