prettydeep

Check the requirements.txt: opencv-python>=4.1.2

Do the training and regularization images need to be 512x512, like the original SD model is based on?

The script [here](https://github.com/basujindal/stable-diffusion/issues/88#issuecomment-1234182086) works well for 512x512 images. But for other resolutions I get an error.

@kiriri The script you provide works well for 512x512 images (init-img and mask), although it appears the masked areas must be inverted compared to the original optimized inpainting gradio script...

I find it doesn't work on Ubuntu 18.04. As referenced[ here](https://www.digitalocean.com/community/questions/how-install-glibc-2-29-or-higher-in-ubuntu-18-04). You need to install libvulkan-dev and GLIBC_2.29, which can be done manually, but can easily break your system if...

Turns out the realesrgan binary was not running b/c of [this](https://github.com/jquesnelle/txt2imghd/issues/18#issuecomment-1242833861). If there is any way to run your script with python libraries instead of vulkan binaries, it would be...

@prusnak Can you provide hashes for the 13B files?

I have pdf and docx forms downloaded by Autogpt and saved in the workspace. But when I try to open them they are all corrupted. The only success I've had...