V5: 

-New training images, modified the tags to make the Lora much more varied.

-(I think) 3D will now be applied consistently; the backgrounds should look better, like my recent Snudd Lora.

-Added more characters.

-Use the "Short" or shortstack tag.