Photorealistic Asian female character LoRA trained on SDXL base.
Optimized for consistent character generation across varied poses, settings, and outfits.
Training: 42 images, 20 epochs
Best results: Epoch 15-16
Trigger: Use with base prompts, no specific trigger word required
Recommended settings:
- Base Model: EpicRealism XL or similar SDXL checkpoint
- LoRA Weight: 0.7-1.0
- CFG Scale: 4-7
- Sampler: DPM++ 2M Karras
- Steps: 30-50
For personal content creation only.
