If you figure this out please let us know. I was just completing your course and using it as the foundation for my SaaS and literally had the same problem today. I was going to prompt hero trying to find better prompts to create higher quality images but haven’t had much luck yet. Definitely open to swapping out Dall E for something else
But running a stable diffusion model on your own infrastructure could cost alot because it need powerful like T4 graphic card to generate the image, if not the time it take to generate 1 image would be too long. Please tell us how to overcome this issue 😃
If you figure this out please let us know. I was just completing your course and using it as the foundation for my SaaS and literally had the same problem today. I was going to prompt hero trying to find better prompts to create higher quality images but haven’t had much luck yet. Definitely open to swapping out Dall E for something else
Sure thing. The replicate training seems to work ok. It can just quickly become expensive to train new models.
Have you tried looking up a better model on huggingface? there are some dedicated ones for icons (e.g. low-poly-hd-logos-icons)
You’re doing great babe!!!
Thanks my love! 😘 💋
how do you ensure color palette?
how bout if i have 2 pictures: a shirt and a person. using stable diffusion, how can i make that person wears that shirt?
ik
inpainting
Hey have you made a course for the icon generator app. I wanted to see an end to end sass building process. 😊
I have a course that builds out some of it, link in description
But running a stable diffusion model on your own infrastructure could cost alot because it need powerful like T4 graphic card to generate the image, if not the time it take to generate 1 image would be too long. Please tell us how to overcome this issue 😃
You either need your own machine with an awesome graphics card which costs money, or you use a third party service and rent their machines
Try using Vim keybindings, everyone seems using it... Try making a video while practising it...
I've been thinking about trying vim.. my computer seems to lag up during streams when I use vscode
try to train dot-to-dot model :)) I was testing, but model fails to generate correct numbers..
what is a dot-to-dot model?
its a simple dot to dot art :X
Need connect numbers, to see final image. Art for kids
@@pastuh oh ok, elaborate on what you’re doing? You pass an image and it converts it to a dot to dot image?
The best stable diffusion model I’ve seen is automatic 1111
I don’t think that’s the model. Automatic 1111 is just a Ui and python scripts which run models you download. The SD model is either 1.5 or 2.1