For full functionality and improved performance, duplicate this Space or clone the repository and run it locally (preferably with a GPU-enabled environment). This ensures unrestricted execution, optimal inference speed, and persistent storage of generated outputs.

🎨 Super Resolution Anime Diffusion

Generate high-quality anime images with automatic super resolution enhancement. Combines Stable Diffusion models with advanced upscaling techniques (RealESRGAN & Waifu2x). [Duplicte](https://huggingface.co/spaces/yangheng95/SuperResolutionAnimeDiffusion/duplicate) this Space to run locally for full functionality and performance.

🖥️ Device: NVIDIA H200 MIG 3g.71gb | 💾 RAM: 1644.0GB

👤 Personal Info

HENG YANG —

AI Researcher and Developer with a focus on generative models and computer vision.

🚀 How to Use

  1. Select Mode: Choose Text-to-Image or Image-to-Image tab
  2. Pick Model: Select from available models or use custom path
  3. Create Prompt: Describe your desired image (use negative prompt to avoid elements)
  4. Upload Image: For img2img mode, provide base image
  5. Adjust Settings: Fine-tune resolution, steps, and guidance
  6. Set Upscaling: Choose super-resolution method and scale
  7. Generate: Click the generate button and wait for results!

🤝 Enhanced Features

This interface improves upon the original work with:

  • Modern UI: Clean tabbed interface with Gradio Blocks
  • Better UX: Progress tracking and real-time feedback
  • Enhanced Parameters: Organized controls with descriptions
  • Gallery View: Browse and manage generated images
  • Error Handling: Comprehensive error reporting and recovery

Generated Images