Are you tired of using the same old profile picture or avatar on your social media accounts?
Do you want to stand out and have a unique, personal touch to your online presence?
Look no further, because in this tutorial, we will show you how to create your own AI-generated avatars for free, without any coding.
With just a few simple steps, you can have personalized and one-of-a-kind avatar that represents you online. Let’s get started!
How to create your AI-generated avatars for free?
Time needed: 1 hour and 15 minutes.
We are creating AI-generated avatars using the following 3 tools.
1. Stable Diffusion is an AI model that can help you make pictures using words.
2. Dreambooth is a technique that helps you give Stable Diffusion some extra information so it can make even better pictures.
Together, Stable Diffusion and Dreambooth can help you create unique and personalized images from text and photos. It’s a fun way to use technology to express your creativity and imagination!
3. Google Colab. Google Colab is a tool that is kind of like Google Docs, but instead of writing words, you can write and run Python code right in your web browser.
Don’t worry, you won’t have to code; you need to just run some code.
That’s it. Let’s get to it
- Make your photos ready (~ 10 min)
Select 20 photos of yourself.
The better you follow these guidelines, the better your avatars:
✅ No other people in your photos
✅ Photos with different expression
✅ Photos in different locations, backgrounds, and angles
✅ Photos at different times of day
✅ Look into the camera and also away
Then, go to this app called Birme and set 512×512 dimensions for every photo. Save as zip, and download it locally.
Finally, change the names of the photos with unique keyword and add numbers to them, in an orderly manner. For example: john(1).jpg, john(2).jpg… up to john(20).jpg
Note: This unique keyword will be your instance name when you generate a prompt to the AI. In this case, the prompts will be: “a portrait of john …”. So, make sure it’s unique!.
We will use the photos in Step 6.
- Open Google Colab document (~ 1 min)
Open Colab document .
You will be generating your AI avatars using this Colab document.
- Connect your Google Drive with the Colab document (~ 1 min)
Click on the Play icon here.
Click on “Run Anyway”
An authentication message will be displayed. Click on “Connect to Google Drive“.
This will take just 1 minute to run. You can see a message like “Mounted at /content/gdrive“.
- Set up your environment ( ~ 1 min )
Just click on the play button to run this code to set up your environment.
Once it is completed you can see the message “DONE !”
- Download Stable Diffusion (~ 5 mins)
Follow the below steps to download Stable Diffusion.
Signup to Hugging Face.
Click on the Profile image on the top right.
Click on Settings.
Click on Access Tokens.
Click on New Token. Role should be Read.
Copy the token.
Paste your token in your Colab notebook under “Huggingface token”.
Run the model by clicking on the Play icon.
- Set up DreamBooth (~ 5 mins)
You need to enter a session name. You can enter any name you want. Since we have named the image as John, we can keep that name here as well.
Then run the cell without changing anything.
Scroll down. Now you will get a request to upload your images.
Upload the 20 images.
- Train your own model (~ 45 mins)
Set the training steps to 1600.
If you set training steps to 1600 and the number of photos as 20. It will take around 45 minutes to finish the training process.
- Test your model
First, run the code in the cell by clicking the Play icon. This will start the testing process.
When it’s done, you’ll see a link next to “Running on local URL“.
Click on this link to open a new window where you can try out your model.
This notebook will continue to run while you test your prompts, because it provides connection with the web UI.
- Generate photos
In this new window, you can type in a prompt and the model will try to generate an image based on your prompt.
You can check Open Art for prompt inspiration.
If you didn’t get any prompts, try the following prompts.
portrait of john, gta san andreas cover art
cybernetic john photorealistic portrait
john in a space suit, 8k, white background, octane render, 4k, ultra realistic
Since our picture name is john in Step 1, we gave that prompt.
Click on Generate and you can see your ai avatar.
Creating your own AI-generated avatars is a fun and easy way to personalize your online presence and add a unique touch to your social media accounts.
By following the steps outlined in this tutorial, you can create personalized avatars without any coding experience, using the free and user-friendly Stable Diffusion and Dreambooth tools.
Give it a try and see what you can create with AI and your imagination!
Jasper Reed is the founder of heyhowtodoit.com with 8+ years of experience in Search Engine Optimization.He has researched, tested, and written hundreds of articles ranging from social media platforms to messaging apps.
With a passion for technology and a natural aptitude for understanding consumer behavior, Jasper has built a successful career in blogging.He has been quoted and referenced by major publications and media companies.
An avid traveler and tech enthusiast, Jasper loves to explore new places and experience different cultures. With his thick-rimmed glasses and geeky demeanor, Jasper is a true tech nerd at heart.
You can read more about the team here.