Nodes Browser
ComfyDeploy: How Clothing Migration Kit works in ComfyUI?
What is Clothing Migration Kit?
This is an experimental project focused on Stable Diffusion (SD) models. In a single generated image, the same object or character consistently maintains a very high level of consistency. I had already attempted to address this issue in the SDXL model.
How to install it in ComfyDeploy?
Head over to the machine page
- Click on the "Create a new machine" button
- Select the
Edit
build steps - Add a new step -> Custom Node
- Search for
Clothing Migration Kit
and select it - Close the build step dialig and then click on the "Save" button to rebuild the machine
Migration Kit Overview
Contact Me
| Platform | Details | |---------------|-------------------------------------------------------------------------| | QQ Group | 571587838 | | Bilibili | Homepage | | Civitai | ttplanet | | WeChat | tangtuanzhuzhu |
Project Introduction
This experimental project focuses on leveraging Stable Diffusion (SD) models for high-consistency object and character rendering. The methodology integrates advanced workflows such as ControlNet, DIT model, and latent-guided processes for superior control and consistency.
Modules
1. Clothing Migration Kit
Model Download: Clothing Migration LoRA
workflow Download: Cloth workflow
| Feature | Description | |---------------------------------|-----------------------------------------------------------------------------| | Consistent Clothing Migration | Transfer clothing styles across reference images with high accuracy. | | Cartoon to Realism Conversion | Seamlessly convert cartoon clothing to realistic styles and vice versa. | | Creative Design Control | Adjust clothing similarity via weights to inspire creativity. |
Image Examples:
| | |
|------------------------------------------------------------------------------------------------|-----------------------------------------------------------------------------------------------|
2. 3D Figure Kit
Model Download: 3D Figure LoRA
workflow Download: Figure workflow
| Feature | Description | |----------------------------------|---------------------------------------------------------------------------------| | 3D Conversion | Transform 2D character designs into 3D printable figurines. | | Pose Customization | Adjust poses, expressions, and other dynamic features to suit your design needs.| | Material and Texture Enhancements | Ensure high fidelity in textures and details for professional-grade outputs. |
Image Examples:
| | |
|------------------------------------------------------------------------------------------------|-----------------------------------------------------------------------------------------------|
Future Plans
| Task | Details | |--------------------------------|-------------------------------------------------------| | Transfer Distinct Features | Improve the migration of objects with unique attributes.| | Complex Pattern Handling | Develop models to manage intricate designs. | | Face Transfer | Implement face style migrations. | | 3D Enhancements | Introduce automated rigging and advanced pose control for 3D figures. | | Community Ideas | Incorporate suggestions and use cases shared by users. |
How to Use
| Step | Details | |----------------|----------------------------------------------------------------------------------------------| | 1. Install | Install ComfyUI and required custom nodes. | | 2. Download | Obtain the models from the Hugging Face Project. | | 3. Load | Load the selected model into ComfyUI. | | 4. Workflow | Use the provided workflow examples for your application. |
Dependencies
| Node | Link | |-----------------|------------------------------------------------------------------------------------------| | TTP Toolset | ComfyUI_TTP_Toolset | | Tag Node | ComfyUI_JC2 | | Alimama Flux| Alimama_flux_inpaint |
Note: Ensure sufficient VRAM for high-complexity workflows. Use optimization tools like FluxExt-MZ if needed.
Contribute
Contributions and suggestions are welcome! Open an issue or submit a pull request with your ideas.