nicer formatting of the header section that introduces the libary
collapse the three steps of the example into a single snippet that can be directly copy-pasted into a notebook
add a second example (first one is quick and only between two transformers, second chooses between 17 suggested models)
add a line about TransformerRanker being useful for layerwise analysis
Things still missing that would be helpful:
Some tutorials that explain the options in more details (examples with other datasets, explain downsampling, explain the transferability estimation options)
Some suggestions on how to organize the README:
Things still missing that would be helpful: