facebookresearch / ContrastiveSceneContexts

Code for CVPR 2021 oral paper "Exploring Data-Efficient 3D Scene Understanding with Contrastive Scene Contexts"
MIT License
218 stars 29 forks source link

Training data split #26

Closed ahme0307 closed 2 years ago

ahme0307 commented 2 years ago

Thank you for the interesting work. Would it be possible for you to share the training data split filenames or script for generating for ScanNet and Shapenet?

image image Thanks in advance,

Sekunde commented 2 years ago

Hi, thanks for your interest in our work!

Please refer to scannet data-efficient benchmark for the splits: http://kaldir.vc.in.tum.de/scannet_benchmark/data_efficient/documentation

Unfortunately, I don't have the splits for ShapeNet.

ahme0307 commented 2 years ago

Hi Ji,

Thanks for the response. I will close the issue.