Prerequisites:
- Python3.10
- CUDA 11.7
First, download Synthetic NeRF dataset and unzip it under the root dir.
Then, setup pip venv environment.
pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu117
python -m pip install -U pip
pip install -i https://pypi.taichi.graphics/simple/ taichi-nightly
pip install -r requirements.txt
run train.py with parameters:
python train.py --root_dir ./Synthetic_NeRF/Lego --exp_name Lego --perf --num_epochs 20 --batch_size 512 --lr 1e-2 --no_save_test --gui