AI

Midjourney releases V7, its first new AI image model in nearly a year

Midjourney, one of the first AI-based image generators on the internet, has released his first new AI image model in almost a year.

Dubbed V7The model started rolling on Thursday around Midnight Est in Alpha, a week after OpenAi comes a new image generator in Chatgpt who quickly went viral because of his ability to take photos in Ghibli style.

Midjourney’s, however, is not a ghibli-optimized model-aldhans not officially but it can nevertheless generate aesthetically pleasant work, at least for the dilettant eye of this reporter.

To use it, you must first assess around 200 images to build a Midjourney “Personalization” profile if you haven’t done that yet. This profile adjusts the model to your individual visual preferences; V7 is the first Midjourney model that has enabled the personalization as standard.

Once you have done that, you can switch V7 on the website of Midjourney and, if you are a member of Midjourney’s Discord Server, on his Discord Chatbot. In the web app you can quickly select the model in the drop-down menu next to the “Version” label.

Midjourney V7
Image Credits:Midjourney

Midjourney CEO David Holz described V7 as a “totally different architecture” in one Post on X.

“V7 is […] Much smarter with text prompts, “Holz continued in an announcement on Discord.”[I]Mage prompts look fantastic, the image quality is noticeably higher with beautiful textures, and bodies, hands and all kinds of objects have a considerably better coherence on all details. “

See also  Coding Camps vs. Online Classes: What Is the Best Option?

V7 is available in two flavors, turbo (more expensive to run) and to relax, and feeds a new tool called Draft mode that displays images with 10x the speed and half the costs of the standard mode. Dress images are of lower quality than standard mode images, but they can be improved and re -displayed with a click.

A number of standard Midjourney functions are not yet available for V7, according to Holz, including images upscaling and retexuring. They will arrive in the near future, he said, possibly within two months.

“This is a completely new model with unique strengths and probably a few weaknesses,” wrote Holz op Discord. ‘[W]E wants to learn from you what it is good at, but make sure that it might need different styles of promptly. So play around a bit. “

Midjourney V7
My prompt: “A dragon with a spiky tail.”Image Credits:Midjourney

In my short tests, V7 kept my prompts reasonably well, although I didn’t have time Real Put the model through the ringtone.

Midjourney V7
My prompt: “A character in Ghibli style.”Image Credits:Midjourney

Midjourney is an unusual operation. Started in 2022 by Holz, who co-founder of the PC-perifere company protection Motion, did not raise a cent from outside money.

At the end of 2023 was Midjourney Reportedly Expect to yield around $ 200 million in income. Recently the company established in San Francisco said it founded a hardware team to work on some projects that it has not detailed, and it continues to train models for Video and 3D object generation.

The company is confronted multiple lawsuits This accuses the infringement of the rights of millions of artists by training AI tools on images that have been scraped from the internet without the permission of the makers of the images.

See also  Synthetic Data: A Double-Edged Sword for the Future of AI



Source link

Related Articles

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top button