Ced-NeRF: A Compact and Efficient Method for Dynamic Neural Radiance Fields

This project is based on Nerfacc.

Usage

For HyperNeRF dataset:

python train_real.py --data_root /home/loyot/workspace/Datasets/NeRF/HyberNeRF/  --scene vrig_3dprinter

For DyNeRF dataset:

python train_real.py --data_root /home/loyot/workspace/Datasets/NeRF/3d_vedio_datasets  --scene flame_salmon_1

TODO before release

• Add DNeRF dataset support
• Fuse encoder and MLP after training
• Support visualization on camera pose and radiance field
• Add taichi-ngp-renderer support