Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

https://github.com/limuloo/MIGC

[CVPR 2024 Highlight] "MIGC: Multi-Instance Generation Controller for Text-to-Image Synthesis" (Official Implementation)
https://github.com/limuloo/MIGC

aigc computer-vision cvpr cvpr2024 stable-diffusion text-to-image

Last synced: about 2 months ago
JSON representation

[CVPR 2024 Highlight] "MIGC: Multi-Instance Generation Controller for Text-to-Image Synthesis" (Official Implementation)

Lists

README 

        

# [CVPR2024 Highlight] MIGC: Multi-Instance Generation Controller for Text-to-Image Synthesis



**COCO-MIG Bench:**  [![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/migc-multi-instance-generation-controller-for/conditional-text-to-image-synthesis-on-coco-1)](https://paperswithcode.com/sota/conditional-text-to-image-synthesis-on-coco-1?p=migc-multi-instance-generation-controller-for)



**Online Demo on Colab:** [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/drive/1rkhi7EylHXACbzfXvWiblM4m1BCGOX5-?usp=sharing)

### [[Paper]](https://arxiv.org/pdf/2402.05408.pdf)     [[Project Page]](https://migcproject.github.io/)    [[ZhiHu(知乎)]](https://zhuanlan.zhihu.com/p/686367982)

**MIGC: Multi-Instance Generation Controller for Text-to-Image Synthesis**


_Dewei Zhou, You Li, Fan Ma, Xiaoting Zhang, Yi Yang_


## To Do List

- [x] Project Page

- [x] Code

- [x] COCO-MIG Benchmark

- [x] Pretrained Weights on SD1.4

- [x] WebUI

- [x] Colab Demo

- [ ] Pretrained Weights on SD1.5, SD2, SDXL (Note that MIGC_SD14.ckpt can be used directly for the SD1.5 model.)

- [ ] The training code will be released before May 2024



## Gallery

![attr_control](figures/attr_control.png)

![quantity_control](figures/quantity_control.png)

![animation_creation](figures/animation_creation.png)





## Installation



### Conda environment setup

```

conda create -n MIGC_diffusers python=3.9 -y

conda activate MIGC_diffusers

pip install -r requirement.txt

pip install -e .

```

### Checkpoints

Download the [MIGC_SD14.ckpt (219M)](https://drive.google.com/file/d/1v5ik-94qlfKuCx-Cv1EfEkxNBygtsz0T/view?usp=sharing) and put it under the 'pretrained_weights' folder.

```

├── pretrained_weights

│   ├── MIGC_SD14.ckpt

├── migc

│   ├── ...

├── bench_file

│   ├── ...

```



## Single Image Generation

By using the following command, you can quickly generate an image with MIGC.

```

CUDA_VISIBLE_DEVICES=0 python inference_single_image.py

```

The following is an example of the generated image based on stable diffusion v1.4.

 



  example

  example_annotation




🚀 **Enhanced Attribute Control**: For those seeking finer control over attribute management, consider exploring the `python inferencev2_single_image.py` script. This advanced version, `InferenceV2`, offers a significant improvement in mitigating attribute leakage issues. By accepting a slight increase in inference time, it enhances the Instance Success Ratio from 66% to an impressive 68% on COCO-MIG Benchmark. It is worth mentioning that increasing the `NaiveFuserSteps` in `inferencev2_single_image.py` can also gain stronger attribute control.





  example




💡 **Versatile Image Generation**: MIGC stands out as a plug-and-play controller, enabling the creation of images with unparalleled variety and quality. By simply swapping out different base generator weights, you can achieve results akin to those showcased in our [Gallery](#Gallery). For instance:



- 🌆 **[RV60B1](https://civitai.com/models/4201/realistic-vision-v60-b1)**: Ideal for those seeking lifelike detail, RV60B1 specializes in generating images with stunning realism.

- 🎨 **[Cetus-Mix](https://civitai.com/models/6755/cetus-mix)** and **[Ghost](https://civitai.com/models/36520/ghostmix)**: These robust base models excel in crafting animated content.



  example




## COCO-MIG Bench



To validate the model's performance in position and attribute control, we designed the [COCO-MIG](https://github.com/LeyRio/MIG_Bench) benchmark for evaluation and validation.



By using the following command, you can quickly run inference on our method on the COCO-MIG bench:

```

CUDA_VISIBLE_DEVICES=0 python inference_mig_benchmark.py

```

We sampled [800 images](https://drive.google.com/drive/folders/1UyhNpZ099OTPy5ILho2cmWkiOH2j-FrB?usp=sharing) and compared MIGC with InstanceDiffusion, GLIGEN, etc. On COCO-MIG Benchmark, the results are shown below.



  

    

      Method

      MIOU↑

      Instance Success Rate↑

	  Model Type

    Publication

    

	

      L2

      L3

      L4

      L5

      L6

      Avg

	  L2

      L3

      L4

      L5

      L6

      Avg

    

  

  

	

      Box-Diffusion

      0.37

      0.33

      0.25

      0.23

      0.23

      0.26

	  0.28

      0.24

      0.14

      0.12

      0.13

      0.16

	  Training-free

    ICCV2023

    

	

      Gligen

      0.37

      0.29

      0.253

      0.26

      0.26

      0.27

	0.42

      0.32

      0.27

      0.27

      0.28

      0.30

	  Adapter

    CVPR2023

    

	

      ReCo

      0.55

      0.48

      0.49

      0.47

      0.49

      0.49

	  0.63

      0.53

      0.55

      0.52

      0.55

      0.55

	  Full model tuning

    CVPR2023

    

	

      InstanceDiffusion

      0.52

      0.48

      0.50

      0.42

      0.42

      0.46

	  0.58

      0.52

      0.55

      0.47

      0.47

      0.51

	  Adapter

    CVPR2024

    

	

      Ours

      0.64

      0.58

      0.57

      0.54

      0.57

      0.56

	  0.74

      0.67

      0.67

      0.63

      0.66

      0.66

	  Adapter

    CVPR2024

    

  



## MIGC-GUI

We have combined MIGC and [GLIGEN-GUI](https://github.com/mut-ex/gligen-gui) to make art creation more convenient for users. 🔔This GUI is still being optimized. If you have any questions or suggestions, please contact me at [email protected].



![Demo1](videos/video1.gif)



### Start with MIGC-GUI

**Step 1**: Download the [MIGC_SD14.ckpt](https://drive.google.com/file/d/1v5ik-94qlfKuCx-Cv1EfEkxNBygtsz0T/view?usp=drive_link) and place it in `pretrained_weights/MIGC_SD14.ckpt`. 🚨If you have already completed this step during the [Installation](#Installation) phase, feel free to skip it.



**Step 2**: Download the [CLIPTextModel](https://drive.google.com/file/d/1Z_BFepTXMbe-cib7Lla5A224XXE1mBcS/view?usp=sharing) and place it in `migc_gui_weights/clip/text_encoder/pytorch_model.bin`.



**Step 3**: Download the [CetusMix](https://drive.google.com/file/d/1cmdif24erg3Pph3zIZaUoaSzqVEuEfYM/view?usp=sharing) model and place it in `migc_gui_weights/sd/cetusMix_Whalefall2.safetensors`. Alternatively, you can visit [civitai](https://civitai.com/) to download other models of your preference and place them in `migc_gui_weights/sd/`.



```

├── pretrained_weights

│   ├── MIGC_SD14.ckpt

├── migc_gui_weights

│   ├── sd

│   │   ├── cetusMix_Whalefall2.safetensors

│   ├── clip

│   │   ├── text_encoder

│   │   │   ├── pytorch_model.bin

├── migc_gui

│   ├── app.py

```



**Step 4**: `cd migc_gui`



**Step 5**: Launch the application by running `python app.py --port=3344`. You can now access the MIGC GUI through http://localhost:3344/. Feel free to switch the port as per your convenience.



## MIGC + LoRA

MIGC can achieve powerful attribute-and-position control capabilities while combining with LoRA. 🚀 We will integrate this function into MIGC-GUI in the future, so stay tuned! 🌟👀



  migc_lora_id

  migc_lora

  migc_lora_anno

  migc_lora_gui_creation




## Ethical Considerations

The broad spectrum of image creation possibilities offered by MIGC might present comparable ethical dilemmas to those encountered with numerous other methods of generating images from text.



## Contact us

If you have any questions, feel free to contact me via email [email protected] 



## Acknowledgements

Our work is based on [stable diffusion](https://github.com/Stability-AI/StableDiffusion), [diffusers](https://github.com/huggingface/diffusers), [CLIP](https://github.com/openai/CLIP), and [GLIGEN-GUI](https://github.com/mut-ex/gligen-gui). We appreciate their outstanding contributions.



## Citation

If you find this repository useful, please use the following BibTeX entry for citation.

```

@misc{zhou2024migc,

      title={MIGC: Multi-Instance Generation Controller for Text-to-Image Synthesis}, 

      author={Dewei Zhou and You Li and Fan Ma and Xiaoting Zhang and Yi Yang},

      year={2024},

      eprint={2402.05408},

      archivePrefix={arXiv},

      primaryClass={cs.CV}

}

```