Hugging Face's logo

Datasets:
DanqingZ
/
test-merged-dataset-modular

Dataset card Data Studio Files
xet
 Community
test-merged-dataset-modular / README.md
DanqingZ's picture
DanqingZ
Upload merged LeRobot dataset (0-1 scoring system)
1f38755 verified
preview code
|
raw
history blame contribute delete
1.43 kB

test-merged-dataset-modular

This is a merged LeRobot dataset created using a modular 0-1 scoring system. Selected episodes from multiple source datasets were combined based on checkbox selection.

Dataset Information

  • Total Episodes: 4
  • Total Frames: 2079
  • Source Repositories: DanqingZ/eval_act_so100_test, DanqingZ/so100_test_6
  • Scoring System: 0-1 checkbox selection
  • Format Version: 1.0
  • Created With: pathonai-backend-modular

Scoring System

This dataset uses a 0-1 scoring system:

  • Score 1: Episodes selected via checkbox interface (good quality)
  • Score 0: Episodes not selected (filtered out)

Only episodes with score 1 are included in this merged dataset.

Usage 

from lerobot.common.datasets.lerobot_dataset import LeRobotDataset

# Load the dataset
dataset = LeRobotDataset("DanqingZ/test-merged-dataset-modular")

# Access episode data
episode_0 = dataset.get_episode(0)

License

This dataset combines data from multiple sources. Please check the licenses of the original datasets:

  • DanqingZ/eval_act_so100_test
  • DanqingZ/so100_test_6

Pipeline Steps

  1. Generate Judge: Created judge file with 0-1 scoring based on checkbox selection
  2. Data Cleaning: Filtered episodes with score 1 from each source repository
  3. Validation: Verified data integrity and episode structure
  4. Upload: Combined and uploaded to Hugging Face Hub