Version names includes the base model it was trained from, use appropriate version for each model. Rouwei version wont work well with noob, the opposite is also true. Old noob versions would be too weak on a fresh noob checkpoints also, eps and vpred should also be used with their dedicated versions. Don't forget to check version descriptions for more specific info and list of supported style tags.
V5 versions was trained with the help of Minthy's outstanding 2xGTX710 rig hardware
V6 version should be one of the cleanest from artifacts due to edm2, simple example of that. Should work fine with cfg++ and classic samplers. Some of previous version styles was dropped for various reasons.
BIO - best in one versions, in particular v1.337 version was forked from v4.02 version of the dataset with further cleaning of useless, ugly or anatomically incorrect styles and adding more styles from recent versions. It's 90% filled with naiv3 or styles from local models with very distinctive look only. Pretty lucky version, it doesn't have any problems with artifacts or darker scenes, from first limited tests.
BIO v2.217 is a massive dataset update from v1.337 version, most of it still naiv3 or distinctive local xl models styles.
Description
Version inspired by this artstyle lora. Half of the dataset was kindly provided by the author of that lora, the other half was additionally added from nai gens. This version was trained from noob-0.5-epsilon checkpoint, so should be compatible with most illustrious epsilon models.