This repository contains a Docker configuration for running a ComfyUI workflow with the flux-schnell model and RealESRGAN_x2 upscaler in a serverless environment.
- Base Image: RunPod ComfyUI worker (5.1.0-base)
- Custom Nodes: KJNodes
- Models:
- flux1-schnell-fp8.safetensors (FLUX1 checkpoint)
- RealESRGAN_x2.pth (upscaler)
This Docker image is designed to run the "flux-schnell-RealESRGAN_x2" ComfyUI workflow. The workflow leverages the FLUX1 model for image generation and the RealESRGAN_x2 upscaler for enhancing the output quality.
To build and run the Docker image locally:
# Build the Docker image
docker build -t comfyui-flux-schnell .
# Run the container
docker run -p 8188:8188 comfyui-flux-schnell
- Push this repository to your GitHub account
- Create a new RunPod template using this repository URL
- Deploy a new pod using the template
- Access the ComfyUI interface through the exposed port (typically 8188)
The Dockerfile:
- Uses the RunPod ComfyUI worker as a base image
- Installs KJNodes custom nodes
- Downloads the flux1-schnell-fp8 model
- Downloads the RealESRGAN_x2 upscaler model
To modify this setup:
- Update the Dockerfile to include additional models or custom nodes
- Build and test locally
- Push changes to this repository
- Redeploy your RunPod instance
Please respect the licenses of all included models and software.
- FLUX1 model: License Information
- RealESRGAN: License Information