VITA-Group / NeRF-SOS

[ICLR2023] "NeRF-SOS: Any-View Self-supervised Object Segmentation from Complex Real-World Scenes", Zhiwen Fan, Peihao Wang, Xinyu Gong, Yifan Jiang, Dejia Xu, Zhangyang Wang
MIT License
131 stars 13 forks source link

Training #2

Closed ziyuz-vision closed 2 years ago

ziyuz-vision commented 2 years ago

Hi, I just try to train with other scenes in llff datasets. How to make a dataset like flower as I noticed that there is a masks and segments subfolder under the flower folder.

zhiwenfan commented 2 years ago

segments and masks are not required for training NeRF-SOS, but are required for semantic-NeRF. You can safely remove the requirement of masks and segments in gen_dataset.py