Midjourney V5 – the next generation of software that creates images from text is coming. The effects are impressive

Probably most of us have already heard of Midjourney. The program itself “powered” by artificial intelligence algorithms has been with us for a short time. As in the case of DALL-E and Stable Diffusion, we can describe in text what image we would like to see. So far, the program has handled this task quite well, but in some aspects it required improvements. The next generation is not only to solve the existing problems, but also to raise images to a completely different level.

Soon, the next generation of Midjourney, marked as V5, will be made available to the public. The samples of possibilities shown are truly amazing. The show just got even better.

Increasingly human communication with ChatGPT thanks to Chat D-ID. We can now talk to the AI ​​face to face

Anyone who has used Midjourney so far knows how good images this tool can generate. Of course, like any that is still developing and is in the testing phase, it has its problems. Sometimes the hand or other parts of the body don’t look quite right. However, in most cases, you could count on really interesting paintings, which sometimes looked like real works of art. More tests have just begun, this time showing the capabilities of the upcoming generation. In the new phase of development, multiple groups consisting of users who have an active subscription have been selected. Each of them had to evaluate two similar and completely different images, and then choose the one they liked more. This selection not only indicated the tastes of users, but also somehow fine-tuned the program to generate images that would be visually more attractive to people. Midjourney itself has also undergone a lot of changes.

Microsoft ND H100 v5 – New Virtual Machines for AI Algorithms Coming Soon with NVIDIA H100 Accelerators

From now on, what we enter will be better understood by the software, and we ourselves will have more control over the final effect. The big change will affect the generation of faces, eyes and individual details. We no longer have to worry about crooked fingers or limbs. The images are to be much more detailed, and by the way, a higher output resolution than before will be made available. The time needed to generate them will be significantly reduced. We can also expect the availability of other proportions, although this function can already be obtained by other methods. For example, if you want an aspect ratio of 16:10, just enter “–ar 16:10” at the end of your query. The developers said that the new installment is expected to be available later this week. In turn, all the images that have already been created by the fifth version of Midjourney can literally knock you off your feet. You can see the effects in each of the photos.

Source: Twitter, Reddit