Embodied Episode Data Schema v0.1
A small, documented format for task metadata, robot state, actions, timing, environment context, and terminal outcomes.
ArtiBot Cloud publishes small, inspectable technical resources: an embodied episode data schema, a fully synthetic example, capture guidance, and explicit public-release boundaries.
A small, documented format for task metadata, robot state, actions, timing, environment context, and terminal outcomes.
A compact JSON and CSV example with an offline validator. It contains no real robot, customer, scene, video, or model data.
Practical guidance for stable visual coverage, authorization, privacy, sensitive backgrounds, and public-release review.
A clear description of material that may be published and material that must remain out of public repositories.
This page is a public documentation index. It is not a research-paper archive, a benchmark leaderboard, a customer-data gallery, a production-source mirror, or a remote robot-control service.
| May be published after review | Must not be published here |
|---|---|
| Public documentation, fully synthetic examples, properly licensed standalone utilities, and documented public interfaces. | Production source code, credentials, internal infrastructure, customer data, private robot-control endpoints, or unverified research claims. |
| Public task descriptions and datasets with clear provenance and redistribution rights. | Third-party assets, videos, models, or meshes without verified permission to redistribute. |
ArtiBot Cloud 在此发布小而可检查的技术资料:具身 episode 数据规范、完全合成的最小样例、场景采集指南,以及清晰的公开边界。
这些材料帮助开发者讨论数据结构、时间同步、动作语义和公开数据治理,但不公开生产代码、客户数据、部署基础设施、私有机器人控制能力或未验证研究结论。
最小样例用 JSON 与 CSV 表示一条合成的“观测—动作—结果”交互 episode,并提供仅使用 Python 标准库的离线校验器。
后续独立组件、数据集、接口或基准任务只有在来源、许可、隐私与安全边界明确后才会发布。