Node packVerifiedExternal

comfyui_hunyuanvideo_1.5_plugin

Sourcegithub.com/yuanyuan-spec/comfyui_hunyuanvideo_1.5_plugin

Maintainer@yuanyuan-specB · adoptiongrade

Package profile

Status: Verified
Publisher: @yuanyuan-spec
Source: GitHub
Stars: 28

README

中文文档

ComfyUI HunyuanVideo-1.5 Plugin

A ComfyUI plugin based on HunyuanVideo-1.5, offering both simplified and complete node sets for quick usage or deep workflow customization.

✨ Features

Simplified Nodes: Includes only the core HyVideo15ModelLoader and HyVideo15I2VSampler nodes, ideal for quick testing and result preview.
Complete Nodes: Provides more finely split nodes for detailed workflow adjustments and replacements.
Auto-Download Models: Built-in automatic model download; no need to manually prepare model files (manual download also supported).

📦 Installation

Step 1: Install Dependencies

Install required libraries from requirements.txt: pip install -r requirements.txt
Flash Attention: It's recommended to install Flash Attention for faster inference and reduced GPU memory consumption. Detailed installation instructions are available at Flash Attention.

Step 2: Download Models

Choose one of the following methods to download the model files (including hunyuanvideo-1.5 model, text_encoder, and vision_encoder):

Method 1: Auto-Download (Recommended)

Enable the Auto-Download option in the plugin when running a workflow. Models will be automatically downloaded to the default path.When using the model's auto-download feature, please set the path of the model loading node to "None." The corresponding model will be automatically downloaded to the default directory (if it already exists, it will not be downloaded again). The next time you run the workflow, you can see the auto-downloaded model in the node's path options.

Method 2: Manual Download

Manually download the model file and place it in the model directory specified by the plugin. For detailed instructions, please refer to checkpoints-download.md. (From the HunyuanVideo-1.5 open source project).

The directory structure for model placement is as follows:


models/
├── clip_vision
│   └── hyvideo15
│       └── siglip
│           ├── feature_extractor
│           │   └── preprocessor_config.json
│           ├── flux1-redux-dev.safetensors
│           ├── image_embedder
│           │   ├── config.json
│           │   └── diffusion_pytorch_model.safetensors
│           ├── image_encoder
│           │   ├── config.json
│           │   └── model.safetensors
│           ├── LICENSE.md
│           ├── model_index.json
│           ├── README.md
│           └── redux.png
├── diffusion_models
│   └──hyvideo15
│       ├── 1080p_sr_distilled
│       │   ├── config.json
│       │   └── diffusion_pytorch_model.safetensors
│       ├── 480p_i2v
│       │   ├── config.json
│       │   └── diffusion_pytorch_model.safetensors
│       ├── 480p_i2v_distilled
│       │   ├── config.json
│       │   └── diffusion_pytorch_model.safetensors
│       ├── 480p_t2v
│       │   ├── config.json
│       │   └── diffusion_pytorch_model.safetensors
│       ├── 480p_t2v_distilled
│       │   ├── config.json
│       │   └── diffusion_pytorch_model.safetensors
│       ├── 720p_i2v
│       │   ├── config.json
│       │   └── diffusion_pytorch_model.safetensors
│       ├── 720p_i2v_distilled
│       │   ├── config.json
│       │   └── diffusion_pytorch_model.safetensors
│       ├── 720p_i2v_distilled_sparse
│       │   ├── config.json
│       │   └── diffusion_pytorch_model.safetensors
│       ├── 720p_sr_distilled
│       │   ├── config.json
│       │   └── diffusion_pytorch_model.safetensors
│       ├── 720p_t2v
│       │   ├── config.json
│       │   └── diffusion_pytorch_model.safetensors
│       ├── 720p_t2v_distilled
│       │   ├── config.json
│       │   └── diffusion_pytorch_model.safetensors
│       └── 720p_t2v_distilled_sparse
│           ├── config.json
│           └── diffusion_pytorch_model.safetensors
│   
├── text_encoders
│   ├── byt5-small
│   │   ├── config.json
│   │   ├── generation_config.json
│   │   ├── pytorch_model.bin
│   │   ├── special_toke

Nodes in this pack

23 nodes