search

Vegeta2020 / SE-SSD

SE-SSD: Self-Ensembling Single-Stage Object Detector From Point Cloud, CVPR 2021.
Apache License 2.0
811 stars 128 forks source link

some confusion about data augmentation QAQ #70

Closed OuyangJunyuan closed 2 years ago

OuyangJunyuan commented 2 years ago

different catagories object, like 'Car' and 'Pedestrain' , part points cloud will be swapped ?

Vegeta2020 commented 2 years ago

The code was wrritten for a single class, as we trained SE-SSD on each class separately for better performance. If you want to train it on all classes, you may use the gt_names to separate different classes and perform swap on each class.

OuyangJunyuan commented 2 years ago

many thanks for your rapid reply!