The official documentation for Caffe is maintained and hosted by BVLC here. This documentation is part of the Caffe source code repository and is meant to be an accurate reflection of the code.

This wiki hosts the unofficial part of the Caffe documentation, such as user-contributed examples and tips, links to blog posts and tutorials, a list of referencing publications, etc.

It is publicly-editable! Anyone logged into Github can change the content, so use this power only for good. BVLC maintainers will periodically review the contents of the wiki. If you spot something that's wrong, feel free to make a note or adjust it yourself. Thanks!

The Model Zoo is a public clearinghouse for Caffe models to promote progress and reproduce research. The official model zoo documentation lists the BVLC reference models and explains the model format.

Development hints help explain day-to-day Caffe hacking. The official developing & contributing guide covers the license, copyright, and development workflow of the project.

BVLC and community guides to deep learning and the Caffe framework help with learning the ideas behind deep learning and the practical how-to knowledge of Caffe.

• DIY Deep Learning for Vision: a Hands-On Tutorial with Caffe is the official tutorial by Evan Shelhamer, Jeff Donahue, Yangqing Jia, and Ross Girshick.
• Intel / TAU Summer Workshop on Deep Learning and Caffe is a short course on deep learning and Caffe by Boris Ginzburg. It is detailed and complementary to the official tutorial, but not up-to-date with the current Caffe version. Refer to this for a close look at the forward and backward passes of common layer types.

As a framework Caffe can be extended to many ends. Check out related projects that make use of Caffe including research pipelines, interfaces, and visualization.

Help translate the Caffe documentation into your language!

• Farsi Translation thanks to Seyyed Hossein Hasan Pour
• A Japanese translation of the Caffe documentation is in the works.