首个完整封装、一体化训练stable diffusion dreambooth的autodl镜像环境,可训练定制自己的独特大模型风格、人物,开箱即用,内含详细教程。
StreamDiffusion: A Pipeline-Level Solution for Real-Time Interactive Generation
Dreambooth (LoRA) with well-organized code structure. Naive adaptation from 🤗Diffusers.
attempt to use TensorRT with ComfyUI
Bamboo: 4 times larger than ImageNet; 2 time larger than Object365; Built by active learning.