This LORA works with XL, PNY, ILL and even NAI (eps) checkpoints. NAI support is more limited.

It affects many aspects of image. It adds several new concepts (surrealism, cyberpunk, mixed creatures, nsfw). It improves anatomy, composition, texture details, shading realism, general aesthetics and narrative aspects.

It's hard to tell in simple sentence what it does. Look at 1Pass samples I've posted on this page (pay attention to captions with CKPT name). This LORA is a part of CinEro ILL v6 R6 training (upcoming). This is the public LORa trained on CivitAI. One user asked me about how I train full checkpoint on CivitAI. This is the partial reveal of the whole process.

Feel free to add this LORA into your usual pipeline and give a feedback. PROs and CONs (constructive) will help me to progress further.

PS: BUZZ donations or participation in Bid investments are welcome and very helpful way to support my work. This helps a lot to find the resource and motivation for this very time consuming activity.