The first image with the stained glass windows is very pretty. I also like the 4th photo in fact I like them all. With such a pretty rendering it becomes difficult to sort. I would like to keep them all. I would go to discord to try your AI which looks magnificent. 10,000 times better than the stable first generation broadcast installed on my computer.
@cyboule I'm glad you liked these. I think number four, which was one of my favs, looks a lot like @alexa.art.
The renderings do turn out so nice... it's hard to sort through which ones to like best or which ones to try and improve on. I don't know what kind of hardware you are using... but Ausbit and a few of the others in there may be able to help you run one of these bots locally on your own setup like you did with the 1st gen. I wish I could run one of these bots on my own.
@castleberry In fact I installed Linux on an SSD on which I also installed "stablediffusion" first generation. I cut off internet access on my SSD because the software was updated and buggy. I'm not a programmer and had a lot of difficulty installing the software. I have a 3070 video card with only 8 GB of RAM dedicated to the card and its problem. I cannot have a resolution higher than 512. Afterwards, when an image suits me, I rework it and try to increase its resolution to be able to have an image that can be used in my creations. I use software that works a bit like a PDF file, it allows an image to enlarge to make posters. With this long method I manage to create my artistic works which feed my blog. The solution would be to have another video card with 16 GB of dedicated RAM. At the moment, I don't make enough money to be able to buy one. I'm amazed to see the images generated with the new AI model. The positive thing about this business is that I process my images myself and this has allowed me to acquire skills that I would not have had if I had had a better card. Subsequently, if you would like to give me several prompt lines, I would like to generate them for you and send them back to you so that you can compare them with your software. But I already know the result in 512.
The standard preset for size of image output is 512x512... but there are settings to change size (height and width separately) or to change the orientation with preset sizes, and to upscale as well as high resolution settings to get even more interesting and quality results. So, the size isn't a limiting factor for the quality with the Arty bot. That's the name of the main one.
I don't know all the specs about what resources are required to run the bot locally... but you could use the bot for a while and see if it's something that will give you better and hopefully faster results to achieve what you are aiming to do. I commend you on learning all that you have to make it this far in your art journey and applying the new technology.
When I have some time this afternoon or evening I will try and round up a few prompts for you to run on your setup and we can compare and contrast the results together, if you like. Is there anything in particular that you would like me to do theme or subject wise?
I am really looking forward to trying out the collage and digital arts while also supplementing the designs with some AI infused aspects. It's been a seed of an idea that's been germinating for quite a while. Fortunately, I have quite a bit of material already generated to draw from.
No, nothing special, you can choose what you like.
Yeah, you are right. They all look so good. Pleasing to the eyes.