Datasets:
SO101 Chess Test 5 (TsFile)
Converted from zacapa/SO101_chess_test5 at pinned revision 94cff45ac1ebc7ff9c3053430f2833d1fa7be887. Modalities: Time-series.
Dataset description
This LeRobot v2.1 SO-101 dataset records demonstrations of starting a chess game by moving the e-pawn.
The pinned task instruction is: “start chess game with move the E pawn to E4”.
- Source repository owner/publisher: zacapa
- License: apache-2.0 (declared by the source card).
- Paper/homepage/citation: no completed paper, homepage, or citation is documented in the pinned source card unless linked above.
Dataset Scale
| Split | Episodes | Tasks | Source trajectory Parquets | TsFile rows | Sampling rate | TsFile files/shards |
|---|---|---|---|---|---|---|
train |
50 | 1 | 50 | 26,999 | 30 Hz | 1 |
The staged Parquet has 26,999 rows and 17 columns including Time; TsFile chunk metadata independently reports the same 26,999 rows.
TsFile schema
| Column | Role | TsFile type | Observed/source range |
|---|---|---|---|
Time |
TIME | INT64 | 0–17,967 ms; restarts per episode |
episode_index |
TAG | STRING | source integer 0–49 |
task_index |
TAG | STRING | source integer 0–0 |
frame_index |
FIELD | INT64 | 0–539 within an episode |
sample_index |
FIELD | INT64 | 0–26,998 globally |
Exact remaining FIELD names/ranges and imported types:
action_0–action_5(FLOAT): shoulder pan/lift, elbow flex, wrist flex/roll, and gripper actionobservation_state_0–observation_state_5(FLOAT): the corresponding SO-101 state
Conversion
- All source episodes in the train split are merged into
data/so101_chess_test5_train.tsfile;episode_indexandtask_indexare TAG dimensions. Time = round(timestamp * 1000)in milliseconds. The sourcetimestampcolumn is omitted because it is redundant withTime / 1000seconds.frame_indexis retained. Sourceindexis retained assample_index.- Every vector is fully flattened: the complete source name is kept,
.becomes_, and element indices are appended. Float vectors are imported as single-precision FLOAT fields. - Other scalar source columns shown above are retained; no trajectory rows are intentionally dropped.
- Source metadata is mirrored for publication, with copied
meta/info.jsonrewritten to describe the converted data path and conversion semantics.
Video policy
The pinned metadata declares 50 source videos for observation.images.front. MP4 files were not downloaded or uploaded; they remain in the pinned source videos/ tree.
Minimal read example
from tsfile import TsFileReader
reader = TsFileReader("data/so101_chess_test5_train.tsfile")
print(reader.get_all_table_schemas().keys())
reader.close()
Source and provenance
- Source dataset:
zacapa/SO101_chess_test5 - Pinned source revision:
94cff45ac1ebc7ff9c3053430f2833d1fa7be887 - Converted artifact:
data/so101_chess_test5_train.tsfile
- Downloads last month
- 41