Stable Diffusion 2.0 is the newest release from Stability.AI and has been out for couple of weeks. This latest version is out there so I put this post together to help you find related information about it in one place to get started quickly.
Important Links
Here are a bunch of links that you want to know about to access the related information quickly.
- GIT REPO: https://github.com/Stability-AI/stablediffusion
- Stable Diffusion v2 Model 768×768 px: https://huggingface.co/stabilityai/stable-diffusion-2
- Stable Diffusion v2 Model 512×512 px: https://huggingface.co/stabilityai/stable-diffusion-2-base
- Stable Diffusion v2 Depth Model: https://huggingface.co/stabilityai/stable-diffusion-2-depth
- Stable Diffusion v2 In-Painting Model: https://huggingface.co/stabilityai/stable-diffusion-2-inpainting
- Stable Diffusion v2 Upscaler Model: https://huggingface.co/stabilityai/stable-diffusion-x4-upscaler/tree/main
Download Models CKPT files
It can be confusing navigating the huggingface site to find the link of the model file, so here are the links for all the models related to Stable Diffusion.
- 768 model download link on HuggingFace: https://huggingface.co/stabilityai/stable-diffusion-2/blob/main/768-v-ema.ckpt
- 512 base model download link: https://huggingface.co/stabilityai/stable-diffusion-2-base/blob/main/512-base-ema.ckpt
- 512 depth model download link: https://huggingface.co/stabilityai/stable-diffusion-2-depth/blob/main/512-depth-ema.ckpt
- 512 inpainting model download link: https://huggingface.co/stabilityai/stable-diffusion-2-inpainting/blob/main/512-inpainting-ema.ckpt
- new x4 upscaler download link: https://huggingface.co/stabilityai/stable-diffusion-x4-upscaler/blob/main/x4-upscaler-ema.ckpt
Amongst these you will also need the configuration file for Stable Diffusion 2.0: https://raw.githubusercontent.com/Stability-AI/stablediffusion/main/configs/stable-diffusion/v2-inference-v.yaml. Put this directory in the same folder as your CKPT file and rename it to the same filename as the model you are using, for example for model file 768-v-ema.ckpt rename the configuration file to 768-v-ema.yaml and so on.
Running Stable Diffusion 2.0
I am running mostly using Web UI of Automatic1111 on Paperspace platform (Start from as little as $8 per month – no longer using Google Colab) but as I find resources that will provide ability to run this in Google Colab, I will share these here. If you are already using a Google Colab notebook for Stable Diffusion, you can place the model files and configuration in the same folders as existing and you should be able to run Stable Diffusion 2.0 (I’ve not tested this but this is what I have read across my twitter feeds and people I’m following).
Hey, is it necessary to download all the model files as in is there a model file that combines all the other models? If not, what do you recommend?
Thanks a lot for the Information!
No you can download one either the 512 or 768 to get started. This will allow you to generate the images.
If I do want a secondary model (ckpt), do I need to run the bat file again after adding it to the directory? tyvm
If its in the same directory as the other model then you can click Refresh on in the model drop down menu in Automatic1111. It should load without having to run it again. Otherwise you need to run the batch file again. I’m not 100% on this as normally I download all my models before I run the UI script.