https://github.com/junjie-shentu/AttenCraft

Implementation of AttenCraft
https://github.com/junjie-shentu/AttenCraft

Last synced: 3 months ago
JSON representation

Implementation of AttenCraft

• Host: GitHub
• URL: https://github.com/junjie-shentu/AttenCraft
• Owner: junjie-shentu
• Created: 2024-05-22T14:05:53.000Z (8 months ago)
• Default Branch: main
• Last Pushed: 2024-06-26T16:03:37.000Z (7 months ago)
• Last Synced: 2024-08-01T18:34:55.294Z (6 months ago)
• Language: Python
• Size: 33.9 MB
• Stars: 5
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

• awesome-diffusion-categorized - [Code

README

        
# AttenCraft: Attention-guided Disentanglement of Multiple Concepts for Text-to-Image Customization

This is the implementation of the paper "AttenCraft: Attention-guided Disentanglement of Multiple Concepts for Text-to-Image Customization". [Paper Link](https://arxiv.org/abs/2405.17965)

## Getting Started

Intsall environment:

```

conda create --name attencraft --file environment.yml

conda activate attencraft

```

## Training AttenCraft

```

bash train.sh

```

Note that the `--output_dir` flag specifies the output directory where the checkpoints will be saved, and should contain 'wkwv' since this will be used in the inference script.

## Inference

Input the chackpoint path, output path, and the text ptompt for the image generation in the `inference.py` file and run the python script.

## Citation

If you find this work helpful, please consider citing the following BibTeX entry:

```

@article{shentu2024attencraft,

  title={AttenCraft: Attention-guided Disentanglement of Multiple Concepts for Text-to-Image Customization},

  author={Shentu, Junjie and Watson, Matthew and Moubayed, Noura Al},

  journal={arXiv preprint arXiv:2405.17965},

  year={2024}

}

```