Leather Jacket LORA | Generate detailed leather outfits

Short description

• Leather Jacket LORA— A leather-focused LoRA trained on 500 curated images for 50,000 steps on the Perfect Pony XL checkpoint. Adds realistic glossy leather reflections, hardware details (zippers, studs), cropped jacket silhouettes, and tight leather pants/skirts to prompts while preserving fine facial and skin detail when combined with appropriate checkpoints and prompt tokens.

Key metadata

• Main trigger: fnxleather

• Trigger phrases: black cropped leather jacket, metallic zipper hardware, leather, side zippers, jacket zipper, glossy reflections

• Trained on: 500 images

• Training steps: 50,000

• Best-tested checkpoint: Perfect Pony XL — /model/439889/prefect-pony-xl

Quick usage guide

• Basic usage: Add the token fnxleather inside your prompt where you want the leather styling applied. Example placement: "fnxleather, black cropped leather jacket, metallic zipper hardware, glossy reflections, ...".

• Weighting: start with LoRA weight 0.6–1.0. decrease (0.2–0.5) for subtle accents.

• Combining: works best with Perfect Pony XL. When combining with other LoRAs, place fnxleather near the clothing/texture descriptors and adjust weights to avoid conflicts.

• Resolution & sampling: try 768×1024 or 1024×1536 for more detail; 25–50 sampling steps (or follow checkpoint recommendations); CFG 6–8 as a starting point; adjust sampler to your workflow (Euler a, DPM++ etc. depending on your UI).

Strengths

• Highly consistent with leather clothing aesthetics: cropped jackets, zippers, side zippers, hardware and gloss.

• Preserves silhouette and clothing form well when prompt placed near clothing descriptors.

• Works well in semi-realistic anime and photorealistic portrait pipelines (especially Perfect Pony XL).