I am aware that there are AI generals in virtually every board on this site, but I don't like asking technical questions like this there because I don't want to interrupt the flow of creativity they got going. That being said, I hope someone can help me.
So I followed a simple, retard proof guide to get a Stable Diffusion working in a Google Collab by using this notebook:
https://colab.research.google.com/github/TheLastBen/fast-stable-diffusion/blob/main/fast_stable_diffusion_AUTOMATIC1111.ipynb?authuser=0&pli=1#scrollTo=PjzwxTkPSPHfAnd it worked fine. I got it up and running but then I wanted to try different models, which the script itself allows you to do by specifying the URL of a model to download. That is also working fine. I got it to download the waifu diffusion model. All pretty cool.
However changing models like this requires that you download the model every time, since the script renames every downloaded model to model.ckpt, essentially deleting whatever previous model you had. I read on some guide that you could just paste whatever model you wanted in the models/ folder and when I did so, the model names appear in the dropdown list but if you try to select it you get an error about some size being mismatched or something like that.
I also tried this guide
https://invoke-ai.github.io/InvokeAI/installation/050_INSTALLING_MODELS/ that says to paste the model and edit a config file named models.yaml but no such file exists in the Colab install. So I guess that guide doesn't work with that.
Any help appreciated.