ComfyUI - RunPod: Installation and use tutorial Cheap rental GPU ComfyUI Manager Stable Diffusion

Introduction

This tutorial will guide you through the process of setting up a RunPod GPU rental machine with permanent disk storage and installing ComfyUI, including its Manager. You'll also learn how to start using ComfyUI by installing custom nodes and models. This setup is ideal for those looking to leverage powerful GPU capabilities at a reduced cost.

Step 1: Setting Up Your RunPod GPU Rental Machine

1. Visit RunPod: Go to the RunPod website using this link.
2. Create an Account: Sign up for a new account if you don't have one already.

Select a Template

• For PyTorch 2.4, use this template.
• For PyTorch 2.2, use this template.
3. Choose Your GPU: Select a GPU option that suits your needs based on performance and pricing.
4. Configure Storage: Ensure you set up permanent disk storage to retain your data across sessions.
5. Launch the Instance: Start your GPU instance and wait for it to initialize.

Step 2: Installing ComfyUI

1. Access the Instance: Once the GPU instance is running, connect to it via SSH or the provided interface.

Download ComfyUI

• Run the following command to clone the ComfyUI repository:

  git clone <ComfyUI repository URL>
  

2. Navigate to the Directory: Change into the ComfyUI directory:

   cd ComfyUI
   

3. Install Dependencies: Ensure you have all necessary dependencies:

   pip install -r requirements.txt
   

4. Run ComfyUI: Start the ComfyUI application:

   python app.py
   

Step 3: Configuring ComfyUI Manager

1. Access the Manager: Open your web browser and go to the ComfyUI Manager URL, typically http://localhost:port (replace port with the designated port number).
2. Customize Settings: Adjust settings as needed for your workflow.

Install Custom Nodes

• Navigate to the nodes section.
• Search for and install any additional nodes you wish to use.
3. Load Models: You can upload or link to models that you want to use in ComfyUI.

Step 4: Starting with ComfyUI

1. Explore Interface: Familiarize yourself with the layout and tools available in ComfyUI.
2. Test Functionality: Create a simple project to test the functionality of installed nodes and models.
3. Save Your Work: Regularly save your projects to the permanent storage to avoid data loss.

Conclusion

You have successfully set up a RunPod GPU rental machine and installed ComfyUI along with its Manager. This setup allows you to utilize powerful GPU resources for your projects efficiently. As next steps, consider exploring more advanced features of ComfyUI, experimenting with custom nodes, and applying your new skills to real-world projects. Enjoy your journey with ComfyUI!