{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/videogpt-video-generation-using-vq-vae-and","title":"VideoGPT: Video Generation using VQ-VAE and Transformers","arxiv_id":"2104.10157","date":"2021-04-20","proceeding":null,"authors":["Wilson Yan","Yunzhi Zhang","Pieter Abbeel","Aravind Srinivas"],"abstract":"We present VideoGPT: a conceptually simple architecture for scaling likelihood based generative modeling to natural videos. VideoGPT uses VQ-VAE that learns downsampled discrete latent representations of a raw video by employing 3D convolutions and axial self-attention. A simple GPT-like architecture is then used to autoregressively model the discrete latents using spatio-temporal position encodings. Despite the simplicity in formulation and ease of training, our architecture is able to generate samples competitive with state-of-the-art GAN models for video generation on the BAIR Robot dataset, and generate high fidelity natural videos from UCF-101 and Tumbler GIF Dataset (TGIF). We hope our proposed architecture serves as a reproducible reference for a minimalistic implementation of transformer based video generation models. Samples and code are available at https://wilson1yan.github.io/videogpt/index.html","url_abs":"https://arxiv.org/abs/2104.10157v2","url_pdf":"https://arxiv.org/pdf/2104.10157v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"videogpt-video-generation-using-vq-vae-and","repo_url":"https://github.com/wilson1yan/VideoGPT","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"videogpt-video-generation-using-vq-vae-and","repo_url":"https://github.com/Alescontrela/viper_rl","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"jax","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"videogpt-video-generation-using-vq-vae-and","repo_url":"https://github.com/alescontrela/viper","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"jax","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":null,"task_name":"Position"},{"task_slug":"video-generation","task_name":"Video Generation"}],"methods":[{"method_slug":"vq-vae","method_name":"VQ-VAE"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/video-generation-on-bair-robot-pushing","task":"Video Generation","dataset":"BAIR Robot Pushing","model":"VideoGPT","rank_in_archive_order":9,"of":31,"metrics":{"Cond":"1","FVD score":"103.3","Pred":"15","Train":"15"},"uses_additional_data":false},{"leaderboard":"/sota/video-generation-on-ucf-101-16-frames-128x128","task":"Video Generation","dataset":"UCF-101 16 frames, 128x128, Unconditional","model":"VideoGPT","rank_in_archive_order":3,"of":6,"metrics":{"Inception Score":"24.69"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/2104.10157","atlas_url":"https://app.syntology.ai/?focus=2104.10157","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}