You can view documentation here on the website, but it is also included within the toolbox. If you have already downloaded the toolbox, you will find a set of Markdown files in the code repository, under the "documentation" sub-folder. Those Markdown files' contents will be equivalent to what you find on this portion of the site, so that if you have the repository you will always have a local copy of the documentation. Both the documentation in Markdown format and the HTML versions on this site will be updated with each release, and previous releases' documentations will be preserved below.
(Also, if you prefer viewing HTML versions of the documentation to Markdown but don't have Internet access at a particular moment, note that all the pages on this website are also part of the code repository. Just look in the "website" folder of the repository and the "web_files" subdirectory within that, and open any of the HTML files inside there with your favorite browser.)
As noted on the main page, if you find the documentation lacking or just aren't able to figure out how to do something, please feel free to get in touch with the dev team and we can give you a hand. Your questions will also help inform which parts of the docs and website need fleshing out most.
And now, without further do, click here to see documentation for version 0.4b, the current release.
Our previous releases were 0.1b, 0.2b, and 0.3b! But the documentation system was not complete in those. In the future, this section will contain links to older versions of the documentation so people can make sure they keep things straight with whatever toolbox version they are using. For now, we doubt anyone using early releases needs web-ified documentation but if you are still using that release, your best bet for documentation is going to be inside the corresponding folder in that release.
If you just want a (semi-)succinct intro to using the toolbox, setting up a machine for deep learning, and/or doing TensorFlow-based analyses with it, see one or more of the videos below. Remember that things change rapidly in the machine learning world, so some of this might go slightly out of date -- but the basic principles should still apply! If anything goes horribly out of date, we'll try to redo the videos. In the meantime, videos are chronological, so those at the bottom will be the most recent.