NEW Midjourney UPDATE - Niji V6 Anime Generator
TLDRMidjourney's new update, Niji V6, is an anime-focused algorithm that enhances image generation. Users can access it via the Midjourney bot or the Niji server, with the ability to add simple text to images. Despite being in alpha, the update impresses with its subtle upscaler and creative capabilities. The video showcases examples of generated images, highlighting the AI's advanced capabilities in character and text integration, with tips on optimizing results using variation and stylize values.
Takeaways
- 🌟 Midjourney has released an alpha update introducing Niji V6, an anime-focused image generation algorithm.
- 🔍 Users can access Niji V6 through the Midjourney bot by typing 'D-nii nji 6' at the end of their prompt.
- 🎨 The update showcases impressive visual results, such as Batman in a yellow bat suit eating pizza in the Batcave.
- 🚧 The current version is in Alpha, lacking features like inpainting, panning, and zooming out, but these are expected in the full launch.
- 🔑 There are two ways to access Niji V6: through the Midjourney bot or via the Niji server, with an invite link provided in the description.
- 📝 Users can add simple text to images, but it's recommended to keep it brief and not too complex.
- 🔄 A 'variation strong' feature allows for fine-tuning of image results by generating text and selecting the closest match for further enhancement.
- 🤖 The model is trained on anime, so prompts related to anime will yield the best results, but style adjustments can be made for different aesthetics.
- 🎭 The 'style raw' parameter can be included in prompts to reduce the cartoonish anime aesthetic if desired.
- 🔢 The 'stylize' value doesn't significantly impact the results in Niji V6, except at the extremes of 1000 and 0.
- 🔄 Keeping the 'chaos' value low (between 0 and 6) is recommended for better image coherence, with higher values leading to less recognizable images.
Q & A
What was the significant update to Mid Journey last night?
-The significant update was the alpha launch of Niji V6, an anime-focused algorithm.
How can users access the Niji V6 version through the Mid Journey bot?
-Users can access Niji V6 by typing 'D-nii nji 6' at the end of their prompt in the Mid Journey bot.
What is the visual example given in the script to demonstrate the capabilities of Niji V6?
-The visual example given is Batman wearing a yellow bat suit and eating pizza in the Batcave.
What features are currently not available in the alpha version of Niji V6 but are expected in the full launch?
-Inpainting, panning, and zooming out are features not available in the alpha version but are expected in the full launch.
How can users access the Niji bot through the Niji server?
-Users can access the Niji bot through the Niji server by joining an image generation channel and reacting to their created image with an envelope emoji.
What is the recommended way to add text to images generated by Niji V6?
-Users should add simple text, a word or two, or maybe three, within quotation marks to the image prompts.
What is the 'variation strong' feature used for in the image generation process?
-The 'variation strong' feature is used to refine the image generation when the initial result is close but not perfect, allowing for better results.
What is the main focus of the Niji V6 generator, and how is this reflected in the generated images?
-Niji V6 is an anime-focused generator, and this is reflected in the style and aesthetic of the generated images, which have a strong anime influence.
How can the cartoonish anime aesthetic be toned back in the generated images?
-The cartoonish anime aesthetic can be toned back by including 'style raw' in the prompt.
What is the role of the 'stylize' value in the image generation process, and how does it affect the results?
-The 'stylize' value affects the style intensity of the generated images. However, in Niji V6, values other than the extremes (1000 and 0) do not make significant differences.
What advice is given regarding the use of 'chaos' values in image generation?
-It is recommended to keep 'chaos' values low, between 0 and 6, to avoid nonsensical or undesirable results.
Outlines
🚀 Mid Journey's Nii Version 6 Alpha Launch
Mid Journey has launched the alpha version of its anime-focused algorithm, Nii Version 6. Users can access it through the Mid Journey bot by typing 'D- nii n ji 6' at the end of their prompt. The update brings impressive visuals, such as Batman in a yellow bat suit eating pizza. However, inpainting, panning, and zooming features are not yet available in the alpha version. Access is also possible through the Nii server, with instructions provided for joining and using the bot. Users can add simple text to images using quotation marks, and the alpha offers various upscalers, although it's still limited in functionality.
🔍 Exploring Nii Version 6 Features
Nii Version 6 offers significant improvements in generating anime-style images. Users can now create images of two characters together, a feature not possible in previous versions. The model's strength in producing detailed and accurate images is highlighted with examples like Diva and Tracer from Overwatch. For those preferring less anime-like aesthetics, the 'style raw' option can tone down the cartoonish elements. Stylize values generally don't affect the prompt significantly, except at the extremes of 1000 and 0, which produce distinct differences. Keeping chaos values low is recommended for better results, with the option to experiment for varied outcomes. Users are encouraged to explore and utilize these features to their advantage.
Mindmap
Keywords
💡Mid Journey
💡Niji V6
💡Anime
💡Upscale
💡Inpainting
💡Panning
💡Zooming Out
💡Stylize Value
💡Chaos Value
💡Variation Strong
💡Text Generation
Highlights
Mid Journey has released a new update with the alpha launch of Niji V6, an anime-focused algorithm.
There are two methods to access Niji V6: through the Mid Journey bot or the Niji server.
Users can type 'D-nii nji 6' at the end of their prompt to access this version.
Example of generated image: Batman wearing a yellow bat suit eating pizza in the Batcave.
The update is currently in Alpha, with features like inpainting, panning, and zooming out expected in the full launch.
A subtle upscaler and a creative upscaler are the only available features at the moment.
To access the Niji bot through the server, an invite link is provided in the description.
Users can interact with the bot by reacting to their generated images with an envelope emoji.
Adding simple text to images is possible, with a word or two recommended for best results.
The method to generate text in images involves inputting words within quotation marks in the prompt.
Example of text integration in images: 'Batman saying nii till I die'.
Variation strong button can be used to refine and improve the generated images.
Nolan's goal is to make learning AI straightforward, and he encourages viewers to like the video for more content.
Niji V6 is powerful for generating images of characters sitting together, such as Diva and Tracer from Overwatch.
The model was trained on anime, but the style raw parameter can tone down the cartoonish aesthetic.
Stylize values do not significantly change the prompt, except for the extremes of 1000 and 0.
Chaos values should be kept low for better image generation, with a recommendation not to exceed 6.
Niji V6 offers a playground for users to experiment with various parameters and generate unique images.