From 42ecf48d9093c292131ae5f57e29d525c6790795 Mon Sep 17 00:00:00 2001
From: Pablo Montalvo <pablo.montalvo.leroux@gmail.com>
Date: Thu, 25 Jan 2024 11:02:14 +0000
Subject: [PATCH 01/29] expand kwargs from align

---
 .../models/align/processing_align.py          | 88 +++++++++++++++++--
 1 file changed, 82 insertions(+), 6 deletions(-)

diff --git a/src/transformers/models/align/processing_align.py b/src/transformers/models/align/processing_align.py
index 0863c11310e318..ab8671c495fb4d 100644
--- a/src/transformers/models/align/processing_align.py
+++ b/src/transformers/models/align/processing_align.py
@@ -17,8 +17,11 @@
 """
 
 
+from typing import Dict, List, Optional, Union
+
 from ...processing_utils import ProcessorMixin
-from ...tokenization_utils_base import BatchEncoding
+from ...tokenization_utils_base import BatchEncoding, PreTokenizedInput, TextInput, TruncationStrategy
+from ...utils import PaddingStrategy, TensorType
 
 
 class AlignProcessor(ProcessorMixin):
@@ -42,11 +45,49 @@ class AlignProcessor(ProcessorMixin):
     def __init__(self, image_processor, tokenizer):
         super().__init__(image_processor, tokenizer)
 
-    def __call__(self, text=None, images=None, padding="max_length", max_length=64, return_tensors=None, **kwargs):
+    def __call__(
+        self,
+        text=None,
+        images=None,
+        do_crop_margin: bool = None,
+        do_resize: bool = None,
+        size: Dict[str, int] = None,
+        resample: "PILImageResampling" = None,  # noqa: F821
+        do_thumbnail: bool = None,
+        do_align_long_axis: bool = None,
+        do_pad: bool = None,
+        do_rescale: bool = None,
+        rescale_factor: Union[int, float] = None,
+        do_normalize: bool = None,
+        image_mean: Optional[Union[float, List[float]]] = None,
+        image_std: Optional[Union[float, List[float]]] = None,
+        data_format: Optional["ChannelDimension"] = "channels_first",  # noqa: F821
+        input_data_format: Optional[Union[str, "ChannelDimension"]] = None,  # noqa: F821
+        text_pair: Optional[Union[TextInput, PreTokenizedInput, List[TextInput], List[PreTokenizedInput]]] = None,
+        text_target: Union[TextInput, PreTokenizedInput, List[TextInput], List[PreTokenizedInput]] = None,
+        text_pair_target: Optional[
+            Union[TextInput, PreTokenizedInput, List[TextInput], List[PreTokenizedInput]]
+        ] = None,
+        add_special_tokens: bool = True,
+        padding: Union[bool, str, PaddingStrategy] = "max_length",
+        truncation: Union[bool, str, TruncationStrategy] = None,
+        max_length: Optional[int] = 64,
+        stride: int = 0,
+        is_split_into_words: bool = False,
+        pad_to_multiple_of: Optional[int] = None,
+        return_tensors: Optional[Union[str, TensorType]] = None,
+        return_token_type_ids: Optional[bool] = None,
+        return_attention_mask: Optional[bool] = None,
+        return_overflowing_tokens: bool = False,
+        return_special_tokens_mask: bool = False,
+        return_offsets_mapping: bool = False,
+        return_length: bool = False,
+        verbose: bool = True,
+    ):
         """
         Main method to prepare text(s) and image(s) to be fed as input to the model. This method forwards the `text`
-        and `kwargs` arguments to BertTokenizerFast's [`~BertTokenizerFast.__call__`] if `text` is not `None` to encode
-        the text. To prepare the image(s), this method forwards the `images` and `kwargs` arguments to
+        arguments to BertTokenizerFast's [`~BertTokenizerFast.__call__`] if `text` is not `None` to encode
+        the text. To prepare the image(s), this method forwards the `images` arguments to
         EfficientNetImageProcessor's [`~EfficientNetImageProcessor.__call__`] if `images` is not `None`. Please refer
         to the doctsring of the above two methods for more information.
 
@@ -87,11 +128,46 @@ def __call__(self, text=None, images=None, padding="max_length", max_length=64,
 
         if text is not None:
             encoding = self.tokenizer(
-                text, padding=padding, max_length=max_length, return_tensors=return_tensors, **kwargs
+                text,
+                text_pair=text_pair,
+                text_target=text_target,
+                text_pair_target=text_pair_target,
+                add_special_tokens=add_special_tokens,
+                padding=padding,
+                truncation=truncation,
+                max_length=max_length,
+                stride=stride,
+                is_split_into_words=is_split_into_words,
+                pad_to_multiple_of=pad_to_multiple_of,
+                return_tensors=return_tensors,
+                return_token_type_ids=return_token_type_ids,
+                return_attention_mask=return_attention_mask,
+                return_overflowing_tokens=return_overflowing_tokens,
+                return_special_tokens_mask=return_special_tokens_mask,
+                return_offsets_mapping=return_offsets_mapping,
+                return_length=return_length,
+                verbose=verbose,
             )
 
         if images is not None:
-            image_features = self.image_processor(images, return_tensors=return_tensors, **kwargs)
+            image_features = self.image_processor(
+                images,
+                do_crop_margin=do_crop_margin,
+                do_resize=do_resize,
+                size=size,
+                resample=resample,
+                do_thumbnail=do_thumbnail,
+                do_align_long_axis=do_align_long_axis,
+                do_pad=do_pad,
+                do_rescale=do_rescale,
+                rescale_factor=rescale_factor,
+                do_normalize=do_normalize,
+                image_mean=image_mean,
+                image_std=image_std,
+                return_tensors=return_tensors,
+                data_format=data_format,
+                input_data_format=input_data_format,
+            )
 
         if text is not None and images is not None:
             encoding["pixel_values"] = image_features.pixel_values

From ccb214774e9e7653d7af540f0447286b17ea8e60 Mon Sep 17 00:00:00 2001
From: Pablo Montalvo <pablo.montalvo.leroux@gmail.com>
Date: Thu, 25 Jan 2024 11:16:22 +0000
Subject: [PATCH 02/29] remove kwargs from altclip processor

---
 .../models/altclip/processing_altclip.py      | 94 +++++++++++++++++--
 1 file changed, 85 insertions(+), 9 deletions(-)

diff --git a/src/transformers/models/altclip/processing_altclip.py b/src/transformers/models/altclip/processing_altclip.py
index e9b4f45269ca76..8388b9c416e79e 100644
--- a/src/transformers/models/altclip/processing_altclip.py
+++ b/src/transformers/models/altclip/processing_altclip.py
@@ -16,9 +16,11 @@
 Image/Text processor class for AltCLIP
 """
 import warnings
+from typing import Dict, List, Optional, Union
 
 from ...processing_utils import ProcessorMixin
-from ...tokenization_utils_base import BatchEncoding
+from ...tokenization_utils_base import BatchEncoding, PreTokenizedInput, TextInput, TruncationStrategy
+from ...utils import PaddingStrategy, TensorType
 
 
 class AltCLIPProcessor(ProcessorMixin):
@@ -34,22 +36,21 @@ class AltCLIPProcessor(ProcessorMixin):
             The image processor is a required input.
         tokenizer ([`XLMRobertaTokenizerFast`], *optional*):
             The tokenizer is a required input.
+        feature_extractor ([`CLIPFeatureExtractor`], *optional*):
+            The feature extractor is a deprecated input.
     """
 
     attributes = ["image_processor", "tokenizer"]
     image_processor_class = "CLIPImageProcessor"
     tokenizer_class = ("XLMRobertaTokenizer", "XLMRobertaTokenizerFast")
 
-    def __init__(self, image_processor=None, tokenizer=None, **kwargs):
-        feature_extractor = None
-        if "feature_extractor" in kwargs:
+    def __init__(self, image_processor=None, tokenizer=None, feature_extractor=None):
+        if "feature_extractor":
             warnings.warn(
                 "The `feature_extractor` argument is deprecated and will be removed in v5, use `image_processor`"
                 " instead.",
                 FutureWarning,
             )
-            feature_extractor = kwargs.pop("feature_extractor")
-
         image_processor = image_processor if image_processor is not None else feature_extractor
         if image_processor is None:
             raise ValueError("You need to specify an `image_processor`.")
@@ -58,7 +59,45 @@ def __init__(self, image_processor=None, tokenizer=None, **kwargs):
 
         super().__init__(image_processor, tokenizer)
 
-    def __call__(self, text=None, images=None, return_tensors=None, **kwargs):
+    def __call__(
+        self,
+        text=None,
+        images=None,
+        do_crop_margin: bool = None,
+        do_resize: bool = None,
+        size: Dict[str, int] = None,
+        resample: "PILImageResampling" = None,  # noqa: F821
+        do_thumbnail: bool = None,
+        do_align_long_axis: bool = None,
+        do_pad: bool = None,
+        do_rescale: bool = None,
+        rescale_factor: Union[int, float] = None,
+        do_normalize: bool = None,
+        image_mean: Optional[Union[float, List[float]]] = None,
+        image_std: Optional[Union[float, List[float]]] = None,
+        data_format: Optional["ChannelDimension"] = "channels_first",  # noqa: F821
+        input_data_format: Optional[Union[str, "ChannelDimension"]] = None,  # noqa: F821
+        text_pair: Optional[Union[TextInput, PreTokenizedInput, List[TextInput], List[PreTokenizedInput]]] = None,
+        text_target: Union[TextInput, PreTokenizedInput, List[TextInput], List[PreTokenizedInput]] = None,
+        text_pair_target: Optional[
+            Union[TextInput, PreTokenizedInput, List[TextInput], List[PreTokenizedInput]]
+        ] = None,
+        add_special_tokens: bool = True,
+        padding: Union[bool, str, PaddingStrategy] = False,
+        truncation: Union[bool, str, TruncationStrategy] = None,
+        max_length: Optional[int] = None,
+        stride: int = 0,
+        is_split_into_words: bool = False,
+        pad_to_multiple_of: Optional[int] = None,
+        return_tensors: Optional[Union[str, TensorType]] = None,
+        return_token_type_ids: Optional[bool] = None,
+        return_attention_mask: Optional[bool] = None,
+        return_overflowing_tokens: bool = False,
+        return_special_tokens_mask: bool = False,
+        return_offsets_mapping: bool = False,
+        return_length: bool = False,
+        verbose: bool = True,
+    ):
         """
         Main method to prepare for the model one or several sequences(s) and image(s). This method forwards the `text`
         and `kwargs` arguments to XLMRobertaTokenizerFast's [`~XLMRobertaTokenizerFast.__call__`] if `text` is not
@@ -98,10 +137,47 @@ def __call__(self, text=None, images=None, return_tensors=None, **kwargs):
             raise ValueError("You have to specify either text or images. Both cannot be none.")
 
         if text is not None:
-            encoding = self.tokenizer(text, return_tensors=return_tensors, **kwargs)
+            encoding = self.tokenizer(
+                text,
+                text_pair=text_pair,
+                text_target=text_target,
+                text_pair_target=text_pair_target,
+                add_special_tokens=add_special_tokens,
+                padding=padding,
+                truncation=truncation,
+                max_length=max_length,
+                stride=stride,
+                is_split_into_words=is_split_into_words,
+                pad_to_multiple_of=pad_to_multiple_of,
+                return_tensors=return_tensors,
+                return_token_type_ids=return_token_type_ids,
+                return_attention_mask=return_attention_mask,
+                return_overflowing_tokens=return_overflowing_tokens,
+                return_special_tokens_mask=return_special_tokens_mask,
+                return_offsets_mapping=return_offsets_mapping,
+                return_length=return_length,
+                verbose=verbose,
+            )
 
         if images is not None:
-            image_features = self.image_processor(images, return_tensors=return_tensors, **kwargs)
+            image_features = self.image_processor(
+                images,
+                do_crop_margin=do_crop_margin,
+                do_resize=do_resize,
+                size=size,
+                resample=resample,
+                do_thumbnail=do_thumbnail,
+                do_align_long_axis=do_align_long_axis,
+                do_pad=do_pad,
+                do_rescale=do_rescale,
+                rescale_factor=rescale_factor,
+                do_normalize=do_normalize,
+                image_mean=image_mean,
+                image_std=image_std,
+                return_tensors=return_tensors,
+                data_format=data_format,
+                input_data_format=input_data_format,
+            )
 
         if text is not None and images is not None:
             encoding["pixel_values"] = image_features.pixel_values

From f999e0c40f144621e0051a96c3a615ede062ec17 Mon Sep 17 00:00:00 2001
From: Pablo Montalvo <pablo.montalvo.leroux@gmail.com>
Date: Thu, 25 Jan 2024 13:04:46 +0000
Subject: [PATCH 03/29] add explicit args for donut processor

---
 .../models/donut/image_processing_donut.py    | 12 +++-
 .../models/donut/processing_donut.py          | 62 +++++++++++++++----
 2 files changed, 60 insertions(+), 14 deletions(-)

diff --git a/src/transformers/models/donut/image_processing_donut.py b/src/transformers/models/donut/image_processing_donut.py
index 2a1672e22041fb..6298a65d70aaf1 100644
--- a/src/transformers/models/donut/image_processing_donut.py
+++ b/src/transformers/models/donut/image_processing_donut.py
@@ -84,6 +84,13 @@ class DonutImageProcessor(BaseImageProcessor):
             channels in the image. Can be overridden by the `image_mean` parameter in the `preprocess` method.
         image_std (`float` or `List[float]`, *optional*, defaults to `IMAGENET_STANDARD_STD`):
             Image standard deviation.
+        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.
+                - `'pt'`: Return PyTorch `torch.Tensor` objects.
+                - `'np'`: Return NumPy `np.ndarray` objects.
+                - `'jax'`: Return JAX `jnp.ndarray` objects.
     """
 
     model_input_names = ["pixel_values"]
@@ -101,7 +108,9 @@ def __init__(
         do_normalize: bool = True,
         image_mean: Optional[Union[float, List[float]]] = None,
         image_std: Optional[Union[float, List[float]]] = None,
-        **kwargs,
+        return_tensors: Optional[Union[str, TensorType]] = None,
+        **kwargs
+        
     ) -> None:
         super().__init__(**kwargs)
 
@@ -122,6 +131,7 @@ def __init__(
         self.do_normalize = do_normalize
         self.image_mean = image_mean if image_mean is not None else IMAGENET_STANDARD_MEAN
         self.image_std = image_std if image_std is not None else IMAGENET_STANDARD_STD
+        self.return_tensors = return_tensors
 
     def align_long_axis(
         self,
diff --git a/src/transformers/models/donut/processing_donut.py b/src/transformers/models/donut/processing_donut.py
index 5636ecb9435cf3..91de6768f8b782 100644
--- a/src/transformers/models/donut/processing_donut.py
+++ b/src/transformers/models/donut/processing_donut.py
@@ -19,8 +19,12 @@
 import warnings
 from contextlib import contextmanager
 
-from ...processing_utils import ProcessorMixin
+from typing import Dict, List, Optional, Union
+
+from transformers.tokenization_utils_base import PreTokenizedInput, TextInput, TruncationStrategy
 
+from ...processing_utils import ProcessorMixin
+from ...utils import PaddingStrategy, TensorType
 
 class DonutProcessor(ProcessorMixin):
     r"""
@@ -36,21 +40,21 @@ class DonutProcessor(ProcessorMixin):
             An instance of [`DonutImageProcessor`]. The image processor is a required input.
         tokenizer ([`XLMRobertaTokenizer`/`XLMRobertaTokenizerFast`], *optional*):
             An instance of [`XLMRobertaTokenizer`/`XLMRobertaTokenizerFast`]. The tokenizer is a required input.
+        feature_extractor ([`CLIPFeatureExtractor`], *optional*):
+            The feature extractor is a deprecated input.
     """
 
     attributes = ["image_processor", "tokenizer"]
     image_processor_class = "AutoImageProcessor"
     tokenizer_class = "AutoTokenizer"
 
-    def __init__(self, image_processor=None, tokenizer=None, **kwargs):
-        feature_extractor = None
-        if "feature_extractor" in kwargs:
+    def __init__(self, image_processor=None, tokenizer=None, feature_extractor=None):
+        if "feature_extractor":
             warnings.warn(
                 "The `feature_extractor` argument is deprecated and will be removed in v5, use `image_processor`"
                 " instead.",
                 FutureWarning,
             )
-            feature_extractor = kwargs.pop("feature_extractor")
 
         image_processor = image_processor if image_processor is not None else feature_extractor
         if image_processor is None:
@@ -62,22 +66,54 @@ def __init__(self, image_processor=None, tokenizer=None, **kwargs):
         self.current_processor = self.image_processor
         self._in_target_context_manager = False
 
-    def __call__(self, *args, **kwargs):
+    def __call__(self,
+        text=None,
+        images=None,
+        do_crop_margin: bool = None,
+        do_resize: bool = None,
+        size: Dict[str, int] = None,
+        resample: "PILImageResampling" = None,  # noqa: F821
+        do_thumbnail: bool = None,
+        do_align_long_axis: bool = None,
+        do_pad: bool = None,
+        do_rescale: bool = None,
+        rescale_factor: Union[int, float] = None,
+        do_normalize: bool = None,
+        image_mean: Optional[Union[float, List[float]]] = None,
+        image_std: Optional[Union[float, List[float]]] = None,
+        data_format: Optional["ChannelDimension"] = "channels_first",  # noqa: F821
+        input_data_format: Optional[Union[str, "ChannelDimension"]] = None,  # noqa: F821
+        text_pair: Optional[Union[TextInput, PreTokenizedInput, List[TextInput], List[PreTokenizedInput]]] = None,
+        text_target: Union[TextInput, PreTokenizedInput, List[TextInput], List[PreTokenizedInput]] = None,
+        text_pair_target: Optional[
+            Union[TextInput, PreTokenizedInput, List[TextInput], List[PreTokenizedInput]]
+        ] = None,
+        add_special_tokens: bool = True,
+        padding: Union[bool, str, PaddingStrategy] = False,
+        truncation: Union[bool, str, TruncationStrategy] = None,
+        max_length: Optional[int] = None,
+        stride: int = 0,
+        is_split_into_words: bool = False,
+        pad_to_multiple_of: Optional[int] = None,
+        return_tensors: Optional[Union[str, TensorType]] = None,
+        return_token_type_ids: Optional[bool] = None,
+        return_attention_mask: Optional[bool] = None,
+        return_overflowing_tokens: bool = False,
+        return_special_tokens_mask: bool = False,
+        return_offsets_mapping: bool = False,
+        return_length: bool = False,
+        verbose: bool = True,
+        ):
         """
         When used in normal mode, this method forwards all its arguments to AutoImageProcessor's
         [`~AutoImageProcessor.__call__`] and returns its output. If used in the context
         [`~DonutProcessor.as_target_processor`] this method forwards all its arguments to DonutTokenizer's
-        [`~DonutTokenizer.__call__`]. Please refer to the doctsring of the above two methods for more information.
+        [`~DonutTokenizer.__call__`]. Please refer to the docstring of the above two methods for more information.
         """
         # For backward compatibility
         if self._in_target_context_manager:
-            return self.current_processor(*args, **kwargs)
+            return self.current_processor(*args)
 
-        images = kwargs.pop("images", None)
-        text = kwargs.pop("text", None)
-        if len(args) > 0:
-            images = args[0]
-            args = args[1:]
 
         if images is None and text is None:
             raise ValueError("You need to specify either an `images` or `text` input to process.")

From 8fb3a6bf0d6b5c498d4d36be0155988b9c2c9c0e Mon Sep 17 00:00:00 2001
From: Pablo Montalvo <pablo.montalvo.leroux@gmail.com>
Date: Thu, 25 Jan 2024 13:20:04 +0000
Subject: [PATCH 04/29] add explicit call to current processor for in context
 manager

---
 .../models/donut/processing_donut.py          | 56 ++++++++++++++++++-
 1 file changed, 53 insertions(+), 3 deletions(-)

diff --git a/src/transformers/models/donut/processing_donut.py b/src/transformers/models/donut/processing_donut.py
index 91de6768f8b782..8c31fe88127aa4 100644
--- a/src/transformers/models/donut/processing_donut.py
+++ b/src/transformers/models/donut/processing_donut.py
@@ -112,16 +112,66 @@ def __call__(self,
         """
         # For backward compatibility
         if self._in_target_context_manager:
-            return self.current_processor(*args)
+            return self.current_processor(
+                                images,
+                do_crop_margin=do_crop_margin,
+                do_resize=do_resize,
+                size=size,
+                resample=resample,
+                do_thumbnail=do_thumbnail,
+                do_align_long_axis=do_align_long_axis,
+                do_pad=do_pad,
+                do_rescale=do_rescale,
+                rescale_factor=rescale_factor,
+                do_normalize=do_normalize,
+                image_mean=image_mean,
+                image_std=image_std,
+                return_tensors=return_tensors,
+                data_format=data_format,
+                input_data_format=input_data_format
+            )
 
 
         if images is None and text is None:
             raise ValueError("You need to specify either an `images` or `text` input to process.")
 
         if images is not None:
-            inputs = self.image_processor(images, *args, **kwargs)
+            inputs = self.image_processor(images,
+                do_crop_margin=do_crop_margin,
+                do_resize=do_resize,
+                size=size,
+                resample=resample,
+                do_thumbnail=do_thumbnail,
+                do_align_long_axis=do_align_long_axis,
+                do_pad=do_pad,
+                do_rescale=do_rescale,
+                rescale_factor=rescale_factor,
+                do_normalize=do_normalize,
+                image_mean=image_mean,
+                image_std=image_std,
+                return_tensors=return_tensors,
+                data_format=data_format,
+                input_data_format=input_data_format,)
         if text is not None:
-            encodings = self.tokenizer(text, **kwargs)
+            encodings = self.tokenizer(text,
+                text_pair=text_pair,
+                text_target=text_target,
+                text_pair_target=text_pair_target,
+                add_special_tokens=add_special_tokens,
+                padding=padding,
+                truncation=truncation,
+                max_length=max_length,
+                stride=stride,
+                is_split_into_words=is_split_into_words,
+                pad_to_multiple_of=pad_to_multiple_of,
+                return_tensors=return_tensors,
+                return_token_type_ids=return_token_type_ids,
+                return_attention_mask=return_attention_mask,
+                return_overflowing_tokens=return_overflowing_tokens,
+                return_special_tokens_mask=return_special_tokens_mask,
+                return_offsets_mapping=return_offsets_mapping,
+                return_length=return_length,
+                verbose=verbose,)
 
         if text is None:
             return inputs

From a90c766d002e074311a80ba06c26af75d3c3ff58 Mon Sep 17 00:00:00 2001
From: Pablo Montalvo <pablo.montalvo.leroux@gmail.com>
Date: Thu, 25 Jan 2024 15:06:04 +0000
Subject: [PATCH 05/29] format

---
 .../models/donut/image_processing_donut.py    |  3 +--
 .../models/donut/processing_donut.py          | 24 +++++++++++--------
 2 files changed, 15 insertions(+), 12 deletions(-)

diff --git a/src/transformers/models/donut/image_processing_donut.py b/src/transformers/models/donut/image_processing_donut.py
index 6298a65d70aaf1..7b24e3fe7e3b71 100644
--- a/src/transformers/models/donut/image_processing_donut.py
+++ b/src/transformers/models/donut/image_processing_donut.py
@@ -109,8 +109,7 @@ def __init__(
         image_mean: Optional[Union[float, List[float]]] = None,
         image_std: Optional[Union[float, List[float]]] = None,
         return_tensors: Optional[Union[str, TensorType]] = None,
-        **kwargs
-        
+        **kwargs,
     ) -> None:
         super().__init__(**kwargs)
 
diff --git a/src/transformers/models/donut/processing_donut.py b/src/transformers/models/donut/processing_donut.py
index 8c31fe88127aa4..ae2588b85e2040 100644
--- a/src/transformers/models/donut/processing_donut.py
+++ b/src/transformers/models/donut/processing_donut.py
@@ -18,7 +18,6 @@
 import re
 import warnings
 from contextlib import contextmanager
-
 from typing import Dict, List, Optional, Union
 
 from transformers.tokenization_utils_base import PreTokenizedInput, TextInput, TruncationStrategy
@@ -26,6 +25,7 @@
 from ...processing_utils import ProcessorMixin
 from ...utils import PaddingStrategy, TensorType
 
+
 class DonutProcessor(ProcessorMixin):
     r"""
     Constructs a Donut processor which wraps a Donut image processor and an XLMRoBERTa tokenizer into a single
@@ -66,7 +66,8 @@ def __init__(self, image_processor=None, tokenizer=None, feature_extractor=None)
         self.current_processor = self.image_processor
         self._in_target_context_manager = False
 
-    def __call__(self,
+    def __call__(
+        self,
         text=None,
         images=None,
         do_crop_margin: bool = None,
@@ -103,7 +104,7 @@ def __call__(self,
         return_offsets_mapping: bool = False,
         return_length: bool = False,
         verbose: bool = True,
-        ):
+    ):
         """
         When used in normal mode, this method forwards all its arguments to AutoImageProcessor's
         [`~AutoImageProcessor.__call__`] and returns its output. If used in the context
@@ -113,7 +114,7 @@ def __call__(self,
         # For backward compatibility
         if self._in_target_context_manager:
             return self.current_processor(
-                                images,
+                images,
                 do_crop_margin=do_crop_margin,
                 do_resize=do_resize,
                 size=size,
@@ -128,15 +129,15 @@ def __call__(self,
                 image_std=image_std,
                 return_tensors=return_tensors,
                 data_format=data_format,
-                input_data_format=input_data_format
+                input_data_format=input_data_format,
             )
 
-
         if images is None and text is None:
             raise ValueError("You need to specify either an `images` or `text` input to process.")
 
         if images is not None:
-            inputs = self.image_processor(images,
+            inputs = self.image_processor(
+                images,
                 do_crop_margin=do_crop_margin,
                 do_resize=do_resize,
                 size=size,
@@ -151,9 +152,11 @@ def __call__(self,
                 image_std=image_std,
                 return_tensors=return_tensors,
                 data_format=data_format,
-                input_data_format=input_data_format,)
+                input_data_format=input_data_format,
+            )
         if text is not None:
-            encodings = self.tokenizer(text,
+            encodings = self.tokenizer(
+                text,
                 text_pair=text_pair,
                 text_target=text_target,
                 text_pair_target=text_pair_target,
@@ -171,7 +174,8 @@ def __call__(self,
                 return_special_tokens_mask=return_special_tokens_mask,
                 return_offsets_mapping=return_offsets_mapping,
                 return_length=return_length,
-                verbose=verbose,)
+                verbose=verbose,
+            )
 
         if text is None:
             return inputs

From 49cb6cc44f5102b8f59889b2b9f808c3488e6639 Mon Sep 17 00:00:00 2001
From: Pablo Montalvo <pablo.montalvo.leroux@gmail.com>
Date: Thu, 25 Jan 2024 15:07:27 +0000
Subject: [PATCH 06/29] remove unused kwargs

---
 src/transformers/models/bit/image_processing_bit.py | 1 -
 1 file changed, 1 deletion(-)

diff --git a/src/transformers/models/bit/image_processing_bit.py b/src/transformers/models/bit/image_processing_bit.py
index 7aa49145ae0527..6676cd4d37c822 100644
--- a/src/transformers/models/bit/image_processing_bit.py
+++ b/src/transformers/models/bit/image_processing_bit.py
@@ -188,7 +188,6 @@ def preprocess(
         return_tensors: Optional[Union[str, TensorType]] = None,
         data_format: Optional[ChannelDimension] = ChannelDimension.FIRST,
         input_data_format: Optional[Union[str, ChannelDimension]] = None,
-        **kwargs,
     ) -> PIL.Image.Image:
         """
         Preprocess an image or batch of images.

From 3ac1c7ed66c62c82a501d7c1f13b3aa9db69af1e Mon Sep 17 00:00:00 2001
From: Pablo Montalvo <pablo.montalvo.leroux@gmail.com>
Date: Thu, 25 Jan 2024 15:40:47 +0000
Subject: [PATCH 07/29] move conditions for encodings

---
 .../models/blip/processing_blip.py            | 26 +++----------------
 .../models/blip_2/processing_blip_2.py        | 26 +++----------------
 2 files changed, 8 insertions(+), 44 deletions(-)

diff --git a/src/transformers/models/blip/processing_blip.py b/src/transformers/models/blip/processing_blip.py
index 3b9d5c369a4412..8919bf2aefcdcd 100644
--- a/src/transformers/models/blip/processing_blip.py
+++ b/src/transformers/models/blip/processing_blip.py
@@ -79,6 +79,10 @@ def __call__(
         # Get only text
         if images is None:
             self.current_processor = self.tokenizer
+        if text is None: 
+            text_encoding = None
+
+        if images is None or text is not None:
             text_encoding = self.tokenizer(
                 text=text,
                 add_special_tokens=add_special_tokens,
@@ -102,28 +106,6 @@ def __call__(
         # add pixel_values
         encoding_image_processor = self.image_processor(images, return_tensors=return_tensors)
 
-        if text is not None:
-            text_encoding = self.tokenizer(
-                text=text,
-                add_special_tokens=add_special_tokens,
-                padding=padding,
-                truncation=truncation,
-                max_length=max_length,
-                stride=stride,
-                pad_to_multiple_of=pad_to_multiple_of,
-                return_attention_mask=return_attention_mask,
-                return_overflowing_tokens=return_overflowing_tokens,
-                return_special_tokens_mask=return_special_tokens_mask,
-                return_offsets_mapping=return_offsets_mapping,
-                return_token_type_ids=return_token_type_ids,
-                return_length=return_length,
-                verbose=verbose,
-                return_tensors=return_tensors,
-                **kwargs,
-            )
-        else:
-            text_encoding = None
-
         if text_encoding is not None:
             encoding_image_processor.update(text_encoding)
 
diff --git a/src/transformers/models/blip_2/processing_blip_2.py b/src/transformers/models/blip_2/processing_blip_2.py
index ff7044c82aedb6..f1df2229d9c8b5 100644
--- a/src/transformers/models/blip_2/processing_blip_2.py
+++ b/src/transformers/models/blip_2/processing_blip_2.py
@@ -81,6 +81,10 @@ def __call__(
         # Get only text
         if images is None:
             self.current_processor = self.tokenizer
+        if text is None: 
+            text_encoding = None
+            
+        if images is None or text is not None:
             text_encoding = self.tokenizer(
                 text=text,
                 add_special_tokens=add_special_tokens,
@@ -104,28 +108,6 @@ def __call__(
         # add pixel_values
         encoding_image_processor = self.image_processor(images, return_tensors=return_tensors)
 
-        if text is not None:
-            text_encoding = self.tokenizer(
-                text=text,
-                add_special_tokens=add_special_tokens,
-                padding=padding,
-                truncation=truncation,
-                max_length=max_length,
-                stride=stride,
-                pad_to_multiple_of=pad_to_multiple_of,
-                return_attention_mask=return_attention_mask,
-                return_overflowing_tokens=return_overflowing_tokens,
-                return_special_tokens_mask=return_special_tokens_mask,
-                return_offsets_mapping=return_offsets_mapping,
-                return_token_type_ids=return_token_type_ids,
-                return_length=return_length,
-                verbose=verbose,
-                return_tensors=return_tensors,
-                **kwargs,
-            )
-        else:
-            text_encoding = None
-
         if text_encoding is not None:
             encoding_image_processor.update(text_encoding)
 

From 7a819fd77e3b2f860ee7fadc3519eae88ec37bf2 Mon Sep 17 00:00:00 2001
From: Pablo Montalvo <pablo.montalvo.leroux@gmail.com>
Date: Thu, 25 Jan 2024 16:00:59 +0000
Subject: [PATCH 08/29] improve flow over text/image

---
 .../models/blip/processing_blip.py            | 26 +++++++++----------
 .../models/blip_2/processing_blip_2.py        | 26 +++++++++----------
 2 files changed, 24 insertions(+), 28 deletions(-)

diff --git a/src/transformers/models/blip/processing_blip.py b/src/transformers/models/blip/processing_blip.py
index 8919bf2aefcdcd..713d1b8c194b82 100644
--- a/src/transformers/models/blip/processing_blip.py
+++ b/src/transformers/models/blip/processing_blip.py
@@ -76,13 +76,10 @@ def __call__(
         if images is None and text is None:
             raise ValueError("You have to specify either images or text.")
 
-        # Get only text
-        if images is None:
-            self.current_processor = self.tokenizer
-        if text is None: 
-            text_encoding = None
+        text_encoding = None
 
-        if images is None or text is not None:
+        if text is not None:
+            self.current_processor = self.tokenizer
             text_encoding = self.tokenizer(
                 text=text,
                 add_special_tokens=add_special_tokens,
@@ -101,15 +98,16 @@ def __call__(
                 return_tensors=return_tensors,
                 **kwargs,
             )
-            return text_encoding
-
-        # add pixel_values
-        encoding_image_processor = self.image_processor(images, return_tensors=return_tensors)
-
-        if text_encoding is not None:
-            encoding_image_processor.update(text_encoding)
 
-        return encoding_image_processor
+        # add pixel_values encoding. If we also have text_encoding, update image encoding and return it.
+        # else, return the text encoding.
+        if images is not None:
+            encoding_image_processor = self.image_processor(images, return_tensors=return_tensors)
+            if text_encoding is not None:
+                encoding_image_processor.update(text_encoding)
+            return encoding_image_processor
+        
+        return text_encoding
 
     def batch_decode(self, *args, **kwargs):
         """
diff --git a/src/transformers/models/blip_2/processing_blip_2.py b/src/transformers/models/blip_2/processing_blip_2.py
index f1df2229d9c8b5..54a4bdb3bc982f 100644
--- a/src/transformers/models/blip_2/processing_blip_2.py
+++ b/src/transformers/models/blip_2/processing_blip_2.py
@@ -78,13 +78,10 @@ def __call__(
         if images is None and text is None:
             raise ValueError("You have to specify either images or text.")
 
-        # Get only text
-        if images is None:
+        text_encoding = None
+
+        if text is not None:
             self.current_processor = self.tokenizer
-        if text is None: 
-            text_encoding = None
-            
-        if images is None or text is not None:
             text_encoding = self.tokenizer(
                 text=text,
                 add_special_tokens=add_special_tokens,
@@ -103,15 +100,16 @@ def __call__(
                 return_tensors=return_tensors,
                 **kwargs,
             )
-            return text_encoding
-
-        # add pixel_values
-        encoding_image_processor = self.image_processor(images, return_tensors=return_tensors)
-
-        if text_encoding is not None:
-            encoding_image_processor.update(text_encoding)
 
-        return encoding_image_processor
+        # add pixel_values encoding. If we also have text_encoding, update image encoding and return it.
+        # else, return the text encoding.
+        if images is not None:
+            encoding_image_processor = self.image_processor(images, return_tensors=return_tensors)
+            if text_encoding is not None:
+                encoding_image_processor.update(text_encoding)
+            return encoding_image_processor
+        
+        return text_encoding
 
     # Copied from transformers.models.blip.processing_blip.BlipProcessor.batch_decode with BertTokenizerFast->PreTrainedTokenizer
     def batch_decode(self, *args, **kwargs):

From 9cc38b78c41521afa4df47a325bcf284256298ef Mon Sep 17 00:00:00 2001
From: Pablo Montalvo <pablo.montalvo.leroux@gmail.com>
Date: Thu, 25 Jan 2024 17:37:23 +0000
Subject: [PATCH 09/29] [breaking] pass explicit args to bridgetower

---
 .../image_processing_bridgetower.py           | 12 +++---
 .../bridgetower/processing_bridgetower.py     | 39 +++++++++++++++----
 2 files changed, 39 insertions(+), 12 deletions(-)

diff --git a/src/transformers/models/bridgetower/image_processing_bridgetower.py b/src/transformers/models/bridgetower/image_processing_bridgetower.py
index 1e2b8ea40b0703..0ffdf27e5931cf 100644
--- a/src/transformers/models/bridgetower/image_processing_bridgetower.py
+++ b/src/transformers/models/bridgetower/image_processing_bridgetower.py
@@ -161,6 +161,9 @@ class BridgeTowerImageProcessor(BaseImageProcessor):
         do_pad (`bool`, *optional*, defaults to `True`):
             Whether to pad the image to the `(max_height, max_width)` of the images in the batch. Can be overridden by
             the `do_pad` parameter in the `preprocess` method.
+        pad_and_return_pixel_mask (`bool`, *optional*, defaults to `True`):
+            Deprecated. Whether to pad the image to the `(max_height, max_width)` of the images in the batch.
+            Sets do_pad.
     """
 
     model_input_names = ["pixel_values"]
@@ -178,12 +181,12 @@ def __init__(
         image_std: Optional[Union[float, List[float]]] = None,
         do_center_crop: bool = True,
         do_pad: bool = True,
-        **kwargs,
+        pad_and_return_pixel_mask: bool = None,
+        max_text_len: int = 512 # TODO (Molbap): find original values in model config 
     ) -> None:
-        if "pad_and_return_pixel_mask" in kwargs:
-            do_pad = kwargs.pop("pad_and_return_pixel_mask")
+        if pad_and_return_pixel_mask:
+            do_pad = pad_and_return_pixel_mask
 
-        super().__init__(**kwargs)
         size = size if size is not None else {"shortest_edge": 288}
         size = get_size_dict(size, default_to_square=False)
 
@@ -381,7 +384,6 @@ def preprocess(
         return_tensors: Optional[Union[str, TensorType]] = None,
         data_format: ChannelDimension = ChannelDimension.FIRST,
         input_data_format: Optional[Union[str, ChannelDimension]] = None,
-        **kwargs,
     ) -> PIL.Image.Image:
         """
         Preprocess an image or batch of images.
diff --git a/src/transformers/models/bridgetower/processing_bridgetower.py b/src/transformers/models/bridgetower/processing_bridgetower.py
index 7718c3bf833fec..e73a64db98d26e 100644
--- a/src/transformers/models/bridgetower/processing_bridgetower.py
+++ b/src/transformers/models/bridgetower/processing_bridgetower.py
@@ -16,7 +16,7 @@
 Processor class for BridgeTower.
 """
 
-from typing import List, Optional, Union
+from typing import List, Optional, Union, Dict
 
 from ...processing_utils import ProcessorMixin
 from ...tokenization_utils_base import BatchEncoding, PaddingStrategy, PreTokenizedInput, TextInput, TruncationStrategy
@@ -48,13 +48,27 @@ def __init__(self, image_processor, tokenizer):
 
     def __call__(
         self,
-        images,
-        text: Union[TextInput, PreTokenizedInput, List[TextInput], List[PreTokenizedInput]] = None,
+        images=None,
+        text=None,
+        do_resize: Optional[bool] = None,
+        size: Optional[Dict[str, int]] = None,
+        size_divisor: Optional[int] = None,
+        resample: "PILImageResampling" = None,
+        do_rescale: Optional[bool] = None,
+        rescale_factor: Optional[float] = None,
+        do_normalize: Optional[bool] = None,
+        image_mean: Optional[Union[float, List[float]]] = None,
+        image_std: Optional[Union[float, List[float]]] = None,
+        do_pad: Optional[bool] = None,
+        pad_and_return_pixel_mask: Optional[bool] = None,
+        do_center_crop: Optional[bool] = None,
+        return_tensors: Optional[Union[str, TensorType]] = None,
         add_special_tokens: bool = True,
         padding: Union[bool, str, PaddingStrategy] = False,
         truncation: Union[bool, str, TruncationStrategy] = None,
         max_length: Optional[int] = None,
         stride: int = 0,
+        is_split_into_words: bool = False,
         pad_to_multiple_of: Optional[int] = None,
         return_token_type_ids: Optional[bool] = None,
         return_attention_mask: Optional[bool] = None,
@@ -63,8 +77,6 @@ def __call__(
         return_offsets_mapping: bool = False,
         return_length: bool = False,
         verbose: bool = True,
-        return_tensors: Optional[Union[str, TensorType]] = None,
-        **kwargs,
     ) -> BatchEncoding:
         """
         This method uses [`BridgeTowerImageProcessor.__call__`] method to prepare image(s) for the model, and
@@ -79,6 +91,7 @@ def __call__(
             truncation=truncation,
             max_length=max_length,
             stride=stride,
+            is_split_into_words=is_split_into_words,
             pad_to_multiple_of=pad_to_multiple_of,
             return_token_type_ids=return_token_type_ids,
             return_attention_mask=return_attention_mask,
@@ -88,11 +101,23 @@ def __call__(
             return_length=return_length,
             verbose=verbose,
             return_tensors=return_tensors,
-            **kwargs,
         )
         # add pixel_values + pixel_mask
+
         encoding_image_processor = self.image_processor(
-            images, return_tensors=return_tensors, do_normalize=True, do_center_crop=True, **kwargs
+            images,
+            do_resize=do_resize,
+            size=size,
+            size_divisor=size_divisor,
+            resample=resample,
+            do_rescale=do_rescale,
+            rescale_factor=rescale_factor,
+            do_normalize=do_normalize,
+            image_mean=image_mean,
+            image_std=image_std,
+            do_center_crop=do_center_crop,
+            do_pad=do_pad,
+            pad_and_return_pixel_mask=pad_and_return_pixel_mask
         )
         encoding.update(encoding_image_processor)
 

From 7db64a06cdca3e6507d5c00e2dd444788fb9fc8c Mon Sep 17 00:00:00 2001
From: Pablo Montalvo <pablo.montalvo.leroux@gmail.com>
Date: Fri, 26 Jan 2024 13:23:06 +0000
Subject: [PATCH 10/29] add default kwargs for BC

---
 .../models/bridgetower/image_processing_bridgetower.py    | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/src/transformers/models/bridgetower/image_processing_bridgetower.py b/src/transformers/models/bridgetower/image_processing_bridgetower.py
index 0ffdf27e5931cf..75ffe4a0f8b498 100644
--- a/src/transformers/models/bridgetower/image_processing_bridgetower.py
+++ b/src/transformers/models/bridgetower/image_processing_bridgetower.py
@@ -182,7 +182,12 @@ def __init__(
         do_center_crop: bool = True,
         do_pad: bool = True,
         pad_and_return_pixel_mask: bool = None,
-        max_text_len: int = 512 # TODO (Molbap): find original values in model config 
+        max_text_len: int = 50, # From here on, kwargs are kept for backwards compatibility to load existing processors.
+        cache_dir: str = "/tmp",
+        downstream_fusion: bool = False,
+        downstream_fusion_layers: int = 1,
+        downstream_fusion_method: str = "elmo",
+        
     ) -> None:
         if pad_and_return_pixel_mask:
             do_pad = pad_and_return_pixel_mask
@@ -384,6 +389,7 @@ def preprocess(
         return_tensors: Optional[Union[str, TensorType]] = None,
         data_format: ChannelDimension = ChannelDimension.FIRST,
         input_data_format: Optional[Union[str, ChannelDimension]] = None,
+        pad_and_return_pixel_mask: Optional[bool] = None,
     ) -> PIL.Image.Image:
         """
         Preprocess an image or batch of images.

From 41674d9cb037df747995c8b9a6c905a4f7d97df7 Mon Sep 17 00:00:00 2001
From: Pablo Montalvo <pablo.montalvo.leroux@gmail.com>
Date: Fri, 26 Jan 2024 17:43:19 +0000
Subject: [PATCH 11/29] fix bridgetower

---
 .../image_processing_bridgetower.py           | 44 ++++++++++++++-----
 .../bridgetower/processing_bridgetower.py     | 20 ++++++---
 2 files changed, 46 insertions(+), 18 deletions(-)

diff --git a/src/transformers/models/bridgetower/image_processing_bridgetower.py b/src/transformers/models/bridgetower/image_processing_bridgetower.py
index 75ffe4a0f8b498..ac5ed176a9b98d 100644
--- a/src/transformers/models/bridgetower/image_processing_bridgetower.py
+++ b/src/transformers/models/bridgetower/image_processing_bridgetower.py
@@ -180,15 +180,32 @@ def __init__(
         image_mean: Optional[Union[float, List[float]]] = None,
         image_std: Optional[Union[float, List[float]]] = None,
         do_center_crop: bool = True,
-        do_pad: bool = True,
-        pad_and_return_pixel_mask: bool = None,
-        max_text_len: int = 50, # From here on, kwargs are kept for backwards compatibility to load existing processors.
-        cache_dir: str = "/tmp",
-        downstream_fusion: bool = False,
-        downstream_fusion_layers: int = 1,
-        downstream_fusion_method: str = "elmo",
-        
+        do_pad: Optional[bool] = True,
+        pad_and_return_pixel_mask: Optional[bool] = None,
+        return_tensors: Optional[bool] = None,
+        **kwargs,
     ) -> None:
+        valid_processor_keys = {
+            "do_resize",
+            "size",
+            "size_divisor",
+            "resample",
+            "do_rescale",
+            "rescale_factor",
+            "do_normalize",
+            "image_mean",
+            "image_std",
+            "do_center_crop",
+            "do_pad",
+            "pad_and_return_pixel_mask",
+            "return_tensors"
+        }
+
+        unused_keys = set(kwargs.keys()) - valid_processor_keys
+        if unused_keys:
+            unused_key_str = ", ".join(unused_keys)
+            logger.warning_once(f"Unused or unrecognized configuration parameters: {unused_key_str}.")
+
         if pad_and_return_pixel_mask:
             do_pad = pad_and_return_pixel_mask
 
@@ -206,6 +223,7 @@ def __init__(
         self.image_std = image_std if image_std is not None else OPENAI_CLIP_STD
         self.do_pad = do_pad
         self.do_center_crop = do_center_crop
+        self.return_tensors = return_tensors
 
     # Copied from transformers.models.vilt.image_processing_vilt.ViltImageProcessor.resize
     def resize(
@@ -371,7 +389,7 @@ def pad(
             data["pixel_mask"] = masks
 
         return BatchFeature(data=data, tensor_type=return_tensors)
-
+    
     def preprocess(
         self,
         images: ImageInput,
@@ -443,6 +461,9 @@ def preprocess(
                 - `"channels_first"` or `ChannelDimension.FIRST`: image in (num_channels, height, width) format.
                 - `"channels_last"` or `ChannelDimension.LAST`: image in (height, width, num_channels) format.
                 - `"none"` or `ChannelDimension.NONE`: image in (height, width) format.
+            pad_and_return_pixel_mask (`bool`, *optional*, deprecated, defaults to `self.do_pad`):
+                Whether to pad the image to the (max_height, max_width) in the batch. If `True`, a pixel mask is also
+                created and returned. Deprecated version of do_pad.
         """
         do_resize = do_resize if do_resize is not None else self.do_resize
         size_divisor = size_divisor if size_divisor is not None else self.size_divisor
@@ -452,8 +473,9 @@ def preprocess(
         do_normalize = do_normalize if do_normalize is not None else self.do_normalize
         image_mean = image_mean if image_mean is not None else self.image_mean
         image_std = image_std if image_std is not None else self.image_std
-        do_pad = do_pad if do_pad is not None else self.do_pad
-        do_center_crop if do_center_crop is not None else self.do_center_crop
+        do_pad = do_pad if (do_pad is not None or pad_and_return_pixel_mask) else self.do_pad
+        do_center_crop = do_center_crop if do_center_crop is not None else self.do_center_crop
+        return_tensors = return_tensors if return_tensors is not None else self.return_tensors
 
         size = size if size is not None else self.size
         size = get_size_dict(size, default_to_square=False)
diff --git a/src/transformers/models/bridgetower/processing_bridgetower.py b/src/transformers/models/bridgetower/processing_bridgetower.py
index e73a64db98d26e..9062b80f578880 100644
--- a/src/transformers/models/bridgetower/processing_bridgetower.py
+++ b/src/transformers/models/bridgetower/processing_bridgetower.py
@@ -16,10 +16,11 @@
 Processor class for BridgeTower.
 """
 
-from typing import List, Optional, Union, Dict
+from typing import Dict, List, Optional, Union
 
+from ...image_utils import ChannelDimension
 from ...processing_utils import ProcessorMixin
-from ...tokenization_utils_base import BatchEncoding, PaddingStrategy, PreTokenizedInput, TextInput, TruncationStrategy
+from ...tokenization_utils_base import BatchEncoding, PaddingStrategy, TruncationStrategy, TextInput, PreTokenizedInput
 from ...utils import TensorType
 
 
@@ -48,8 +49,8 @@ def __init__(self, image_processor, tokenizer):
 
     def __call__(
         self,
-        images=None,
-        text=None,
+        images,
+        text:Union[TextInput, PreTokenizedInput, List[TextInput], List[PreTokenizedInput]]=None,
         do_resize: Optional[bool] = None,
         size: Optional[Dict[str, int]] = None,
         size_divisor: Optional[int] = None,
@@ -63,6 +64,8 @@ def __call__(
         pad_and_return_pixel_mask: Optional[bool] = None,
         do_center_crop: Optional[bool] = None,
         return_tensors: Optional[Union[str, TensorType]] = None,
+        data_format: Optional[ChannelDimension] = ChannelDimension.FIRST,
+        input_data_format: Optional[Union[str, ChannelDimension]] = None,
         add_special_tokens: bool = True,
         padding: Union[bool, str, PaddingStrategy] = False,
         truncation: Union[bool, str, TruncationStrategy] = None,
@@ -99,11 +102,11 @@ def __call__(
             return_special_tokens_mask=return_special_tokens_mask,
             return_offsets_mapping=return_offsets_mapping,
             return_length=return_length,
-            verbose=verbose,
             return_tensors=return_tensors,
+            verbose=verbose,
         )
         # add pixel_values + pixel_mask
-
+        print(size)
         encoding_image_processor = self.image_processor(
             images,
             do_resize=do_resize,
@@ -116,8 +119,11 @@ def __call__(
             image_mean=image_mean,
             image_std=image_std,
             do_center_crop=do_center_crop,
+            data_format=data_format,
+            input_data_format=input_data_format,
             do_pad=do_pad,
-            pad_and_return_pixel_mask=pad_and_return_pixel_mask
+            pad_and_return_pixel_mask=pad_and_return_pixel_mask,
+            return_tensors=return_tensors,
         )
         encoding.update(encoding_image_processor)
 

From 618a687ff568ba7f9292bd54316f2b422a70bc9c Mon Sep 17 00:00:00 2001
From: Pablo Montalvo <pablo.montalvo.leroux@gmail.com>
Date: Fri, 26 Jan 2024 17:47:36 +0000
Subject: [PATCH 12/29] debug bridgetower image proc

---
 tests/models/bridgetower/test_image_processing_bridgetower.py | 1 +
 1 file changed, 1 insertion(+)

diff --git a/tests/models/bridgetower/test_image_processing_bridgetower.py b/tests/models/bridgetower/test_image_processing_bridgetower.py
index f8837fdc964a76..5665248d77af0e 100644
--- a/tests/models/bridgetower/test_image_processing_bridgetower.py
+++ b/tests/models/bridgetower/test_image_processing_bridgetower.py
@@ -143,6 +143,7 @@ def image_processor_dict(self):
         return self.image_processor_tester.prepare_image_processor_dict()
 
     def test_image_processor_properties(self):
+        print(self.image_processor_dict)
         image_processing = self.image_processing_class(**self.image_processor_dict)
         self.assertTrue(hasattr(image_processing, "image_mean"))
         self.assertTrue(hasattr(image_processing, "image_std"))

From f39cdc1b76b1ae10d70ffc009dbec09ff0f4d6ef Mon Sep 17 00:00:00 2001
From: Pablo Montalvo <pablo.montalvo.leroux@gmail.com>
Date: Fri, 26 Jan 2024 17:48:11 +0000
Subject: [PATCH 13/29] format

---
 src/transformers/models/blip/processing_blip.py     | 2 +-
 src/transformers/models/blip_2/processing_blip_2.py | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/src/transformers/models/blip/processing_blip.py b/src/transformers/models/blip/processing_blip.py
index 713d1b8c194b82..605a11a154d391 100644
--- a/src/transformers/models/blip/processing_blip.py
+++ b/src/transformers/models/blip/processing_blip.py
@@ -106,7 +106,7 @@ def __call__(
             if text_encoding is not None:
                 encoding_image_processor.update(text_encoding)
             return encoding_image_processor
-        
+
         return text_encoding
 
     def batch_decode(self, *args, **kwargs):
diff --git a/src/transformers/models/blip_2/processing_blip_2.py b/src/transformers/models/blip_2/processing_blip_2.py
index 54a4bdb3bc982f..0b11db42e53095 100644
--- a/src/transformers/models/blip_2/processing_blip_2.py
+++ b/src/transformers/models/blip_2/processing_blip_2.py
@@ -108,7 +108,7 @@ def __call__(
             if text_encoding is not None:
                 encoding_image_processor.update(text_encoding)
             return encoding_image_processor
-        
+
         return text_encoding
 
     # Copied from transformers.models.blip.processing_blip.BlipProcessor.batch_decode with BertTokenizerFast->PreTrainedTokenizer

From 9a6f97d8b27cba43f2ece9722f58da5cb405f6f6 Mon Sep 17 00:00:00 2001
From: Pablo Montalvo <pablo.montalvo.leroux@gmail.com>
Date: Fri, 26 Jan 2024 17:58:32 +0000
Subject: [PATCH 14/29] move kwargs message to info level

---
 .../models/bridgetower/image_processing_bridgetower.py          | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/src/transformers/models/bridgetower/image_processing_bridgetower.py b/src/transformers/models/bridgetower/image_processing_bridgetower.py
index ac5ed176a9b98d..4510c9b852700b 100644
--- a/src/transformers/models/bridgetower/image_processing_bridgetower.py
+++ b/src/transformers/models/bridgetower/image_processing_bridgetower.py
@@ -204,7 +204,7 @@ def __init__(
         unused_keys = set(kwargs.keys()) - valid_processor_keys
         if unused_keys:
             unused_key_str = ", ".join(unused_keys)
-            logger.warning_once(f"Unused or unrecognized configuration parameters: {unused_key_str}.")
+            logger.info(f"Unused or unrecognized configuration parameters: {unused_key_str}.")
 
         if pad_and_return_pixel_mask:
             do_pad = pad_and_return_pixel_mask

From 380f82f515da9637374788d8bf2c62a3e34b0d54 Mon Sep 17 00:00:00 2001
From: Pablo Montalvo <pablo.montalvo.leroux@gmail.com>
Date: Fri, 26 Jan 2024 18:14:32 +0000
Subject: [PATCH 15/29] add debug messages

---
 .../image_processing_bridgetower.py           | 31 +++++++++++++++++--
 1 file changed, 29 insertions(+), 2 deletions(-)

diff --git a/src/transformers/models/bridgetower/image_processing_bridgetower.py b/src/transformers/models/bridgetower/image_processing_bridgetower.py
index 4510c9b852700b..1d083d61ebaae4 100644
--- a/src/transformers/models/bridgetower/image_processing_bridgetower.py
+++ b/src/transformers/models/bridgetower/image_processing_bridgetower.py
@@ -171,7 +171,7 @@ class BridgeTowerImageProcessor(BaseImageProcessor):
     def __init__(
         self,
         do_resize: bool = True,
-        size: Dict[str, int] = 288,
+        size: Dict[str, int] = {"shortest_edge": 288},
         size_divisor: int = 32,
         resample: PILImageResampling = PILImageResampling.BICUBIC,
         do_rescale: bool = True,
@@ -183,8 +183,12 @@ def __init__(
         do_pad: Optional[bool] = True,
         pad_and_return_pixel_mask: Optional[bool] = None,
         return_tensors: Optional[bool] = None,
+        data_format: Optional["ChannelDimension"] = "channels_first",  # noqa: F821
+        input_data_format: Optional[Union[str, "ChannelDimension"]] = None,  # noqa: F821
         **kwargs,
     ) -> None:
+        print('DEBUG', kwargs)
+
         valid_processor_keys = {
             "do_resize",
             "size",
@@ -224,6 +228,8 @@ def __init__(
         self.do_pad = do_pad
         self.do_center_crop = do_center_crop
         self.return_tensors = return_tensors
+        self.data_format = data_format
+        self.input_data_format = input_data_format
 
     # Copied from transformers.models.vilt.image_processing_vilt.ViltImageProcessor.resize
     def resize(
@@ -476,10 +482,31 @@ def preprocess(
         do_pad = do_pad if (do_pad is not None or pad_and_return_pixel_mask) else self.do_pad
         do_center_crop = do_center_crop if do_center_crop is not None else self.do_center_crop
         return_tensors = return_tensors if return_tensors is not None else self.return_tensors
-
+        input_data_format = input_data_format if input_data_format is not None else self.input_data_format
+        data_format = data_format if data_format is not None else self.data_format
+        
         size = size if size is not None else self.size
+
         size = get_size_dict(size, default_to_square=False)
 
+        print('DEBUG')
+        print(f'images: {images}')
+        print(f'do_resize: {do_resize}')
+        print(f'size: {size}')
+        print(f'size_divisor: {size_divisor}')
+        print(f'resample: {resample}')
+        print(f'do_rescale: {do_rescale}')
+        print(f'rescale_factor: {rescale_factor}')
+        print(f'do_normalize: {do_normalize}')
+        print(f'image_mean: {image_mean}')
+        print(f'image_std: {image_std}')
+        print(f'do_pad: {do_pad}')
+        print(f'do_center_crop: {do_center_crop}')
+        print(f'return_tensors: {return_tensors}')
+        print(f'data_format: {data_format}')
+        print(f'input_data_format: {input_data_format}')
+        print(f'pad_and_return_pixel_mask: {pad_and_return_pixel_mask}')
+
         if not is_batched(images):
             images = [images]
 

From 75f15d37ef1c7fe5f70d1bf5c9ea3cd279ec7fee Mon Sep 17 00:00:00 2001
From: Pablo Montalvo <pablo.montalvo.leroux@gmail.com>
Date: Fri, 26 Jan 2024 18:38:17 +0000
Subject: [PATCH 16/29] fix arguments not being passed in bridgetower

---
 .../image_processing_bridgetower.py           | 21 +------------------
 .../test_image_processing_bridgetower.py      |  2 +-
 2 files changed, 2 insertions(+), 21 deletions(-)

diff --git a/src/transformers/models/bridgetower/image_processing_bridgetower.py b/src/transformers/models/bridgetower/image_processing_bridgetower.py
index 1d083d61ebaae4..6033eab37ba6a3 100644
--- a/src/transformers/models/bridgetower/image_processing_bridgetower.py
+++ b/src/transformers/models/bridgetower/image_processing_bridgetower.py
@@ -179,7 +179,7 @@ def __init__(
         do_normalize: bool = True,
         image_mean: Optional[Union[float, List[float]]] = None,
         image_std: Optional[Union[float, List[float]]] = None,
-        do_center_crop: bool = True,
+        do_center_crop: Optional[bool] = None,
         do_pad: Optional[bool] = True,
         pad_and_return_pixel_mask: Optional[bool] = None,
         return_tensors: Optional[bool] = None,
@@ -187,7 +187,6 @@ def __init__(
         input_data_format: Optional[Union[str, "ChannelDimension"]] = None,  # noqa: F821
         **kwargs,
     ) -> None:
-        print('DEBUG', kwargs)
 
         valid_processor_keys = {
             "do_resize",
@@ -489,24 +488,6 @@ def preprocess(
 
         size = get_size_dict(size, default_to_square=False)
 
-        print('DEBUG')
-        print(f'images: {images}')
-        print(f'do_resize: {do_resize}')
-        print(f'size: {size}')
-        print(f'size_divisor: {size_divisor}')
-        print(f'resample: {resample}')
-        print(f'do_rescale: {do_rescale}')
-        print(f'rescale_factor: {rescale_factor}')
-        print(f'do_normalize: {do_normalize}')
-        print(f'image_mean: {image_mean}')
-        print(f'image_std: {image_std}')
-        print(f'do_pad: {do_pad}')
-        print(f'do_center_crop: {do_center_crop}')
-        print(f'return_tensors: {return_tensors}')
-        print(f'data_format: {data_format}')
-        print(f'input_data_format: {input_data_format}')
-        print(f'pad_and_return_pixel_mask: {pad_and_return_pixel_mask}')
-
         if not is_batched(images):
             images = [images]
 
diff --git a/tests/models/bridgetower/test_image_processing_bridgetower.py b/tests/models/bridgetower/test_image_processing_bridgetower.py
index 5665248d77af0e..fee9041447fddd 100644
--- a/tests/models/bridgetower/test_image_processing_bridgetower.py
+++ b/tests/models/bridgetower/test_image_processing_bridgetower.py
@@ -39,7 +39,7 @@ def __init__(
         do_rescale: bool = True,
         rescale_factor: Union[int, float] = 1 / 255,
         do_normalize: bool = True,
-        do_center_crop: bool = True,
+        do_center_crop: bool = False,
         image_mean: Optional[Union[float, List[float]]] = [0.48145466, 0.4578275, 0.40821073],
         image_std: Optional[Union[float, List[float]]] = [0.26862954, 0.26130258, 0.27577711],
         do_pad: bool = True,

From 3df5faad79e975e7e7695c9a9399677b3183ad4f Mon Sep 17 00:00:00 2001
From: Pablo Montalvo <pablo.montalvo.leroux@gmail.com>
Date: Thu, 1 Feb 2024 17:03:04 +0000
Subject: [PATCH 17/29] keep backwards compat for processing + modify testing
 args dict

---
 .../models/bridgetower/image_processing_bridgetower.py          | 2 +-
 tests/models/bridgetower/test_image_processing_bridgetower.py   | 1 +
 2 files changed, 2 insertions(+), 1 deletion(-)

diff --git a/src/transformers/models/bridgetower/image_processing_bridgetower.py b/src/transformers/models/bridgetower/image_processing_bridgetower.py
index 6033eab37ba6a3..38f8ab6b2e13c9 100644
--- a/src/transformers/models/bridgetower/image_processing_bridgetower.py
+++ b/src/transformers/models/bridgetower/image_processing_bridgetower.py
@@ -179,7 +179,7 @@ def __init__(
         do_normalize: bool = True,
         image_mean: Optional[Union[float, List[float]]] = None,
         image_std: Optional[Union[float, List[float]]] = None,
-        do_center_crop: Optional[bool] = None,
+        do_center_crop: Optional[bool] = True,
         do_pad: Optional[bool] = True,
         pad_and_return_pixel_mask: Optional[bool] = None,
         return_tensors: Optional[bool] = None,
diff --git a/tests/models/bridgetower/test_image_processing_bridgetower.py b/tests/models/bridgetower/test_image_processing_bridgetower.py
index fee9041447fddd..8a8a94c8ea6b69 100644
--- a/tests/models/bridgetower/test_image_processing_bridgetower.py
+++ b/tests/models/bridgetower/test_image_processing_bridgetower.py
@@ -70,6 +70,7 @@ def prepare_image_processor_dict(self):
             "image_std": self.image_std,
             "do_normalize": self.do_normalize,
             "do_resize": self.do_resize,
+            "do_center_crop": self.do_center_crop,
             "size": self.size,
             "size_divisor": self.size_divisor,
         }

From 69e5a2d58f86fa7814e60f64263384f5c0edb0d9 Mon Sep 17 00:00:00 2001
From: Pablo Montalvo <pablo.montalvo.leroux@gmail.com>
Date: Thu, 1 Feb 2024 17:31:53 +0000
Subject: [PATCH 18/29] fix quality

---
 .../image_processing_bridgetower.py           | 29 +++++++++++++++----
 .../bridgetower/processing_bridgetower.py     |  8 ++---
 .../test_image_processing_bridgetower.py      |  1 -
 3 files changed, 27 insertions(+), 11 deletions(-)

diff --git a/src/transformers/models/bridgetower/image_processing_bridgetower.py b/src/transformers/models/bridgetower/image_processing_bridgetower.py
index 38f8ab6b2e13c9..dce25aab84fa28 100644
--- a/src/transformers/models/bridgetower/image_processing_bridgetower.py
+++ b/src/transformers/models/bridgetower/image_processing_bridgetower.py
@@ -128,7 +128,7 @@ class BridgeTowerImageProcessor(BaseImageProcessor):
         do_resize (`bool`, *optional*, defaults to `True`):
             Whether to resize the image's (height, width) dimensions to the specified `size`. Can be overridden by the
             `do_resize` parameter in the `preprocess` method.
-        size (`Dict[str, int]` *optional*, defaults to 288):
+        size (`Dict[str, int]` *optional*, defaults to `{'shortest_edge': 288}`):
             Resize the shorter side of the input to `size["shortest_edge"]`. The longer side will be limited to under
             `int((1333 / 800) * size["shortest_edge"])` while preserving the aspect ratio. Only has an effect if
             `do_resize` is set to `True`. Can be overridden by the `size` parameter in the `preprocess` method.
@@ -161,9 +161,27 @@ class BridgeTowerImageProcessor(BaseImageProcessor):
         do_pad (`bool`, *optional*, defaults to `True`):
             Whether to pad the image to the `(max_height, max_width)` of the images in the batch. Can be overridden by
             the `do_pad` parameter in the `preprocess` method.
-        pad_and_return_pixel_mask (`bool`, *optional*, defaults to `True`):
+        pad_and_return_pixel_mask (`bool`, *optional*):
             Deprecated. Whether to pad the image to the `(max_height, max_width)` of the images in the batch.
             Sets do_pad.
+        return_tensors (`str` or `TensorType`, *optional*):
+            The type of tensors to return. Can be one of:
+            - Unset: Return a list of `np.ndarray`.
+            - `TensorType.TENSORFLOW` or `'tf'`: Return a batch of type `tf.Tensor`.
+            - `TensorType.PYTORCH` or `'pt'`: Return a batch of type `torch.Tensor`.
+            - `TensorType.NUMPY` or `'np'`: Return a batch of type `np.ndarray`.
+            - `TensorType.JAX` or `'jax'`: Return a batch of type `jax.numpy.ndarray`.
+        data_format (`ChannelDimension` or `str`, *optional*, defaults to `"channels_first"`):
+            The channel dimension format for the output image. Can be one of:
+            - `"channels_first"` or `ChannelDimension.FIRST`: image in (num_channels, height, width) format.
+            - `"channels_last"` or `ChannelDimension.LAST`: image in (height, width, num_channels) format.
+            - Unset: Use the channel dimension format of the input image.
+        input_data_format (`ChannelDimension` or `str`, *optional*):
+            The channel dimension format for the input image. If unset, the channel dimension format is inferred
+            from the input image. Can be one of:
+            - `"channels_first"` or `ChannelDimension.FIRST`: image in (num_channels, height, width) format.
+            - `"channels_last"` or `ChannelDimension.LAST`: image in (height, width, num_channels) format.
+            - `"none"` or `ChannelDimension.NONE`: image in (height, width) format.
     """
 
     model_input_names = ["pixel_values"]
@@ -187,7 +205,6 @@ def __init__(
         input_data_format: Optional[Union[str, "ChannelDimension"]] = None,  # noqa: F821
         **kwargs,
     ) -> None:
-
         valid_processor_keys = {
             "do_resize",
             "size",
@@ -201,7 +218,7 @@ def __init__(
             "do_center_crop",
             "do_pad",
             "pad_and_return_pixel_mask",
-            "return_tensors"
+            "return_tensors",
         }
 
         unused_keys = set(kwargs.keys()) - valid_processor_keys
@@ -394,7 +411,7 @@ def pad(
             data["pixel_mask"] = masks
 
         return BatchFeature(data=data, tensor_type=return_tensors)
-    
+
     def preprocess(
         self,
         images: ImageInput,
@@ -483,7 +500,7 @@ def preprocess(
         return_tensors = return_tensors if return_tensors is not None else self.return_tensors
         input_data_format = input_data_format if input_data_format is not None else self.input_data_format
         data_format = data_format if data_format is not None else self.data_format
-        
+
         size = size if size is not None else self.size
 
         size = get_size_dict(size, default_to_square=False)
diff --git a/src/transformers/models/bridgetower/processing_bridgetower.py b/src/transformers/models/bridgetower/processing_bridgetower.py
index 9062b80f578880..681af3f5cb4db0 100644
--- a/src/transformers/models/bridgetower/processing_bridgetower.py
+++ b/src/transformers/models/bridgetower/processing_bridgetower.py
@@ -18,9 +18,9 @@
 
 from typing import Dict, List, Optional, Union
 
-from ...image_utils import ChannelDimension
+from ...image_utils import ChannelDimension, PILImageResampling
 from ...processing_utils import ProcessorMixin
-from ...tokenization_utils_base import BatchEncoding, PaddingStrategy, TruncationStrategy, TextInput, PreTokenizedInput
+from ...tokenization_utils_base import BatchEncoding, PaddingStrategy, PreTokenizedInput, TextInput, TruncationStrategy
 from ...utils import TensorType
 
 
@@ -50,11 +50,11 @@ def __init__(self, image_processor, tokenizer):
     def __call__(
         self,
         images,
-        text:Union[TextInput, PreTokenizedInput, List[TextInput], List[PreTokenizedInput]]=None,
+        text: Union[TextInput, PreTokenizedInput, List[TextInput], List[PreTokenizedInput]] = None,
         do_resize: Optional[bool] = None,
         size: Optional[Dict[str, int]] = None,
         size_divisor: Optional[int] = None,
-        resample: "PILImageResampling" = None,
+        resample: PILImageResampling = None,
         do_rescale: Optional[bool] = None,
         rescale_factor: Optional[float] = None,
         do_normalize: Optional[bool] = None,
diff --git a/tests/models/bridgetower/test_image_processing_bridgetower.py b/tests/models/bridgetower/test_image_processing_bridgetower.py
index 8a8a94c8ea6b69..78b96ff3f75753 100644
--- a/tests/models/bridgetower/test_image_processing_bridgetower.py
+++ b/tests/models/bridgetower/test_image_processing_bridgetower.py
@@ -144,7 +144,6 @@ def image_processor_dict(self):
         return self.image_processor_tester.prepare_image_processor_dict()
 
     def test_image_processor_properties(self):
-        print(self.image_processor_dict)
         image_processing = self.image_processing_class(**self.image_processor_dict)
         self.assertTrue(hasattr(image_processing, "image_mean"))
         self.assertTrue(hasattr(image_processing, "image_std"))

From 68c2f40f6a885f05924c544834ba8cd9e67415b6 Mon Sep 17 00:00:00 2001
From: Pablo Montalvo <pablo.montalvo.leroux@gmail.com>
Date: Thu, 1 Feb 2024 18:18:56 +0000
Subject: [PATCH 19/29] log kwargs mismatch to info level

---
 .../image_processing_bridgetower.py           | 49 +++++++++++--------
 .../image_processing_chinese_clip.py          | 24 +++++++++
 2 files changed, 52 insertions(+), 21 deletions(-)

diff --git a/src/transformers/models/bridgetower/image_processing_bridgetower.py b/src/transformers/models/bridgetower/image_processing_bridgetower.py
index dce25aab84fa28..e639f80a79efec 100644
--- a/src/transformers/models/bridgetower/image_processing_bridgetower.py
+++ b/src/transformers/models/bridgetower/image_processing_bridgetower.py
@@ -205,27 +205,7 @@ def __init__(
         input_data_format: Optional[Union[str, "ChannelDimension"]] = None,  # noqa: F821
         **kwargs,
     ) -> None:
-        valid_processor_keys = {
-            "do_resize",
-            "size",
-            "size_divisor",
-            "resample",
-            "do_rescale",
-            "rescale_factor",
-            "do_normalize",
-            "image_mean",
-            "image_std",
-            "do_center_crop",
-            "do_pad",
-            "pad_and_return_pixel_mask",
-            "return_tensors",
-        }
-
-        unused_keys = set(kwargs.keys()) - valid_processor_keys
-        if unused_keys:
-            unused_key_str = ", ".join(unused_keys)
-            logger.info(f"Unused or unrecognized configuration parameters: {unused_key_str}.")
-
+        
         if pad_and_return_pixel_mask:
             do_pad = pad_and_return_pixel_mask
 
@@ -430,6 +410,7 @@ def preprocess(
         data_format: ChannelDimension = ChannelDimension.FIRST,
         input_data_format: Optional[Union[str, ChannelDimension]] = None,
         pad_and_return_pixel_mask: Optional[bool] = None,
+        **kwargs,
     ) -> PIL.Image.Image:
         """
         Preprocess an image or batch of images.
@@ -487,6 +468,32 @@ def preprocess(
                 Whether to pad the image to the (max_height, max_width) in the batch. If `True`, a pixel mask is also
                 created and returned. Deprecated version of do_pad.
         """
+        valid_processor_keys = {
+            "images",
+            "do_resize",
+            "size",
+            "size_divisor",
+            "resample",
+            "do_rescale",
+            "rescale_factor",
+            "do_normalize",
+            "image_mean",
+            "image_std",
+            "do_pad",
+            "do_center_crop",
+            "return_tensors",
+            "data_format",
+            "input_data_format",
+            "pad_and_return_pixel_mask",
+        }
+
+        unused_keys = set(kwargs.keys()) - valid_processor_keys
+        if unused_keys:
+            unused_key_str = ", ".join(unused_keys)
+            logger.info(f"Unused or unrecognized configuration parameters: {unused_key_str}.")
+
+
+
         do_resize = do_resize if do_resize is not None else self.do_resize
         size_divisor = size_divisor if size_divisor is not None else self.size_divisor
         resample = resample if resample is not None else self.resample
diff --git a/src/transformers/models/chinese_clip/image_processing_chinese_clip.py b/src/transformers/models/chinese_clip/image_processing_chinese_clip.py
index 4f1048a45e6ac6..c49027d4bd3e10 100644
--- a/src/transformers/models/chinese_clip/image_processing_chinese_clip.py
+++ b/src/transformers/models/chinese_clip/image_processing_chinese_clip.py
@@ -230,6 +230,30 @@ def preprocess(
                 - `"channels_last"` or `ChannelDimension.LAST`: image in (height, width, num_channels) format.
                 - `"none"` or `ChannelDimension.NONE`: image in (height, width) format.
         """
+        valid_processor_keys = {
+            "images",
+            "do_resize",
+            "size",
+            "resample",
+            "do_center_crop",
+            "crop_size",
+            "do_rescale",
+            "rescale_factor",
+            "do_normalize",
+            "image_mean",
+            "image_std",
+            "do_convert_rgb",
+            "return_tensors",
+            "data_format",
+            "input_data_format",
+        }
+
+        unused_keys = set(kwargs.keys()) - valid_processor_keys
+        if unused_keys:
+            unused_key_str = ", ".join(unused_keys)
+            logger.info(f"Unused or unrecognized configuration parameters: {unused_key_str}.")
+
+
         do_resize = do_resize if do_resize is not None else self.do_resize
         size = size if size is not None else self.size
         size = get_size_dict(size, default_to_square=False)

From e1e408421aecca34f4bb2b00635eb712b034d422 Mon Sep 17 00:00:00 2001
From: Pablo Montalvo <pablo.montalvo.leroux@gmail.com>
Date: Thu, 1 Feb 2024 18:24:21 +0000
Subject: [PATCH 20/29] fix quality

---
 .../models/bridgetower/image_processing_bridgetower.py         | 3 ---
 .../models/chinese_clip/image_processing_chinese_clip.py       | 1 -
 2 files changed, 4 deletions(-)

diff --git a/src/transformers/models/bridgetower/image_processing_bridgetower.py b/src/transformers/models/bridgetower/image_processing_bridgetower.py
index e639f80a79efec..ee53893e06b3e6 100644
--- a/src/transformers/models/bridgetower/image_processing_bridgetower.py
+++ b/src/transformers/models/bridgetower/image_processing_bridgetower.py
@@ -205,7 +205,6 @@ def __init__(
         input_data_format: Optional[Union[str, "ChannelDimension"]] = None,  # noqa: F821
         **kwargs,
     ) -> None:
-        
         if pad_and_return_pixel_mask:
             do_pad = pad_and_return_pixel_mask
 
@@ -492,8 +491,6 @@ def preprocess(
             unused_key_str = ", ".join(unused_keys)
             logger.info(f"Unused or unrecognized configuration parameters: {unused_key_str}.")
 
-
-
         do_resize = do_resize if do_resize is not None else self.do_resize
         size_divisor = size_divisor if size_divisor is not None else self.size_divisor
         resample = resample if resample is not None else self.resample
diff --git a/src/transformers/models/chinese_clip/image_processing_chinese_clip.py b/src/transformers/models/chinese_clip/image_processing_chinese_clip.py
index c49027d4bd3e10..fa62db7139de87 100644
--- a/src/transformers/models/chinese_clip/image_processing_chinese_clip.py
+++ b/src/transformers/models/chinese_clip/image_processing_chinese_clip.py
@@ -253,7 +253,6 @@ def preprocess(
             unused_key_str = ", ".join(unused_keys)
             logger.info(f"Unused or unrecognized configuration parameters: {unused_key_str}.")
 
-
         do_resize = do_resize if do_resize is not None else self.do_resize
         size = size if size is not None else self.size
         size = get_size_dict(size, default_to_square=False)

From 4b557b0ac30da058b38f7e97f328a9eb6fc60c82 Mon Sep 17 00:00:00 2001
From: Pablo Montalvo <pablo.montalvo.leroux@gmail.com>
Date: Thu, 15 Feb 2024 15:04:01 +0000
Subject: [PATCH 21/29] address comments

---
 src/transformers/models/blip/processing_blip.py               | 1 -
 .../models/bridgetower/image_processing_bridgetower.py        | 4 ++--
 2 files changed, 2 insertions(+), 3 deletions(-)

diff --git a/src/transformers/models/blip/processing_blip.py b/src/transformers/models/blip/processing_blip.py
index 605a11a154d391..3a9f9332afae25 100644
--- a/src/transformers/models/blip/processing_blip.py
+++ b/src/transformers/models/blip/processing_blip.py
@@ -79,7 +79,6 @@ def __call__(
         text_encoding = None
 
         if text is not None:
-            self.current_processor = self.tokenizer
             text_encoding = self.tokenizer(
                 text=text,
                 add_special_tokens=add_special_tokens,
diff --git a/src/transformers/models/bridgetower/image_processing_bridgetower.py b/src/transformers/models/bridgetower/image_processing_bridgetower.py
index 26d17f5045faf8..996d856de35336 100644
--- a/src/transformers/models/bridgetower/image_processing_bridgetower.py
+++ b/src/transformers/models/bridgetower/image_processing_bridgetower.py
@@ -197,8 +197,8 @@ def __init__(
         do_normalize: bool = True,
         image_mean: Optional[Union[float, List[float]]] = None,
         image_std: Optional[Union[float, List[float]]] = None,
-        do_center_crop: Optional[bool] = True,
-        do_pad: Optional[bool] = True,
+        do_center_crop: bool = True,
+        do_pad: bool = True,
         pad_and_return_pixel_mask: Optional[bool] = None,
         return_tensors: Optional[bool] = None,
         data_format: Optional["ChannelDimension"] = "channels_first",  # noqa: F821

From b7fc377f823749dda234b33a5f384da8be3cc8ba Mon Sep 17 00:00:00 2001
From: Pablo Montalvo <pablo.montalvo.leroux@gmail.com>
Date: Thu, 15 Feb 2024 15:58:19 +0000
Subject: [PATCH 22/29] fix typo

---
 utils/check_docstrings.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/utils/check_docstrings.py b/utils/check_docstrings.py
index 7c895163d95988..32d57b2976289c 100644
--- a/utils/check_docstrings.py
+++ b/utils/check_docstrings.py
@@ -827,7 +827,7 @@ def find_indent(line: str) -> int:
 def stringify_default(default: Any) -> str:
     """
     Returns the string representation of a default value, as used in docstring: numbers are left as is, all other
-    objects are in backtiks.
+    objects are in backticks.
 
     Args:
         default (`Any`): The default value to process
@@ -857,7 +857,7 @@ def stringify_default(default: Any) -> str:
 def eval_math_expression(expression: str) -> Optional[Union[float, int]]:
     # Mainly taken from the excellent https://stackoverflow.com/a/9558001
     """
-    Evaluate (safely) a mathematial expression and returns its value.
+    Evaluate (safely) a mathematical expression and returns its value.
 
     Args:
         expression (`str`): The expression to evaluate.

From 270bb9e2f2935af8d05909a5fb801a55a2ec0053 Mon Sep 17 00:00:00 2001
From: Pablo Montalvo <pablo.montalvo.leroux@gmail.com>
Date: Fri, 16 Feb 2024 14:04:13 +0000
Subject: [PATCH 23/29] fix expected tests for bridgetower

---
 src/transformers/models/blip_2/processing_blip_2.py          | 1 -
 .../models/bridgetower/processing_bridgetower.py             | 1 -
 .../models/bridgetower/test_image_processing_bridgetower.py  | 5 +++--
 3 files changed, 3 insertions(+), 4 deletions(-)

diff --git a/src/transformers/models/blip_2/processing_blip_2.py b/src/transformers/models/blip_2/processing_blip_2.py
index 0b11db42e53095..f160e3431d3e0e 100644
--- a/src/transformers/models/blip_2/processing_blip_2.py
+++ b/src/transformers/models/blip_2/processing_blip_2.py
@@ -81,7 +81,6 @@ def __call__(
         text_encoding = None
 
         if text is not None:
-            self.current_processor = self.tokenizer
             text_encoding = self.tokenizer(
                 text=text,
                 add_special_tokens=add_special_tokens,
diff --git a/src/transformers/models/bridgetower/processing_bridgetower.py b/src/transformers/models/bridgetower/processing_bridgetower.py
index 681af3f5cb4db0..56d5a756ca97ee 100644
--- a/src/transformers/models/bridgetower/processing_bridgetower.py
+++ b/src/transformers/models/bridgetower/processing_bridgetower.py
@@ -106,7 +106,6 @@ def __call__(
             verbose=verbose,
         )
         # add pixel_values + pixel_mask
-        print(size)
         encoding_image_processor = self.image_processor(
             images,
             do_resize=do_resize,
diff --git a/tests/models/bridgetower/test_image_processing_bridgetower.py b/tests/models/bridgetower/test_image_processing_bridgetower.py
index 78b96ff3f75753..84350d34575546 100644
--- a/tests/models/bridgetower/test_image_processing_bridgetower.py
+++ b/tests/models/bridgetower/test_image_processing_bridgetower.py
@@ -39,7 +39,7 @@ def __init__(
         do_rescale: bool = True,
         rescale_factor: Union[int, float] = 1 / 255,
         do_normalize: bool = True,
-        do_center_crop: bool = False,
+        do_center_crop: bool = False,  # Current expected shape are not center-cropped.
         image_mean: Optional[Union[float, List[float]]] = [0.48145466, 0.4578275, 0.40821073],
         image_std: Optional[Union[float, List[float]]] = [0.26862954, 0.26130258, 0.27577711],
         do_pad: bool = True,
@@ -66,9 +66,10 @@ def __init__(
 
     def prepare_image_processor_dict(self):
         return {
+            "do_normalize": self.do_normalize,
             "image_mean": self.image_mean,
             "image_std": self.image_std,
-            "do_normalize": self.do_normalize,
+            "do_pad": self.do_pad,
             "do_resize": self.do_resize,
             "do_center_crop": self.do_center_crop,
             "size": self.size,

From 004c961d6b167431d17e23140281c681143ea216 Mon Sep 17 00:00:00 2001
From: Pablo Montalvo <pablo.montalvo.leroux@gmail.com>
Date: Mon, 26 Feb 2024 09:57:03 +0000
Subject: [PATCH 24/29] fix valid processor keys

---
 .../image_processing_bridgetower.py           | 25 +------------------
 1 file changed, 1 insertion(+), 24 deletions(-)

diff --git a/src/transformers/models/bridgetower/image_processing_bridgetower.py b/src/transformers/models/bridgetower/image_processing_bridgetower.py
index 2a80978eebe02a..ef62949745fef4 100644
--- a/src/transformers/models/bridgetower/image_processing_bridgetower.py
+++ b/src/transformers/models/bridgetower/image_processing_bridgetower.py
@@ -246,6 +246,7 @@ def __init__(
             "return_tensors",
             "data_format",
             "input_data_format",
+            "pad_and_return_pixel_mask"
         ]
 
     # Copied from transformers.models.vilt.image_processing_vilt.ViltImageProcessor.resize
@@ -493,30 +494,6 @@ def preprocess(
                 Whether to pad the image to the (max_height, max_width) in the batch. If `True`, a pixel mask is also
                 created and returned. Deprecated version of do_pad.
         """
-        valid_processor_keys = {
-            "images",
-            "do_resize",
-            "size",
-            "size_divisor",
-            "resample",
-            "do_rescale",
-            "rescale_factor",
-            "do_normalize",
-            "image_mean",
-            "image_std",
-            "do_pad",
-            "do_center_crop",
-            "return_tensors",
-            "data_format",
-            "input_data_format",
-            "pad_and_return_pixel_mask",
-        }
-
-        unused_keys = set(kwargs.keys()) - valid_processor_keys
-        if unused_keys:
-            unused_key_str = ", ".join(unused_keys)
-            logger.info(f"Unused or unrecognized configuration parameters: {unused_key_str}.")
-
         do_resize = do_resize if do_resize is not None else self.do_resize
         size_divisor = size_divisor if size_divisor is not None else self.size_divisor
         resample = resample if resample is not None else self.resample

From c2e49f542ca5000c4506d2bfb5ff4c29ddd59afd Mon Sep 17 00:00:00 2001
From: Pablo Montalvo <pablo.montalvo.leroux@gmail.com>
Date: Mon, 26 Feb 2024 10:00:18 +0000
Subject: [PATCH 25/29] remove unused arg list

---
 .../image_processing_chinese_clip.py          | 22 -------------------
 tests/test_image_processing_common.py         |  1 +
 2 files changed, 1 insertion(+), 22 deletions(-)

diff --git a/src/transformers/models/chinese_clip/image_processing_chinese_clip.py b/src/transformers/models/chinese_clip/image_processing_chinese_clip.py
index 98495e04c6558e..d92caa79923e92 100644
--- a/src/transformers/models/chinese_clip/image_processing_chinese_clip.py
+++ b/src/transformers/models/chinese_clip/image_processing_chinese_clip.py
@@ -249,28 +249,6 @@ def preprocess(
                 - `"channels_last"` or `ChannelDimension.LAST`: image in (height, width, num_channels) format.
                 - `"none"` or `ChannelDimension.NONE`: image in (height, width) format.
         """
-        valid_processor_keys = {
-            "images",
-            "do_resize",
-            "size",
-            "resample",
-            "do_center_crop",
-            "crop_size",
-            "do_rescale",
-            "rescale_factor",
-            "do_normalize",
-            "image_mean",
-            "image_std",
-            "do_convert_rgb",
-            "return_tensors",
-            "data_format",
-            "input_data_format",
-        }
-
-        unused_keys = set(kwargs.keys()) - valid_processor_keys
-        if unused_keys:
-            unused_key_str = ", ".join(unused_keys)
-            logger.info(f"Unused or unrecognized configuration parameters: {unused_key_str}.")
 
         do_resize = do_resize if do_resize is not None else self.do_resize
         size = size if size is not None else self.size
diff --git a/tests/test_image_processing_common.py b/tests/test_image_processing_common.py
index 90c1a4e7e12708..074da8a2bbab38 100644
--- a/tests/test_image_processing_common.py
+++ b/tests/test_image_processing_common.py
@@ -291,6 +291,7 @@ def test_call_numpy_4_channels(self):
         )
 
     def test_image_processor_preprocess_arguments(self):
+        # Test that an instantiated image processor is called with the correct arg spec
         image_processor = self.image_processing_class(**self.image_processor_dict)
         if hasattr(image_processor, "_valid_processor_keys") and hasattr(image_processor, "preprocess"):
             preprocess_parameter_names = inspect.getfullargspec(image_processor.preprocess).args

From 79958b5802b98caf6c93561c428d479eac1b4cf2 Mon Sep 17 00:00:00 2001
From: Pablo Montalvo <pablo.montalvo.leroux@gmail.com>
Date: Mon, 26 Feb 2024 10:25:44 +0000
Subject: [PATCH 26/29] quality

---
 src/transformers/models/bit/image_processing_bit.py             | 1 +
 .../models/bridgetower/image_processing_bridgetower.py          | 2 +-
 2 files changed, 2 insertions(+), 1 deletion(-)

diff --git a/src/transformers/models/bit/image_processing_bit.py b/src/transformers/models/bit/image_processing_bit.py
index d91c050662ba0c..c9d5c7a7594a49 100644
--- a/src/transformers/models/bit/image_processing_bit.py
+++ b/src/transformers/models/bit/image_processing_bit.py
@@ -207,6 +207,7 @@ def preprocess(
         return_tensors: Optional[Union[str, TensorType]] = None,
         data_format: Optional[ChannelDimension] = ChannelDimension.FIRST,
         input_data_format: Optional[Union[str, ChannelDimension]] = None,
+        **kwargs,
     ) -> PIL.Image.Image:
         """
         Preprocess an image or batch of images.
diff --git a/src/transformers/models/bridgetower/image_processing_bridgetower.py b/src/transformers/models/bridgetower/image_processing_bridgetower.py
index ef62949745fef4..bf3c37b9b7cd2c 100644
--- a/src/transformers/models/bridgetower/image_processing_bridgetower.py
+++ b/src/transformers/models/bridgetower/image_processing_bridgetower.py
@@ -246,7 +246,7 @@ def __init__(
             "return_tensors",
             "data_format",
             "input_data_format",
-            "pad_and_return_pixel_mask"
+            "pad_and_return_pixel_mask",
         ]
 
     # Copied from transformers.models.vilt.image_processing_vilt.ViltImageProcessor.resize

From 3238dd35cd82d686a6eeffa25d025f4de62ad7e2 Mon Sep 17 00:00:00 2001
From: Pablo Montalvo <pablo.montalvo.leroux@gmail.com>
Date: Thu, 18 Apr 2024 16:35:03 +0000
Subject: [PATCH 27/29] skeleton draft - uniform processor call

---
 .../models/donut/processing_donut.py          | 139 +++++++-----------
 1 file changed, 57 insertions(+), 82 deletions(-)

diff --git a/src/transformers/models/donut/processing_donut.py b/src/transformers/models/donut/processing_donut.py
index ae2588b85e2040..38e9b9426f8fd7 100644
--- a/src/transformers/models/donut/processing_donut.py
+++ b/src/transformers/models/donut/processing_donut.py
@@ -64,46 +64,61 @@ def __init__(self, image_processor=None, tokenizer=None, feature_extractor=None)
 
         super().__init__(image_processor, tokenizer)
         self.current_processor = self.image_processor
+
+        self.processing_kwargs = {
+            'common_kwargs':{
+                'return_tensors': 'pt'
+                },
+            'text_kwargs': {
+                'text_pair': None,
+                'text_target': None,
+                'text_pair_target': None,
+                'add_special_tokens': True,
+                'padding': 'max_length',
+                'truncation': True,
+                'max_length': 512,
+                'stride': 0,
+                'is_split_into_words': False,
+                'pad_to_multiple_of': None,
+                'return_token_type_ids': True,
+                'return_attention_mask': True,
+                'return_overflowing_tokens': False,
+                'return_special_tokens_mask': False,
+                'return_offsets_mapping': False,
+                'return_length': False,
+                'verbose': True
+            },
+            'images_kwargs': {
+                'do_crop_margin': False,
+                'do_resize': True,
+                'size': {'height': 256, 'width': 256},
+                'resample': 'bilinear',
+                'do_thumbnail': False,
+                'do_align_long_axis': False,
+                'do_pad': False,
+                'do_rescale': False,
+                'rescale_factor': 1.0,
+                'do_normalize': True,
+                'image_mean': [0.485, 0.456, 0.406],
+                'image_std': [0.229, 0.224, 0.225],
+                'data_format': 'channels_first',
+                'input_data_format': None
+            },
+            'audio_kwargs': {
+            },
+            'videos_kwargs': {
+            },
+        }
+
         self._in_target_context_manager = False
 
     def __call__(
         self,
         text=None,
         images=None,
-        do_crop_margin: bool = None,
-        do_resize: bool = None,
-        size: Dict[str, int] = None,
-        resample: "PILImageResampling" = None,  # noqa: F821
-        do_thumbnail: bool = None,
-        do_align_long_axis: bool = None,
-        do_pad: bool = None,
-        do_rescale: bool = None,
-        rescale_factor: Union[int, float] = None,
-        do_normalize: bool = None,
-        image_mean: Optional[Union[float, List[float]]] = None,
-        image_std: Optional[Union[float, List[float]]] = None,
-        data_format: Optional["ChannelDimension"] = "channels_first",  # noqa: F821
-        input_data_format: Optional[Union[str, "ChannelDimension"]] = None,  # noqa: F821
-        text_pair: Optional[Union[TextInput, PreTokenizedInput, List[TextInput], List[PreTokenizedInput]]] = None,
-        text_target: Union[TextInput, PreTokenizedInput, List[TextInput], List[PreTokenizedInput]] = None,
-        text_pair_target: Optional[
-            Union[TextInput, PreTokenizedInput, List[TextInput], List[PreTokenizedInput]]
-        ] = None,
-        add_special_tokens: bool = True,
-        padding: Union[bool, str, PaddingStrategy] = False,
-        truncation: Union[bool, str, TruncationStrategy] = None,
-        max_length: Optional[int] = None,
-        stride: int = 0,
-        is_split_into_words: bool = False,
-        pad_to_multiple_of: Optional[int] = None,
-        return_tensors: Optional[Union[str, TensorType]] = None,
-        return_token_type_ids: Optional[bool] = None,
-        return_attention_mask: Optional[bool] = None,
-        return_overflowing_tokens: bool = False,
-        return_special_tokens_mask: bool = False,
-        return_offsets_mapping: bool = False,
-        return_length: bool = False,
-        verbose: bool = True,
+        audio=None,
+        videos=None, # end of supported modalities in call
+        **kwargs
     ):
         """
         When used in normal mode, this method forwards all its arguments to AutoImageProcessor's
@@ -113,68 +128,28 @@ def __call__(
         """
         # For backward compatibility
         if self._in_target_context_manager:
+            image_kwargs = {**self.processing_kwargs.get('images_kwargs', {}), **self.processing_kwargs.get('common_kwargs'), **kwargs}
             return self.current_processor(
                 images,
-                do_crop_margin=do_crop_margin,
-                do_resize=do_resize,
-                size=size,
-                resample=resample,
-                do_thumbnail=do_thumbnail,
-                do_align_long_axis=do_align_long_axis,
-                do_pad=do_pad,
-                do_rescale=do_rescale,
-                rescale_factor=rescale_factor,
-                do_normalize=do_normalize,
-                image_mean=image_mean,
-                image_std=image_std,
-                return_tensors=return_tensors,
-                data_format=data_format,
-                input_data_format=input_data_format,
+                **image_kwargs,
             )
 
         if images is None and text is None:
             raise ValueError("You need to specify either an `images` or `text` input to process.")
 
         if images is not None:
+            image_kwargs = {**self.processing_kwargs.get('images_kwargs', {}), **self.processing_kwargs.get('common_kwargs'), **kwargs}
             inputs = self.image_processor(
                 images,
-                do_crop_margin=do_crop_margin,
-                do_resize=do_resize,
-                size=size,
-                resample=resample,
-                do_thumbnail=do_thumbnail,
-                do_align_long_axis=do_align_long_axis,
-                do_pad=do_pad,
-                do_rescale=do_rescale,
-                rescale_factor=rescale_factor,
-                do_normalize=do_normalize,
-                image_mean=image_mean,
-                image_std=image_std,
-                return_tensors=return_tensors,
-                data_format=data_format,
-                input_data_format=input_data_format,
+                **image_kwargs,
+
             )
         if text is not None:
+            text_kwargs = {**self.processing_kwargs.get('text_kwargs', {}), **self.processing_kwargs.get('common_kwargs'), **kwargs}
+
             encodings = self.tokenizer(
                 text,
-                text_pair=text_pair,
-                text_target=text_target,
-                text_pair_target=text_pair_target,
-                add_special_tokens=add_special_tokens,
-                padding=padding,
-                truncation=truncation,
-                max_length=max_length,
-                stride=stride,
-                is_split_into_words=is_split_into_words,
-                pad_to_multiple_of=pad_to_multiple_of,
-                return_tensors=return_tensors,
-                return_token_type_ids=return_token_type_ids,
-                return_attention_mask=return_attention_mask,
-                return_overflowing_tokens=return_overflowing_tokens,
-                return_special_tokens_mask=return_special_tokens_mask,
-                return_offsets_mapping=return_offsets_mapping,
-                return_length=return_length,
-                verbose=verbose,
+                **text_kwargs,
             )
 
         if text is None:

From 3afde2299dde1c47acc7195780720fdaf9eb9bae Mon Sep 17 00:00:00 2001
From: Pablo Montalvo <pablo.montalvo.leroux@gmail.com>
Date: Thu, 18 Apr 2024 16:48:01 +0000
Subject: [PATCH 28/29] fix quality

---
 .../models/donut/processing_donut.py          | 103 +++++++++---------
 1 file changed, 53 insertions(+), 50 deletions(-)

diff --git a/src/transformers/models/donut/processing_donut.py b/src/transformers/models/donut/processing_donut.py
index 38e9b9426f8fd7..809e3ddf57e68e 100644
--- a/src/transformers/models/donut/processing_donut.py
+++ b/src/transformers/models/donut/processing_donut.py
@@ -18,12 +18,8 @@
 import re
 import warnings
 from contextlib import contextmanager
-from typing import Dict, List, Optional, Union
-
-from transformers.tokenization_utils_base import PreTokenizedInput, TextInput, TruncationStrategy
 
 from ...processing_utils import ProcessorMixin
-from ...utils import PaddingStrategy, TensorType
 
 
 class DonutProcessor(ProcessorMixin):
@@ -66,48 +62,44 @@ def __init__(self, image_processor=None, tokenizer=None, feature_extractor=None)
         self.current_processor = self.image_processor
 
         self.processing_kwargs = {
-            'common_kwargs':{
-                'return_tensors': 'pt'
-                },
-            'text_kwargs': {
-                'text_pair': None,
-                'text_target': None,
-                'text_pair_target': None,
-                'add_special_tokens': True,
-                'padding': 'max_length',
-                'truncation': True,
-                'max_length': 512,
-                'stride': 0,
-                'is_split_into_words': False,
-                'pad_to_multiple_of': None,
-                'return_token_type_ids': True,
-                'return_attention_mask': True,
-                'return_overflowing_tokens': False,
-                'return_special_tokens_mask': False,
-                'return_offsets_mapping': False,
-                'return_length': False,
-                'verbose': True
-            },
-            'images_kwargs': {
-                'do_crop_margin': False,
-                'do_resize': True,
-                'size': {'height': 256, 'width': 256},
-                'resample': 'bilinear',
-                'do_thumbnail': False,
-                'do_align_long_axis': False,
-                'do_pad': False,
-                'do_rescale': False,
-                'rescale_factor': 1.0,
-                'do_normalize': True,
-                'image_mean': [0.485, 0.456, 0.406],
-                'image_std': [0.229, 0.224, 0.225],
-                'data_format': 'channels_first',
-                'input_data_format': None
+            "common_kwargs": {"return_tensors": "pt"},
+            "text_kwargs": {
+                "text_pair": None,
+                "text_target": None,
+                "text_pair_target": None,
+                "add_special_tokens": True,
+                "padding": "max_length",
+                "truncation": True,
+                "max_length": 512,
+                "stride": 0,
+                "is_split_into_words": False,
+                "pad_to_multiple_of": None,
+                "return_token_type_ids": True,
+                "return_attention_mask": True,
+                "return_overflowing_tokens": False,
+                "return_special_tokens_mask": False,
+                "return_offsets_mapping": False,
+                "return_length": False,
+                "verbose": True,
             },
-            'audio_kwargs': {
-            },
-            'videos_kwargs': {
+            "images_kwargs": {
+                "do_crop_margin": False,
+                "do_resize": True,
+                "size": {"height": 256, "width": 256},
+                "resample": "bilinear",
+                "do_thumbnail": False,
+                "do_align_long_axis": False,
+                "do_pad": False,
+                "do_rescale": False,
+                "rescale_factor": 1.0,
+                "do_normalize": True,
+                "image_mean": [0.485, 0.456, 0.406],
+                "image_std": [0.229, 0.224, 0.225],
+                "data_format": "channels_first",
+                "input_data_format": None,
             },
+            "audio_kwargs": {},
+            "videos_kwargs": {},
         }
 
         self._in_target_context_manager = False
@@ -117,8 +109,8 @@ def __call__(
         text=None,
         images=None,
         audio=None,
-        videos=None, # end of supported modalities in call
-        **kwargs
+        videos=None,  # end of supported modalities in call
+        **kwargs,
     ):
         """
         When used in normal mode, this method forwards all its arguments to AutoImageProcessor's
@@ -128,7 +120,11 @@ def __call__(
         """
         # For backward compatibility
         if self._in_target_context_manager:
-            image_kwargs = {**self.processing_kwargs.get('images_kwargs', {}), **self.processing_kwargs.get('common_kwargs'), **kwargs}
+            image_kwargs = {
+                **self.processing_kwargs.get("images_kwargs", {}),
+                **self.processing_kwargs.get("common_kwargs"),
+                **kwargs,
+            }
             return self.current_processor(
                 images,
                 **image_kwargs,
@@ -138,14 +134,21 @@ def __call__(
             raise ValueError("You need to specify either an `images` or `text` input to process.")
 
         if images is not None:
-            image_kwargs = {**self.processing_kwargs.get('images_kwargs', {}), **self.processing_kwargs.get('common_kwargs'), **kwargs}
+            image_kwargs = {
+                **self.processing_kwargs.get("images_kwargs", {}),
+                **self.processing_kwargs.get("common_kwargs"),
+                **kwargs,
+            }
             inputs = self.image_processor(
                 images,
                 **image_kwargs,
-
             )
         if text is not None:
-            text_kwargs = {**self.processing_kwargs.get('text_kwargs', {}), **self.processing_kwargs.get('common_kwargs'), **kwargs}
+            text_kwargs = {
+                **self.processing_kwargs.get("text_kwargs", {}),
+                **self.processing_kwargs.get("common_kwargs"),
+                **kwargs,
+            }
 
             encodings = self.tokenizer(
                 text,

From eb99e290081f43001b2fe8746e8c4cf28976a206 Mon Sep 17 00:00:00 2001
From: Pablo Montalvo <pablo.montalvo.leroux@gmail.com>
Date: Thu, 25 Apr 2024 12:30:25 +0000
Subject: [PATCH 29/29] add broken wav2vec audio processing

---
 .../models/wav2vec2/processing_wav2vec2.py    | 62 +++++++++++++++----
 1 file changed, 49 insertions(+), 13 deletions(-)

diff --git a/src/transformers/models/wav2vec2/processing_wav2vec2.py b/src/transformers/models/wav2vec2/processing_wav2vec2.py
index dc6e9d14ee66cd..47d020343fbd4d 100644
--- a/src/transformers/models/wav2vec2/processing_wav2vec2.py
+++ b/src/transformers/models/wav2vec2/processing_wav2vec2.py
@@ -45,6 +45,37 @@ def __init__(self, feature_extractor, tokenizer):
         super().__init__(feature_extractor, tokenizer)
         self.current_processor = self.feature_extractor
         self._in_target_context_manager = False
+        self.processing_kwargs = {
+            "common_kwargs": {"return_tensors": "pt"},
+            "text_kwargs": {
+                "text_pair": None,
+                "text_target": None,
+                "text_pair_target": None,
+                "add_special_tokens": True,
+                "padding": "max_length",
+                "truncation": True,
+                "max_length": 512,
+                "stride": 0,
+                "is_split_into_words": False,
+                "pad_to_multiple_of": None,
+                "return_token_type_ids": True,
+                "return_attention_mask": True,
+                "return_overflowing_tokens": False,
+                "return_special_tokens_mask": False,
+                "return_offsets_mapping": False,
+                "return_length": False,
+                "verbose": True,
+            },
+            "images_kwargs": {
+            },
+            "audio_kwargs": {
+                "sampling_rate": None,
+                "raw_speech": None,
+            },
+            "videos_kwargs": {
+            },
+        }
+
 
     @classmethod
     def from_pretrained(cls, pretrained_model_name_or_path, **kwargs):
@@ -65,7 +96,15 @@ def from_pretrained(cls, pretrained_model_name_or_path, **kwargs):
 
             return cls(feature_extractor=feature_extractor, tokenizer=tokenizer)
 
-    def __call__(self, *args, **kwargs):
+    def __call__(
+        self,
+        audio=None,
+        text=None,
+        images=None,
+        videos=None,  # end of supported modalities in call
+        *deprecated_args,
+        **deprecated_kwargs,
+    ):
         """
         When used in normal mode, this method forwards all its arguments to Wav2Vec2FeatureExtractor's
         [`~Wav2Vec2FeatureExtractor.__call__`] and returns its output. If used in the context
@@ -74,26 +113,23 @@ def __call__(self, *args, **kwargs):
         """
         # For backward compatibility
         if self._in_target_context_manager:
-            return self.current_processor(*args, **kwargs)
-
-        if "raw_speech" in kwargs:
+            return self.current_processor(audio, *deprecated_args, **self.processing_kwargs['audio_kwargs'], **deprecated_kwargs)
+        print(deprecated_args, deprecated_kwargs)
+        if "raw_speech" in deprecated_kwargs:
+            breakpoint()
             warnings.warn("Using `raw_speech` as a keyword argument is deprecated. Use `audio` instead.")
-            audio = kwargs.pop("raw_speech")
+            audio = deprecated_kwargs.pop("raw_speech")
         else:
-            audio = kwargs.pop("audio", None)
-        sampling_rate = kwargs.pop("sampling_rate", None)
-        text = kwargs.pop("text", None)
-        if len(args) > 0:
-            audio = args[0]
-            args = args[1:]
+            audio = deprecated_kwargs.pop("audio", None)
+        sampling_rate = deprecated_kwargs.pop("sampling_rate", None)
 
         if audio is None and text is None:
             raise ValueError("You need to specify either an `audio` or `text` input to process.")
 
         if audio is not None:
-            inputs = self.feature_extractor(audio, *args, sampling_rate=sampling_rate, **kwargs)
+            inputs = self.feature_extractor(audio, *deprecated_args, sampling_rate=sampling_rate, **deprecated_kwargs)
         if text is not None:
-            encodings = self.tokenizer(text, **kwargs)
+            encodings = self.tokenizer(text, *deprecated_args, **deprecated_kwargs)
 
         if text is None:
             return inputs