Reviews
Animagine XL 4.0
Overwhelmingly Positive
99% of 3,529 reviews recommend it · 3,529 reviews
yokinarudesu351
VexenAI
This is one of my favourite checkpoints. Great for multiple character LoRAs in one image
schneesturmx91988
Fontose
Lilium_Album
When compared to 3.1 it gets significantly better backgrounds and characters can be generated with incredible accuracy and minimal description. However I don't like that it is seemingly extremely dependent of following the recommended CFG and sampling method, any variation on these parameters will result in significantly worse results. The updated Dataset is much welcome, but I am not much of a fan of the seemingly strong bias towards generating pictures with a cell shading technique, this can be worked around with enough prompt expertise but one of the reasons I loved 3.1 was what they call the "2.5D aesthetic" which was actually a side effect of over-using the aesthetic-score keywords, but the fact is I still love it, and it is much more difficult to get with this version. That being said, this one definitively feels like an improvement and a technological advancement. I would like to Thank the kind people of Cagliostro Lab for all their hard work and for sharing this lovely checkpoint to the public for no cost at all
ShinyLilly
impressive model for characters and art!
thamngapwei546
GreatMonster
MintXil

S27

plsdonotdisturb
maxnami517
"No need" for Loras.
Using the correct tags that they used from training would deliver a nice generation without any loras help.






.jpeg)




