Elektrine lite

← Feed

@naturepoker@genomic.social

Post #3588557

2026-05-19 09:20 UTC

Main difference being that weights trained on complete genome reference keeps on looking toward other distant signals to make a call on current input, while contig level genome reference trained model constantly focuses on immediate signal neighbors more often than not.

Replies (1)

  • @naturepoker@genomic.social 2026-05-19 09:21

    By the time it gets to layer 18 (the output from which gets fed into direct signal-to-sequence translation CTC), the two models end up with a very different looking output.

    Open ##3588558