Pix2Struct support

### Model description

Pix2Struct is an image encoder - text decoder model that is trained on image-text pairs for various tasks, including image captionning and visual question answering.

https://huggingface.co/google/pix2struct-textcaps-base
https://huggingface.co/google/pix2struct-docvqa-base
https://huggingface.co/google/pix2struct-screen2words-base

### Prerequisites

- [X] The model is supported in Transformers (i.e., listed [here](https://huggingface.co/docs/transformers/index#supported-models-and-frameworks))
- [X] The model can be exported to ONNX with Optimum (i.e., listed [here](https://huggingface.co/docs/optimum/main/en/exporters/onnx/overview))

### Additional information

```Python
import requests
from PIL import Image
from transformers import Pix2StructForConditionalGeneration, Pix2StructProcessor

url = "https://www.ilankelman.org/stopsigns/australia.jpg"
image = Image.open(requests.get(url, stream=True).raw)

model = Pix2StructForConditionalGeneration.from_pretrained("google/pix2struct-textcaps-base")
processor = Pix2StructProcessor.from_pretrained("google/pix2struct-textcaps-base")

# image only
inputs = processor(images=image, return_tensors="pt")

predictions = model.generate(**inputs)
print(processor.decode(predictions[0], skip_special_tokens=True))
>>> A stop sign is on a street corner.
```

This example is from the textcaps model page. Note that the processor also can take in text to ask questions about the image, like in the Docvqa version.

### Your contribution

I do not believe I have the knowledge to help make the interference logic.