microsoft / Semi-supervised-learning

A Unified Semi-Supervised Learning Codebase (NeurIPS'22)
https://usb.readthedocs.io
MIT License
1.37k stars 182 forks source link

How to decide the number of labels in experiments #196

Closed ZahraaHM closed 6 months ago

ZahraaHM commented 8 months ago

In general how you decide the number of labels for each dataset? would you please provide the intuition behind the chosen numbers. and would you please guide me on how I should choose these numbers if I want to run the code on datasets other than used in this repository, I am working on image classification, any guidance is appreciated.

Capture

github-actions[bot] commented 6 months ago

Stale issue message