TransBody2026 is an SDXL LoRA focused on stable trans anatomy, natural facial proportions, and consistent full-body rendering.
Designed for realism-based checkpoints, it performs reliably across portrait and full-body compositions, including clothed and neutral scenes.
If you test it, I’d love to hear your feedback – especially about anatomy consistency, lighting behavior and flexibility in prompts.
Recommended usage:
Sampler: DPM++ 2M SDE
Steps: 28–32
CFG: 5–6
LoRA weight: 0.6–0.7
For optimal full-body detail, use resolutions of 1024×1536 or higher.
This LoRA responds best to concise, clean prompts.
Description
• First public release of TransBody2026 (SDXL)
• Focus on stable trans anatomy and consistent proportions
• Tested primarily on realism-based checkpoints (EpicRealism, RealismByStableYogi)
• Reliable portrait and full-body performance
• Indoor and clothed compositions validated
Recommended Settings
This LoRA performs best with controlled settings and clean prompts.
Sampler: DPM++ 2M SDE
Steps: 28–32
CFG: 5–6
LoRA weight: 0.6–0.7
For full-body compositions, higher resolution significantly improves anatomical detail.
Recommended resolution: 1024×1536 or even higher.
This LoRA responds well to concise prompts. Overly long prompt stacks are not necessary.
If you test it, I’d love to hear your feedback – especially about anatomy consistency, lighting behavior and flexibility in prompts.
Future updates will expand dataset diversity and background robustness.
FAQ
Comments (1)
I would appreciate comments and your expierence for necessary improvement