Add image-to-text task category

This PR adds the `image-to-text` task category to the dataset card.

README.md

@@ -20,9 +20,10 @@ configs:
   data_files:
   - split: train
     path: data/train-*
+task_categories:
+- image-to-text
 ---
 
-
 <h1 align="center"> Text-Based Reasoning About Vector Graphics </h1>
 
 <p align="center">