Update README.md
Browse files
README.md
CHANGED
|
@@ -18,6 +18,18 @@ library_name: diffusers
|
|
| 18 |
|
| 19 |
# **LZO-1-Preview**
|
| 20 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 21 |
|
| 22 |
## Trigger words
|
| 23 |
|
|
|
|
| 18 |
|
| 19 |
# **LZO-1-Preview**
|
| 20 |
|
| 21 |
+
> LZO-1-Preview (Lossless-Zoom-Operator) is an experimental adapter for black-forest-lab’s FLUX.1-Kontext-dev. It is an experimental LoRA designed to zoom into a defined object frame within an image without altering the object's position, maintaining strict center-staged positioning. The model was trained on 550 image pairs (225 original “start” images and 225 “end” images). Synthetic result nodes were generated using Gemini 2.5 Flash Image Preview from Google and annotated with DeepCaption-VLA-7B.
|
| 22 |
+
|
| 23 |
+
> [!note]
|
| 24 |
+
[photo content], zoom in on the specified [face/object/region close-up], enhancing resolution and detail while preserving sharpness, realism, and original context. Maintain natural proportions and background continuity around the zoomed area.
|
| 25 |
+
|
| 26 |
+
## Sample Inference
|
| 27 |
+
|
| 28 |
+
> [photo content], zoom in on the specified [face close-up], enhancing resolution and detail while preserving sharpness, realism, and original context. Maintain natural proportions and background continuity around the zoomed area.
|
| 29 |
+
|
| 30 |
+
|
| 31 |
+
|
| 32 |
+
|
| 33 |
|
| 34 |
## Trigger words
|
| 35 |
|