internet has been something to integrate Many years now, decades even. However, during this period, we have seen how its use, the trends, have changed. The era of blogs gave way to the era of social networks, and this one began its decline (although in Europe networks like Instagram are still in full swing) at the same time that another emerges, the age of artificial intelligence. More and more services are using systems of this type to help create texts, images, execute processes faster…
While ChatGPT has taken on the task of making this type of application massively known, the image creation AIs are not left behind, with two options that stood out at the start, Stable Diffusion and Midjourney. Others like LeonardoAi or CatBirdAI were launched later. The global consensus was that Midjourney was the best of them all, but a few months ago it changed its policy, preventing its free use, even in a limited way.
Now the reference in open source imaging, Stable Diffusion, released a new version which significantly improves image creation, and places itself on an equal footing with other alternatives. The current version is SDXL 0.9, a version that has been in beta since April and is the prelude to version 1.0 which will arrive in July.
The new SDXL version Image quality has improved significantly. (although sometimes things like hands still show problems). In addition to this, we have seen how other options have been made possiblesuch as creating an image from another image, the ability to reconstruct missing parts of an image, or creating an extension for an existing image.
How can you test
The best thing about Stable Diffusion is that being an open source system, anyone can use it in their projects, leading to mass adoption. This can be done with the API, which will be available in the next few days, but it can already be used in Clipdrop, a free service that allows you to use it even without registering
Example of a stable XL0.9 broadcast with cinematic and origami style
The only thing to do is write in the text box what we want to appear in the image and, optionally, choose a style
Also It is possible to install it on a computer to create images independently. In this case, it is necessary to meet certain requirements. It will be mandatory to have a computer with a Windows 10 or 11, or Linux operating system, 16 GB of RAM, an Nvidia GeForce RTX 20 graphics card (equivalent or higher) equipped with a minimum of 8 GB of VRAM. Linux users can also use an AMD-compatible card with 16 GB of VRAM.
Image created with Stable Diffusion XL09
At the end of July, Stability AI, the company behind Stable Diffusion, confirmed the arrival of version 1.0, although he did not advance the improvements he might have had. The SDXL0.9 beta itself has come a long way in the last couple of months, so in two months when the 1.0 version is released (possibly also in beta) we’ll see another big leap in quality .
10 keys to Artificial Intelligence
You may be interested
Follow the topics that interest you
