SETFIT - HYPER Parameter Optimization for SBERT Text Classification (SBERT 45)

Поділитися
Вставка
  • Опубліковано 6 лис 2024

КОМЕНТАРІ • 8

  • @mfarhan5917
    @mfarhan5917 Рік тому

    how can we use multiple evaluation metrics like f-score,precision and recall?

  • @nilshorn769
    @nilshorn769 Рік тому

    Your videos are brilliant! Thank you so much :)

    • @code4AI
      @code4AI  Рік тому

      Thank you for such a positive feedback.

  • @sanjayrao8406
    @sanjayrao8406 Рік тому

    Is there a similar video explaining hyperparameter optimization for binary text classification or multiclass text classification? I tried working through multilabel code for sst2 dataset but couldn't get through.

    • @damwyn
      @damwyn Рік тому

      Same here. I didn't see which "multi_target_strategy" I should use for multi-class text classification ? :/

  • @pravallikaa3306
    @pravallikaa3306 Рік тому

    Hi , I'd like to improve the performance of my SBERT model by training it on my data, but i do not have labelled data. Any help on how to proceed?

    • @code4AI
      @code4AI  Рік тому +1

      If you have no labelled data available, we call the multiple methods you can apply UNSUPERVISED LEARNING. This includes methods like TSDAE (see my SBERT 15 video) or you go for an advanced "Domain Adaptation" methodology (I explain it and show you how to code it in my video SBERT 25, or you just search in my channel with these keywords). Have fun!

    • @pravallikaa3306
      @pravallikaa3306 Рік тому

      @@code4AI okay, thank a lot!