Zero-Shot Learning (ZSL)

In Zero-Shot Learning a model should detect previously unseen data during training.


Notes:

  • Zero-Shot Learning is often used to evaluate pre-trained models and test it's ability to identify new categories without prior information.