Downloading and Adding Models

There are many places where you can download Stable Diffusion models freely online. These include repositories like Hugging Face, Civitai, Tensor.art, and others.

While you could use the original base models including Stable Diffusion 1.5, 2.1, etc. fine-tuned models offer superior results for specific tasks or art styles. Fine-tuned models are trained on a specific subset of data, allowing them to excel in generating images within a particular niche or theme.

Generally speaking, models fall into categories which may include the following:

• Realism
• Cartoons/Anime
• General
• Fashion/Beauty
• Fantasy/Sci-Fi
• Architectural

Before downloading a model, it's important to review the model card as it has information about the model's specifications, intended uses, and any known limitations or potential issues. It also provides details about the model's creator and licensing information, which are crucial to understand before incorporating the model into your project.

Always download the .safetensors version of the model - never .ckpt as it could contain malicious code.

Depending on your project's requirements or personal preferences, you would select a model from one of these categories. Once you have downloaded the desired model, you can add it to your Stable Diffusion installation.

For Automatic1111 WebUI, this will be in the stable-diffusion-webui/models/Stable-diffusion directory. In ComfyUI, it will be in the ComfyUI/models/checkpoints directory.

Once a model is added, you can refresh the application to see it available within the dropdown menu.

‍