Recently, a netizen asked the editor, although Stable Diffusion is open source, but there is a certain barrier to installation, the requirements of the computer configuration is relatively high, is there a free platform that can not be installed, but also like Stable Diffusion to train their own models?
In the spirit of sharing everything, I finally found a platform that meets the requirements of netizens - leonardo AI. This platform is based on Stable Diffusion and at the same time deeply integrates various plug-ins of Stable Diffusion, such as The platform is based on Stable Diffusion and deeply integrates various plug-ins of Stable Diffusion, such as ControlNET's openpose pose reference, local redrawing, prompt tips and so on, and even provides a foolproof function of online training of their own models.
The platform is particularly well suited for video game material, allowing users to use pre-trained models or train their own to create various types of artwork such as objects, environments, helmets, architecture and concept art.Leonardo AI also offers a wide range of features such as real-time suggestions and adjustments, intelligent color selection, creative suggestions, autocomplete functionality, diverse brushes and materials, online collaboration, and cross-platform compatibility, designed to help artists create great work more easily.
Leonardo's official website:https://www.1ai.net/3864.html
After registering and logging in to Leonardo AI, after logging in we can see that the left side of the homepage shows that we are given 150 token coins by default. These token coins are enough for us to use every day, and we need about 1-2 token coins for each image.
It's the menu bar, and the left side is divided into two main panels, which are the AI tools and the advanced settings for the model in question. The row above the middle panel is some default models, and the following waterfall way to show images are recent creations and excellent examples.
After clicking on AI Image Generation to enter
I gave you a rough description of the main features, then we will try to generate a picture down, we first go back to the main page, find a favorite image. I just randomly found a rabbit as an example.
Click on the selected image
You can see the detailed parameters for making this image, we just need to click on the colored bar below, and all the information of this image is copied to our making interface.
Then we take HIPPO (hippo) as an example. In order to demonstrate, I will be inside the rabbit keywords replaced by "hippo", to generate images as follows
How about this simple replacement can also realize the picture you want, the small icon below the image can be appropriate to adjust the image, of course, you can try to try more statements, or borrow from other people's cue words or use the same model for the picture.
Leonardo has a large community of trained models, and of course you can train your own models for a more personalized creative effect. In addition, Leonardo has many other features such as local redraws, prompt hints, pose references, etc. that allow users to create more efficiently.
Finally, some students who like woolgathering are going to ask what to do when the 150 TOKEN coins run out? The reason why I recommend this, because it will provide 150 TOKEN every day, enough to like to study Leonardo students play. Of course, the paid version of the function is more powerful, the model selection will be more, but also can generate video, this will wait for the users to develop their own.
OK, that's it for today.