ComfyUI_DyPE

ComfyUI_DyPE
★ 29

超高分辨率位置外推ComfyUI集成z-image/qwen/klein兼容
在ComfyUI中封装DyPE（Dynamic Position Extrapolation），用于超高分辨率扩散图像的动态位置外推，支持z-image、qwen-image与Flux2/Klein等模型，便于高分辨率推理集成。
💡 在ComfyUI中对超大尺寸图像进行高质量位置外推与扩散生成。
🍴 3 Forks💻 Python🔄 2026-02-26
📦 网盘链接待填入
📦 requirements.txt
torch
torchvision
diffusers
transformers
accelerate
sentencepiece
📄 README

ComfyUI_DyPE

DyPE: Dynamic Position Extrapolation for Ultra High Resolution Diffusion ,you can use a wrapper node it in comfyUI

Upadte

  • add z image,qwen-image,flux2-klein9/4B support，klein need diffuser>0.37
  • 新增z-image，千问，克莱因9/4B的dype支持, klein需要最新版的diffuser

    • 1.Installation

    In the ./ComfyUI/custom_nodes directory, run the following:

    git clone https://github.com/smthemex/ComfyUI_DyPE

    2.requirements

    pip install -r requirements.txt

    3.checkpoints

  • 3.1 Krea-dev or fp8 don’t support scaled model / 不支持scaled 模型
  • 3.2 ae/T5/clip_l / normal flux dev
  • 3.3 turbo lora alimama-creative
  • 3.4 ip adapter xflux
  • 3.5 clip_vision open
  • qwen/zimage/Klein use comfyUI split models，only klein vae need diffuser version
    • ├── ComfyUI/models/diffusion_models
|     ├── flux1-krea-dev-fp8.safetensors   #  or  flux1-krea-dev.safetensors # flux1
|     ├── FLUX.2-klein-4B.safetensors  #  or 9B or gguf   #flux2
|     ├──Qwen-Image-2512-bf16.safetensors  or gguf  # qwen
|     ├──z_image_turbo_bf16.safetensors or gguf  #zimage
├── ComfyUI/models/vae
|        ├──ae.safetensors or UltraFlux.safetensors  ##flux1 and z image
|        ├──flux2_4b.safetensors  # flux2 klein need diffuser version,don't  support  comfyUI version 需要diffuser的模型
|        ├──qwen_image_vae.safetensors #qwen
├── ComfyUI/models/clip
|        ├──clip_l.safetensors  #flux1
|        ├──t5xxl_fp8_e4m3fn.safetensors #flux1
|        ├──qwen_3_4b.safetensors # z image flux2 klein4b
|        ├──qwen_3_8b_fp8mixed.safetensors # flux2 klein9b
|        ├──qwen_2.5_vl_7b.safetensors #qwen
├── ComfyUI/models/lora
|        ├──flux_turbo.safetensors # flux1
|        ├──flux_real.safetensors #flux1
|        ├──Qwen-Image-Lightning-4steps-V1.0-bf16.safetensors  #qwen
├── ComfyUI/models/photomaker  #ip only flux1
|        ├──ip_adapter.safetensors 
├── ComfyUI/models/clip_vision #ip only flux1
|        ├──clip-vit-large-patch14.safetensors

    4.Example

  • z-image
  • 9B Klein
  • 4B Klein
  • qwen
  • t2i flux1
  • i2i flux2
  • example flux1

    • 5.License and Commercial Use

    This work is patent pending. For commercial use or licensing inquiries, please contact the authors.

    6.Citation

    @misc{issachar2025dype,
      title={DyPE: Dynamic Position Extrapolation for Ultra High Resolution Diffusion}, 
      author={Noam Issachar and Guy Yariv and Sagie Benaim and Yossi Adi and Dani Lischinski and Raanan Fattal},
      year={2025},
      eprint={2510.20766},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2510.20766}, 
}