suven-p / efficientnerf Goto Github PK

$ conda create -n EfficientNeRF python=3.8
$ conda activate EfficientNeRF
$ pip install torch==1.9.1+cu111 torchvision==0.10.1+cu111 torchaudio==0.9.1 -f https://download.pytorch.org/whl/torch_stable.html
$ pip install -r requirements.txt

$ DATA_DIR=/path/to/lego
$ python train.py \
   --dataset_name blender \
   --root_dir $DATA_DIR \
   --N_samples 128 \
   --N_importance 5 --img_wh 800 800 \
   --num_epochs 16 --batch_size 4096 \
   --optimizer radam --lr 2e-3 \
   --lr_scheduler poly \
   --coord_scope 3.0 \
   --warmup_step 5000\
   --sigma_init 30.0 \
   --weight_threashold 1e-5 \
   --exp_name lego_coarse128_fine5_V384

$ tensorboard --logdir=./logs

@InProceedings{Hu_2022_CVPR,
    author    = {Hu, Tao and Liu, Shu and Chen, Yilun and Shen, Tiancheng and Jia, Jiaya},
    title     = {EfficientNeRF  Efficient Neural Radiance Fields},
    booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
    month     = {June},
    year      = {2022},
    pages     = {12902-12911}
}