SD3 - Local Install Guide! FASTEST Way to run the new Model - Stable Diffusion 3

Olivio Sarikas
12 Jun 202406:15

TLDRThis video tutorial guides viewers on how to download, install, and run Stable Diffusion 3 Medium for creating AI-generated images. The host highlights the importance of signing a free license for non-commercial use and choosing the correct model file. The guide covers setting up the software in comu I, updating it, and loading workflows for image generation. It also includes a demonstration of the model's capabilities with a sample prompt, showcasing the model's understanding and creativity.

Takeaways

  • 😀 Stable Diffusion 3 is a new model for image rendering that has been released.
  • 📝 To use Stable Diffusion 3, you need to sign a free license for non-commercial use, or contact Stability AI for a commercial license.
  • 🔍 Download the 'sd3 medium including clip save tensor' file, which is around 6 GB for optimal use without the text encoder.
  • 📁 The model can be downloaded into the models folder for automatic 1111 or for com UI, which is linked to the automatic 1111 models folder.
  • 🛠️ Update com UI to ensure compatibility with the new model, using the manager extension to update all components and restart com UI.
  • 💡 If the torch Cuda model breaks, fix it by updating com UI and Python dependencies from the update folder.
  • 📚 Download and try out different workflows available in the com UI example workflows folder, such as basic, multi-prompt, and upscaling workflows.
  • 📝 Test the model with the provided 'sd3 demo prompts txt' file containing multiple different prompts.
  • 🔧 Set up the workflow in com UI by loading the checkpoint and customizing settings like the scheduler, steps, and sampler.
  • 🎨 The model makes creative decisions, as demonstrated by the 'cat holding a sign with the text I love you' prompt, which resulted in a heart in the middle.
  • 👍 The video provides a guide on how to download, update, and run Stable Diffusion 3 for rendering images, and offers advice for better image quality.

Q & A

  • What is the Stable Diffusion 3 medium model?

    -The Stable Diffusion 3 medium model is a version of the AI model that does not include the text encoder and is designed for non-commercial use. It is one of the versions available for download on Hugging Face.

  • Why is signing a license necessary for using Stable Diffusion 3?

    -Signing a license is necessary because it grants the user permission to use the model for non-commercial purposes. For commercial use, one must contact Stability AI to obtain a commercial license.

  • What is the difference between the 'sd3 medium safe tensor' and 'sd3 medium including clip save tensor' models?

    -The 'sd3 medium safe tensor' model does not include the text encoder, while the 'sd3 medium including clip save tensor' model does, making it more suitable for users who require text-to-image functionality.

  • How large are the different versions of the Stable Diffusion 3 model?

    -The 'sd3 medium including clip save tensor' file is around 6 GB, and the version that includes clip and T5 XXL fp8 is approximately 11 GB.

  • What is the purpose of downloading the example workflows from Hugging Face?

    -The example workflows provide different methods for using the Stable Diffusion 3 model, such as basic, multi-prompt, and upscaling workflows, which users can try out within their software like comu.

  • Why is it recommended to update comu before using the Stable Diffusion 3 model?

    -Updating comu ensures that the software is compatible with the new model and can utilize its features effectively. It also updates any custom notes within comu.

  • What should you do if the torch Cuda model breaks after updating comu?

    -If the torch Cuda model breaks, navigate to the comu windows portable folder, find the 'update comu and python dependencies' file in the update folder, and run it to fix the issue.

  • How can you load the workflows provided by comu Anonymous?

    -To load the workflows, download the images provided by comu Anonymous and drag them into the comu canvas.

  • What settings are suggested by comu Anonymous for the Stable Diffusion 3 model?

    -The suggested settings include using the sgm uniform scheduler with 30 steps, a CFG value of 5.5, and the uler sampler.

  • What is an example of a prompt that could be used with the Stable Diffusion 3 model?

    -An example prompt is 'cat holding a sign with the text I love you', which the model interprets creatively while still understanding the text content.

  • How can users test the model with different prompts?

    -Users can refer to the 'sd3 demo prompts.txt' file available on Hugging Face, which contains multiple different prompts to test the model's capabilities.

Outlines

00:00

🖼️ Downloading and Setting Up Stable Diffusion 3 Medium

This paragraph provides a step-by-step guide on how to download and set up Stable Diffusion 3 Medium, a new AI model for generating images. It starts with the necessity of visiting Hugging Face to sign a free license for non-commercial use, with an option for a commercial license upon contacting Stability AI. The user is directed to select the 'sd3 medium including clip save tensor' file, which is around 6 GB, for download. The paragraph also mentions the availability of different workflows and demo prompts for testing the model. The speaker plans to demonstrate the use of the model in Comfy UI and discusses potential issues with updating Comfy UI, offering a solution to fix the torch Cuda model if it fails to start after the update.

05:03

🎨 Testing Stable Diffusion 3 Medium with Comfy UI Workflows

The second paragraph focuses on the practical application of Stable Diffusion 3 Medium using Comfy UI. It details the process of updating Comfy UI and loading the workflows provided by the developer, Comy Anonymous. The paragraph describes a specific workflow for the 'medium including clip save tensor' model and another for the 'including clip and T5 XXL fp8' models. The user is shown how to drag and drop the workflows into Comfy UI and customize settings such as the scheduler and sampler. An example prompt is given, 'cat holding a sign with the text I love you,' which results in a creative image that includes a heart, demonstrating the model's understanding of text input. The paragraph concludes with an invitation for viewers to like, subscribe, and look forward to more videos.

Mindmap

Keywords

💡Stable Diffusion 3

Stable Diffusion 3 is a new model in the field of AI image generation. It is a successor to previous versions and is designed to produce high-quality images from textual descriptions. In the video, the host is focused on demonstrating how to download and run this model on a personal computer, making it a central theme of the tutorial.

💡Hugging Face

Hugging Face is a platform that hosts a variety of AI models, including Stable Diffusion 3. In the script, the host instructs viewers to visit Hugging Face to sign a license agreement, which is necessary for the use of the Stable Diffusion 3 model. It is an essential step in the installation process.

💡License

A license in this context refers to a legal agreement that allows users to use the Stable Diffusion 3 model for non-commercial purposes. The video mentions that for commercial use, one must contact Stability AI to obtain a commercial license. The license is a key concept as it governs the legal use of the AI model.

💡Model Versions

The script refers to different versions of the Stable Diffusion 3 model available for download, such as 'sd3 medium safe tensor' and 'sd3 medium including clip save tensor'. These versions vary in size and features, with the latter including a text encoder for more advanced image generation capabilities.

💡Comfy UI

Comfy UI, often abbreviated as comUI, is a user interface for running AI models like Stable Diffusion 3. The video script details how to use comUI to run the Stable Diffusion 3 model, including updating the software and loading workflows, making it a crucial tool in the local installation process.

💡Workflows

Workflows in the context of the video are pre-configured sets of operations within comUI that guide users through the process of generating images with Stable Diffusion 3. The script mentions different types of workflows such as basic, multi-prompt, and upscaling, which are designed to enhance the user's experience with the model.

💡Prompts

Prompts are textual descriptions that users provide to the AI model to guide the generation of images. The script discusses the use of prompts in the Stable Diffusion 3 model, including a demo file of prompts provided for testing purposes.

💡Update

Updating refers to the process of ensuring that comUI and its associated components are up-to-date, which is necessary for running the latest AI models like Stable Diffusion 3. The script provides instructions on how to update comUI to be compatible with the new model.

💡Checkpoint

In the context of AI models, a checkpoint is a snapshot of the model's training progress that can be loaded for further training or inference. The script instructs viewers to load a specific checkpoint of the Stable Diffusion 3 model for use in comUI.

💡Scheduler

A scheduler in AI image generation refers to an algorithm that controls the process of generating an image over multiple steps. The script mentions the use of an 'sgm uniform scheduler' with specific settings, which is part of the process of generating images with Stable Diffusion 3 in comUI.

💡Sampler

A sampler in the context of AI models is a method used to generate samples from a probability distribution, which in this case is used to create images. The script refers to the 'uler sampler' as the recommended sampler for generating images with the Stable Diffusion 3 model.

Highlights

Introduction to Stable Diffusion 3 and its capabilities.

Instructions to download Stable Diffusion 3 from Hugging Face.

Explanation of the free license for non-commercial use and how to obtain a commercial use license.

Details on choosing the correct model file for Stable Diffusion 3.

Recommendation to use the 'sd3 medium including clip save tensor' file.

Guidance on downloading the model into the models folder for automatic 1111 or for com UI.

Information about different workflows available for com UI.

How to access and download example workflows and demo prompts.

Steps to update com UI to use the new Stable Diffusion 3 model.

Advice on fixing the torch Cuda model if com UI fails to start.

Loading workflows in com UI after updates and fixes.

Introduction to workflows developed by com UI Anonymous.

How to load and customize workflows for different model versions.

Explanation of the Tex to image workflow and its settings.

Demonstration of a creative prompt resulting in an image with a cat holding a sign.

Discussion on the model's ability to understand and creatively interpret text prompts.

Encouragement to like, subscribe, and watch more videos for further insights.