README.md

---
license: mit
---
# Unique3d-MVImage-Diffuser Model Card 

[🌟GitHub](https://github.com/TingtingLiao/unique3d_diffuser) | [🦸 Project Page](https://wukailu.github.io/Unique3D/) | [🔋Normal Diffuser](https://huggingface.co/Luffuly/unique3d-normal-diffuser)</a> 


## Example  
Note the input image is required to be **white background**. 

![mv-boy](https://github.com/user-attachments/assets/65558519-dbfe-40de-ac13-5d78527541c5)

```bash 
import torch 
import numpy as np 
from PIL import Image 
from pipeline import StableDiffusionImage2MVCustomPipeline

pipe = Unique3dDiffusionPipeline.from_pretrained( 
    "Luffuly/unique3d-mvimage-diffuser", 
    torch_dtype=torch.float16, 
    trust_remote_code=True,  
    class_labels=torch.tensor(range(4)),
).to("cuda")

seed = -1    
generator = torch.Generator(device='cuda').manual_seed(-1)


image = Image.open('data/boy.png') 
forward_args = dict(
    width=256,
    height=256,
    num_images_per_prompt=4, 
    num_inference_steps=50, 
    width_cond=256,
    height_cond=256, 
    generator=generator,
    guidance_scale=1.5,  
) 

out = pipe(image, **forward_args).images
rgb_np = np.hstack([np.array(img) for img in out])
Image.fromarray(rgb_np).save(f"mv-boy.png")

```



## Citation
```bash 
@misc{wu2024unique3d,
      title={Unique3D: High-Quality and Efficient 3D Mesh Generation from a Single Image}, 
      author={Kailu Wu and Fangfu Liu and Zhihan Cai and Runjie Yan and Hanyang Wang and Yating Hu and Yueqi Duan and Kaisheng Ma},
      year={2024},
      eprint={2405.20343},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}
```