An open-source text-to-video and image-to-video model family with weights, inference code, and fine-tuning examples.