Two custom character LoRAs in one image? ComfyUI user seeks fix for feature bleed
badgaab · reddit · 2026-10-19
A ComfyUI beginner on Reddit asks for help: they trained two custom character LoRAs on Civitai — one female, one male — using WAI-Illustrious-SDXL as the checkpoint. Each LoRA works well when generating its character alone, but combining both in one image causes feature bleed, duplicated characters (2 girls, 2 boys, or even 3 people), and reduced fidelity.
Their goal is natural interaction between the two characters (standing close, hugging, touching) with each LoRA affecting only its own character. Not using ControlNet for now, they're looking for workflow examples, recommended custom nodes, or approaches for dual-character LoRA generation on Illustrious/SDXL.
More from Multimodal
- AVC-Reward: an event-centric framework for evaluating audio-video generation consistency — ZhejiangUniversity · 2026-10-19
- Tsinghua and Z.AI study shows image-generation training expands unified multimodal models' knowledge — zai-org · 2026-10-19
- New benchmark exposes widespread attribute-object binding failures in multimodal retrievers — Netanel Y. Tamir · 2026-10-19
- Midjourney v8.2 Sketch-Watercolor Look via --sref 4094042060 — azed_ai · 2026-10-19
- A Minute of 30fps Video Is 7M Tokens: Why Video LMs Still Need Visual Encoders — HildeKuehne · 2026-10-19
- CyBot-19: stylized robot girl art made with ComfyUI and Z-Image Turbo — No-While1332 · 2026-10-19