Add a PART_INDEX_DELIVERY mode to the package operator (default off): a part delivers a single part.json carrying both split levels, the task-level slice boundaries and the action-level segments, plus a classification for every reviewed interval. Kept intervals land in slices[], discarded intervals in discardedSlices[] with their discard evidence, which deliberately makes the discard reason part of the customer deliverable. Media, PTS, IMU, pose and calibration bytes are no longer published; downstream packaging resolves them from the pinned materialization revision. The worker publishes the index through the same frozen-destination path as the directory form. Extract prepareDeliveryTarget so single-file and directory publication share one owner marker, overlap guards, delegated-scope check, conditional create, sha256 verification, idempotent retry and the refusal to overwrite another delivery. A part index also refuses to adopt a destination that holds files without an owner marker, so additive migration cannot claim a legacy directory by accident. The API and download control plane recognise the new schema, project the split classification for the browser, and treat a part index as a first-class final delivery. Cross-language contract tests read the operator's own schema and field literals so a one-sided rename fails a test. Verification: 10 operator tests, 8 worker tests and 3 API tests for the new behaviour pass. Failure sets match the HEAD baseline name-for-name for internal/api (75), internal/worker (4), tests/backend and the frontend suite; no regression. The end-to-end local replica stack was not exercised: it needs image builds and the opt-in Go materialization fixture, so the Go to operator handoff is covered by the contract tests instead.
内建算子分册
当前目录按“每个内建算子一份 Markdown”组织,内容基于当前平台真实算子定义按算子名称去重生成。
- 当前收录数量:
115 - 文档定位:单算子详细手册,重点覆盖输入输出契约、运行时路径、配置示例、排障与验收。
- 配套总览:
../builtin-operators-manual.zh-CN.md
使用建议
- 先阅读总手册理解平台级规则,再进入本目录阅读单算子分册。
- 设计 DAG 时优先确认端口类型、artifact 语义、外部副作用和验收口径,而不是只看算子名称。
分类索引
AI
- 6DRepNet HeadPose 预标(带渲染)
- ACT 策略训练
- Diffusion Policy 训练
- Hand Removal 视频去手
- HoloBrain RoboTwin 评测
- HoloBrain 完整训练
- HoloBrain 模型导出
- HoloBrain 模型测试
- OpenVLA 微调训练
- OpenVLA 推理
- Qwen3-VL 200秒时间区间定位(真实)
- Qwen3-VL 22类持续训练(真实 LoRA)
- Qwen3-VL 语义行为粗切(PTS)
- WiLoR 3D HandPose(自动)
- YOLO 具身目标检测训练
- 有效行为检测(真实 Qwen3-VL)
- 模型推理
- 轻量策略评测
- 轻量线性策略推理
- 轻量线性策略训练
DATA
DATASET
- COCO 标注校验
- HF 数据集探测
- HoloBrain 数据准备
- Orbbec Ego 左右相机去畸变
- Orbbec Ego 数据标准化
- Parquet 表摘要
- WebDataset 分片概览
- X 平台数据集索引
- YOLO 检测标注校验
- 媒体引用可读性校验
- 数据集Schema合约校验
- 数据集切分泄漏校验
- 数据集来源许可审计
- 训练切分清单生成
- 采集目录校验
HUMAN
IO
LOGIC
MATH
QA
ROBOTICS
- ALOHA HDF5 示教校验
- AgiBotWorld2026 数据集校验
- BOP 6D姿态数据集校验
- BridgeData TFDS-RLDS清单校验
- CALVIN Episode 校验
- DROID Raw Episode校验
- DexYCB 数据集校验
- Episode 边界校验
- FurnitureBench PKL示教校验
- GraspNet 抓取数据集校验
- HDF5 示教数据概览
- HO3D/H2O3D 数据集校验
- LIBERO HDF5任务校验
- LeRobot v3 Episode索引校验
- LeRobot 元数据一致性校验
- LeRobot 数据集校验
- LeRobot 数据集概览
- LeRobot 视频对齐校验
- MCAP 主题概览
- MJCF 执行器校验
- ManiSkill 轨迹校验
- PartNet-Mobility 资产校验
- RGB-D 帧对齐校验
- RH20T 场景结构校验
- RLBench 数据集校验
- RLDS 轨迹概览
- ROS JointState 校验
- ROSBag 数据概览
- RT-X动作空间Schema校验
- RoboCasa HDF5示教校验
- RoboMIND HDF5质量校验
- RoboSet HDF5轨迹校验
- RoboTurk Real HDF5校验
- TF 坐标树校验
- TFRecord 结构校验
- UMI Zarr数据集校验
- URDF 关节限位校验
- VLA 样本清单校验
- Zarr 轨迹数组概览
- robomimic HDF5 示教校验
- 分割 Mask 校验
- 力扭矩传感器数据校验
- 动作状态数值质量校验
- 多传感器时间同步校验
- 多机器人Embodiment元信息校验
- 手-物体姿态标注校验
- 机器人模型资产校验
- 点云数据概览
- 相机标定校验
- 语言指令校验
- 轨迹多模态完整性校验