[Bug]: Improving on-boarding experience

[laserkelvin]

Expected behavior

There should be a clear "landing zone" for when new people looking at the repository: what exact script should I run to try out a specific model and dataset?

The next grade up from this would likely be how do I train the same model on a bigger dataset?

This will help ease users into ramping up on the project and lead them into development from laptop to workstation to cluster.

Actual behavior

N/A

Steps to reproduce the problem

Scan the README, and it might be a little like drinking from the fire hose right now as we get a sensory overload.

Specifications

N/A