CoolFace
Datasetpublic

THULab/berkeley_cable_routing

berkeley_cable_routing This dataset is a TsFile conversion of the Hugging Face dataset lerobot/berkeley_cable_routing. The source dataset was created using LeRobot. Modalities: Time-series. The original repository also contains video streams; videos are not included in this converted repository. Source Dataset Original dataset: lerobot/berkeley_cable_routing License: apache-2.0 LeRobot codebase version: v3.0 Robot type: unknown Task: route cable Split: train… See the full description on the dataset page: https://huggingface.co/datasets/THULab/berkeley_cable_routing.

sourceHugging Faceapache-2.0updated 1mo agoView on Hugging Face
0likes28downloads
README.md111 linesDownload Raw Back to root
1---2license: apache-2.03task_categories:4- robotics5tags:6- LeRobot7- tsfile8- timeseries9- format:tsfile10configs:11- config_name: default12  data_files:13  - split: train14    path: data/berkeley_cable_routing.tsfile15size_categories:16- 10K<n<100K17modality:18- tabular19- timeseries20---21 22# berkeley_cable_routing23 24This dataset is a TsFile conversion of the Hugging Face dataset25[`lerobot/berkeley_cable_routing`](https://huggingface.co/datasets/lerobot/berkeley_cable_routing).26The source dataset was created using [LeRobot](https://github.com/huggingface/lerobot).27 28Modalities: Time-series. The original repository also contains video streams; videos are not29included in this converted repository.30 31## Source Dataset32 33- Original dataset: [`lerobot/berkeley_cable_routing`](https://huggingface.co/datasets/lerobot/berkeley_cable_routing)34- License: `apache-2.0`35- LeRobot codebase version: `v3.0`36- Robot type: `unknown`37- Task: `route cable`38- Split: `train` (`0:1647`)39- Source scale from `meta/info.json`: `1,647` episodes, `42,328` frames, `1` task40- Sampling rate: `10` fps41- Source data layout: `data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet`42- Source video layout: `videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4`43 44## Converted Files45 46- TsFile: `data/berkeley_cable_routing.tsfile`47- Converted rows: `42,328`48- TsFile table: `berkeley_cable_routing`49- Time precision: milliseconds50- TAG columns: `episode_index`, `task_index`51 52The conversion follows the LeRobot v3 data path in `meta/info.json` and uses the53`file-*.parquet` frame data layout.54 55## Schema56 57`Time` is synthesized as `round(timestamp * 1000)` in milliseconds. The source58`timestamp` column is dropped because it is redundant with `Time / 1000` seconds.59 60TAG columns:61 62- `episode_index`63- `task_index`64 65FIELD columns:66 67- `frame_index`68- `sample_index` (renamed from source `index`)69- `next_reward` (renamed from source `next.reward`)70- `next_done` (renamed from source `next.done`, stored as `0/1`)71- `observation_state_0` to `observation_state_7`72- `action_0` to `action_6`73 74Vector features are flattened by preserving the source feature name and replacing75`.` with `_`. For example, `observation.state` becomes76`observation_state_0` to `observation_state_7`, and `action` becomes77`action_0` to `action_6`.78 79## Video Policy80 81The following source video features are not converted into TsFile and are not82uploaded here:83 84- `observation.images.top_image`85- `observation.images.wrist225_image`86- `observation.images.wrist45_image`87- `observation.images.image`88 89Use the original dataset for videos:90[`lerobot/berkeley_cable_routing/videos`](https://huggingface.co/datasets/lerobot/berkeley_cable_routing/tree/main/videos).91 92## Validation93 94The converted TsFile was validated with the project pipeline and read back using95the TsFile Python SDK:96 97- staged Parquet rows: `42,328`98- TsFile metadata rows: `42,328`99- TsFile query rows: `42,328`100 101## Usage102 103```python104from tsfile import TsFileReader105 106path = "data/berkeley_cable_routing.tsfile"107with TsFileReader(path) as reader:108    schemas = reader.get_all_table_schemas()109    print(schemas.keys())110```111