Imagen 3

We introduce Imagen 3, a latent diffusion model that generates high quality images from text prompts. We describe our quality and responsibility evaluations. Imagen 3 is preferred over other state-of-the-art (SOTA) models at the time of evaluation. In addition, we discuss issues around safety and re...

Full description

Saved in:
Bibliographic Details
Published in:arXiv.org 2024-12
Main Authors: Baldridge, Jason, Bauer, Jakob, Brichtova, Nicole, Bunner, Andrew, Castrejon, Lluis, Chan, Kelvin, Chen, Yichang, Dieleman, Sander, Hongliang Fei, Guo, Mandy, Haig, Alex, Yelin, Kim, Konyushkova, Ksenia, Langner, Karol, Luo, Shixin, Onoe, Yasumasa, Strudel, Robin, Waters, Austin, Wolff, Chris, Xiong, Hao, Xu, Keyang, Marc van Zee, Zolna, Konrad, Akerlund, Oscar, Bariach, Ben, Berman, Dana, Biles, Courtney, Blok, Irina, Caine, Ben, Carpenter, Viral, Casagrande, Norman, Chen, Tony, Choi, John, Churbanau, Dmitry, rester, Cole, Dektiarev, Mikhail, Du, Vincent, Dutta, Praneet, Feden, Ashley, Fruchter, Shlomi, Garcia, Frankie, Garg, Roopal, Goodman, Andrew, Gowal, Sven, Halpern, Yoni, Hertz, Amir, Hirst, Ed, Ibrahim, Mohamed, Iljazi, Joana, Isaac, William, Jennings, Gemma, Jia, Xuhui, Ju, Xiaoen, Ayan, Burcu Karagol, Kothawade, Suraj, Lagun, Dmitry, Mosseri, Inbar, Oliveira, João Gabriel, Ortiz-Jimenez, Guillermo, Tom Le Paine, Paiss, Roni, Pirozhenko, Alex, Poplin, Ryan, Rahtz, Matthew, Rastogi, Charvi, Amit, Raul, Sylvestre-Alvise Rebuffi, Robinson, Dirk, Rosgen, Bill, Rumbley, Sarah, Ryu, Moonkyung, Shah, Tanmay, Shillingford, Brendan, Sokolov, Valerii, Stimberg, Florian, Stone, Brad, Yu-Chuan, Su, Tabellion, Eric, Toor, Andeep, Udrescu, Cristian, Vasconcelos, Cristina, Vasiloff, Alex, Wang, Simon, Wang, Stanley, Wang, Qifei, Wang, Yuxiao, Wiles, Olivia, Wu, Chenxia, Xue, Andrew, Zhaxybay, Adilet, Zhou, Miao, Zhu, Shengqi, Zhu, Zhenkai, Hassabis, Demis, Hume, Tom, Kavukcuoglu, Koray, Orban, Andras, Subramanya, Amar, Vinyals, Oriol, Zwols, Yori
Format: Article
Language:English
Subjects:
Image quality
State-of-the-art reviews
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:We introduce Imagen 3, a latent diffusion model that generates high quality images from text prompts. We describe our quality and responsibility evaluations. Imagen 3 is preferred over other state-of-the-art (SOTA) models at the time of evaluation. In addition, we discuss issues around safety and representation, as well as methods we used to minimize the potential harm of our models.
ISSN:2331-8422