Papers With Code 2 | ML Benchmarks, SotA Results & Code

The ComTQA dataset is a visual table question answering benchmark. It includes images collected from FinTabNet and PubTables-1M, comprising a total of 9,070 QA pairs with 1,591 images. The dataset is designed to address tasks related to table question answering and is available in English. It falls under the size category of 1K<n<10K and is licensed under cc-by-nc-4.0¹.

This dataset is particularly useful for developing and testing algorithms that can interpret and answer questions based on tabular data. It's a valuable resource for researchers and practitioners in the field of machine learning and natural language processing, especially those focusing on the intersection of visual data interpretation and question answering systems¹.

(1) ByteDance/ComTQA · Datasets at Hugging Face. https://huggingface.co/datasets/ByteDance/ComTQA. (2) ComQA Dataset | Papers With Code. https://paperswithcode.com/dataset/comqa. (3) COMETA: the corpus of online medical entities - GitHub. https://github.com/cambridgeltl/cometa. (4) COMETA Dataset | Papers With Code. https://paperswithcode.com/dataset/cometa.