Stable Diffusion Ultimate Guide pt. 3: High Resolution

Umberto Grando
9 min readMar 26, 2023

Hello World!

In the third part of this series, we are going to have a look at how to work with higher resolution images. Stable Diffusion v1.5 is trained on 512x512 images (while v2 is also trained on 768x768) so it can be difficult for it to output images with a much higher resolution than that. To do this consistently, we’ll need to do a few tricks.

For today’s example, I’m going to use my own custom checkpoint (model) that you can find here:

Requirements

To follow this example, you’ll need:

  • The Automatic1111 Stable Diffusion UI. If you don’t know how to install it, check out the first part of this series here: https://medium.com/p/bd7dbcd5ce4b
  • A GPU that can handle images of 768x1152 size. I’m using a RTX 4080 with 16 GB of RAM. In case you have access to less VRAM, you’ll probably be able to work with a lower resolution.
  • An image editing program. I’m using Krita (which is an awesome free digital painting application), but you…

--

--

Umberto Grando

I’m Umberto Grando, an IoT Specialist with a passion for programming, gaming and technology in general.