[Feature request] Visual Question Answering Support

**Name of the feature**
*In general, the feature you want added should be supported by HuggingFace's [transformers](https://github.com/huggingface/transformers) library:*
- *If requesting a **model**, it must be listed [here](https://huggingface.co/docs/transformers/index#supported-models).*
- *If requesting a **pipeline**, it must be listed [here](https://huggingface.co/docs/transformers/main_classes/pipelines).*
- *If requesting a **task**, it must be listed [here](https://huggingface.co/tasks).*

- Model: https://huggingface.co/dandelin/vilt-b32-finetuned-vqa
- Pipeline: Visual Question answering

**Reason for request**
*Why is it important that we add this feature? What is your intended use case? Remember, we are more likely to add support for models/pipelines/tasks that are popular (e.g., many downloads), or contain functionality that does not exist (e.g., new input type).*

**Additional context**
*Add any other context or screenshots about the feature request here.*