Cosmos-Transfer1: Conditional World Generation with Adaptive Multimodal Control

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: NVIDIA, :, Alhaija, Hassan Abu, Alvarez, Jose, Bala, Maciej, Cai, Tiffany, Cao, Tianshi, Cha, Liz, Chen, Joshua, Chen, Mike, Ferroni, Francesco, Fidler, Sanja, Fox, Dieter, Ge, Yunhao, Gu, Jinwei, Hassani, Ali, Isaev, Michael, Jannaty, Pooya, Lan, Shiyi, Lasser, Tobias, Ling, Huan, Liu, Ming-Yu, Liu, Xian, Lu, Yifan, Luo, Alice, Ma, Qianli, Mao, Hanzi, Ramos, Fabio, Ren, Xuanchi, Shen, Tianchang, Sun, Xinglong, Tang, Shitao, Wang, Ting-Chun, Wu, Jay, Xu, Jiashu, Xu, Stella, Xie, Kevin, Ye, Yuchong, Yang, Xiaodong, Zeng, Xiaohui, Zeng, Yu
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866908295438008320
author NVIDIA
:
Alhaija, Hassan Abu
Alvarez, Jose
Bala, Maciej
Cai, Tiffany
Cao, Tianshi
Cha, Liz
Chen, Joshua
Chen, Mike
Ferroni, Francesco
Fidler, Sanja
Fox, Dieter
Ge, Yunhao
Gu, Jinwei
Hassani, Ali
Isaev, Michael
Jannaty, Pooya
Lan, Shiyi
Lasser, Tobias
Ling, Huan
Liu, Ming-Yu
Liu, Xian
Lu, Yifan
Luo, Alice
Ma, Qianli
Mao, Hanzi
Ramos, Fabio
Ren, Xuanchi
Shen, Tianchang
Sun, Xinglong
Tang, Shitao
Wang, Ting-Chun
Wu, Jay
Xu, Jiashu
Xu, Stella
Xie, Kevin
Ye, Yuchong
Yang, Xiaodong
Zeng, Xiaohui
Zeng, Yu
author_facet NVIDIA
:
Alhaija, Hassan Abu
Alvarez, Jose
Bala, Maciej
Cai, Tiffany
Cao, Tianshi
Cha, Liz
Chen, Joshua
Chen, Mike
Ferroni, Francesco
Fidler, Sanja
Fox, Dieter
Ge, Yunhao
Gu, Jinwei
Hassani, Ali
Isaev, Michael
Jannaty, Pooya
Lan, Shiyi
Lasser, Tobias
Ling, Huan
Liu, Ming-Yu
Liu, Xian
Lu, Yifan
Luo, Alice
Ma, Qianli
Mao, Hanzi
Ramos, Fabio
Ren, Xuanchi
Shen, Tianchang
Sun, Xinglong
Tang, Shitao
Wang, Ting-Chun
Wu, Jay
Xu, Jiashu
Xu, Stella
Xie, Kevin
Ye, Yuchong
Yang, Xiaodong
Zeng, Xiaohui
Zeng, Yu
contents We introduce Cosmos-Transfer, a conditional world generation model that can generate world simulations based on multiple spatial control inputs of various modalities such as segmentation, depth, and edge. In the design, the spatial conditional scheme is adaptive and customizable. It allows weighting different conditional inputs differently at different spatial locations. This enables highly controllable world generation and finds use in various world-to-world transfer use cases, including Sim2Real. We conduct extensive evaluations to analyze the proposed model and demonstrate its applications for Physical AI, including robotics Sim2Real and autonomous vehicle data enrichment. We further demonstrate an inference scaling strategy to achieve real-time world generation with an NVIDIA GB200 NVL72 rack. To help accelerate research development in the field, we open-source our models and code at https://github.com/nvidia-cosmos/cosmos-transfer1.
format Preprint
id arxiv_https___arxiv_org_abs_2503_14492
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Cosmos-Transfer1: Conditional World Generation with Adaptive Multimodal Control
NVIDIA
:
Alhaija, Hassan Abu
Alvarez, Jose
Bala, Maciej
Cai, Tiffany
Cao, Tianshi
Cha, Liz
Chen, Joshua
Chen, Mike
Ferroni, Francesco
Fidler, Sanja
Fox, Dieter
Ge, Yunhao
Gu, Jinwei
Hassani, Ali
Isaev, Michael
Jannaty, Pooya
Lan, Shiyi
Lasser, Tobias
Ling, Huan
Liu, Ming-Yu
Liu, Xian
Lu, Yifan
Luo, Alice
Ma, Qianli
Mao, Hanzi
Ramos, Fabio
Ren, Xuanchi
Shen, Tianchang
Sun, Xinglong
Tang, Shitao
Wang, Ting-Chun
Wu, Jay
Xu, Jiashu
Xu, Stella
Xie, Kevin
Ye, Yuchong
Yang, Xiaodong
Zeng, Xiaohui
Zeng, Yu
Computer Vision and Pattern Recognition
Artificial Intelligence
Machine Learning
Robotics
We introduce Cosmos-Transfer, a conditional world generation model that can generate world simulations based on multiple spatial control inputs of various modalities such as segmentation, depth, and edge. In the design, the spatial conditional scheme is adaptive and customizable. It allows weighting different conditional inputs differently at different spatial locations. This enables highly controllable world generation and finds use in various world-to-world transfer use cases, including Sim2Real. We conduct extensive evaluations to analyze the proposed model and demonstrate its applications for Physical AI, including robotics Sim2Real and autonomous vehicle data enrichment. We further demonstrate an inference scaling strategy to achieve real-time world generation with an NVIDIA GB200 NVL72 rack. To help accelerate research development in the field, we open-source our models and code at https://github.com/nvidia-cosmos/cosmos-transfer1.
title Cosmos-Transfer1: Conditional World Generation with Adaptive Multimodal Control
topic Computer Vision and Pattern Recognition
Artificial Intelligence
Machine Learning
Robotics
url https://arxiv.org/abs/2503.14492