If you want to run WAN 2.1 on Paperspace Gradient using ComfyUI, this guide will help you set it up step by step. Paperspace Gradient provides an easy-to-use cloud environment for running AI models, while ComfyUI is a powerful tool for working with Stable Diffusion models. Let’s get started!
Why Use /tmp Storage?
We will use the /tmp storage because, for now, it has no storage limit. However, keep in mind that every time you start the environment, it will take longer to load. The advantage is that we can store large WAN 2.1 14B models without any restrictions.
Prerequisites
Before starting, ensure you have the following:
A Paperspace account (sign up at Paperspace if you don’t have one).
Access to Gradient Notebooks (a free tier is available, but GPU usage may be limited).
Step 1: Create a Paperspace Gradient Notebook
Log in to your Paperspace account.
Navigate to Gradient Notebooks and click Create Notebook.
Choose a GPU instance (e.g., A4000 or any available GPU with at least 16GB VRAM).
Select the template: Start from Scratch.
Click Start Notebook to launch the instance.
Step 2: Install ComfyUI
Download the Jupyter Notebook file: ConfyUI.ipynb.
Upload ConfyUI.ipynb to your Paperspace Notebook.
Open ConfyUI.ipynb and run each code block one by one.
In Step 4 of the notebook, you can add your own model download URLs if needed.
Step 3: Open the ComfyUI Web Portal
Once all the notebook cells have been executed, open a terminal from the left-side menu.
Run the following command to create a public URL for accessing ComfyUI in your browser:
ssh -R 80:localhost:8188 serveo.netCopy and paste the generated URL into your browser to access the ComfyUI interface.
That’s it! You now have ComfyUI running on Paperspace Gradient with WAN 2.1. If you have any issues, feel free to ask in the community forums or check the official documentation. Happy experimenting!