Add task_categories metadata to the dataset card

This pull request adds the `task_categories: image-text-to-text` metadata to the dataset card. This accurately reflects the dataset's use case in evaluating multimodal large language models on scientific understanding, which involves both image-to-text and text-to-image tasks. This addition will improve the discoverability of the dataset on the Hugging Face Hub.

README.md

@@ -1,4 +1,9 @@
 ---
+language:
+- en
+license: cc-by-nc-sa-4.0
+size_categories:
+- 1K<n<10K
 configs:
 - config_name: image2text_info
   data_files: image2text_info.csv
@@ -8,18 +13,16 @@ configs:
   data_files: text2image_info.csv
 - config_name: text2image_option
   data_files: text2image_option.csv
-license: cc-by-nc-sa-4.0
-language:
-- en
 tags:
 - benchmark
 - mllm
 - scientific
 - cover
 - live
-size_categories:
-- 1K<n<10K
+task_categories:
+- image-text-to-text
 ---
+
 # MAC: A Live Benchmark for Multimodal Large Language Models in Scientific Understanding
 
 [![arXiv](https://img.shields.io/badge/arXiv-2508.15802-b31b1b.svg)](https://arxiv.org/abs/2508.15802)