search

Julie-tang00 / Point-BERT

[CVPR 2022] Pre-Training 3D Point Cloud Transformers with Masked Point Modeling
MIT License
542 stars 65 forks source link

log issue #25

Closed antoalli closed 2 years ago

antoalli commented 2 years ago

Hi! Thanks for the interesting work and for publishing the code. I've encountered a little bug/typo when training the DVAE with DistributedDataParallel, it affects only the loss logging: https://github.com/lulutang0608/Point-BERT/blob/c1df1e85a53c4fecbba518e3adc3cd3f165aaab1/tools/runner.py#L146 Should be: loss_2 = dist_utils.reduce_tensor(loss_2, args) Antonio

yuxumin commented 2 years ago

Hi,

Thanks for pointing it out. Sorry for the typo. We will correct it. :)