Add task category

Since this is a multimodal (VLM) dataset, it's appropriate to add the "image-text-to-text" task category.

Btw would be cool to make your dataset compatible with the Datasets library so that the viewer works: https://huggingface.co/docs/dataset-viewer/en/index.

We recently added native support for the PDF feature in Datasets: https://huggingface.co/docs/datasets/main/en/document_load

Let me know if you need any help.

README.md

@@ -2,6 +2,7 @@
 license: mit
 task_categories:
 - table-question-answering
+- image-text-to-text
 ---
 
 This repository contains the dataset for [MDocAgent: A Multi-Modal Multi-Agent Framework for Document Understanding](https://huggingface.co/papers/2503.13964).