Stable Diffusion Web UI for AMD GPUs

Stable Diffusion Web UI for AMD GPUs

Category: Computer Vision
License: Other
Model Type: Image Generation
This repository provides a web interface for Stable Diffusion, optimized for AMD GPUs. It offers a user-friendly interface for generating images from text prompts, with support for various features like inpainting, outpainting, and upscaling. The interface is built using the Gradio library and includes enhancements for AMD hardware compatibility.

Key Features

  • Original txt2img and img2img modes
  • Outpainting and inpainting capabilities
  • Color sketch and prompt matrix tools
  • Stable Diffusion upscale functionality
  • Attention mechanisms for fine-tuning model focus
  • Loopback for multiple img2img processing
  • 3D plot generation (X/Y/Z plot)
  • Textual inversion support
  • Face restoration tools: GFPGAN and CodeFormer
  • Image upscaling: RealESRGAN, ESRGAN, SwinIR, Swin2SR, LDSR
  • Aspect ratio resizing options
  • Sampling method selection with advanced noise settings
  • Interrupt processing at any time
  • 4GB video card support (with reports of 2GB working)
  • Correct seeds for batches
  • Live prompt token length validation
  • Generation parameters saved with images
  • Settings page for customization
  • Running arbitrary Python code from UI (with --allow-code)
  • Mouseover hints for most UI elements
  • Tiling support for texture-like images
  • Progress bar and live image generation preview
  • Negative prompt field
  • Styles and variations for prompt management
  • Seed resizing for image variation
  • CLIP interrogator for prompt guessing
  • Prompt editing during generation
  • Batch processing support
  • Highres fix for high-resolution images
  • Reloading checkpoints on the fly
  • Checkpoint merger for combining models
  • Custom scripts and community extensions
  • Composable-Diffusion for multi-prompt generation
  • No token limit for prompts
  • DeepDanbooru integration for anime prompts
  • xformers support for speed improvement
  • History tab via extension
  • Training tab for hypernetworks and embeddings
  • Clip skip functionality
  • Hypernetworks and LoRAs support
  • VAE selection from settings
  • Estimated completion time in progress bar
  • API support
  • Inpainting model support by RunwayML
  • Aesthetic Gradients via extension
  • Stable Diffusion 2.0 and Alt-Diffusion support
  • Safetensors format for checkpoints
  • Eased resolution restriction
  • Reorder UI elements via settings
  • Segmind Stable Diffusion support
  • DirectML support for GPUs with DirectX 12
  • ZLUDA support for AMD GPUs
  • ONNX Runtime support
  • Olive model optimization support
GitHub

Project Screenshots

Project Screenshot

Similar Projects

Awesome GPT-4o Images – Curated Collection of AI-Generated Visuals and Prompts

Awesome GPT-4o Images – Curated Collection of AI-Generated Visuals and Prompts

Computer Vision
MultiDiffusion Upscaler – Tiled Diffusion & VAE Extension for Stable Diffusion WebUI

MultiDiffusion Upscaler – Tiled Diffusion & VAE Extension for Stable Diffusion WebUI

Computer Vision
OmniGen – Unified Image Generation Model

OmniGen – Unified Image Generation Model

Computer Vision
Kandinsky 2 — Multilingual Text-to-Image Latent Diffusion Model

Kandinsky 2 — Multilingual Text-to-Image Latent Diffusion Model

Computer Vision
Image-Generation-CoT: Chain-of-Thought Reasoning for Complex Text-to-Image Generation

Image-Generation-CoT: Chain-of-Thought Reasoning for Complex Text-to-Image Generation

Computer Vision
DragGAN – Interactive Point-Based Image Editing on the Generative Image Manifold

DragGAN – Interactive Point-Based Image Editing on the Generative Image Manifold

Computer Vision