export_to_lerobot_mapper¶
Export processed video data to LeRobot v2.0 dataset format (LIBERO-style).
Designed for Ray distributed execution: each actor writes files
independently using UUID-based names (no cross-process coordination).
After all actors finish, call finalize_dataset() once to assign
sequential episode indices, rename files, and generate metadata.
Processing phase (parallel, per actor): staging/ ├── data/{uuid}.parquet ├── videos/{uuid}.mp4 └── meta/episodes_{uuid}.jsonl
After finalize_dataset() (single-threaded): dataset_dir/ ├── data/chunk-{NNN}/episode_XXXXXX.parquet ├── videos/chunk-{NNN}/observation.images.image/episode_XXXXXX.mp4 └── meta/ ├── info.json ├── tasks.jsonl ├── episodes.jsonl └── modality.json
将处理后的视频数据导出为 LeRobot v2.0 数据集格式（LIBERO 风格）。
专为 Ray 分布式执行设计：每个 actor 使用基于 UUID 的名称独立写入文件（无需跨进程协调）。所有 actor 完成后，调用一次 finalize_dataset() 以分配连续的 episode 索引、重命名文件并生成元数据。
处理阶段（并行，每个 actor）： staging/ ├── data/{uuid}.parquet ├── videos/{uuid}.mp4 └── meta/episodes_{uuid}.jsonl
调用 finalize_dataset() 后（单线程）： dataset_dir/ ├── data/chunk-{NNN}/episode_XXXXXX.parquet ├── videos/chunk-{NNN}/observation.images.image/episode_XXXXXX.mp4 └── meta/ ├── info.json ├── tasks.jsonl ├── episodes.jsonl └── modality.json
Type 算子类型: mapper
Tags 标签: cpu, multimodal
🔧 Parameter Configuration 参数配置¶
name 参数名 |
type 类型 |
default 默认值 |
desc 说明 |
|---|---|---|---|
|
<class 'str'> |
|
Root directory for the LeRobot dataset output. |
|
<class 'str'> |
|
Meta field with action/state data. Used in whole-video mode (segment_field=None). |
|
<class 'int'> |
|
Frames per second for the dataset. |
|
<class 'str'> |
|
Robot type identifier for info.json. |
|
<class 'int'> |
|
Max episodes per chunk directory (default 1000). |
|
<class 'str'> |
|
Meta field storing atomic action segments. When set, each segment becomes a separate episode with its own caption as task description. When None (default), falls back to whole-video export via hand_action_field. |
|
<class 'str'> |
|
Sample field with extracted frame image paths. Used in segment mode to create per-segment videos. |
|
`` |
||
|
`` |