steam data large validation set

[Li-Muyang]

Hi, all, when I loaded Steam dataset and started training, for some reason the validation batch got suspiciously large

I set the split ratio to be [0.8,0.1,0,1], so why this happen still confuses me, any help will be much appreciated.