File size: 803 Bytes
55da75b |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 |
---
library_name: diffusers
license: mit
---
This is the official released checkpoint of **an aligned SD v2.1** of SafetyDPO: Scalable Safety Alignment for Text-to-Image Generation, designed to generate more safe images from our Safe-SD2.1.
Our project page is [🏠SafetyDPO HomePage](https://safetydpo.github.io/) and the GitHub repo is [⚙️SafetyDPO GitHub](https://github.com/Visualignment/SafetyDPO) where we released all the code and the data.
In the future, we will release additional safe models.
# Usage
A simple use case of our model is:
```
from diffusers import DiffusionPipeline
pipe = DiffusionPipeline.from_pretrained("Visualignment/safe-stable-diffusion-v2-1")
prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k"
image = pipe(prompt).images[0]
``` |