Show HN: Autofit2 – End-to-end pipeline for multilingual text classification(github.com)
28 pointsby leschak6 days ago2 comments
  • nmstoker5 days ago
    How does this differ from SetFit? Is it just an alternative implementation?

    I found the HF version pretty effective and it often works well for multilingual classification. I've used it for intent matching and was pleasantly surprised that Polish, German and other translations of our intents tended to work "for free" when training with just English training data!

    https://github.com/huggingface/setfit

    • leschak4 days ago
      Yes, this is an alternative original implementation, from four years ago, when the concept of SetFit was still new and HF's project didn't exist. I guess its value nowadays lies in its simplicity. It is really simple. And practicable if you use TorchServe, because embeddings and classification model get serialized into one object.
  • 5 days ago
    undefined