[SigLIP] Don't pad by default (#28578)

First draft

docs/source/en/model_doc/siglip.md

@@ -28,7 +28,7 @@ The abstract from the paper is the following:
 
 - Usage of SigLIP is similar to [CLIP](clip). The main difference is the training loss, which does not require a global view of all the pairwise similarities of images and texts within a batch. One needs to apply the sigmoid activation function to the logits, rather than the softmax.
 - Training is not yet supported. If you want to fine-tune SigLIP or train from scratch, refer to the loss function from [OpenCLIP](https://github.com/mlfoundations/open_clip/blob/73ad04ae7fb93ede1c02dc9040a828634cb1edf1/src/open_clip/loss.py#L307), which leverages various `torch.distributed` utilities.
-- When using the standalone [`SiglipTokenizer`], make sure to pass `padding="max_length"` as that's how the model was trained. The multimodal [`SiglipProcessor`] takes care of this behind the scenes.
+- When using the standalone [`SiglipTokenizer`] or [`SiglipProcessor`], make sure to pass `padding="max_length"` as that's how the model was trained.
 
 <img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/transformers/model_doc/siglip_table.jpeg"
 alt="drawing" width="600"/>
@@ -82,7 +82,8 @@ If you want to do the pre- and postprocessing yourself, here's how to do that:
 >>> image = Image.open(requests.get(url, stream=True).raw)
 
 >>> texts = ["a photo of 2 cats", "a photo of 2 dogs"]
->>> inputs = processor(text=texts, images=image, return_tensors="pt")
+>>> # important: we pass `padding=max_length` since the model was trained with this
+>>> inputs = processor(text=texts, images=image, padding="max_length", return_tensors="pt")
 
 >>> with torch.no_grad():
 ...     outputs = model(**inputs)

src/transformers/models/siglip/modeling_siglip.py

@@ -1123,7 +1123,8 @@ def forward(
         >>> image = Image.open(requests.get(url, stream=True).raw)
 
         >>> texts = ["a photo of 2 cats", "a photo of 2 dogs"]
-        >>> inputs = processor(text=texts, images=image, return_tensors="pt")
+        >>> # important: we pass `padding=max_length` since the model was trained with this
+        >>> inputs = processor(text=texts, images=image, padding="max_length", return_tensors="pt")
 
         >>> with torch.no_grad():
         ...     outputs = model(**inputs)

src/transformers/models/siglip/processing_siglip.py

@@ -50,9 +50,9 @@ def __call__(
         self,
         text: Union[TextInput, PreTokenizedInput, List[TextInput], List[PreTokenizedInput]] = None,
         images: ImageInput = None,
-        padding: Union[bool, str, PaddingStrategy] = "max_length",
+        padding: Union[bool, str, PaddingStrategy] = False,
         truncation: Union[bool, str, TruncationStrategy] = None,
-        max_length=None,
+        max_length: int = None,
         return_tensors: Optional[Union[str, TensorType]] = TensorType.PYTORCH,
     ) -> BatchFeature:
         """
@@ -71,7 +71,7 @@ def __call__(
                 The image or batch of images to be prepared. Each image can be a PIL image, NumPy array or PyTorch
                 tensor. In case of a NumPy array/PyTorch tensor, each image should be of shape (C, H, W), where C is a
                 number of channels, H and W are image height and width.
-            padding (`bool`, `str` or [`~utils.PaddingStrategy`], *optional*, defaults to `max_length`):
+            padding (`bool`, `str` or [`~utils.PaddingStrategy`], *optional*, defaults to `False`):
                 Select a strategy to pad the returned sequences (according to the model's padding side and padding
                 index) among:
                 - `True` or `'longest'`: Pad to the longest sequence in the batch (or no padding if only a single