How to get the same validation/testing set as reported in the paper? And will the evaluation tools be publicly available?

[chixma]

With the same validation/testing set and evaluation tools, we can make a fair comparison between other methods and the baselines. Thank you!

[liminghao1630]

We re-train both of the Table Detection and Table Structure Recognition models using Detectron2 and OpenNMT tools, update the results. We provide the validation and testing set for the new result in the MS COCO format.

[chixma]

@liminghao1630 Thanks for you reply! BTW, will the evaluation tools be publicly available?

[liminghao1630]

@chixma
https://github.com/liminghao1630/dataset-tools
This is a version I used, for reference only.