Welcome to the documentation of Convis

Convis lets you build vision models using PyTorch.

Want to jump right in? Have a look at these quickstart guides:

Get it on github!

When using convis for your scientific publications, please cite:

  • Huth J, Masquelier T and Arleo A (2018) Convis: A Toolbox to Fit and Simulate Filter-Based Models of Early Visual Processing. Front. Neuroinform. 12:9. doi: 10.3389/fninf.2018.00009 link

Installation

Installing convis and PyTorch itself is not complicated. Go to http://pytorch.org and follow the installation instructions, either for pip or conda.

Requirements for the base installation are: Python 2.7 or Python 3.6 and PyTorch. All other requirements are installed when running:

pip install convis

or for the most recent version:

pip install git+https://github.com/jahuth/convis.git

I recommend installing opencv, and jupyter notebook, if you do not already have it installed:

pip install convis notebook
# eg. for ubuntu:
sudo apt-get install python-opencv

Found a bug or want to contribute?

Bug reports and feature requests are always welcome! The best place to put them is the github issue tracker. If you have questions about usage of functions and classes and can not find an answer in the documentation and docstrings, this is considered a bug and I appreciate it if you open an issue for that!

If you want, you can flag your issue already with one of the labels:

If you have fixed a bug or added a feature and you would like to see the change included in the main repository, the preferred method is for you to commit the change to a fork on your own github account and submit a pull request.

General discussion is encouraged on the two mailing lists:

Contents:

Get Started

Features

API