datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
figofigofigofigoRoVid-X
Rethinking Video Generation Model for the Embodied World
If you like our project, please give us a star ⭐ on GitHub for the latest update.
Key features
4M robotic video clips(10K+ hours) for large-scale video generation training.
1300+ fine-grained robotic skills, covering diverse actions and task primitives.
Multi-modal physical annotations, including RGB, depth, and optical flow.
Multi-robot and multi-task diversity… See the full description on the dataset page: https://huggingface.co/datasets/DAGroup-PKU/RoVid-X.dAgger_build_block_tower_1.1.0-advantages
Advantage Values for villekuosmanen/dAgger_build_block_tower_1.1.0
Pre-computed advantage values for offline RL training.
Source
Dataset: villekuosmanen/dAgger_build_block_tower_1.1.0
Value Model: villekuosmanen/rewact_build_block_tower_all_3
N-step lookahead: 50
Files
This dataset contains per-episode parquet files with advantage values for each frame.
Usage
from pathlib import Path
import pandas as pd
# Load advantages for a specific episode… See the full description on the dataset page: https://huggingface.co/datasets/villekuosmanen/dAgger_build_block_tower_1.1.0-advantages.dAgger_build_block_tower_1.0.0-advantages
Advantage Values for villekuosmanen/dAgger_build_block_tower_1.0.0
Pre-computed advantage values for offline RL training.
Source
Dataset: villekuosmanen/dAgger_build_block_tower_1.0.0
Value Model: villekuosmanen/rewact_build_block_tower_all_3
N-step lookahead: 50
Files
This dataset contains per-episode parquet files with advantage values for each frame.
Usage
from pathlib import Path
import pandas as pd
# Load advantages for a specific episode… See the full description on the dataset page: https://huggingface.co/datasets/villekuosmanen/dAgger_build_block_tower_1.0.0-advantages.dagger_final_1_21This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "openarms_follower",
"total_episodes": 1,
"total_frames": 841,
"total_tasks": 1,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 200,
"fps": 30,
"splits": {
"train": "0:1"
},
"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/lerobot-data-collection/dagger_final_1_21.dAgger_build_block_tower_1.4.0-advantages
Advantage Values for villekuosmanen/dAgger_build_block_tower_1.4.0
Pre-computed advantage values for offline RL training.
Source
Dataset: villekuosmanen/dAgger_build_block_tower_1.4.0
Value Model: villekuosmanen/rewact_build_block_tower_all_3
N-step lookahead: 50
Files
This dataset contains per-episode parquet files with advantage values for each frame.
Usage
from pathlib import Path
import pandas as pd
# Load advantages for a specific episode… See the full description on the dataset page: https://huggingface.co/datasets/villekuosmanen/dAgger_build_block_tower_1.4.0-advantages.TS-Arena-Archive
TS-Arena Archive
This dataset contains historical time-series data from the TS-Arena. It is provided for offline model evaluation and backtesting. Additional domains (such as urban mobility) and regions will be added as they become available on the TS-Arena platform.
Evaluation Rules
Researchers using this data must follow these guidelines:
Self-Reported Status: It must be clearly stated in any publication that results are based on self-obtained evaluation using… See the full description on the dataset page: https://huggingface.co/datasets/DAG-UPB/TS-Arena-Archive.navigation-corpus-dagbani-speech
This dataset is shared under CC BY-NC 4.0, which means you are free to use, share, and adapt it for non-commercial research and educational purposes with attribution. You can read the full license at https://creativecommons.org/licenses/by-nc/4.0/.
Dag Speech Segments (sentence splitting)
52799 speech-text pairs split from long recordings.
Processing pipeline
Source audio from ghananlpcommunity/navigation-corpus-speech-full-dagbani
Full-file CTC forced… See the full description on the dataset page: https://huggingface.co/datasets/ghanaopenai/navigation-corpus-dagbani-speech.RBench
Rethinking Video Generation Model for the Embodied World
🔍 Benchmark Overview
The benchmark is constructed from two complementary perspectives: task categories and robot embodiment types, covering a total of 650 image-text evaluation cases.
🧩 Task-Oriented Evaluation Set (5 Tasks)
The task-oriented split contains 250 image-text pairs, with 50 samples per task, spanning five representative robotic task categories:
Common… See the full description on the dataset page: https://huggingface.co/datasets/DAGroup-PKU/RBench.dagongbaodAgger_build_block_tower_1.3.0-advantages
Advantage Values for villekuosmanen/dAgger_build_block_tower_1.3.0
Pre-computed advantage values for offline RL training.
Source
Dataset: villekuosmanen/dAgger_build_block_tower_1.3.0
Value Model: villekuosmanen/rewact_build_block_tower_all_3
N-step lookahead: 50
Files
This dataset contains per-episode parquet files with advantage values for each frame.
Usage
from pathlib import Path
import pandas as pd
# Load advantages for a specific episode… See the full description on the dataset page: https://huggingface.co/datasets/villekuosmanen/dAgger_build_block_tower_1.3.0-advantages.dAgger_pack_toothbrush_Nov26-advantages
Advantage Values for villekuosmanen/dAgger_pack_toothbrush_Nov26
Pre-computed advantage values for offline RL training.
Source
Dataset: villekuosmanen/dAgger_pack_toothbrush_Nov26
Value Model: villekuosmanen/rewact_toothbrush_pistar_1.5.0
N-step lookahead: 50
Files
This dataset contains per-episode parquet files with advantage values for each frame.
Usage
from pathlib import Path
import pandas as pd
# Load advantages for a specific episode… See the full description on the dataset page: https://huggingface.co/datasets/villekuosmanen/dAgger_pack_toothbrush_Nov26-advantages.dAgger_build_block_tower_1.2.0-advantages
Advantage Values for villekuosmanen/dAgger_build_block_tower_1.2.0
Pre-computed advantage values for offline RL training.
Source
Dataset: villekuosmanen/dAgger_build_block_tower_1.2.0
Value Model: villekuosmanen/rewact_build_block_tower_all_3
N-step lookahead: 50
Files
This dataset contains per-episode parquet files with advantage values for each frame.
Usage
from pathlib import Path
import pandas as pd
# Load advantages for a specific episode… See the full description on the dataset page: https://huggingface.co/datasets/villekuosmanen/dAgger_build_block_tower_1.2.0-advantages.cucumber-peel-DAgger-iter2-v2-trim
cucumber-peel-DAgger-iter2-v2-trim
Materialized collection — 188 episodes · 58,994 frames @ 20 fps (~49 min of demonstration).
Collection cucumber-peel-DAgger-iter2@v2 (frozen 2026-07-22), mode trim_rebuild — built by Vibe Data Studio; the machine-readable recipe in meta/vibedata/collection.json makes this dataset reproducible from its pinned components.
Tasks
Instruction
Episodes
peel the skin of the cucumber with several strokes, starting close to the… See the full description on the dataset page: https://huggingface.co/datasets/VibeCuisine/cucumber-peel-DAgger-iter2-v2-trim.cucumber-place-DAgger-iter2-doris070926-v1-trim
cucumber-place-DAgger-iter2-doris070926-v1-trim
Materialized collection — 117 episodes · 4,677 frames @ 20 fps (~4 min of demonstration).
Collection cucumber-place-DAgger-iter2-doris070926@v1 (frozen 2026-07-09), mode trim_rebuild — built by Vibe Data Studio; the machine-readable recipe in meta/vibedata/collection.json makes this dataset reproducible from its pinned components.
Tasks
Instruction
Episodes
place the cucumber on the middle of the cutting… See the full description on the dataset page: https://huggingface.co/datasets/VibeCuisine/cucumber-place-DAgger-iter2-doris070926-v1-trim.cucumber-place-DAgger-iter1-doris070726-v1-trim
cucumber-place-DAgger-iter1-doris070726-v1-trim
Materialized collection — 93 episodes · 3,546 frames @ 20 fps (~3 min of demonstration).
Collection cucumber-place-DAgger-iter1-doris070726@v1 (frozen 2026-07-08), mode trim_rebuild — built by Vibe Data Studio; the machine-readable recipe in meta/vibedata/collection.json makes this dataset reproducible from its pinned components.
Tasks
Instruction
Episodes
place the cucumber on the middle of the cutting… See the full description on the dataset page: https://huggingface.co/datasets/VibeCuisine/cucumber-place-DAgger-iter1-doris070726-v1-trim.cucumber-place-DAgger-iter0-trim-doris070126
cucumber-place-DAgger-iter0-trim-doris070126
Materialized collection — 87 episodes · 6,098 frames @ 20 fps (~5 min of demonstration).
Collection cucumber-place-DAgger-iter0-trim-doris070126@v1 (frozen 2026-07-01), mode trim_rebuild — built by Vibe Data Studio; the machine-readable recipe in meta/vibedata/collection.json makes this dataset reproducible from its pinned components.
Tasks
Instruction
Episodes
place the cucumber on the middle of the cutting… See the full description on the dataset page: https://huggingface.co/datasets/VibeCuisine/cucumber-place-DAgger-iter0-trim-doris070126.cucumber-subtask-grab-DAgger-iter2-task-unified_aux
cucumber-subtask-grab-DAgger-iter2-task-unified_aux
Materialized collection — 150 episodes · 13,484 frames @ 20 fps (~11 min of demonstration).
Collection cucumber-subtask-grab-DAgger-iter2@v1 (frozen 2026-06-25), mode trim_rebuild — built by Vibe Data Studio; the machine-readable recipe in meta/vibedata/collection.json makes this dataset reproducible from its pinned components.
Tasks
Instruction
Episodes
grab the cucumber close to one of the cucumber's… See the full description on the dataset page: https://huggingface.co/datasets/VibeCuisine/cucumber-subtask-grab-DAgger-iter2-task-unified_aux.cucumber-place-DAgger-iter0-trim-doris063026
cucumber-place-DAgger-iter0-v1-trim
Materialized collection — 88 episodes · 12,149 frames @ 20 fps (~10 min of demonstration).
Collection cucumber-place-DAgger-iter0@v1 (frozen 2026-06-29), mode trim_rebuild — built by Vibe Data Studio; the machine-readable recipe in meta/vibedata/collection.json makes this dataset reproducible from its pinned components.
Tasks
Instruction
Episodes
place the cucumber on the middle of the cutting board, aligned… See the full description on the dataset page: https://huggingface.co/datasets/VibeCuisine/cucumber-place-DAgger-iter0-trim-doris063026.cucumber-peel-DAgger-iter0-v1-trim
cucumber-peel-DAgger-iter0-v1-trim
Materialized collection — 42 episodes · 15,870 frames @ 20 fps (~13 min of demonstration).
Collection cucumber-peel-DAgger-iter0@v1 (frozen 2026-07-17), mode trim_rebuild — built by Vibe Data Studio; the machine-readable recipe in meta/vibedata/collection.json makes this dataset reproducible from its pinned components.
Tasks
Instruction
Episodes
peel the skin of the cucumber with several strokes, starting close to the… See the full description on the dataset page: https://huggingface.co/datasets/VibeCuisine/cucumber-peel-DAgger-iter0-v1-trim.cucumber-place-DAgger-iter4-doris072326-v2-trim
cucumber-place-DAgger-iter4-doris072326-v2-trim
Materialized collection — 264 episodes · 10,775 frames @ 20 fps (~9 min of demonstration).
Collection cucumber-place-DAgger-iter4-doris072326@v2 (frozen 2026-07-24), mode trim_rebuild — built by Vibe Data Studio; the machine-readable recipe in meta/vibedata/collection.json makes this dataset reproducible from its pinned components.
Tasks
Instruction
Episodes
place the cucumber on the middle of the cutting… See the full description on the dataset page: https://huggingface.co/datasets/VibeCuisine/cucumber-place-DAgger-iter4-doris072326-v2-trim.cucumber-subtask-grab-DAgger-iter1-v1-trim
cucumber-subtask-grab-DAgger-iter1-v1-trim
Materialized collection — 100 episodes · 9,700 frames @ 20 fps (~8 min of demonstration).
Collection cucumber-subtask-grab-DAgger-iter1@v1 (frozen 2026-06-24), mode trim_rebuild — built by Vibe Data Studio; the machine-readable recipe in meta/vibedata/collection.json makes this dataset reproducible from its pinned components.
Tasks
Instruction
Episodes
grab
55
grab the cucumber close to one of the cucumber's… See the full description on the dataset page: https://huggingface.co/datasets/VibeCuisine/cucumber-subtask-grab-DAgger-iter1-v1-trim.cucumber-grab-DAgger-iter4-v3-trim
cucumber-grab-DAgger-iter4-v3-trim
Materialized collection — 201 episodes · 16,387 frames @ 20 fps (~14 min of demonstration).
Collection cucumber-grab-DAgger-iter4@v3 (frozen 2026-07-24), mode trim_rebuild — built by Vibe Data Studio; the machine-readable recipe in meta/vibedata/collection.json makes this dataset reproducible from its pinned components.
Tasks
Instruction
Episodes
grab the cucumber close to one of the cucumber's end
201
Episodes… See the full description on the dataset page: https://huggingface.co/datasets/VibeCuisine/cucumber-grab-DAgger-iter4-v3-trim.smoke-dagger-20260729b
smoke-dagger-20260729b
Recorded dataset — captured on jetson1 — 2 episodes · 876 frames @ 20 fps (~1 min of demonstration).
Tasks
Instruction
Episodes
grab the cucumber
2
Recording
Rig
jetson1 (calibration sidecar)
Recorded
2026-07-29
Operator
jeremy
Episode sources
5 policy · 1 teleop
Assisting policies
VibeCuisine/jetson1-act-grab-DAgger-iter3-062926@65edddcc, code:vibedata_core.recording.skills:NudgeSkill@v1… See the full description on the dataset page: https://huggingface.co/datasets/VibeCuisine/smoke-dagger-20260729b.cucumber-peel-DAgger-iter1-adaptive1-v1-trimThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "vibeboard_follower_tilt",
"total_episodes": 78,
"total_frames": 28053,
"total_tasks": 1,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 200,
"fps": 20,
"splits": {
"train": "0:78"
},
"data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet"… See the full description on the dataset page: https://huggingface.co/datasets/VibeCuisine/cucumber-peel-DAgger-iter1-adaptive1-v1-trim.cucumber-subtask-grab-DAgger-iter2-task-unified-aux-collection-v1-flatThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 20,
"features": {
"action": {
"dtype": "float32",
"names": [
"shoulder_pan.pos",
"shoulder_lift.pos",
"elbow_flex.pos",
"wrist_flex.pos",
"wrist_roll.pos",
"gripper.pos",
"tilt.pos"
]… See the full description on the dataset page: https://huggingface.co/datasets/VibeCuisine/cucumber-subtask-grab-DAgger-iter2-task-unified-aux-collection-v1-flat.cucumber-peel-DAgger-iter1-v1-trimThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "vibeboard_follower_tilt",
"total_episodes": 70,
"total_frames": 26265,
"total_tasks": 1,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 200,
"fps": 20,
"splits": {
"train": "0:70"
},
"data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet"… See the full description on the dataset page: https://huggingface.co/datasets/VibeCuisine/cucumber-peel-DAgger-iter1-v1-trim.cucumber-subtask-grab-DAgger-iter2-task-unifiedThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 20,
"features": {
"action": {
"dtype": "float32",
"names": [
"shoulder_pan.pos",
"shoulder_lift.pos",
"elbow_flex.pos",
"wrist_flex.pos",
"wrist_roll.pos",
"gripper.pos",
"tilt.pos"
]… See the full description on the dataset page: https://huggingface.co/datasets/VibeCuisine/cucumber-subtask-grab-DAgger-iter2-task-unified.cucumber-place-DAgger-iter0-v1-flatThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "vibeboard_follower_tilt",
"total_episodes": 113,
"total_frames": 15688,
"total_tasks": 1,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 200,
"fps": 20,
"splits": {
"train": "0:113"
},
"data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet"… See the full description on the dataset page: https://huggingface.co/datasets/VibeCuisine/cucumber-place-DAgger-iter0-v1-flat.cucumber-subtask-grab-DAgger-iter1-v1-trim_auxThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 20,
"features": {
"action": {
"dtype": "float32",
"names": [
"shoulder_pan.pos",
"shoulder_lift.pos",
"elbow_flex.pos",
"wrist_flex.pos",
"wrist_roll.pos",
"gripper.pos",
"tilt.pos"
]… See the full description on the dataset page: https://huggingface.co/datasets/VibeCuisine/cucumber-subtask-grab-DAgger-iter1-v1-trim_aux.