03/11/25: Accidently trained it also for SDXL (i wanted to train it on chroma). But works on SDXL pretty nicely too!

------------------------------------------------

My aim with this LoRa was to replicate the aesthetics of post punk doomer playlists on YT with these 2d anime girls edited on eastern european photorealistic moody backgrounds. Trained on Illustrious-XL v0.1 but upon checking with some models I have to say that this LoRa works aesthetics wise very well with my own checkpoint Synta-XL Illustrious Mix. I recomment using it with this LoRa.

CLIP Skip; 1 - I feel like the realistic backgrounds come out better if CLIP Skip is set at 1 instead of 2)

Triggerword: s0lnos, photo background,

Additional prompts: city, scenery, heavy film grain, distortion, lowres

Indoor scenes tend to work better as close-ups but you can play around and see for yourself.

REMOVE ALL QUALITY TAGS IN POSITIVE AND NEGATIVE PROMPT BOX FOR MAXIMUM VIBE

Of course you can use the LoRa also purely as a sort of texture and colors filter, bomb your prompt with all kind of topics, use quality tags etc. This will work on its own too, but if you want that specific doomer aesthetics you should follow the advice given. Enjoy!