VGen provides a comprehensive video generation ecosystem with implementations of multiple diffusion-based video synthesis methods including I2VGen-xl, VideoComposer, and text-to-video models. It supports generating videos from text, images, motion, and subject inputs, and includes tools for training, inference, visualization, and acceleration.