Hugging Face's logo

Spaces:
jatingocodeo
/
StableDiffusion
Running

App Files Community
StableDiffusion / README.md
jatingocodeo's picture
jatingocodeo
Update README.md
9417338 verified
preview code
|
raw
history blame
2.56 kB

Stable Diffusion Style Explorer

This Hugging Face Space demonstrates the use of textual inversion concepts from the Stable Diffusion concepts library with an optional car guidance feature.

🎨 Features

  • Multiple Concepts: Choose from various pre-trained concepts including:

    • Canna Lily (Object)
    • Samurai Jack Style
    • Babies Poster Style
    • Animal Toy (Object)
    • Sword Lily (Object)

  • Custom Prompts: Create your own base prompts or use the default landscape prompt

  • Car Guidance: Optional feature to guide the image generation towards including cars using CLIP-based semantic guidance

  • Reproducible Results: Set custom seeds for consistent results

  • Example Gallery: Browse through pre-generated examples showing different combinations of:

    • Concepts and prompts
    • With and without car guidance
    • Various scene types (landscapes, cityscapes, fantasy scenes)

🚀 How to Use

  1. Select a Concept: Choose from the dropdown menu of available concepts
  2. Enter a Prompt: Write your own prompt or use the default
  3. Set a Seed: Enter a number for reproducible results
  4. Toggle Car Guidance: Choose whether to include car-based semantic guidance
  5. Generate: Click the button and wait for your image!

📚 Example Combinations

  1. Samurai Jack Landscape

    • With and without car guidance
    • See how the animation style affects the scene

  2. Urban Canna Lily

    • Modern cityscape with the flower concept
    • Compare regular vs car-guided versions

  3. Magical Forest

    • Using the babies poster style
    • Creates a whimsical fantasy scene

  4. Playful Scene

    • Animal toy in a playground setting
    • See how car guidance changes the scene

  5. Garden Scene

    • Sword lily in a traditional garden
    • Demonstrates object concept integration

🛠️ Technical Details

  • Base Model: Stable Diffusion 2.0
  • Guidance Method: CLIP-based semantic guidance for car integration
  • Concepts Source: SD Concepts Library

📝 Notes

  • Generation time may vary depending on server load
  • Car guidance takes longer as it involves multiple refinement steps
  • Results may vary depending on the concept and prompt combination
  • Examples are cached for quick viewing

🔗 Links