BIG Update - Midjourney Will Never Be the Same Again

Future Tech Pilot
12 Jun 202412:16

TLDRMidjourney introduces a groundbreaking personalization feature that tailors image generation to individual user preferences. By using a unique code, users can now influence the style of generated images, enhancing the creative process. The video demonstrates how to activate and adjust this feature, emphasizing its potential to transform short and long prompts, and even style raw images. It also discusses the importance of experimenting with aspect ratios and chaos values for diverse outcomes. The personalization is optional and part of an evolving journey that encourages users to explore and have fun with the tool.

Takeaways

  • 😀 Midjourney has introduced a new personalization feature that customizes image generation based on user preferences.
  • 🔄 To use personalization, add '-dP' at the end of your prompt, or enable it in the settings to apply the personalization code to all prompts.
  • 💡 Personalization is turned off by default but can be easily activated to influence the style of generated images.
  • 🛠️ The personalization code is developed by the system based on around 200 images that the user likes or ranks on the explore page.
  • 🔍 Users can test their personalization code by using a simple prompt and adding the code within curly brackets, e.g., '{dP}'.
  • 🎨 The strength of personalization can be adjusted with the '-dS' parameter, which has a range from 0 to 1000, with 100 being the default.
  • 🌟 High stylized values (above 200) may not significantly change the outcome, suggesting diminishing returns on increasing the strength.
  • 🔄 Experimenting with the 'chaos' value can also yield interesting results when personalization is enabled.
  • 🖼️ Aspect ratio can influence the perception of personal preferences in image generation, with wider ratios potentially altering the style.
  • 📝 Longer prompts can also benefit from personalization, leading to significantly different image outcomes compared to non-personalized versions.
  • 🐍 Personal codes do not currently work with 'nii journey' and are specific to the mid Journey platform.

Q & A

  • What is the new feature introduced by Midjourney?

    -Midjourney has introduced a personalization feature that allows the bot to take into account the user's personal preferences when creating a generation.

  • How does the personalization feature work in Midjourney?

    -The personalization feature works by adding a personalization code (e.g., '-d-p') at the end of the prompt, which enables the bot to generate images based on the user's preferences.

  • What is the purpose of the personalization code in Midjourney?

    -The personalization code helps the Midjourney bot to generate images that align with the user's personal style and preferences, making the generated content more tailored to the individual.

  • How can users enable the personalization feature in Midjourney?

    -Users can enable the personalization feature by adding '-d-p' at the end of their prompts or by turning it on in their settings.

  • What is the default setting for the personalization feature in Midjourney?

    -The personalization feature is turned off by default, but users can easily turn it on through their settings.

  • How does Midjourney learn the user's personal style?

    -Midjourney learns the user's personal style based on the images the user likes on the explore page and the images they choose in the rank pairs. The Midjourney team suggests that around 200 images need to be chosen for the bot to develop the user's personal preference.

  • What is the process of developing a personalization code in Midjourney?

    -To develop a personalization code, users need to go to the Midjourney website and rank images they like or dislike. The bot learns from these selections and develops a personal preference code.

  • How can users test their personalization code in Midjourney?

    -Users can test their personalization code by writing a prompt, adding a curly bracket with the '-d-p' parameter, and then inputting their personalization code. The bot will generate images based on this code.

  • What is the range of the stylized value in Midjourney's personalization feature?

    -The stylized value in Midjourney's personalization feature ranges from 0 to 1000, with the default being 100. Users can adjust this value to control the strength of the personalization.

  • How does the aspect ratio affect the personalization feature in Midjourney?

    -The aspect ratio can significantly change the idea of the user's preference in the generated images. Different aspect ratios might result in different styles and vibes in the generated content.

Outlines

00:00

🌟 Introduction to Mid Journey's Personalization Feature

Mid Journey has introduced a new personalization feature that tailors image generation based on user preferences. The bot learns from the images users like and dislike, and this preference is encoded into a personalization code. Users can activate this feature by adding '-dP' to their prompts, which will generate images that align with their style. The video demonstrates how to use this feature, including how to enable it in settings or manually adding the '-dP' parameter to prompts. The speaker also shares their personal code, 'wml1 AI', and shows the difference in generated images with and without personalization.

05:02

🔧 Developing and Testing Your Personalization Code

The speaker guides viewers on how to develop their personalization code by visiting the Mid Journey website and ranking pairs of images. The bot learns the user's preferences through this process, which can take around 200 image selections. Once the code is developed, viewers can test it by adding '-dP' to their prompts and observing the changes in the generated images. The speaker also explains how to adjust the strength of personalization using the '-dS' parameter, which ranges from 0 to 1000, and demonstrates the effects of different stylized values on image generation.

10:04

🎨 Exploring Advanced Features and Limitations

The video continues by discussing how the aspect ratio can influence the personalization effect, showing that wider ratios might alter the generated images significantly. The speaker also explores the impact of the personalization code on longer prompts, revealing that even detailed descriptions can be influenced by the user's style. Additionally, the speaker demonstrates how the personalization code can be used with the 'style raw' feature to achieve different visual effects. The video concludes with a reminder that personal codes do not work with 'nii journey' and encourages viewers to experiment with the chaos value to see how it interacts with personalization.

Mindmap

Keywords

💡Personalization

Personalization in the context of the video refers to a new feature introduced by Midjourney, which allows the AI to tailor its image generation based on individual user preferences. This is achieved by analyzing the images a user likes and dislikes through a ranking system. The video demonstrates how adding a personalization code to a prompt can significantly alter the generated output, making it more in line with the user's aesthetic tastes.

💡Midjourney

Midjourney is the AI-driven image generation platform discussed in the video. It uses machine learning to create images based on textual prompts provided by users. The platform is highlighted for its recent update that includes a personalization feature, which enhances user experience by generating images that reflect individual preferences.

💡Prompt

A prompt in the video script is a textual description provided by the user to guide the AI in generating an image. The script explains how to use prompts effectively, especially in conjunction with the new personalization feature, by adding specific codes to the end of the prompt to influence the style and content of the generated images.

💡Personalization Code

The personalization code is a unique identifier that represents a user's preferences in the Midjourney platform. It is used in the video to demonstrate how users can customize their image generation by including this code in their prompts. The code is developed through a process of ranking images that the user likes or dislikes, allowing the AI to learn and adapt to the user's style.

💡Settings

Settings in the video refer to the user interface within the Midjourney platform where users can adjust their preferences, including turning on the personalization feature. The script mentions navigating to the settings to enable personalization, which automatically appends the user's personalization code to all prompts.

💡Rank Pairs

Rank pairs is a process described in the video where users choose between two images presented by the Midjourney platform to indicate their preferences. This selection helps the AI learn the user's style over time. The video emphasizes the importance of this process in developing a personalization code that accurately reflects the user's aesthetic choices.

💡Stylized Value

The stylized value is a parameter that users can adjust in their prompts to control the strength of the personalization in their image generation. The video script explains that this value ranges from 0 to 1000, with higher values potentially leading to more pronounced personalization effects in the generated images.

💡Chaos Value

Chaos value is another parameter mentioned in the video that can be used in conjunction with the personalization code. It affects the randomness and diversity in the image generation process. The script shows examples of how varying the chaos value can lead to different visual outcomes, even when using the same personalization code.

💡Aspect Ratio

Aspect ratio in the video is discussed in relation to how it can influence the style and composition of the generated images. The script suggests that different aspect ratios can significantly change the look and feel of the images, even when using the same personalization code and prompt.

💡Style Raw

Style raw is a feature mentioned in the video that allows users to tap into their personal preferences even when using longer, more descriptive prompts. The script demonstrates how including the personalization code in a prompt with a style raw parameter can result in images that are more aligned with the user's aesthetic preferences.

Highlights

Midjourney introduces a new personalization feature that adapts to user preferences.

Personalization is applied to image generation based on user prompts.

Example provided showing the difference in image generation with and without personalization.

Personalization code 'd-p' can be added to prompts for customized results.

Personalization can be enabled or disabled through settings.

Developing a personalization code requires ranking images on the explore page.

The model learns from around 200 images to develop personal preferences.

A live example demonstrates the effect of personalization on image prompts.

Personalization can enhance short prompts by filling in description gaps.

Testing the personalization code involves using specific parameters in prompts.

The strength of personalization can be adjusted with the 'd-s' stylized value.

Experimentation with the 'chaos' value can yield unexpected results.

Aspect ratio can significantly influence the outcome of personalization.

Personalization also impacts longer, more complex image prompts.

Personal codes can be used with 'style raw' to tap into user preferences.

Personalization codes are not currently compatible with 'nii journey'.

The feature is off by default, offering users the choice to opt-in.

The Midjourney team encourages users to have fun and not get too attached to a specific style.