Pre-processing script and train-val-test for fair comparison

[Evm7]

Hi! Thanks a lot for the nice work as the dataset looks great.
In the paper, it is mentioned that some preprocessing is done (mirroring all motions as well as augmenting the descriptions). Would it be possible to share this script so everyone can work with similar datasets? Same for the train-val-test split.

[tr3e]

Thanks for the kind suggestions! We are releasing that part as soon as possible.

[Evm7]

Hi again! Would it be able to share the visualization code for your data? How do you plot the images in characters?