SpectraNet-SO(3): Learning Satellite Orientation from Optical Spectra By Implicitly Modeling Mutually Exclusive Probability Distributions on the Rotation Manifold (Official Repo)
- Create conda environment
conda create -n ipdf python=3.9 pip
conda activate ipdf
- Install torch and dependencies. Uses mlflow for logging artifacts/metrics and pyrallis for easy config management
pip install -U pip
# cuda version >= 11.0
pip install torch torchvision --extra-index-url https://download.pytorch.org/whl/cu113
# cuda version < 11.0
pip install torch torchvision
pip install mlflow pyrallis pandas tqdm pillow
- Install implicit-pdf repo
git clone https://github.com/phelps-matthew/implicit-pdf.git
cd implicit-pdf
pip install -e .
- Train implicit pdf model on GPU 0
python train.py --gpus [0]
- View train configuration options
python train.py --help
- Train from yaml configuration, with CLI override
python train.py --config_path train_cfg.yaml --lr 0.001 --gpus [4]
- Start mlflow ui to visualize results
# navgiate to implicit_pdf root directory containing `mlruns`
mlflow ui
# to set host and port
mlflow ui --host 0.0.0.0 --port 8080
- Serialize dataclass train config to yaml, outputting
train_cfg.yaml
python cfg.py