AnimateDiff is an innovative tool that utilizes diffusion models to breathe life into your generative art, resulting in captivating animated GIFs. By infusing motion into text or image inputs, AnimateDiff paves the way for a fresh approach to animation, free from the constraints of traditional methods. It represent advancement in the way animation may be approached in future animation workflows, offering fresh creative avenues for a new generation of animators.
In this guide, I'll walk you through how to install the AnimateDiff extension for Automatic1111. Before we get started, please make sure that you've installed all the prerequisites, such as Xformers, and that you have Automatic1111 up and running.
How to Install AnimateDiff in Automatic1111 WebUI
Launch Automatic1111 WebUI.
Go to the Extensions Tab.
Navigate to the Available Tab and click on "Load from."
In the empty space below, type "AnimateDiff" to bring up "sd-webui-animatediff script" in the dropdown.
Restart Automatic1111 or go to Settings and select "Reload UI."
AnimateDiff will now appear as a dropdown menu in both the txt2img & img2img tab, positioned below your ControlNet dropdown menu.
That's it; you've successfully installed the AnimateDiff extension.
Should you prefer an alternative method for installing Stable Diffusion using GitHub Desktop, a cleaner approach with more control over updates to prevent disruptions, a separate guide is provided below.
How to Install AnimateDiff
Reload and Restart Automatic1111
AnimateDiff is in txt2img & img2img
AnimateDiff User Interface
Once installed, this is how your AnimateDiff user interface will look like, but you'll notice that your Motion module is empty. That's because you will need to download and install the motion modules below.
Download and Install AnimateDiff Motion Models
The next important step is downloading the motion models, as AnimateDiff won't function without them.
There's an extensive selection available, and while you can opt to download specific models if you have a particular goal in mind, I recommend obtaining the complete set. This will enable us to explore the full spectrum of possibilities offered by these models in upcoming blogs.
Download everything that's Highlighted
To download the necessary models, follow these steps:
Visit the Hugging Face AnimateDiff Models page at https://huggingface.co/guoyww/animatediff/tree/main.
Once you've completed the model downloads, proceed to the next steps.
Navigate to your Stable Diffusion installation folder.
Find the "Extensions" directory within the installation folder.
Access the "sd-webui-animatediff" directory.
Inside the "sd-webui-animatediff" directory, you'll find a "model" folder. Open it.
Now, copy and paste all the motion models you downloaded into the "model" folder. models > Stable-diffusion
Now, copy and paste all the motion LoRa's you downloaded into this "LoRa" folder. models > Lora
After completing these steps, relaunch Automatic1111 (A1111) or restart the user interface (UI).
By following these instructions, you'll have successfully integrated the motion models into AnimateDiff, ready to be used for the next guide where we'll learn what everything means in the interface.