LoRA: Z-Detail-Slider v1.0
Overview
| Field | Value |
|---|---|
| Category | Detailer |
| NSFW | No |
| Model | Z-Image Turbo |
| File | `zturbo__detailer_z_detail_slider_v1.0_lora_f16.ckpt` |
| Version | v1.0 — released December 7, 2025 |
| File size | 20.33 MB |
| Creator | oosayam |
| Source | https://civitai.com/models/2202638/detail-slider-for-z-image |
| Preference | Love |
(source: lora-z-detail-slider.md)
Best for
- Adds fine-grained detail across the image
- Improves texture clarity in both environments and human subjects
- Enhances perceived sharpness without strongly altering composition
- Can also be run at negative weights to simplify / reduce detail
(source: lora-z-detail-slider.md)
How it works
Trained using the Concept Slider method — fine-tunes the model's understanding of opposing concept pairs ("detail" vs "plain white background") rather than teaching new concepts. This allows continuous control over detail level through a single weight parameter. (source: lora-z-detail-slider.md)
Trigger cues
No trigger word needed. The primary concept is "detail"; the opposing anchor is "plain white background." (source: lora-z-detail-slider.md)
High-detail prompt additions: `extremely detailed, intricate details, high resolution textures, fine details visible, sharp focus, clear photorealistic quality, realistic surface textures`
Strength / weight
The slider's native range is -2.0 to +2.0; typical practical operation is -1.0 to +1.0. In `draw-things-cli` the `--config-json` `loras[].weight` field maps 1:1 to this range.
| Weight | Effect |
|---|---|
| Negative (-1.0 to 0) | Reduces detail, simplifies rendering |
| Low (0–0.2) | Subtle detail boost |
| Medium (~0.4, preferred) | Clear detail and texture improvement |
| High (above 0.6) | Untested / not recommended |
(source: lora-z-detail-slider.md — "Medium (~40, preferred)" on its 0–100% display scale = 0.4 in the CLI)
draw-things-cli usage
```json { "steps": 8, "guidanceScale": 1, "sampler": 15, "loras": [{"file": "zturbo__detailer_z_detail_slider_v1.0_lora_f16.ckpt", "weight": 0.4, "version": "z_image"}] } ```
Note: The general LoRA weight guidance in Z-Image Turbo LoRAs (0.7–0.9) applies to style LoRAs; this detail slider is specifically calibrated at 0.4 for best results.
Compatibility
Works well with: most base styles and subject LoRAs (stack after style LoRAs at ~0.4). (source: lora-z-detail-slider.md)
Conflicts with: not observed (source: lora-z-detail-slider.md)
Failure modes
Secondary color changes: contrast enhancement is an unavoidable side effect noted by the creator — expect minor color shifts at higher weights. (source: lora-z-detail-slider.md)
Notes
- Z-Image Turbo itself has strong inherent detail; this slider may be optional depending on the generation (source: lora-z-detail-slider.md)
- 27.5K downloads, 37.1K favorites on Civitai (source: lora-z-detail-slider.md)