[FR] add documentation about the `loss_class` selection

[PhilipMay]

Can you please add documentation about the loss_class selection?

There seems to be a new loss class implemented called SupConLoss. Maybe you could add some documentation how to select them and which advantages and disadvantages that might have.

Many thanks Philip

[lewtun]

Thanks for raising this @PhilipMay ! Indeed, I should document these options better - will add it to my TODO list :)

[aaronvg]

on that note, when should we use SupConLoss? Seems like all the examples use cosine similarty loss?