120 repos across 6 sub-areas
Model checkpoints, fine-tuned variants, and implementations of Stable Diffusion for text-to-image generation and inpainting tasks. The cluster centers on different community-trained model versions (ranging from base Stable Diffusion v1-4 through specialized variants like Counterfeit and TrinArt) and their supporting infrastructure for inference and deployment. Most repositories here are model weights or minimal wrapper code rather than foundational framework development—the actual diffusers library and core Stable Diffusion research sit outside this cluster, which focuses on the ecosystem of trained models and direct usage examples.
Stable Diffusion text-to-image models
33 repos
Text-to-image generation using the Stable Diffusion architecture and related model variants. This cluster focuses on model checkpoints, fine-tuned versions (including multilingual variants), inference implementations across different frameworks (WebNN, etc.), and tooling around the diffusers library ecosystem. Repositories here include base model releases, specialized inpainting variants, quantized/optimized versions, and community fine-tunes for specific languages or use cases.
Cluster 636924
26 repos
Stable Diffusion Text-to-Image Models
25 repos
Fine-tuned and specialized variants of Stable Diffusion for generating images from text prompts. This cluster contains community-trained model checkpoints optimized for different artistic styles and subjects—including anime/waifu art, specific aesthetic directions (openjourney), and quality refinements (Counterfeit). Most repos focus on model weights, inference compatibility, and safetensor format distribution rather than core framework development.
Cluster 636921
24 repos
Stable Diffusion image generation control
8 repos
Techniques and implementations for controlling and adapting Stable Diffusion text-to-image models through specialized adapters, LoRA fine-tuning, and attention mechanisms. The cluster covers tools for precise output control (IP-Adapter for style/identity injection, ControlLoRA for structured guidance), model optimization (SDXL-Lightning for faster inference), and interactive editing methods (Attend-and-Excite for attention-based manipulation). Most repos are Python implementations or Jupyter notebooks exploring these control methods alongside the core diffusers library ecosystem.
Text-to-Image Diffusion Models
4 repos
Text-to-image generation using diffusion models, centered on the Stable Diffusion family of architectures and their implementations. This cluster covers model checkpoints, fine-tuning frameworks, and inference libraries built around diffusers and safetensors formats. Repositories here enable researchers and practitioners to generate, customize, and deploy image synthesis models across different scales and capabilities.