# Clone the repo.
git clone https://github.com/SuLvXiangXin/zipnerf-pytorch.git
cd zipnerf-pytorch

# Make a conda environment.
conda create --name zipnerf python=3.9
conda activate zipnerf

# Install requirements.
pip install -r requirements.txt

# Install other extensions
pip install ./gridencoder

# Install nvdiffrast (optional, for textured mesh)
git clone https://github.com/NVlabs/nvdiffrast
pip install ./nvdiffrast

# Install a specific cuda version of torch_scatter 
# see more detail at https://github.com/rusty1s/pytorch_scatter
CUDA=cu117
pip install torch-scatter -f https://data.pyg.org/whl/torch-2.0.0+${CUDA}.html

mkdir data
cd data

# e.g. mipnerf360 data
wget http://storage.googleapis.com/gresearch/refraw360/360_v2.zip
unzip 360_v2.zip

# Configure your training (DDP? fp16? ...)
# see https://huggingface.co/docs/accelerate/index for details
accelerate config

# Where your data is 
DATA_DIR=data/360_v2/bicycle
EXP_NAME=360_v2/bicycle

# Experiment will be conducted under "exp/${EXP_NAME}" folder
# "--gin_configs=configs/360.gin" can be seen as a default config 
# and you can add specific config useing --gin_bindings="..." 
accelerate launch train.py \
    --gin_configs=configs/360.gin \
    --gin_bindings="Config.data_dir = '${DATA_DIR}'" \
    --gin_bindings="Config.exp_name = '${EXP_NAME}'" \
    --gin_bindings="Config.factor = 4"

# or you can also run without accelerate (without DDP)
CUDA_VISIBLE_DEVICES=0 python train.py \
    --gin_configs=configs/360.gin \
    --gin_bindings="Config.data_dir = '${DATA_DIR}'" \
    --gin_bindings="Config.exp_name = '${EXP_NAME}'" \
      --gin_bindings="Config.factor = 4" 

# alternatively you can use an example training script 
bash scripts/train_360.sh

# blender dataset
bash scripts/train_blender.sh

# metric, render image, etc can be viewed through tensorboard
tensorboard --logdir "exp/${EXP_NAME}"

accelerate launch render.py \
    --gin_configs=configs/360.gin \
    --gin_bindings="Config.data_dir = '${DATA_DIR}'" \
    --gin_bindings="Config.exp_name = '${EXP_NAME}'" \
    --gin_bindings="Config.render_path = True" \
    --gin_bindings="Config.render_path_frames = 480" \
    --gin_bindings="Config.render_video_fps = 60" \
    --gin_bindings="Config.factor = 4"  

# alternatively you can use an example rendering script 
bash scripts/render_360.sh

# using the same exp_name as in training
accelerate launch eval.py \
    --gin_configs=configs/360.gin \
    --gin_bindings="Config.data_dir = '${DATA_DIR}'" \
    --gin_bindings="Config.exp_name = '${EXP_NAME}'" \
    --gin_bindings="Config.factor = 4"


# alternatively you can use an example evaluating script 
bash scripts/eval_360.sh

# more configuration can be found in internal/configs.py
accelerate launch extract.py \
    --gin_configs=configs/360.gin \
    --gin_bindings="Config.data_dir = '${DATA_DIR}'" \
    --gin_bindings="Config.exp_name = '${EXP_NAME}'" \
    --gin_bindings="Config.factor = 4"
#    --gin_bindings="Config.mesh_radius = 1"  # (optional) smaller for more details e.g. 0.2 in bicycle scene
#    --gin_bindings="Config.isosurface_threshold = 20"  # (optional) empirical value
#    --gin_bindings="Config.mesh_voxels=134217728"  # (optional) number of voxels used to extract mesh, e.g. 134217728 equals to 512**3 . Smaller values may solve OutoFMemoryError
#    --gin_bindings="Config.vertex_color = True"  # (optional) saving mesh with vertex color instead of atlas which is much slower but with more details.
#    --gin_bindings="Config.vertex_projection = True"  # (optional) use projection for vertex color

# or extracting mesh using tsdf method
accelerate launch tsdf.py \
    --gin_configs=configs/360.gin \
    --gin_bindings="Config.data_dir = '${DATA_DIR}'" \
    --gin_bindings="Config.exp_name = '${EXP_NAME}'" \
    --gin_bindings="Config.factor = 4"

# alternatively you can use an example script 
bash scripts/extract_360.sh

DATA_DIR=my_dataset_dir
bash scripts/local_colmap_and_resize.sh ${DATA_DIR}

@misc{barron2023zipnerf,
      title={Zip-NeRF: Anti-Aliased Grid-Based Neural Radiance Fields}, 
      author={Jonathan T. Barron and Ben Mildenhall and Dor Verbin and Pratul P. Srinivasan and Peter Hedman},
      year={2023},
      eprint={2304.06706},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}

@misc{multinerf2022,
      title={{MultiNeRF}: {A} {Code} {Release} for {Mip-NeRF} 360, {Ref-NeRF}, and {RawNeRF}},
      author={Ben Mildenhall and Dor Verbin and Pratul P. Srinivasan and Peter Hedman and Ricardo Martin-Brualla and Jonathan T. Barron},
      year={2022},
      url={https://github.com/google-research/multinerf},
}

@Misc{accelerate,
  title =        {Accelerate: Training and inference at scale made simple, efficient and adaptable.},
  author =       {Sylvain Gugger, Lysandre Debut, Thomas Wolf, Philipp Schmid, Zachary Mueller, Sourab Mangrulkar},
  howpublished = {\url{https://github.com/huggingface/accelerate}},
  year =         {2022}
}

@misc{torch-ngp,
    Author = {Jiaxiang Tang},
    Year = {2022},
    Note = {https://github.com/ashawkey/torch-ngp},
    Title = {Torch-ngp: a PyTorch implementation of instant-ngp}
}