Translated Data: Tencent and the Hong Kong University of Science and Technology have recently collaborated to introduce a high-quality video generation model named VideoCrafter. This model incorporates a video VAE and a 3D denoising U-net, and has been trained on a large-scale dataset, demonstrating excellent performance. VideoCrafter offers two diffusion models, which can be used for text-to-video and image-to-video tasks, opening up new possibilities for video creation. The project encourages research citation and usage of its code, models, and data.