README.md

---
license: mit
---

## How to run

**Note**: This is only a single prior model checkpoint and has to be run with https://huggingface.co/warp-diffusion/wuerstchen

```python
import torch
from diffusers import AutoPipelineForText2Image
from diffusers.pipelines.wuerstchen import WuerstchenPrior

prior_model = WuerstchenPrior.from_pretrained("warp-diffusion/wuerstchen-prior-model-interpolated", torch_dtype=torch.float16)
pipe = AutoPipelineForText2Image.from_pretrained("warp-diffusion/wuerstchen", prior_prior=prior_model, torch_dtype=torch.float16).to("cuda")

prompt = [
    "An old destroyed car standing on a cliff in norway, cinematic photography",
    "Western movie, closeup cinematic photography",
    "Pink nike shoe commercial, closeup cinematic photography",
    "Croatia, closeup cinematic photography",
    "South Tyrol mountains at sunset, closeup cinematic photography",
]
images = pipe(prompt, guidance_scale=8.0, width=1024, height=1024).images
```