ComfyUI Extension: ComfyUI-DyPE

Authored by wildminder

Created 3 days ago

Updated 3 days ago

97 stars

DyPE for FLUX. Artifact-free 4K+ image generation.

Custom Nodes (0)

README

<div align="center"> <h1 align="center">ComfyUI-DyPE</h1> <img src="https://github.com/user-attachments/assets/4f11966b-86f7-4bdb-acd4-ada6135db2f8" alt="ComfyUI-DyPE Banner" width="70%"> <p align="center"> A ComfyUI custom node that implements <strong>DyPE (Dynamic Position Extrapolation)</strong>, enabling FLUX-based models to generate ultra-high-resolution images (4K and beyond) with exceptional coherence and detail. <br /> <br /> <a href="https://github.com/wildminder/ComfyUI-DyPE/issues/new?labels=bug&template=bug-report---.md">Report Bug</a> · <a href="https://github.com/wildminder/ComfyUI-DyPE/issues/new?labels=enhancement&template=feature-request---.md">Request Feature</a> </p> </div>  <div align="center">

</div> <br>

About The Project

DyPE is a novel, training-free method that allows pre-trained diffusion transformers like FLUX to generate images at resolutions far beyond their training data, with no additional sampling cost.

It works by taking advantage of the spectral progression inherent to the diffusion process. By dynamically adjusting the model's positional encodings at each step, DyPE matches their frequency spectrum with the current stage of the generative process—focusing on low-frequency structures early on and resolving high-frequency details in later steps. This prevents the repeating artifacts and structural degradation typically seen when pushing models beyond their native resolution.

<div align="center"> <img alt="ComfyUI-DyPE example workflow" width="70%" src="https://github.com/user-attachments/assets/e5c1d202-b2c4-474b-b52f-9691ab44c47a" /> <p><sub><i>A simple, single-node integration to patch your FLUX model for high-resolution generation.</i></sub></p> </div>

This node provides a seamless, "plug-and-play" integration of DyPE into any FLUX-based workflow.

✨ Key Features:

True High-Resolution Generation: Push FLUX models to 4096x4096 and beyond while maintaining global coherence and fine detail.
Single-Node Integration: Simply place the DyPE for FLUX node after your model loader to patch the model. No complex workflow changes required.
Full Compatibility: Works seamlessly with your existing ComfyUI workflows, samplers, schedulers, and other optimization nodes like Self-Attention or quantization.
Fine-Grained Control: Exposes key DyPE hyperparameters, allowing you to tune the algorithm's strength and behavior for optimal results at different target resolutions.
Zero Inference Overhead: DyPE's adjustments happen on-the-fly with negligible performance impact.

🚀 Getting Started

The easiest way to install is via ComfyUI Manager. Search for ComfyUI-DyPE and click "Install".

Alternatively, to install manually:

Clone the Repository:

Navigate to your ComfyUI/custom_nodes/ directory and clone this repository:
```
git clone https://github.com/wildminder/ComfyUI-DyPE.git
```
Start/Restart ComfyUI: Launch ComfyUI. No further dependency installation is required.

🛠️ Usage

Using the node is straightforward and designed for minimal workflow disruption.

Load Your FLUX Model: Use a standard Load Checkpoint node to load your FLUX model (e.g., FLUX.1-Krea-dev).
Add the DyPE Node: Add the DyPE for FLUX node to your graph (found under model_patches/unet).
Connect the Model: Connect the MODEL output from your loader to the model input of the DyPE node.
Set Resolution: Set the width and height on the DyPE node to match the resolution of your Empty Latent Image.
Connect to KSampler: Use the MODEL output from the DyPE node as the input for your KSampler.
Generate! That's it. Your workflow is now DyPE-enabled.

[!NOTE] This node specifically patches the diffusion model (UNet). It does not modify the CLIP or VAE models. It is designed exclusively for FLUX-based architectures.

Node Inputs

model: The FLUX model to be patched.
width / height: The target image resolution. This must match the resolution set in your Empty Latent Image node.
method: The core position encoding extrapolation method. yarn is the recommended default, as it forms the basis of the paper's best-performing "DY-YaRN" variant.
enable_dype: Enables or disables the dynamic, time-aware component of DyPE.
- Enabled (True): Both the noise schedule and RoPE will be dynamically adjusted throughout sampling. This is the full DyPE algorithm.
- Disabled (False): The node will only apply the dynamic noise schedule shift. The RoPE will use a static extrapolation method (e.g., standard YARN). This can be useful for comparison or if you find it works better at certain moderate resolutions.
dype_exponent: (λt) Controls the "strength" of the dynamic effect over time. This is the most important tuning parameter.
- 2.0 (Exponential): Recommended for 4K+ resolutions. It's an aggressive schedule that transitions quickly.
- 1.0 (Linear): A good starting point for ~2K-3K resolutions.
- 0.5 (Sub-linear): A gentler schedule that may work best for resolutions just above the model's native 1K.
base_shift / max_shift (Advanced): These parameters control the interpolation of the dynamic noise schedule shift (mu). The default values (0.5, 1.15) are taken directly from the FLUX architecture and are generally optimal. Adjust only if you are an advanced user experimenting with the noise schedule.

[!WARNING] It seems the width/height parameters in the node are buggy. Keep the values below 1024x1024; doing so won’t affect your output.

Beyond the code, I believe in the power of community and continuous learning. I invite you to join the 'TokenDiff AI News' and 'TokenDiff Community Hub'

<table border="0" align="center" cellspacing="10" cellpadding="0"> <tr> <td align="center" valign="top"> <h4>TokenDiff AI News</h4> <a href="https://t.me/TokenDiff"> <img width="40%" alt="tokendiff-tg-qw" src="https://github.com/user-attachments/assets/e29f6b3c-52e5-4150-8088-12163a2e1e78" /> </a> <p><sub>🗞️ AI for every home, creativity for every mind!</sub></p> </td> <td align="center" valign="top"> <h4>TokenDiff Community Hub</h4> <a href="https://t.me/TokenDiff_hub"> <img width="40%" alt="token_hub-tg-qr" src="https://github.com/user-attachments/assets/da544121-5f5b-4e3d-a3ef-02272535929e" /> </a> <p><sub>💬 questions, help, and thoughtful discussion.</sub> </p> </td> </tr> </table> <p align="center">══════════════════════════════════</p>

⚠️ Known Issues and Limitations

FLUX Only: This implementation is highly specific to the architecture of the FLUX model and will not work on standard U-Net models (like SD 1.5/SDXL) or other Diffusion Transformers.
Parameter Tuning: The optimal dype_exponent can vary based on your target resolution. Experimentation is key to finding the best setting for your use case. The default of 2.0 is optimized for 4K.

License

The original DyPE project is patent pending. For commercial use or licensing inquiries regarding the underlying method, please contact the original authors.

Acknowledgments

Noam Issachar, Guy Yariv, and the co-authors for their groundbreaking research and for open-sourcing the DyPE project.
The ComfyUI team for creating such a powerful and extensible platform for diffusion model research and creativity.