metadata
language:
- en
tags:
- lora
- z-image
- z_image
- generation
base_model:
- Tongyi-MAI/Z-Image-Turbo
license: apache-2.0
pretty_name: z-index outline test
widget:
- output:
url: img/1.png
text: male in red shirt , front view , with black outline on top
- output:
url: img/2.png
text: cat in red shirt , front view , with black outline on top
- output:
url: img/car.png
text: v2 , no promt , use 1.5-3.0 str
- output:
url: img/3d.png
text: use 0.8-2.0+ str
z-index_outline_test
prompt :
with black outline on topor variation to work better (but it's not needed)

- Prompt
- male in red shirt , front view , with black outline on top

- Prompt
- cat in red shirt , front view , with black outline on top

- Prompt
- v2 , no promt , use 1.5-3.0 str

- Prompt
- use 0.8-2.0+ str
v1 trained on concept slider
v2 works better using pair datasets , one is used for regularization