the details of the dataset split

[slll12]

Dear author,

Hi,
thanks to the author, your work greatly facilitated my research!
I also have some questions about the details of the dataset：
1.For pictures with two masks, whether they need to be merged.
2.Can you provide the random numbers that is split into the dataset in the results of the paper.
The reason is that I found that different data divisions may cause a large difference in the results because of some hard samples.
Looking forward to your reply!

[FengheTan9]

Thanks for using our code.

1. Merged.
2. We also found this problem, we refer to the division of UNeXt. You can use data split random seed of 1337 or 41 on scikit-learn 1.0.2.