Midjourney Update | June 2024: Personalization, V6.5, Website Updates, SREF and More!
TLDRMidjourney has launched model personalization, allowing users to customize their aesthetic preferences. Users can now rank images and use personalization codes to tailor visual outputs. The platform also supports combining multiple SREF codes for varied styles. Upcoming releases include version 6.5 with improved image quality and V7 promising smarter, faster, and better aesthetic models. Midjourney is also exploring demographic aesthetics and offers a yearbook showcasing community and development. Website updates enhance user experience with new features and a chatbot for assistance.
Takeaways
- 🚀 Midjourney has launched a model personalization feature, allowing users to create a customized model based on their aesthetic preferences.
- 🖌️ To use personalization, users must rank at least 200 images and can also use likes on the explore page to teach Midjourney their aesthetics.
- 📸 Users can now add a personalization code (-P) to their prompts to generate images tailored to their personal visual aesthetics.
- 🔧 The Midjourney team is actively working on refining the personalization feature and encourages feedback for improvement.
- 🎨 Multiple personalization codes can be used in a single prompt to combine different users' aesthetic preferences.
- 🔗 Style reference enthusiasts can combine multiple SREF codes in a prompt, with the ability to add relative weights to each code.
- 🆕 Version 6.5 of Midjourney is in training, promising improved image quality, coherence, better hands and skin textures, and possibly an updated image upscaler.
- 🔄 V7 is under development, aiming to be smarter, faster, cheaper, and with enhanced image quality and aesthetics.
- 🎥 Progress is being made on 3D and video capabilities, with no release date set yet to ensure quality.
- 📊 Midjourney has initiated a research project to explore the link between demographics and aesthetics, with a survey available for participation.
- 📚 A 'Year One' book showcasing community explorations and Midjourney's development is available for order.
- 🛠️ The Midjourney website has been updated for improved user experience, including interface changes and the introduction of a help page with a chatbot.
- 👥 Web image creation is being gradually opened to more users, and non-Discord users will soon be able to access the website without needing a Discord account.
Q & A
What is the new feature launched by Midjourney in June 2024?
-Midjourney officially launched its model personalization feature, which allows users to access a customized Midjourney model based on their individual aesthetic preferences.
How can users start using the model personalization feature?
-To start using the model personalization feature, users need to rank at least 200 images on Midjourney's website and can also like images on the explore page to teach Midjourney their preferred aesthetics.
What is the benefit of using the personalized model for users?
-The personalized model creates images tailored to the user's personal visual aesthetics, which most users prefer over the standard model.
Can users provide feedback on the personalized model if they are not satisfied with the results?
-Yes, if users think their personalized model's output is not satisfactory, they are encouraged to contact the Midjourney team so they can check the data behind the user's profile.
What is the new capability for style reference enthusiasts?
-Style reference enthusiasts can now combine multiple SREF codes in a single prompt and use double colon notation to add relative weights to each SREF code.
What is the current status of Midjourney's version 6.5?
-Version 6.5 is currently in training and is expected to be released in the next few weeks, with improvements in image quality, coherence, better hands and skin textures, and possibly an updated image upscaler.
What are the expected improvements in the upcoming V7 model?
-The V7 model is expected to be smarter, faster, cheaper, and have improved image quality and aesthetic enhancements.
Is there any progress on Midjourney's 3D and video capabilities?
-The team is making steady progress on 3D and video, but there is no release date yet as they are taking their time to ensure quality.
What new research project has Midjourney started?
-Midjourney has started a new research project exploring the connection between demographics and aesthetics, for which they have a survey that participants can fill out.
What changes have been made to the Midjourney website to improve user experience?
-The website has seen several interface changes, including merging pan and zoom features into a unified reframe tool, renaming VAR region to repaint, making error messages more descriptive, and adding a dedicated help page with a chatbot. Rooms are now also visible to anyone with an account.
How can users who are not part of Discord access Midjourney's website?
-Midjourney is slowly allowing non-Discord users into the website, which means that soon a Discord account will not be required to use Midjourney.
Outlines
🚀 Mid Journey Model Personalization Launch
The video introduces the launch of Mid Journey's model personalization feature, allowing users to create a customized model based on their aesthetic preferences. Users must rank at least 200 images on the website or like images on the explore page to teach Mid Journey their preferences. Once enough images are ranked, users can add a personalization code to their prompt to generate images tailored to their aesthetic. The video also mentions the possibility of combining multiple personalization codes and style reference (SRF) codes for varied visual outcomes. The team is working on refining the feature, and users are encouraged to provide feedback for improvements.
Mindmap
Keywords
💡Midjourney
💡Model Personalization
💡Aesthetic Preferences
💡SREF
💡Version 6.5
💡V7
💡3D and Video
💡Demographics
💡Website Updates
💡Discord
💡Patreon
Highlights
Midjourney has officially launched its model personalization feature.
Model personalization allows access to a customized Midjourney model based on individual aesthetic preferences.
To start personalization, users need to rank at least 200 images on Midjourney's website.
Users can also teach Midjourney their preferred aesthetics by liking images on the explore page.
Personalized model creation requires adding 'P' to the prompt after ranking enough images.
Most users prefer using their personalized model, but can switch between models depending on the project.
The Midjourney team is working on fine-tuning the personalization feature.
Users can use multiple personalization codes in a prompt to combine different aesthetics.
Combining multiple style reference (SREF) codes in a single prompt is now possible.
Relative weights can be added to each SREF code using double colon notation.
Version 6.5 of Midjourney is in training and expected to bring improvements in image quality and textures.
Version 7 is in development with promises of being smarter, faster, and having aesthetic improvements.
Midjourney is making progress on 3D and video capabilities, with no release date set for quality assurance.
A new research project is exploring the connection between demographics and aesthetics.
The Midjourney Year One book is available for order, showcasing community explorations and development.
Website updates include improved user experience and interface changes for new users.
Non-Discord users will soon be able to access Midjourney's website without needing a Discord account.
The creator is considering starting a Patreon for channel support and community building.