Nimbus: A Unified Embodied Synthetic Data Generation Framework

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: He, Zeyu, Zhang, Yuchang, Zhou, Yuanzhen, Tao, Miao, Li, Hengjie, Wang, Hui, Tian, Yang, Zeng, Jia, Wang, Tai, Cai, Wenzhe, Chen, Yilun, Gao, Ning, Pang, Jiangmiao
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910016212041728
author He, Zeyu
Zhang, Yuchang
Zhou, Yuanzhen
Tao, Miao
Li, Hengjie
Wang, Hui
Tian, Yang
Zeng, Jia
Wang, Tai
Cai, Wenzhe
Chen, Yilun
Gao, Ning
Pang, Jiangmiao
author_facet He, Zeyu
Zhang, Yuchang
Zhou, Yuanzhen
Tao, Miao
Li, Hengjie
Wang, Hui
Tian, Yang
Zeng, Jia
Wang, Tai
Cai, Wenzhe
Chen, Yilun
Gao, Ning
Pang, Jiangmiao
contents Scaling data volume and diversity is critical for generalizing embodied intelligence. While synthetic data generation offers a scalable alternative to expensive physical data acquisition, existing pipelines remain fragmented and task-specific. This isolation leads to significant engineering inefficiency and system instability, failing to support the sustained, high-throughput data generation required for foundation model training. To address these challenges, we present Nimbus, a unified synthetic data generation framework designed to integrate heterogeneous navigation and manipulation pipelines. Nimbus introduces a modular four-layer architecture featuring a decoupled execution model that separates trajectory planning, rendering, and storage into asynchronous stages. By implementing dynamic pipeline scheduling, global load balancing, distributed fault tolerance, and backend-specific rendering optimizations, the system maximizes resource utilization across CPU, GPU, and I/O resources. Our evaluation demonstrates that Nimbus achieves a 2-3X improvement in end-to-end throughput compared to unoptimized baselines and ensuring robust, long-term operation in large-scale distributed environments. This framework serves as the production backbone for the InternData suite, enabling seamless cross-domain data synthesis.
format Preprint
id arxiv_https___arxiv_org_abs_2601_21449
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Nimbus: A Unified Embodied Synthetic Data Generation Framework
He, Zeyu
Zhang, Yuchang
Zhou, Yuanzhen
Tao, Miao
Li, Hengjie
Wang, Hui
Tian, Yang
Zeng, Jia
Wang, Tai
Cai, Wenzhe
Chen, Yilun
Gao, Ning
Pang, Jiangmiao
Robotics
Distributed, Parallel, and Cluster Computing
Scaling data volume and diversity is critical for generalizing embodied intelligence. While synthetic data generation offers a scalable alternative to expensive physical data acquisition, existing pipelines remain fragmented and task-specific. This isolation leads to significant engineering inefficiency and system instability, failing to support the sustained, high-throughput data generation required for foundation model training. To address these challenges, we present Nimbus, a unified synthetic data generation framework designed to integrate heterogeneous navigation and manipulation pipelines. Nimbus introduces a modular four-layer architecture featuring a decoupled execution model that separates trajectory planning, rendering, and storage into asynchronous stages. By implementing dynamic pipeline scheduling, global load balancing, distributed fault tolerance, and backend-specific rendering optimizations, the system maximizes resource utilization across CPU, GPU, and I/O resources. Our evaluation demonstrates that Nimbus achieves a 2-3X improvement in end-to-end throughput compared to unoptimized baselines and ensuring robust, long-term operation in large-scale distributed environments. This framework serves as the production backbone for the InternData suite, enabling seamless cross-domain data synthesis.
title Nimbus: A Unified Embodied Synthetic Data Generation Framework
topic Robotics
Distributed, Parallel, and Cluster Computing
url https://arxiv.org/abs/2601.21449