LeRobotLeRobot
核验标记表示资料完整度,不是产品评分。下列未确认项可能影响选型;请按对应来源逐项确认。The confidence mark describes evidence completeness, not a product score. Unconfirmed items below may affect your choice; check them against the relevant sources.
未确认的信息Information not confirmed
- 国内官网/安装源直达与备案状态本次未实测。
- 上游最近推送/发布节奏本次未通过 API 核验,lifecycle 记为 unknown。
- MCP 接入方式官方 README 未说明。
- Mainland direct access to website/install sources and filing status were not tested in this pass.
- Upstream push/release cadence was not verified via API; lifecycle is unknown.
- MCP integration is not stated in the official README.
English summaries and interface labels are available. Detailed source notes below are currently in Chinese.
LeRobot(LeRobot)是机器人软件栈对象。Hugging Face 机器人学习库,覆盖数据集与策略。
LeRobot is robot learning library in the robotics family; official README was fetched on 2026-11-07 and unknowns are kept explicit.
适合与不适合When to use it, when not to
适合:模仿学习、遥操作数据与真实机器人实验。 不适合:工业安全认证或开箱即用的生产控制。
Fit: robot software integration, simulation or policy experiments under the stated scope. Not fit: unverified production deployment, safety certification or real-hardware guarantees.
中文生态China availability 国产 / 国内可直接用 / 备案状态 · 快照domestic / direct access / CAC filing · snapshot 2026-11-07
- 出品属地Origin
- 国外出品foreign-made 国外主体或国际开源社区;本次未检索网信办公示名单。Foreign entity or international open-source community; the CAC public registry was not searched in this pass.
- 国内使用Direct access
- 直连未核验direct unknown 官方仓库 raw README 直抓 200;官网/安装源国内直达本次未实测。Official repository raw README fetched 200; direct mainland access to the website/install source was not tested here.
- 备案状态CAC filing
- 备案未核验filing unknown:未检索备案名单,记为未知;不推断是否备案。Filing registry not searched; recorded as unknown without inferring filing status.
备案信息以网信办《生成式人工智能服务已备案信息》公示名单为准(按批次滚动更新,本站只抄编号不解读);「大陆可直连」是本站从大陆网络直连实测的结果,不是厂商承诺。GitHub 类站点当日实测有波动,口径写在依据里。Filing status is quoted from the CAC public registry (updated batch by batch; this site copies the number and does not interpret it). "Direct access" is this site's own reachability measurement from a mainland network, not a vendor promise. GitHub-hosted sites fluctuated on the measurement day — the note says so.
固定坐标系Fixed coordinate system
8 个维度dimensions, 与同分区其他对象逐项可比each comparable item by item against others in the same section.
- 模型与开放条件Model & openness
LeRobot 本身不是模型产品;官方 README 未列模型 provider 清单。策略或模型接入通常经用户代码与生态集成,本站记为模型/策略无关(本站判断,核验范围限 README)。
LeRobot is not a model product; the official README lists no model provider list. Policy or model integration usually goes through user code and ecosystem integrations; this site records it as model/policy-agnostic (site judgement, verification limited to README).
- 运行位置Where it runs
LeRobot 以库/框架形态在本机进程中使用,通常经 Python/C++ 或 ROS 生态调用;不是托管云服务。具体安装与平台支持以官方 README/文档为准,本次未实测。
LeRobot is used as a library/framework in local processes, commonly via Python/C++ or the ROS ecosystem; it is not a hosted cloud service. Installation and platform support follow official README/docs and were not tested here.
- 本地文件Local files
LeRobot 会读写本地配置、场景、模型或数据资产;文件权限与资产许可需用户自行核对。它不替用户管理训练数据版权或机器人工作区安全。
LeRobot reads and writes local configuration, scenes, models or data assets; file permissions and asset licences must be checked by the user. It does not manage training-data copyright or robot workspace safety for the user.
- 关机后的任务Tasks after shutdown
节点或训练进程可在后台运行;关机续跑、分布式与故障恢复不是本次核验范围,需按官方部署文档核对。
Nodes or training processes can run in the background; shutdown continuation, distribution and failure recovery are outside this verification scope and must be checked against official deployment docs.
- 工具与扩展Tools & extensions
LeRobot 提供命令行、插件或生态接口(如 ROS 话题/服务、仿真插件、策略训练接口)。MCP 接入方式官方 README 未说明,记为未知。
LeRobot provides command-line, plugin or ecosystem interfaces (such as ROS topics/services, simulation plugins, policy-training interfaces). MCP integration is not stated in the official README and is recorded as unknown.
- 上下文与记忆Context & memory
不适用:LeRobot 不是对话模型,没有 LLM 上下文窗口;状态由配置、场景、节点状态或训练 checkpoint 承载。
Not applicable: LeRobot is not a chat model and has no LLM context window; state is carried by configuration, scenes, node state or training checkpoints.
- 权限与限制Permissions & limits
权限跟随操作系统与运行环境;涉及机器人硬件时还要核对设备访问、网络与急停边界。官方 README 未给统一权限模型,本次未实测。
Permissions follow the operating system and runtime environment; with robot hardware, device access, network and emergency-stop boundaries also need checking. The official README gives no unified permission model, and this was not tested here.
- 适合什么任务Best for
适合:模仿学习、遥操作数据与真实机器人实验。 不适合:工业安全认证或开箱即用的生产控制。
Fit: robot policy training and demonstration-data experiments. Not fit: unverified production deployment, safety certification or real-hardware guarantees.
头号误解Top misconception
- 以为预训练策略通用——任务与硬件差异要重新核对Assuming LeRobot provides model intelligence itself — it provides robot learning library infrastructure; policies or models come from user code and ecosystems.
- 忽略数据集许可与遥操作安全边界Assuming results transfer to real hardware without checks — reality gap, permissions and safety boundaries need separate verification.
证据来源Evidence sources
最近核验记录Latest recorded check 2026-11-07,内容更新日content updated on 2026-11-07。单项复核范围见正文或来源说明,不代表所有字段在同一天重新核验。See the text or source notes for item-specific scope; this does not mean every field was rechecked on that date.
| 类型Type | 名称Name | 链接Link |
|---|---|---|
| repo | huggingface/lerobot · 官方仓库(2026-11-07 README 直抓 200)huggingface/lerobot official repository (README fetched 200 on 2026-11-07) | https://github.com/huggingface/lerobot |
| docs | 官方 README(定位与能力描述,2026-11-07 直抓 200)Official README (positioning and capability description, fetched 200 on 2026-11-07) | https://raw.githubusercontent.com/huggingface/lerobot/main/README.md |
| changelog | Commit history(变更记录入口,2026-11-07 未逐条核验版本)Commit history for huggingface/lerobot (change-record entry, versions not itemised on 2026-11-07) | https://github.com/huggingface/lerobot/commits/ |
官方口径与未核到项Official record and unverified items
本站尚未完成实测。
相关条目Related entries
- llama.cppllama.cpp另一类:本地模型运行器;策略需要本地推理时看它。Another category: local model runner. If a policy needs local inference, read llama.cpp for format and quantisation boundaries.
本页由This page is generated from the ai-handbook 数据层生成(CC BY 4.0)。data layer (CC BY 4.0).
方法论与坐标系定义见仓库内For the methodology and coordinate definitions, see METHODOLOGY.md.