Base model: Anything V5
Training environment: pixai.art LoRA learning
Training images: 512×768 80 images
Trigger word: barbara_a, blonde hair, twintails, drill hair, twin drills, bangs, blue eyes, white headwear, Barbara costume,
80 images used for training are now available on patreon.
→https://www.patreon.com/posts/lora-babara-80-97527765
★If you can't do it well
In my environment, I was able to use a combination of Anylora and DPM++ 3M SDE Exponential, but please adjust according to your own environment.
LoRA strength is also adjusted individually.
If it doesn't work, the base model is Anything v5, so changing the model to an Anyting type may improve the situation.