torchgeometry 0.1.1

differential geometric computer vision for deep learning

Documentation

The PyTorch Geometry package is a geometric computer vision library for PyTorch.

It consists of a set of routines and differentiable modules to solve generic geometry computer vision problems. At its core, the package uses PyTorch as its main backend both for efficiency and to take advantage of the reverse-mode auto-differentiation to define and compute the gradient of complex functions.

Development Setup

Assuming that you are on ubuntu 16.04, with nvidia-drivers installed.

In bash, source the path.bash.inc script. This will install a local conda environment under ./.dev_env, which includes pytorch and some dependencies (no root required).

source ./path.bash.inc
python -c "import torchgeometry; print(torchgeometry.__version__)"

To install, or update the conda environment run setup_dev_env.sh

./setup_dev_env.sh

Quick Usage

import torch
import torchgeometry as tgm

x_rad = tgm.pi * torch.rand(1, 3, 3)
x_deg = tgm.rad2deg(x_rad)

torch.allclose(x_rad, tgm.deg2rad(x_deg))  # True

Examples

Run our Jupyter notebooks examples to learn to use the library.

Installation

From source:

python setup.py install

Testing

python setup.py test

Cite

If you are using torchgeometry in your research-related documents, it is recommended that you cite the poster.

@misc{Arraiy2018,
 author    = {E. Riba, M. Fathollahi, W. Chaney, E. Rublee and G. Bradski}
 title     = {torchgeometry: when PyTorch meets geometry},
 booktitle = {PyTorch Developer Conference},
 year      = {2018},
 url       = {https://drive.google.com/file/d/1xiao1Xj9WzjJ08YY_nYwsthE-wxfyfhG/view?usp=sharing}
}

Future work

The roadmap will add more functions to allow developers to solve geometric problems.

Contributing

We appreciate all contributions. If you are planning to contribute back bug-fixes, please do so without any further discussion. If you plan to contribute new features, utility functions or extensions, please first open an issue and discuss the feature with us.