datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
robomimic_datasetsrobomimic is a framework for robot learning from demonstration.
It offers a broad set of demonstration datasets collected on robot manipulation domains and offline learning algorithms to learn from these datasets.
robomimic aims to make robot learning broadly accessible and reproducible, allowing researchers and practitioners to benchmark tasks and algorithms fairly and to develop the next generation of robot learning algorithms.
This repository contains some of the datasets released with the… See the full description on the dataset page: https://huggingface.co/datasets/robomimic/robomimic_datasets.robomimic_sim_square_noisyThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": null,
"total_episodes": 5000,
"total_frames": 403454,
"total_tasks": 1,
"total_videos": 10000,
"total_chunks": 5,
"chunks_size": 1000,
"fps": 20,
"splits": {
"train": "0:5000"
},
"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/iantc104/robomimic_sim_square_noisy.robomimic-pretrain-data
From Prior to Pro: Efficient Skill Mastery via Distribution Contractive RL Finetuning (DICE-RL)
This repository contains the datasets used in the paper From Prior to Pro: Efficient Skill Mastery via Distribution Contractive RL Finetuning.
Project Website | GitHub Repository
Dataset Description
Distribution Contractive Reinforcement Learning (DICE-RL) is a framework that uses reinforcement learning (RL) to refine pretrained generative robot policies. This repository… See the full description on the dataset page: https://huggingface.co/datasets/wintermelontree/robomimic-pretrain-data.robomimic-mh-square-imageThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "robomimic",
"total_episodes": 300,
"total_frames": 80731,
"total_tasks": 1,
"total_videos": 600,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 20,
"splits": {
"train": "0:300"
},
"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/ankile/robomimic-mh-square-image.robomimic12_absRobomimic-BiInsertrobomimic-ph-lift-imageThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "robomimic",
"total_episodes": 200,
"total_frames": 9666,
"total_tasks": 1,
"total_videos": 400,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 20,
"splits": {
"train": "0:200"
},
"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/ankile/robomimic-ph-lift-image.robomimic-mh-lift-imageThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "robomimic",
"total_episodes": 300,
"total_frames": 31127,
"total_tasks": 1,
"total_videos": 600,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 20,
"splits": {
"train": "0:300"
},
"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/ankile/robomimic-mh-lift-image.robomimic_sim_transportThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": null,
"total_episodes": 164,
"total_frames": 76007,
"total_tasks": 1,
"total_videos": 492,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 20,
"splits": {
"train": "0:164"},
"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/iantc104/robomimic_sim_transport.robomimic_transport_50This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "robomimic",
"total_episodes": 50,
"total_frames": 23720,
"total_tasks": 1,
"total_videos": 200,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 20,
"splits": {
"train": "0:50"
},
"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/wintermelontree/robomimic_transport_50.robomimic-finetune-datarobomimic-ph-can-imageThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "robomimic",
"total_episodes": 200,
"total_frames": 23207,
"total_tasks": 1,
"total_videos": 400,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 20,
"splits": {
"train": "0:200"
},
"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/ankile/robomimic-ph-can-image.robomimic-can-paired-lerobotThis dataset was created using LeRobot.
Dataset Description
robomimic can-paired, converted to the LeRobot v3.0 format with success/failure labels.
The original robomimic can-paired dataset contains 200 teleoperated
demonstrations of the robosuite PickPlaceCan task (Panda robot): 100 paired task initializations,
each with one successful demo (the can is picked up and placed in the correct bin) and one
failed demo (the can is picked up and tossed outside the robot… See the full description on the dataset page: https://huggingface.co/datasets/geonmin-kim/robomimic-can-paired-lerobot.robomimic-mh-can-imageThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "robomimic",
"total_episodes": 300,
"total_frames": 62756,
"total_tasks": 1,
"total_videos": 600,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 20,
"splits": {
"train": "0:300"
},
"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/ankile/robomimic-mh-can-image.robomimic-ph-transport-imageThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "robomimic",
"total_episodes": 200,
"total_frames": 93752,
"total_tasks": 1,
"total_videos": 600,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 20,
"splits": {
"train": "0:200"
},
"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/ankile/robomimic-ph-transport-image.robomimic_tool_hang_50This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "robomimic",
"total_episodes": 50,
"total_frames": 27939,
"total_tasks": 1,
"total_videos": 50,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 20,
"splits": {
"train": "0:50"
},
"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/wintermelontree/robomimic_tool_hang_50.robomimic_sim_square_ph_low_dimrobomimic-mh-transport-imageThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "robomimic",
"total_episodes": 300,
"total_frames": 195800,
"total_tasks": 1,
"total_videos": 600,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 20,
"splits": {
"train": "0:300"
},
"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/ankile/robomimic-mh-transport-image.robomimic_ph__lift_can_square_toolhangThis dataset was created using LeRobot.
Dataset Description
From repo: https://robomimic.github.io/docs/datasets/robomimic_v0.1.html
four tasks included: lift the cube, pick and place can, square and tool hanging (each task has 200 episodes)
The collected trajectories are only from Proficient-Human (PH) operator.
This dataset can be directly used to train VLA models such as openpi. (lerobot v2.1 is more compatible, this dataset is of v3.0)
Homepage: [More Information Needed]
Paper:… See the full description on the dataset page: https://huggingface.co/datasets/yananchen/robomimic_ph__lift_can_square_toolhang.robomimic_sim_can_noisyThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": null,
"total_episodes": 5000,
"total_frames": 328042,
"total_tasks": 1,
"total_videos": 10000,
"total_chunks": 5,
"chunks_size": 1000,
"fps": 20,
"splits": {
"train": "0:5000"},
"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/iantc104/robomimic_sim_can_noisy.robomimic
robomimic
Dataset Summary
This dataset contains demonstrations of robosuite tasks, separated into two directories: images and dino3-embeddings.The demonstrations in both of the directories are exactly the same. They differ by the format in which they store camera observations.
images stores them as images, where dino3-embeddings stores the embeddings of those images, outputed by the small variant of the DINOv3 model.
The dataset uses robotview and eye_in_hand… See the full description on the dataset page: https://huggingface.co/datasets/aboguszewski/robomimic.robomimic-ph-square-imageThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "robomimic",
"total_episodes": 200,
"total_frames": 30154,
"total_tasks": 1,
"total_videos": 400,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 20,
"splits": {
"train": "0:200"
},
"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/ankile/robomimic-ph-square-image.robomimic-ph-square-stateThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "robomimic",
"total_episodes": 200,
"total_frames": 30154,
"total_tasks": 1,
"total_videos": 0,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 20,
"splits": {
"train": "0:200"
},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path": null,
"features": {… See the full description on the dataset page: https://huggingface.co/datasets/hgaurav/robomimic-ph-square-state.robomimic_sim_lift_ph_low_dimrobomimic_tool_hang_50_v2This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "robomimic",
"total_episodes": 50,
"total_frames": 27939,
"total_tasks": 1,
"total_videos": 100,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 20,
"splits": {
"train": "0:50"
},
"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/wintermelontree/robomimic_tool_hang_50_v2.robomimic-ph-can-stateThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "robomimic",
"total_episodes": 200,
"total_frames": 23207,
"total_tasks": 1,
"total_videos": 0,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 20,
"splits": {
"train": "0:200"
},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path": null,
"features": {… See the full description on the dataset page: https://huggingface.co/datasets/hgaurav/robomimic-ph-can-state.robomimic-lift-ph-lerobot-v3
robomimic HDF5 → verified LeRobot v3 episodes
Before → after: raw robomimic HDF5 demonstrations become validated LeRobot v3.0 episodes while original actions and episode boundaries are retained. Convert HDF5 free →
Community conversion produced by ViaCatalyst BYOD. This repository is not an official upstream release and is not affiliated with the robomimic authors.
This is a compact, provenance-complete conversion of the first 10 episodes from the pinned robomimic Lift PH… See the full description on the dataset page: https://huggingface.co/datasets/ViaCatalyst/robomimic-lift-ph-lerobot-v3.robomimic-can-ph-lerobot-v3
RoboTurk Can PH HDF5 → verified LeRobot v3 episodes
Before → after: nested RoboTurk/robomimic HDF5 demonstrations become validated LeRobot v3.0 episodes while original actions and episode boundaries are retained. Convert robomimic HDF5 free →
Community conversion produced by ViaCatalyst BYOD. This repository is not an official upstream release and is not affiliated with the robomimic or RoboTurk authors.
This is a compact, provenance-complete conversion of the first 10… See the full description on the dataset page: https://huggingface.co/datasets/ViaCatalyst/robomimic-can-ph-lerobot-v3.robomimic-mh-square-state-resfit-v22This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "robomimic",
"total_episodes": 300,
"total_frames": 80731,
"total_tasks": 1,
"total_videos": 0,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 20,
"splits": {
"train": "0:300"
},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path": null,
"features": {… See the full description on the dataset page: https://huggingface.co/datasets/ankile/robomimic-mh-square-state-resfit-v22.robomimic_lift_v15_5This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": null,
"total_episodes": 5,
"total_frames": 280,
"total_tasks": 1,
"total_videos": 10,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 20,
"splits": {
"train": "0:5"
},
"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/YSL2683/robomimic_lift_v15_5.