14 repos
LoRA adapters, distillation techniques, and optimization methods for accelerating text-to-image diffusion models like SDXL, SD 1.5, SD3, and Flux. The cluster covers both efficiency improvements (LCM, TCD, flash variants) and creative fine-tuning approaches (style/art augmentation LoRAs), centered around the Hugging Face diffusers ecosystem and safetensors format for model distribution.