DORO EPIC AIRBRUSH — Soft Gradients 1

Versions

v1 DEA_SOFT_GRAD_1 — Initial release. Trained on 15 abstract airbrush gradient crops (768px, 600 steps, dim 32).

---

Compatibility

- Illustrious XL ✅

- Pony XL ✅

- SDXL (base) ❓ not tested

---

Quick Start

Trigger: DEA_SOFT_GRAD_1

Sweet spot:

- 0.7–1.9 → subtle, visible

- 2.0–3.0 → effect kicks in, full style ⭐

- >3.0 → overpowered, style compression

> ⚠️ High-offset LoRA — effective range starts at 2.0+, not the usual 0.5–1.0

Prompt example:

epic clouds over mountains, volumetric light, DEA_SOFT_GRAD_1 <lora:DEA_SOFT_GRAD_1:2.5>

---

Description

📸 Dataset: 15 abstract airbrush gradient crops — no objects, pure tonal transitions and color blending. Trained at 768px, 600 steps, AdamW8bit, cosine_with_restarts.

✨ Emergent effects:

- Gradient surfaces — clouds, smoke, fog, fire, atmosphere: deep analog painterly quality, as if airbrushed on paper

- Smooth surfaces — skin, plastic, metal: surface-blur-like effect, micro-noise and texture artifacts, evens gradients

- Object edges — sharpened and stylized, more "painted" feel

- Background-first — at moderate weights affects mainly background; at high weights touches subjects too

⚠️ Side effect: Smoothing suppresses fine texture (pores, grain, fabric). Not ideal when texture detail is the goal. Workaround: generate smooth, then add noise + slight blur in Photoshop.

💡 Bonus use: Pre-upscale prep — smooths surfaces and reduces artifacts for a cleaner upscale input.

---

What happened under the hood

This LoRA was trained on abstract gradient crops with no recognizable objects — and that turned out to be the key.

The model couldn't learn any specific object, so it learned pure rendering principles: how to blend tones, transitions, and light. When applied, it rewrites the model's rendering language across all volumetric, gradient-by-nature subjects — clouds, smoke, fire — because those subjects are gradients at their core.

This is a case of spontaneous feature disentanglement: an abstract dataset forced style and content to separate. The result is a universal style modifier, not a content LoRA — similar in principle to implicit style-content separation described in B-LoRA (ECCV 2024).

B-LoRA paper:

https://arxiv.org/abs/2403.14572