If your CUDA toolkit is older than 11, then you will need to install CUB as follows:
conda install -c bottler nvidiacub
Since CUDA 11, CUB is shipped with the toolkit. To install the main library, simply run
pip install .
in the root directory.
See opt/opt.py
./launch.sh <exp_name> <GPU_id> <data_dir>
NOTE: can no longer use sh
See opt/render_imgs.py
(in opt/)
python render_imgs.py <CHECKPOINT.npz> <data_dir>
Including evaluation, ablations, and hypertuning (based on the task_manager one from PlenOctrees)
See opt/autotune.py. Configs in opt/tasks/*.json
Automatic eval:
python autotune.py -g '<space delimited GPU ids>' tasks/eval.json. Configs in opt/tasks/*.json
First make sure you have colmap installed. Then
(in opt/)
bash scripts/proc_colmap.sh <img_dir>
Where <img_dir> should be a directory directly containing png/jpg images from a
normal perspective camera.
For custom datasets we adopt a data format similar to that in NSVF
https://github.com/facebookresearch/NSVF
You should be able to use this dataset directly afterwards. The format will be auto-detected.
To view the data use:
python scripts/view_data.py <img_dir>
This should launch a server at localhost:8889
You may notice that this CUDA extension takes forever to install.
A suggestion is using ninja. On Ubuntu,
install it with sudo apt install ninja-build.
Then set the environment variable MAX_JOBS to the number of CPUS to use in parallel (e.g. 12) in your shell startup script.
This will enable parallel compilation and significantly improve iteration speed.