datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
mimicgen_hardThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "panda",
"total_episodes": 11000,
"total_frames": 3063751,
"total_tasks": 9,
"total_videos": 0,
"total_chunks": 11,
"chunks_size": 1000,
"fps": 20,
"splits": {
"train": "0:11000"
},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/mishmish66/mimicgen_hard.dexmimicgen_datasets
DexMimicGen Datasets
This repository contains the official dataset release of simulation environments and datasets for the ICRA 2025 paper "DexMimicGen: Automated Data Generation for Bimanual Dexterous Manipulation via Imitation Learning".
Website: https://dexmimicgen.github.io
For business inquiries, please submit this form: NVIDIA Research Licensing
mimicgen_hard_100This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "panda",
"total_episodes": 1100,
"total_frames": 304988,
"total_tasks": 9,
"total_videos": 0,
"total_chunks": 2,
"chunks_size": 1000,
"fps": 20,
"splits": {
"train": "0:1100"
},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/mishmish66/mimicgen_hard_100.mimicgen_coffee_preparation_d1This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "panda",
"total_episodes": 1000,
"total_frames": 224403,
"total_tasks": 1,
"total_videos": 0,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 20,
"splits": {
"train": "0:1000"
},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/mishmish66/mimicgen_coffee_preparation_d1.DAH_mimicgen_all_12tasks_50demosThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "panda",
"total_episodes": 600,
"total_frames": 213225,
"total_tasks": 12,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 200,
"fps": 20,
"splits": {
"train": "0:600"
},
"data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/JianZhou0420/DAH_mimicgen_all_12tasks_50demos.mimicgen_coffee_d0This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "panda",
"total_episodes": 1000,
"total_frames": 223130,
"total_tasks": 1,
"total_videos": 0,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 20,
"splits": {
"train": "0:1000"
},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/mishmish66/mimicgen_coffee_d0.mimicgen_source_12tasks_v21This dataset was created using LeRobot.
Dataset Description
This dataset is in lerobot format of version v2.1. (feel free to upgrade it to v3.0 using lerobot cli command)
It only includes the part of Core from https://robomimic.github.io/docs/datasets/mimicgen.html.
Therefore, in total, there are 120 human demonstrations / trajectories across 12 tasks.
Upstream files are from mimicgen :
coffee.hdf5 hammer_cleanup.hdf5 mug_cleanup.hdf5 pick_place.hdf5… See the full description on the dataset page: https://huggingface.co/datasets/yananchen/mimicgen_source_12tasks_v21.mimicgen_coffee_d1This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "panda",
"total_episodes": 1000,
"total_frames": 224403,
"total_tasks": 1,
"total_videos": 0,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 20,
"splits": {
"train": "0:1000"
},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/mishmish66/mimicgen_coffee_d1.mimicgen_hard_50This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "panda",
"total_episodes": 550,
"total_frames": 152810,
"total_tasks": 9,
"total_videos": 0,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 20,
"splits": {
"train": "0:550"
},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/mishmish66/mimicgen_hard_50.mimicgen_square_d0_pandaBlockPAP-v1_MimicGen
BlockPAP-v01MimicGenLeRobot v2.1 dataset for BlockPAP-v1 MimicGen trajectories.
❗️ State Alignment between Simulation and Real Robot
In this version of the MimicGen dataset, corresponding biases have been applied to state[2] and state[5] during data generation, unifying the distributions of sim into the states required for a real robot.
stacking_mimicgen