README.md

---
license: other
language:
- en
pipeline_tag: image-to-image
library_name: diffusers
base_model:
- stable-diffusion-v1-5/stable-diffusion-v1-5
---

# BlobCtrl

<a href='http://arxiv.org/abs/2503.13434'><img src="https://img.shields.io/static/v1?label=Arxiv Preprint&message=BlobCtrl&color=red&logo=arxiv"></a>
<a href='https://liyaowei-stu.github.io/project/BlobCtrl/'><img src='https://img.shields.io/badge/Project-Page-green'></a>
<a href='https://huggingface.co/spaces/Yw22/BlobCtrl'><img src='https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-Demo-blue'></a> 
<a href='https://github.com/TencentARC/BlobCtrl'><img src='https://img.shields.io/badge/Github-Repo-blue'></a>

### Overview 

BlobCtrl enables precise, user-friendly element-level visual manipulation.
Main Features: 🦉Element-level Add/Remove/Move/Replace/Enlarge/Shrink.
[<img src="https://github.com/user-attachments/assets/59abefee-9bbd-41c8-a41d-ba18f6bbe2e8" width="600">](https://liyaowei-stu.github.io/project/BlobCtrl)


### Video
[Watch the introduction video](https://liyaowei-stu.github.io/project/BlobCtrl) in our project page or [YouTube](https://youtu.be/rdR4QRR-mbE).

### Code

Please check our [GitHub repository](https://github.com/TencentARC/BlobCtrl) for code.



### Model

Download the model checkpoint using `huggingface_hub` (Version 0.1 as example):

```python
import os
from huggingface_hub import snapshot_download

# download blobctrl models
BlobCtrl_path = "examples/blobctrl/models"
if not (os.path.exists(f"{BlobCtrl_path}/blobnet") and os.path.exists(f"{BlobCtrl_path}/unet_lora")):
    BlobCtrl_path = snapshot_download(
        repo_id="Yw22/BlobCtrl",
        local_dir=BlobCtrl_path,
        token=os.getenv("HF_TOKEN"),
    )
print(f"BlobCtrl checkpoints downloaded to {BlobCtrl_path}")    )
```
 
The downloaded BlobCtrl checkpoint file (blobnet&unet_lora) can be found at `BlobCtrl_path`.



### Demo
You can [try the demo here](https://huggingface.co/spaces/Yw22/BlobCtrl).