search

genforce / insgen

[NeurIPS 2021] Data-Efficient Instance Generation from Instance Discrimination
https://genforce.github.io/insgen/
Other
101 stars 4 forks source link

Training set selection #4

Open ryanwebster90 opened 2 years ago

ryanwebster90 commented 2 years ago

Hello, Thanks for this awesome work, it really pushes the envelope for generation with few data. Would you mind providing the random indices used for training each subset for the ffhq models for example?

limbo0000 commented 2 years ago

Hi @ryanwebster90 ,

The dataset actually samples the subset with a given random seed. We follow the default seed 0 in all experiments.