ComfyUI_JoyAI_Image

JoyAI-Image:Awakening Spatial Intelligence in Unified Multimodal Understanding and Generation

Update

• support te use gguf, use less memory(all need 40G RAM ) / te支持gguf，内存占用更少(共占用40G左右，压缩了24G)，模型在hg或者夸克云
• Support gguf now, use less memory / 支持gguf，内存占用更少，模型在hg或者夸克云

1.Installation

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

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

2.requirements

pip install -r requirements.txt

3.checkpoints

• transformers/gguf/vae/clip links
• or 夸克云
• or hg

├── ComfyUI/models/
|     ├── diffusion_models/
|        ├──joy_image_transformer.safetensors # optional
|     ├── gguf/
|        ├──joy_image_transformer-Q8_0.gguf # optional
|        ├──JoyAI-Image-Und-merger-Q6_K.gguf # optional te
|     ├── vae/
|        ├──Wan2.1_VAE.pth
|     ├── clips
|        ├──JoyAI-Image-Und-merger_bf16.safetensors # te optional

4.Example

• dit gguf
• TE gguf

5.Citation

@jd-opensource