Get all 100+ of my premium LoRA for free by joining my membership including Winter v5 on my kofi or buymeacoffee.
If you want to make your own custom LoRA you can also contact me there.
-------
Trigger word: aespawinter
Use CLIP SKIP 2. You may need to use HIRES FIX with denoise strength between 0.5-0.7, Hires steps 20-30 is good but you can use whatever you want. Upscaler use 4x Foolhardy Remacri or SWINIR 4x.
Best checkpoints/models to use:
Realistic: ChillOutMix, VisionGen, AEros, Snapdd00, Henrealmix
Fanart/painting style: Dreamshaper, LucidMix, SunshineMix, Majicmix
Use it wisely. I don't responsible for anything you generated with this LoRa.
Description
First 2023 Dreambooth version. Net 128. Trained on latest version of Linaqruf's colab using 30 images with HassanBlend 1.4 as its base model.
I use new settings for this, it's better in face similarity than the previous settings but it's harder to stylize it. Best at close-up, pose will be repetitive at some seeds and prompts because its dataset have many similar pose, but it can be solved with Controlnet, the face similarity will be decrease though.
If this is good enough for you all, I'll share guide and tutorial to making this, so let me know.
FAQ
Comments (14)
Hey, I was wondering what new settings you used for this LORA? I've been trying to train some myself to mixed results
I'm looking forward to the tutorial, sir.
There is already a post I wrote about the settings I used but if you never tried training LoRA before it might be confusing you.
Absolutely amazing as always, it works really well in both realistic and anime styles, what about making a chuu lora next?
Thank you. Maybe later, I still have some list and commissions ongoing.
Kim Hyun A，please
I downloaded your winter model and applied it to Korap, but Laura doesn't work. Any idea why?
Details
Files
Available On (2 platforms)
Same model published on other platforms. May have additional downloads or version variants.