The article was created in the framework of paid cooperation.
Artificial intelligence continues every day. Not much time has passed since NVIDIA finally announced the new generation cards, which are still more bets on AI. Whether it is the generation of pictures or scaling. For players and games This is just the futureTo achieve the best possible results and at the same time use the latest graphic technologies, and especially ray tracing or directly tracing of the path.
But this constantly developed path, which could be compared with the river a little better. Sometimes they are divided and controlled in two directions, but this should also not be forgotten about the very first steps that they still work on. So, the named generative AI Although at present he manages to be a very solid communication assistant, he can generate high -quality videos or voices; But how can you remember well It all started with picturesWe want to look at NVIDIA today.
Pictures were probably used by any of you. Through closed platforms, for example, in ChatGPT or New versions of PhotoshopTherefore, despite the fact that open source models, such as necessary Stable diffusionwhich is one of the most popular. Open source models have several advantages. Be local, both confidentiality and modularity. They can be installed more to train better, change in more detailAnd since they are basically not subject to the rules of the company, they do not even tend to censor many things. We all know what it means …
Stable diffusion, as it was already, is the most popular for many reasons. And one is its availability and total speed. To use it and testing, we will use the popular local Web Front Automatic111111111111A field but after installing it through GitHub, we do not finish yet; Now you need to download the model itself. We will use the page for this Hugging face – Collection of everything important in generative AI. Where, by the way, you can easily try not only the models that we will show today, but also to others. And quickly and free, not worrying that you do not have a powerful computer now.
And this is not only a stable diffusion itself. But also a coder of twinsThis will generate a simple application at speed. But about this and other similar things sometimes at a different time. Now let’s focus first of all on the generation of images. For basic testing, let’s find some original versions of stable diffusion. First of all, versions 1.5, and then a little more new SDXL.
Source :Indian TV
