## Dataset
### Phase 1  
For this phase, we build up a comprehensive data suite comprising publically available datasets including DocVQA, FUNSD, IconQA, InfogrpahicsVQA, Tabfact, TextbookVQA, WebSrc, Wildreceipt, WTQ. All these datasets are aligned with the goal of the challenge of image document understanding in very specific domains like table fact-checking, answering questions about receipts, question-answering about infographics, and visual question-answering about documents, etc. This collection consists of a train and test set and can be downloaded from MMFM Data Collection.

### Phase 2  
For this phase, an alien test set will be released. The original intent of this dataset is to prevent people from overfitting on the publicly available datasets. This test set will consist of data that is similar to the distribution of the phase 1 dataset collection but will not consist of any publically available data sources. This data will be shared by the organizers to the teams in due time.