training dataset

[wawpaopao]

1. I am currently working with the HCT116 nanopore data and have successfully run the m6ANet pipeline. The AUROC and AUPR values are within the expected range, which is very encouraging. However, I noticed that after processing the HCT116 data, the final dataset contains approximately 3,200 trainable sites (including both positive and negative samples), with positive samples accounting for about 5% of the total.
   Is this number of trainable sites typical for the HCT116 dataset when processed with m6Anet?

2. I would like to ask whether the fast5 training data for RNA004 has been made public. Also, where can I obtain it?