How to easily generate images in popular neural networks
You can create a unique image, not only in the sensational Midjourney or Stable Diffusion. There are dozens of similar neural networks, if not hundreds. But to use them all, you will either have to install them locally and run them separately, or open separate tabs and applications.
What is the service
In general, it's inconvenient. Therefore, I use the Mage service, in which you can easily generate an image in more than 50 neuronetworks.
All models are divided into three types:
- generating images from text;
- generate GIF from text;
- generating images from images;
How to use
To generate an image, you only need to select the desired model and specify the parameters in the input field:
Additionally, you can specify the aspect ratio, the number of steps, and the degree to which the instructions are followed.
If you really know what you're doing, you can switch to advanced mode. In it, you can more accurately indicate the number of steps and the degree to which the instructions are followed. And also indicate the image seed and negative prompts.
And in about a minute you will get a finished image like this:
The finished image can be slightly edited: make an upscale, and / or correct the face. Or make a new one based on it.
Stable Diffusion 1.5 and 2.1 are available on the free plan. Paid plans start at $4 per month.
When you pay the basic rate, you get access to all neurons, the ability to generate NSFW content, and several active sessions. The professional plan removes ads, adds 50 more networks, and the ability to generate images faster and create GIFs.
Comments powered by Talkyard.