15 repos
Fast, efficient text-to-image diffusion models optimized for various hardware constraints and quantization formats. The cluster centers on FLUX.2-klein variants—compact versions of the FLUX image generation model—with multiple quantized implementations (FP8, NVFP4) designed to run on resource-limited devices while maintaining visual quality. Repositories here focus on model compression, inference optimization, and practical deployment of diffusion-based image synthesis.