ARPG

🚀 Autoregressive Image Generation with Randomized Parallel Decoding

ARPG is a novel autoregressive image generation framework that can perform BERT - style masked modeling with a GPT - style causal architecture.

🚀 Quick Start

ARPG is a novel autoregressive image generation framework capable of performing BERT - style masked modeling with a GPT - style causal architecture. It has the following features: 💪 FID 1.94 🚀 Fast Speed ♻️ Low Memory Usage 🎲 Radnom Order 💡 Zero - shot Inference.

✨ Features

• Capable of BERT - style masked modeling with a GPT - style causal architecture.
• Achieves an FID of 1.94.
• High - speed image generation.
• Low memory usage.
• Supports random order decoding.
• Enables zero - shot inference.

📦 Installation

This section is skipped as no installation steps are provided in the original document.

💻 Usage Examples

Basic Usage

# You can easily load it through the Hugging Face DiffusionPipeline and optionally customize various parameters such as the model type, number of steps, and class labels.
from diffusers import DiffusionPipeline

pipeline = DiffusionPipeline.from_pretrained("hp-l33/ARPG", custom_pipeline="hp-l33/ARPG")

class_labels = [207, 360, 388, 113, 355, 980, 323, 979]

generated_image = pipeline(
    model_type="ARPG-XL",       # choose from 'ARPG-L', 'ARPG-XL', or 'ARPG-XXL'
    seed=0,                     # set a seed for reproducibility
    num_steps=64,               # number of autoregressive steps
    class_labels=class_labels,  # provide valid ImageNet class labels
    cfg_scale=4,                # classifier-free guidance scale
    output_dir="./images",      # directory to save generated images
    cfg_schedule="constant",    # choose between 'constant' (suggested) and 'linear'
    sample_schedule="arccos",   # choose between 'arccos' (suggested) and 'cosine'
)

generated_image.show()

Advanced Usage

No advanced usage code is provided in the original document, so this part is skipped.

📚 Documentation

This section is skipped as no additional detailed documentation is provided in the original document.

🔧 Technical Details

This section is skipped as no technical implementation details are provided in the original document.

📄 License

This project is licensed under the MIT license.

📚 Citation

If this work is helpful for your research, please give it a star or cite it:

@article{li2025autoregressive,
    title={Autoregressive Image Generation with Randomized Parallel Decoding},
    author={Haopeng Li and Jinyue Yang and Guoqi Li and Huan Wang},
    journal={arXiv preprint arXiv:2503.10568},
    year={2025}
}

🙏 Acknowledgement

Thanks to LlamaGen for its open - source codebase. Appreciate RandAR and [RAR](https://github.com/bytedance/1d - tokenizer/blob/main/README_RAR.md) for inspiring this work, and also thank ControlAR.