How To Install Stable Diffusion Models: InvokeAI

All Your Tech AI
17 Apr 202307:03

TLDRThe video script offers a step-by-step guide on how to enhance the capabilities of stable diffusion AI by using specialized models. It suggests resources like Hugging Face and Civic AI for downloading models and explains the process of installing and using them in Invoke AI. The guide covers the difference between checkpoint and safe tensor files, provides tips on selecting models, and demonstrates how to integrate them into Invoke AI for generating images in various styles.


  • 🎨 Exploring new AI art styles can be facilitated by using specialized models beyond the default Stable Diffusion models.
  • πŸ” Websites like Hugging Face and Civic AI offer a variety of specialized models for different art styles, such as anime characters.
  • πŸ“‚ Models are typically in the form of checkpoints or Safe Tensor files, with the latter being more secure and faster to load.
  • 🌐 Civic AI has a wide range of categorized models, and users can download preferred models directly from their website.
  • πŸ”Ž Hugging Face provides a robust search feature for finding specific models like Open Journey, though its interface is less intuitive.
  • πŸ“¦ Downloaded models are usually 2-3 GB in size, and users should be aware of the storage and download times involved.
  • πŸ’» The process of integrating these models into Invoke AI involves moving the downloaded files to the Invoke AI's models directory.
  • πŸ–₯️ Invoke AI's user interface allows for easy addition of new models through the model manager, by simply adding the file path.
  • πŸš€ Once models are added and loaded into GPU memory, users can generate new images using the specialized styles.
  • πŸŽ‰ Successful integration of specialized models into AI platforms can result in high-quality, photo-realistic images tailored to specific styles or themes.

Q & A

  • What are some specialized models available for creating different styles of art in AI?

    -There are specialized models trained on specific styles such as anime characters, women celebrities, and more. These models can be found on platforms like Hugging Face and Civic AI to enhance the quality of generated images in a particular style.

  • What is the difference between a checkpoint and a safe tensor file in AI models?

    -Checkpoints are an older style of models, while safe tensors are a newer format. Safe tensors are preferred as they are more secure, making it harder to inject malicious code, and they also load significantly faster than checkpoint files.

  • How can you find and download specialized AI models?

    -You can find and download specialized AI models by visiting websites like Hugging Face or Civic AI, browsing through categories or using the search function to find specific models like those for anime characters or realistic visions.

  • What is the approximate size of these specialized AI models?

    -Specialized AI models are typically around two to three gigabytes in size.

  • How to install downloaded AI models into Invoke AI?

    -To install the models, you need to move the downloaded files to the 'models' directory in the Invoke AI installation folder, and then use the model manager within the Invoke AI user interface to add the new models.

  • What is the process for adding a new model in Invoke AI?

    -In Invoke AI, click on the model manager icon, select 'Add new writing a checkpoint or safe tensor model', provide the path to the model files, and click 'Find models' to add them to the list of available models.

  • How long does it take to load a new AI model in Invoke AI?

    -The time it takes to load a new model depends on the speed of your computer, as it involves loading a couple of gigabytes of data.

  • What type of images can be generated using the 'Realistic Vision' model?

    -The 'Realistic Vision' model can generate photo-realistic, detailed images, such as chocolate sprinkled donuts on a plate.

  • How can you switch between different models in Invoke AI?

    -You can switch between different models in Invoke AI by clicking on the model dropdown menu and selecting the desired model, then clicking the 'Load' button to load it into your GPU memory for image generation.

  • What is the role of the model manager icon in Invoke AI?

    -The model manager icon in Invoke AI is used to manage and add new models to the software, allowing users to switch between different models for generating images.

  • What is the significance of using specialized models for AI-generated art?

    -Using specialized models allows for more accurate and higher quality generation of images in specific styles or themes, providing artists with more control over the output and enabling the creation of more diverse and complex art pieces.



🎨 Exploring New AI Art Models

This paragraph introduces the viewer to the process of exploring and utilizing new AI art models beyond the default options provided by stable diffusion. It emphasizes the importance of finding specialized models to achieve better results in specific art styles, such as anime characters, and directs the viewer to resources like Hugging Face and Civic AI for model downloads. The speaker shares their experience with realistic vision models and provides a step-by-step guide on how to download, install, and use these models in invoke AI, highlighting the benefits of using safe tensor files over checkpoint files for security and speed.


πŸ“‚ Importing and Managing AI Models in Invoke AI

The second paragraph delves into the technicalities of importing and managing downloaded AI models within the Invoke AI software. It explains how to navigate the user interface to add new models, providing clear instructions on copying model paths and using the model manager to load the desired models into the system. The speaker demonstrates how to switch between different models and generate images using each one, showcasing the versatility of the software and the potential for creating diverse art pieces. The summary ends with a mention of the speaker, Brian Lovett, and an encouragement for viewers to learn more about AI and stable diffusion through his content.



πŸ’‘Stable Diffusion

Stable Diffusion is an AI model that generates images based on text prompts. It is the default model discussed in the video, capable of creating a wide variety of image types. The video mentions that while it's versatile, there might be a need for specialized models for better results in specific styles of art.

πŸ’‘Invoke AI

Invoke AI appears to be the software or platform where users can utilize different AI models to generate images. It is mentioned as the place where downloaded models can be installed and used to produce new images based on the selected model.

πŸ’‘Hugging Face

Hugging Face is a platform mentioned in the video where users can browse and download different AI models. It is described as less user-friendly compared to Civic AI but offers a powerful search ability to find specific models like Open Journey.

πŸ’‘Civic AI

Civic AI is another platform mentioned in the video that provides specialized AI models. It is noted for its wide range of options and categorization, making it easier for users to find models tailored to their needs, such as anime characters or specific styles.

πŸ’‘Checkpoints and Safe Tensors

Checkpoints and Safe Tensors files are terms used to describe different types of AI model files. Checkpoints are an older style of model, while Safe Tensors are newer and considered safer due to their resistance to malicious code and faster loading times.

πŸ’‘Anime Characters

Anime Characters are used as an example in the video to illustrate the need for specialized AI models. Having a model trained specifically on anime would result in better generation of anime-style images.

πŸ’‘Model Management

Model Management refers to the process of organizing, adding, and selecting different AI models within the Invoke AI software. This includes adding new models, finding them in the system, and loading them for use.

πŸ’‘GPU Memory

GPU (Graphics Processing Unit) Memory is the memory used by the GPU to store data for processing. In the context of the video, loading an AI model into GPU memory allows for the generation of new images based on that model, but it requires a significant amount of data.

πŸ’‘Image Generation

Image Generation is the process by which AI models create new images based on input prompts or selected styles. This is the primary function of the AI models discussed in the video and the end goal for users of the Invoke AI software.

πŸ’‘Tech AI

Tech AI seems to be the broader field or community focused on artificial intelligence technologies, including the AI models and platforms discussed in the video. It is the context within which the video's content exists.

πŸ’‘Brian Lovett

Brian Lovett is presumably the speaker or creator of the video content, providing guidance on using AI models for image generation. He introduces himself at the end of the video as the source of the information.


Introduction to using specialized AI models for art generation beyond the default stable diffusion models.

Mention of resources like Hugging Face and Civic Ai for downloading specialized models.

Explanation of the difference between older checkpoint models and newer safe tensor files, with a recommendation for the latter.

Demonstration of how to find and download a specialized model, such as one for anime characters, from Civic Ai.

Discussion on the variety of models available on Civic Ai, including categories like anime characters, women, celebrities, etc.

Personal experience sharing about using the 'Realistic Vision' model and its impressive results.

Introduction to Hugging Face as a platform for finding AI models, including a walkthrough of searching for the popular 'Open Journey' model.

Guidance on navigating the non-intuitive interface of Hugging Face to find and download the desired model files.

Process of adding downloaded models to Invoke AI software by locating them in the 'models' directory.

Instructions on how to add new models to Invoke AI using the model manager within the software's user interface.

Example of loading the 'Open Journey V4' model in Invoke AI and generating an image based on it.

Showcase of the 'Realistic Vision' model's ability to generate detailed and photo-realistic images.

Explanation of the practical application of these AI models for generating art in various styles.

Overview of the entire process from downloading specialized models to generating art with Invoke AI.

Endorsement of Invoke AI for creating amazing art using different models and a call to action for learning more about AI and stable diffusion.