Run ComfyUI workflows across multiple local GPUs or networked machines, allowing for enhanced processing capabilities and efficiency. This tool is particularly useful for users with access to multiple GPUs, enabling them to distribute workloads effectively.
- Supports both simple dual-GPU setups and more complex multi-machine configurations.
- Facilitates the transfer of images and latent data between different ComfyUI instances seamlessly.
- Offers advanced workflow management features, including batch size overrides and the ability to load and save workflows.
Context
This tool, known as ComfyUI_NetDist, is designed to enhance the capabilities of ComfyUI by allowing users to run workflows on multiple GPUs or across different machines. Its primary purpose is to facilitate distributed computing, making it easier for users to leverage additional hardware resources for AI art generation tasks.
Key Features & Benefits
The tool provides a straightforward setup for both dual-GPU and multi-machine configurations, enabling users to efficiently manage their resources. Users can specify local or remote batch sizes and control when nodes trigger, improving workflow flexibility. The ability to load and save workflows adds significant utility, allowing users to customize and streamline their processes.
Advanced Functionalities
For users with more complex setups, the tool offers advanced functionalities such as per-batch overrides and the option to use any workflow JSON as input. This flexibility allows for greater customization in how workflows are executed, accommodating various user needs and preferences. Additionally, the tool includes nodes for saving and loading latents, facilitating seamless data transfer between instances.
Practical Benefits
By using ComfyUI_NetDist, users can significantly improve their workflow efficiency and control over image generation processes. The ability to distribute workloads across multiple GPUs not only speeds up processing times but also enhances the quality of output by allowing for more complex and varied workflows. This tool ultimately empowers users to make the most of their hardware resources in AI art generation.
Credits/Acknowledgments
The tool is developed by contributors to the ComfyUI project, with specific acknowledgments to the original author and community members for their ongoing support and enhancements. The repository is open-source, allowing for collaborative improvements and contributions from users.