How to use models (checkpoints)
Models are large files containing all the weights for your neural network. They are responsible for all the generated anime girls out there. Without a model, you will be unable to generate anything.
If you're serious about your Stable Diffusion works, you would never use the standard SD models. They are more like base models intended to be fine-tuned. The process of fine-tuning a model often called a Dreambooth training. You can train your own model based on a standard one, or you can just choose from dozens existing fine-tuned models on https://civitai.com/.
I would recommend the following models: DreamShaper, Deliberate, RealisticVision. But you can choose any checkpoint you like.
Download it and put to webui/Models/StableDiffusion. You can use subfolders there. In order for the model to appear in the list, press the refresh icon next to it.
In Stable Houdini just put an SD Switch Model node and choose the model from the list. It will be loaded when the node is cooked. If the model is already loaded, it will not take long. Alternatively you can choose the model on the SD Dream node, but be aware, that the model will be loaded immediately after you choose it. If you use different models for different nodes, they will be loaded and unloaded after the cook, it can take time. I recommend to use the SD Switch Model node unless you know what are you doing.
