ComfyUI-Benripack

ComfyUI-Benripack
★ 12

角色管理节点扩展3D模型提示词管理
ComfyUI-Benripack 提供 CharacterPipe 节点，将图片、提示词与模型等素材集中为一体管理，简化角色类图像生成流程并支持本地 FBX 3D 模型加载。
💡 在角色形象生成流程中统一管理素材与模型配置。
🍴 4 Forks💻 JavaScript🔄 2024-09-07
📦
网盘下载
复制链接后前往夸克网盘下载
https://pan.quark.cn/s/a1f1f564f19c
📦 requirements.txt
torch
numpy
Pillow
Example
Example 1
Pixel art walking animation created of the input image
Pixel art jumping animation created of the input image
📄 README

ComfyUI Benripack Nodes

This package contains three custom nodes for ComfyUI:

  • Benri Character Pipe
  • Benri Load 3D Model
  • Benri Animation Extractor

    • 1. Benri Character Pipe

    This node manages character information and prompts for image generation.

    Parameters:

  • character_pipe (optional): An existing CHARACTER_PIPE object
  • face_image (optional): A face image (IMAGE)
  • pose_image (optional): A pose image (IMAGE)
  • add_positive_master_prompt (optional): Positive master prompt (CONDITIONING)
  • add_negative_master_prompt (optional): Negative master prompt (CONDITIONING)
  • add_positive_char_prompt (optional): Positive character prompt (CONDITIONING)
  • add_negative_char_prompt (optional): Negative character prompt (CONDITIONING)
  • model (optional): An AI model (MODEL)
  • clip (optional): A CLIP model (CLIP)
  • vae (optional): A VAE model (VAE)
  • ipadapter (optional): An IPAdapter model (IPADAPTER)

    • Functionality:

    The node combines all inputs into a CharacterPipe object that can be used for further processing in the image generation pipeline.

    Example

    2. Benri Load 3D Model

    This node loads a 3D model (.fbx) from the 3d_models directory in the root of your ComfyUI installation. Before using this node create the

    directory 3d_models. Than put in some .fbx files.

    Parameters:

  • model_name (required): Name of the 3D model to load

    • Functionality:

    The node loads the selected 3D model.

    3. Benri Animation Extractor

    This node extracts animation frames from a 3D model. If you select 4 rows and 4 columns the node will take a snapshot every 1/16 of the

    animation time aligned in a 4×4 grid. You can set the orientation of the model. You can also rotate the model. The created tile-image can be used to create consistent animations of the given object.

    Parameters:

  • cols (required): Number of columns in the output image (INT, default: 4)
  • rows (required): Number of rows in the output image (INT, default: 4)
  • orientation (optional): Orientation of the model (“Front”, “Back”, “Left”, “Right”, “Rotate”)
  • model (optional): A 3D model (MODEL_3D)

    • Functionality:

    The node generates an image with multiple frames of an animation based on the given 3D model and specified parameters.

    Example

    Here are some example GIFs demonstrating what you could do with the output of the nodes:

    Installation

  • Copy the node files into your ComfyUI custom_nodes directory.
  • Ensure you have a 3d_models folder in your ComfyUI main directory.
  • Place your 3D models (.fbx) in the 3d_models folder.

    • Usage

    After installation, you can use the nodes in your ComfyUI workflow. Connect the nodes according to your requirements and experiment with different settings to achieve desired results.

    Dependencies

    This package relies on the following Python libraries:

  • torch
  • numpy
  • PIL (Python Imaging Library)

    • Make sure these dependencies are installed in your ComfyUI environment.