Improving Documentation

[andrewilyas]

Putting this issue here for general documentation improvements. Will add to this issue as more things come up and cross things out as they get done, but pull requests are (very) welcome!

• More examples of FFCV for different applications;
• FFCV v1.0 now has a much more powerful Pipeline system that allows for applications like contrastive learning out-of-the-box, but these features are currently undocumented;
• More robust installation instructions would be helpful, as well as a troubleshooting guide

[numpee]

Willing to contribute (examples and perhaps installation?), but I have a lot of things going on rn so will need a few weeks to start working

[tavisshore]

My current research largely uses contrastive learning, I'll happily contribute to documentation as I mainly use FFCVs now