Open zhanh-he opened 1 week ago
Thanks for your inquiry, here they are: https://drive.google.com/drive/folders/1vyByrctZX9hJ90pz0vtrqLmZ2XPykCIH?usp=drive_link Yes I should have put them in the repository, since we are doing a piece-level split and the result would vary greatly depending on the split level.
Hi, thanks for providing codes and I am current investigating (used it as baseline). My repeat experiments results are a bit away the paper's, and likely due to my composers split difference. Would it possible to provide your metadata.csv files?
In the conf, they are: