Meissonic: Revitalizing Masked Generative Transformers for Efficient High-Resolution Text-to-Image Synthesis
Meissonic is a non-autoregressive mask image modeling text-to-image synthesis model that can generate high-resolution images. It is designed to run on consumer graphics cards.
Key Features:
- ๐ผ๏ธ High-resolution image generation (up to 1024x1024)
- ๐ป Designed to run on consumer GPUs
- ๐จ Versatile applications: text-to-image, image-to-image
git clone https://github.com/viiika/Meissonic/
cd Meissonic
conda create --name meissonic python
conda activate meissonic
pip install -r requirements.txt
git clone https://github.com/huggingface/diffusers.git
cd diffusers
pip install -e .
python app.py
python inference.py --prompt "Your creative prompt here"
python inpaint.py --mode inpaint --input_image path/to/image.jpg
python inpaint.py --mode outpaint --input_image path/to/image.jpg
Optimize performance with FP8 quantization:
Requirements:
- CUDA 12.4
- PyTorch 2.4.1
- TorchAO
Note: Windows users install TorchAO using
pip install --pre torchao --index-url https://download.pytorch.org/whl/nightly/cpu
Command-line inference
python inference_fp8.py --quantization fp8
Gradio for FP8 (Select Quantization Method in Advanced settings)
python app_fp8.py
Precision (Steps=64, Resolution=1024x1024) | Batch Size=1 (Avg. Time) | Memory Usage |
---|---|---|
FP32 | 13.32s | 12GB |
FP16 | 12.35s | 9.5GB |
FP8 | 12.93s | 8.7GB |
If you find this work helpful, please consider citing:
@article{bai2024meissonic,
title={Meissonic: Revitalizing Masked Generative Transformers for Efficient High-Resolution Text-to-Image Synthesis},
author={Bai, Jinbin and Ye, Tian and Chow, Wei and Song, Enxin and Chen, Qing-Guo and Li, Xiangtai and Dong, Zhen and Zhu, Lei and Yan, Shuicheng},
journal={arXiv preprint arXiv:2410.08261},
year={2024}
}
We thank the community and contributors for their invaluable support in developing Meissonic. We thank [email protected] for making Meissonic Demo. We thank @NewGenAI and @้ฃ้ทนใใใ@่ช็งฐๆ็ณปใใญใฐใฉใใฎๅๅผท for making YouTube tutorials. We thank @pprp for making fp8 and int4 quantization. We thank @camenduru for making jupyter tutorial. We thank @chenxwh for making Replicate demo and api. We thank Collov Labs for reproducing Monetico. We thank Shitong et al. for identifying effective design choices for enhancing visual quality.
Made with โค๏ธ by the MeissonFlow Research