question about sampling image pairs in training

magicleap / SuperGluePretrainedNetwork

SuperGlue: Learning Feature Matching with Graph Neural Networks (CVPR 2020, Oral)

Other

3.31k stars 672 forks source link

question about sampling image pairs in training #94

Open chenghegao4556 opened 3 years ago

chenghegao4556 commented 3 years ago

Hi, i have a question about sampling image pairs in training superglew, did you sample 200 image pairs per scene and use these image pairs for all epoches; or at each epoch image pairs are randomly sampled again?

sarlinpe commented 3 years ago

Hi, we randomly resampled the 200 pairs at each epoch, such that after sufficient epochs the network has seen all image pairs in equal proportion for each scene.