search

lgragert / nn-sero-pytorch

PyTorch version of neural network HLA serology prediction
2 stars 1 forks source link

Restore original training and validation sets so that new data is entirely in the testing set #25

Closed gbiagini closed 4 years ago

gbiagini commented 4 years ago

As per NMDP request, restore the original training and validation sets so that any new alleles, even those with serologic assignments, are only present in the testing sets. This will primarily affect the recreation branch. testground remains the testing area, so it wouldn't hurt to keep the reformed datasets as a possibility for comparison, along with the original sets.

gbiagini commented 4 years ago

Completed with commit c8bcc34c358b14b1dbaf4d046e67129be84a0c8e ser_parse_v2.py serves this purpose.