I'm training a model where the samples success rate is low.

I mean how do I tackle such situation - maybe only show the samples which match but then the ones that doesn't may match too.

But on the other hand if I train it with all the samples it seems like it never fully matches when it should.

Any ideas?

  • $\begingroup$ I think you need to specify your question: What do you mean by 'samples success rate', 'samples that match' and what is 'it' referring to in 'seems like it never fully matches when it should'?. $\endgroup$
    – Chillston
    Sep 18, 2022 at 13:48
  • $\begingroup$ Please clarify your specific problem or provide additional details to highlight exactly what you need. As it's currently written, it's hard to tell exactly what you're asking. $\endgroup$
    – Community Bot
    Sep 18, 2022 at 17:52
  • $\begingroup$ Are you referring to something related to class imbalance? $\endgroup$
    – Dave
    Sep 18, 2022 at 19:54

1 Answer 1


You have a situation where class imbalance is quite high. There are two ways to tackle this:

  1. Use Focal loss which focuses on underrepresented classes.
  2. Train a Siamese network and allied techniques to it.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .