Stable Diffusion 1.5 with Fooocus - Yes you can!

Kleebz Tech AI
4 Jun 202408:45

TLDRIn this Kleebz Tech video, Rodney explains how to use Stable Diffusion 1.5 in Fooocus, despite it being designed for SDXL. He guides viewers through setting up the refiner slot for SD 1.5, downloading models from CivitAI, and organizing files for clarity. Rodney demonstrates generating images with SD 1.5 and using LoRAs, emphasizing the importance of experimentation for optimal results.

Takeaways

  • 😀 Stable Diffusion 1.5 can be used in Fooocus, although it's designed for SDXL.
  • 🔍 The results from using SD 1.5 in Fooocus might differ from other interfaces.
  • 📂 In Fooocus, the base model must be SDXL, but the refiner slot can be either SDXL or SD 1.5.
  • 📁 To use SD 1.5, download the model from a trusted source like CivitAI and select the .safe tensors format.
  • 📂 Organize downloaded models by creating a separate folder for SD 1.5 within the checkpoints directory.
  • 🔄 Refresh all files in Fooocus after downloading a new model to make it available for selection.
  • 🔄 The refiner switch allows you to control at what step the model switches from the base to the refiner.
  • 🎨 Experiment with different settings and sliders to find the best results for your needs.
  • 📖 For using LoRAs with SD 1.5, download SD 1.5 compatible LoRAs and place them in the appropriate LoRAs folder.
  • 🔧 Refresh all models in Fooocus to see the newly added LoRAs and select them for use.
  • 🎉 The video encourages viewers to share their tips and experiences with using SD 1.5 in Fooocus.

Q & A

  • What is the main topic of the video?

    -The video discusses how to use Stable Diffusion 1.5 in Fooocus, despite it being designed for SDXL.

  • Why does the presenter prefer SDXL over Stable Diffusion 1.5?

    -The presenter personally prefers SDXL, but does not provide specific reasons for this preference in the script.

  • What is the difference between using SD 1.5 in Fooocus versus other user interfaces?

    -Using SD 1.5 in Fooocus is not the same as using it in other UIs, and the results may vary slightly.

  • Where can one find SD 1.5 models compatible with Fooocus?

    -SD 1.5 models can be found on CivitAI under the models section with the SD 1.5 version filter checked.

  • What file formats are recommended for use with Fooocus?

    -The presenter recommends using the .safe tensors file format for SD 1.5 models in Fooocus.

  • How does one import an SD 1.5 model into Fooocus?

    -After downloading the model, place it in the 'checkpoints' folder within the Fooocus 'models' directory and refresh the files.

  • What is the purpose of the refiner slot in Fooocus?

    -The refiner slot in Fooocus is where you put the SD 1.5 model, allowing you to use it in conjunction with the base SDXL model.

  • How does the 'refiner switch' slider work in Fooocus?

    -The 'refiner switch' slider determines at what point during the generation process the model will switch from the base model to the refiner model.

  • Can you use SD 1.5 LoRAs with SDXL in Fooocus?

    -No, using an SD 1.5 LoRA with SDXL will not work; the LoRA must match the model version being used in the refiner slot.

  • How does one apply a LoRA to an SD 1.5 model in Fooocus?

    -After downloading the appropriate SD 1.5 LoRA, place it in the 'LoRAs' folder within Fooocus and refresh the models.

  • What is the presenter's advice for getting the best results with SD 1.5 in Fooocus?

    -The presenter suggests starting with the recommendations provided in Fooocus and then experimenting with the settings.

Outlines

00:00

🤖 Using Stable Diffusion 1.5 in Fooocus

Rodney from Kleebz Tech explains how to use Stable Diffusion 1.5 in Fooocus, despite it being designed for SDXL. He clarifies that while it can be used, results may vary. Rodney prefers SDXL but demonstrates how to use SD 1.5 by selecting it in the refiner slot of Fooocus. He guides viewers to download a 1.5 model from CivitAI, choosing the full model in safe tensor format for better compatibility. Rodney advises creating a separate folder for 1.5 models within the checkpoints folder for organization. He also explains how to refresh the model list in Fooocus to include the new 1.5 model and adjust the refiner switch to control the influence of the base model versus the refiner model during the generation process.

05:01

📊 Advanced Usage of SD 1.5 with Fooocus and LoRAs

The second paragraph delves into more advanced usage of SD 1.5 with Fooocus, including the integration of LoRAs. Rodney adjusts the refiner switch to demonstrate the influence of SD 1.5 on the generation process, showing how lowering the switch percentage results in a greater reliance on SD 1.5. He then discusses downloading and using LoRAs compatible with SD 1.5, emphasizing the importance of using the safe tensor format. Rodney guides viewers through organizing LoRAs in a separate folder and refreshing the model list in Fooocus. He concludes with a demonstration of generating images using SD 1.5 with a LoRA, adjusting the refiner switch and adding a prompt to generate vector art style images. Rodney invites viewers to share tips and thanks his subscribers and donors, encouraging further exploration and creativity with Fooocus.

Mindmap

Keywords

💡Stable Diffusion 1.5

Stable Diffusion 1.5 is a version of the AI model Stable Diffusion, which is used for generating images from textual descriptions. It is an advancement over previous versions and offers improved capabilities. In the video, Rodney discusses how to use this model within the Fooocus interface, which is not its default environment. The mention of Stable Diffusion 1.5 being designed for SDXL (Stable Diffusion XL) but also compatible with Fooocus highlights its versatility.

💡Fooocus

Fooocus is a user interface that allows users to interact with AI models like Stable Diffusion for image generation. The script explains that while Fooocus is designed for SDXL, it can be adapted to work with Stable Diffusion 1.5, indicating the flexibility of the interface to accommodate different versions of AI models.

💡SDXL

SDXL, or Stable Diffusion XL, refers to an extended version of the Stable Diffusion model that likely offers more features or capabilities than the standard version. The video script mentions that Fooocus is designed for SDXL, suggesting that it may be the preferred or more compatible interface for users looking to leverage the full potential of the XL model.

💡LoRAs

LoRAs (Low-Rank Adaptations) are a type of model adaptation used to fine-tune AI models like Stable Diffusion. They allow for the injection of specific styles or features into the image generation process. In the script, Rodney discusses using LoRAs with Stable Diffusion 1.5 in Fooocus, indicating that these adaptations can be used to customize the output of the AI model.

💡Refiner slot

The refiner slot in Fooocus is a feature that allows users to specify a secondary model that will be used in the image generation process after the base model has completed its initial pass. The script explains that while the base model must be SDXL, the refiner slot can accommodate SD 1.5, demonstrating a layer of customization available to users.

💡CivitAI

CivitAI is mentioned as a preferred source for downloading AI models by Rodney. It is a platform where users can find and download various versions of AI models, including Stable Diffusion 1.5. The mention of CivitAI in the script provides an example of where users can obtain the necessary files to use Stable Diffusion 1.5 in Fooocus.

💡Checkpoint

A checkpoint in the context of AI models refers to a saved state of the model at a particular point in time, which can be used to continue training or to generate images. The script mentions downloading a checkpoint for Stable Diffusion 1.5, which is one of the steps required to use the model in Fooocus.

💡Safe tensors

Safe tensors are a file format used for storing the parameters of AI models. The script suggests that safe tensors are a better and safer file format to use with Fooocus, indicating a preference for this format when downloading and using AI models in the interface.

💡Refiner switch

The refiner switch in Fooocus is a slider that determines at what point in the image generation process the refiner model (SD 1.5 in this case) takes over from the base model. The script explains how adjusting this slider can influence the balance between the base and refiner models, affecting the final output of the image.

💡Prompt

A prompt is the textual description or command that users provide to the AI model to guide the image generation process. The script includes an example prompt 'a cat' and later 'ribbon logo, dear dark background centered', showing how users can specify what they want the AI to create.

💡Negative prompt

A negative prompt is a type of prompt that specifies what the AI model should avoid including in the generated image. The script mentions adding a negative prompt, which can help refine the image generation process by excluding unwanted elements.

Highlights

Stable Diffusion 1.5 can be used in Fooocus despite being designed for SDXL.

Using SD 1.5 in Fooocus will yield slightly different results compared to other UIs.

Rodney prefers SDXL but demonstrates how to use SD 1.5 in Fooocus.

In Fooocus, SD 1.5 should be placed in the refiner slot, not the base model slot.

LoRAs for SD 1.5 can be used in the refiner slot alongside SD 1.5.

CivitAI is recommended as a source for downloading SD 1.5 models.

Download the full model in .safe tensors format for better compatibility with Fooocus.

Create a separate folder for SD 1.5 models in the checkpoints directory for organization.

Refresh all files in Fooocus to detect the newly downloaded SD 1.5 model.

The refiner switch in Fooocus controls the transition from base model to refiner model.

Experiment with the refiner switch to find the best balance between base and refiner models.

The earlier the refiner switch, the more influence SD 1.5 will have on the final output.

Demonstration of generating an image with SD 1.5 showing the switch point in the process.

LoRAs must be compatible with SD 1.5 and should be placed in the correct folder in Fooocus.

Create a separate folder for SD 1.5 LoRAs to keep them organized.

Refresh all models in Fooocus to detect and list the newly downloaded SD 1.5 LoRAs.

Adjust the refiner switch slider to control when the LoRA takes effect during image generation.

Using a vector art style with SD 1.5 and LoRAs to generate themed images.

The video concludes with a call to action for viewers to like, comment, and subscribe.

Casual Browsing

Yes Really! - Get Different Characters with Poses - Stable Diffusion - Fooocus

2024-05-01 06:55:00

Which Should You Choose? Stable Diffusion 1.5 or SDXL?

2024-04-03 01:30:01

Stable Diffusion - Zoom with Outpainting with Fooocus

2024-09-26 22:13:01

Stable Diffusion - Better Images with FreeU - Fooocus

2024-04-04 03:05:00

Stable Diffusion - Inpainting with Fooocus - Don't Regenerate, Fix!

2024-03-27 10:55:01