How to Do Stable Diffusion Face Swap Online

Stable Diffusion is so effective at creating art, and bringing ideas to life that it caused a massive wildfire on Twitter. If you are new to Stable Diffusion, the project is an AI powered generative model which creates realistic images using text prompts as a creative input. Users of the system can define what they want in an image and Stable Diffusion will use existing elements to recreate the prompt. However, the faces used within the generated images will be randomized, and this issue will be tackled in this article. This article is a guide on how you can carry out a Stable Diffusion face swap through the Automatic1111-webui. A bonus alternative to Stable Diffusion is added at the end if you don’t wish to go through all the trouble.

On This Page

• How to Face Swap in Stable Diffusion Online: Step-by-Step
• Stable Diffusion Face Swap Oline Best Alternative - Face Swap in Minutes [100% Free]

How to Face Swap in Stable Diffusion Online: Step-by-Step

This section will guide you through the process of carrying out a face swap using Stable Diffusion. For this guide, we’ll be using the Automatic1111 webui and an extension for the webui called Roop. To carry out this Stable Diffusion face swap tutorial, you will need to have Stable Diffusion pre-installed on your device. Instructions on how you can install Stable Diffusion can be found on the Internet.

Method 1 Stable Diffusion Face Swap with Roop Extension

Step 1: Installing Community Visual Studio

To begin installing Roop, start by installing the Community Visual Studio. Whilst this is being installed, you must include the Python and C++ packages when prompted.

Once the Community Visual Studio has been installed, open your Command Prompt application. Navigate to your Stable Diffusion Web UI’s directory and run the code mentioned below.

pip install insightface==0.7.3

Step 2: Installing Roop

Head on over to the extensions tab in the Web UI and paste the link provided into the “Install from URL” tab’s text box. 

Once the installation is complete, close the Web UI.

Step 3: Download the Inswapper_128 Model

Download the Inswapper_128.onnx and replace the file inside this directory <webui_dir>/models/roop/ directory

Step 4: Start the Face Swap

Open Stable Diffusion and locate the Roop extension under the txt2img tab. Start the Stable Diffusion face swap process by enabling Roop. Check the box provided to enable the extension you just installed. In the restore face options, Click on CodeFormer.

Now it’s time for you to add in the negative and positive prompts to generate the image using Stable Diffusion.

Step 5: Add the portrait image

Add a portrait image to the Roop’s expansion panel. This is the image that will replace the face of the newly generated image.

Step 6: Generate 

Hit generate and wait patiently till the rendering process is completed. Please note that the developers have stopped supporting the creation of NSFW work, therefore if the rendering process creates such a result, it’ll be automatically discarded by the program.

Method 2 Stabel Diffusion by Inputting Celebrity Name

Let’s say that the image you are generating needs to have celebrity deepfakes. In such scenarios, users can utilize the base version of Stable Diffusion as facial elements of these celebrities already exist on the databases used by Stable Diffusion.

Step 1: Open Stable Diffusion

For this face swap Stable Diffusion method, open Stable Diffusion by running the .bat file.

Step2: Prompt

Write a prompt for image generation that includes the name of the celebrity.

Step 3: Generate

Click on generate and Stable Diffusion will create an image using your prompt. This image will have the face of the celebrity that you specified. However, for this to function, elements and images of the celebrity should already exist and be easily available.

Also read: How to Face Swap in Photoshop | Beginner-Friendly Guide>>

Vidnoz AI - Create Free Engaging AI Video with Talking Avatar

• Easily create professional AI videos with realistic avatars.
• Text-to-speech lip sync voices of different languages.
• 800+ video templates for multiple scenarios.

Try It Free Now Sign in with

Stable Diffusion Face Swap Oline Best Alternative - Face Swap in Minutes [100% Free]

Running a face swap Stable Diffusion is hard work as it places enormous amounts of stress on both you and the machine you are using. The rendering process utilizes extreme amounts of CPU power and will make your computer useless till the generation is deemed complete. The time taken to render the image will depend on how powerful your computer is, which forces you with another expensive requirement of having a powerful PC to carry out this method.

This section focuses on an online tool named Vidnoz Face Swap that acts as an alternative to the face swap Stable Diffusion. Vidnoz requires no complex setups, prompts, or massive amounts of computational power. All the user needs is an account on the platform and a browser to access the website!

Vidnoz AI uses an AI engine to identify images uploaded to their platform. Once you upload the two images, the AI will identify different facial features including the facial width and height, structure, and even the location of key identifiers such as the nose and eyebrows. Once these are identified, The features of the target image will be pushed into the main image as they replace these features and identifiers. This process creates a cleaner face swap as it is not a direct cut and paste, but more of an adaptation.

Listed below is a tutorial on how you can use the Vidnoz platform to carry out face swaps that are comparable to the face swap Stable Diffusion. Since the process is extremely fast, Vidnoz AI allows you to carry out face swaps in bulk as batch face swap stable diffusion does not exist.

Step 1 Upload the image

Create an account on the Vidnoz platform, it’s free and easy. Users can even use their Google accounts to register an account. Once you enter the platform, locate the face swap tool.

To start the procedure, upload the base image into the tool. This is the image whose face will be replaced by the target image. When uploading the image, please ensure that it follows the requirements provided.

Step 2: Target Image.

Now, upload the target image onto the tool. This is the image whose features will be lifted and pushed into the base image for a face swap. For the best results, make sure that the image is a clear portrait of the individual.

Step 3: Generate

Hit Generate to start the face swap. Once it’s completed the user can download the image.

Conclusion

The ability to carry out accurate face swaps using Stable Diffusion face swap or Vidnoz AI can be quite scary. Therefore, always be responsible with the images you generate. If you are someone who likes being tech savvy, Stable Diffusion face swap is the way to go. For everyone else, Vidnoz AI is the perfect alternative.