Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,74 @@
|
|
| 1 |
-
---
|
| 2 |
-
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
tags:
|
| 3 |
+
- text-to-image
|
| 4 |
+
- lora
|
| 5 |
+
- diffusers
|
| 6 |
+
- template:diffusion-lora
|
| 7 |
+
|
| 8 |
+
base_model: black-forest-labs/FLUX.1-Kontext-dev
|
| 9 |
+
instance_prompt: null
|
| 10 |
+
license: apache-2.0
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
|
| 14 |
+
# RefControl Flux Kontext – Reference Canny LoRA
|
| 15 |
+
|
| 16 |
+
## 📝 Short description
|
| 17 |
+
A LoRA for **Flux Kontext Dev** that fuses a **reference image** (left) with a **canny edge map** (right).
|
| 18 |
+
It preserves **identity and style** from the reference while following the **edges and outlines** from the control.
|
| 19 |
+
|
| 20 |
+
**Trigger word:** `refcontrolcanny`
|
| 21 |
+
|
| 22 |
+
|
| 23 |
+
[Demo video](https://youtu.be/T0IH0jT2h8M)
|
| 24 |
+
|
| 25 |
+
---
|
| 26 |
+
|
| 27 |
+
|
| 28 |
+
## 📊 Examples
|
| 29 |
+
|
| 30 |
+
| Reference | Result |
|
| 31 |
+
|-----------|--------|
|
| 32 |
+
|  |  |
|
| 33 |
+
|  |  |
|
| 34 |
+
|  |  |
|
| 35 |
+
|
| 36 |
+
---
|
| 37 |
+
|
| 38 |
+
## 📖 Extended description
|
| 39 |
+
This LoRA was primarily trained on **humans and characters**, but it works with objects and scenes as well.
|
| 40 |
+
The main purpose is to **preserve identity** — facial features, clothing, style — from the **reference image**, while using the **canny edge map** to define contours, composition, and layout.
|
| 41 |
+
|
| 42 |
+
---
|
| 43 |
+
|
| 44 |
+
## ⚙️ How to use
|
| 45 |
+
1. Concatenate two images side by side:
|
| 46 |
+
- **Left:** reference image (person, character, or object).
|
| 47 |
+
- **Right:** canny edges (line structure).
|
| 48 |
+
2. Add the trigger word `refcontrolcanny` in your prompt.
|
| 49 |
+
3. Adjust LoRA weight (recommended **0.8–1.0**) depending on how strictly you want to follow edges.
|
| 50 |
+
|
| 51 |
+
### ✅ Example prompt
|
| 52 |
+
refcontrolcanny, change canny edges to photo with reference from left
|
| 53 |
+
|
| 54 |
+
---
|
| 55 |
+
|
| 56 |
+
## 🎯 What it does
|
| 57 |
+
- Preserves **style and identity** of the reference.
|
| 58 |
+
- Uses **canny edges** to control structure, outline, and composition.
|
| 59 |
+
- Produces consistent results across different poses and layouts.
|
| 60 |
+
|
| 61 |
+
---
|
| 62 |
+
|
| 63 |
+
## ⚡ Tips
|
| 64 |
+
- Works best if the canny map captures **clean edges** with minimal noise.
|
| 65 |
+
|
| 66 |
+
---
|
| 67 |
+
|
| 68 |
+
## 📌 Use cases
|
| 69 |
+
- Character consistency in illustrations and comics.
|
| 70 |
+
- Stylized redraws of objects with sharp outlines.
|
| 71 |
+
- Scene composition based on silhouettes.
|
| 72 |
+
- Rapid prototyping of layouts and storyboards.
|
| 73 |
+
|
| 74 |
+
---
|