How to Use Midjourney's Style Tuners Without Spending GPU Credits

In this newsletter, read about:

• 🕵️‍♀️ Using and Combining Styles in Midjourney

• 🗞 News and Top Reads

• 📌 AI Art Tutorial: New Video Generation Tool

• 🎨 Featured Artist: Jonas Peterson

• 🖼 AI-Assisted Artwork of the Week

• 🤓 A Comprehensive Midjourney Guide

🕵️‍♀️ Using and Combining Style Tuners in Midjourney

A great thing about Midjourney’s Style Tuners is that you don’t necessarily need to spend your GPU time on creating your own Style Tuner. You may use – and even fine-tune – the Style Tuners created by other users without spending any additional GPU credits.

So, let’s explore how you can discover, edit, and combine styles created by you or other Midjourney users.

Finding style codes

The new Midjourney website enables you to effortlessly scroll through a collection of user-generated images, allowing you to identify and explore styles and aesthetics that align with your preferences. As you navigate through the image gallery, you will also have access to the specific parameters employed for each image, including the style code.

Fine-tuning styles created by other users

After you have discovered the style you want to experiment with, go to the URL: https://tuner.midjourney.com/code/STYLE_CODE.

For example, while browsing the image gallery, I liked the image with the style code “7W3ZrfSztNiOV2AHDRBGCJHxJxBsDGzrGAY00m9”. So, I used the following URL to get to its Style Tuner page:

https://tuner.midjourney.com/code/7W3ZrfSztNiOV2AHDRBGCJHxJxBsDGzrGAY00m9

On this page, you can see the prompt used to create this Style Tuner. So, let’s run this prompt with the style code to see the original look of the corresponding style.

Looks interesting, but what if we want to tune it a little bit? Maybe adding more colors? It’s totally possible! You can just go through the sample examples and change the preferences as you wish. At the end of the selection process, you’ll get a new style code, reflecting your choices. This fine-tuning doesn’t cost you any GPU credits.

Also, this doesn’t have any effect on the original Style Tuner page of the user who created these samples. Many users can experiment with the same Style Tuner page independently – all of them will just have different style codes at the page bottom, reflecting their selections.

In my case, I went through all 128 generated pairs and updated the selection based on my preferences. Here’s how this new “fine-tuned” style looks.

Much more colorful – great! Let’s continue the experiments.

Combining style codes

As the next step, I decided to combine the style I generated while writing my previous post with this new “woodblock” style.

To refresh your memory, here’s the style from the last week’s newsletter.

Now, I’m going to combine the two styles by using both codes and also triggering phrases from both styles.

We can definitely see how these portraits are lighter and also have a higher focus on eyes, compared to the original “woodblock” style. But what if we want to get even more of our “light” style"?

We can do this by simply adding the corresponding style code one more time. So, in the prompt below, I’ve included the “light”-style code twice and the “woodblock”-style code once. I also kept the “raw” style as I wanted the images to be created on top of a raw style rather than the Midjourney’s default one.

These portraits are in light colors, but keep the bright spots reflecting our fine-tuned “woodblock” style. I like how Midjourney can combine the styles that are so different in a very artistic and aesthetically pleasing way.

I hope this article inspired you for some experiments of your own. Happy prompting!

🗞 News and Top Reads

• OpenAI made DALL-E 3, its powerful image generation tool, accessible through an API.

  • Developers can now seamlessly integrate DALL-E 3's image generation capabilities into their own applications, opening up a plethora of creative possibilities.

  • Also, following the lead of Adobe, Google, and Shutterstock, OpenAI introduced Copyright Shield for its Enterprise customers. This initiative ensures legal protection for customers against copyright infringement claims arising from their use of OpenAI products.

• YouTube will take a more stringent approach to AI-generated content, particularly those that mimic the vocal styles of musicians.

  • Music labels will soon have the ability to request the removal of such content.

  • Additionally, creators will be mandated to label AI-generated content starting next year.

• Hollywood actors reach tentative deal to end their strike over the use of AI in filmmaking.

  • In the corresponding announcement, the union mentioned that the deal includes “unprecedented provisions for consent and compensation that will protect members from the threat of AI”.

  • The exact details haven’t been shared yet.

📌 AI Art Tutorial: New Video Generation Tool

In this video, Matt Wolfe delves into Plaiday, a novel text-to-video tool developed by Plai Labs. The tool's distinguishing feature lies in its ability to personalize any generated video by seamlessly replacing the face with your own or any other face of your choosing. Plaiday doesn’t generate realistic videos (yet?) – so no deepfake concerns – but it definitely offers a new fun way to create gifs and memes.

🎨 Featured Artist: Jonas Peterson

Jonas Peterson is a visual artist, writer, and photographer. Combining fashion, portraiture and highlighting old people as the heroes they are, he used AI to create stunningly striking images. His series “Bon Voyage” was shown to the public in Milan in April and his latest series Guardians has been viewed over 7 million times online. Check out his unique artwork on Instagram @jonaspeterson_ai.

🖼 AI-Assisted Artwork of the Week

🤓 A Comprehensive Midjourney Guide

Share Kiki and Mozart

If you enjoy this newsletter and know someone who might also appreciate it, please feel free to share it with them. Let's spread the word about AI art and introduce more people to this fascinating field!