Confusing about inferencing the single pointcloud

Julie-tang00 / Point-BERT

[CVPR 2022] Pre-Training 3D Point Cloud Transformers with Masked Point Modeling

MIT License

542 stars 65 forks source link

Confusing about inferencing the single pointcloud #40

Open ThomasVercetti opened 2 years ago

ThomasVercetti commented 2 years ago

@yuxumin @lulutang0608 Thanks for sharing the paper and code. One point I'm confused about is whether I can use the model to input a single point cloud and get the lower-dimension representation.

yuxumin commented 2 years ago

yes, inference the point cloud with the pretrained dVAE, and represent the point cloud by a list of int with argmax(logits) https://github.com/lulutang0608/Point-BERT/blob/master/models/dvae.py#L335