datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
PPTAgent-parsed_dataZenodo10K
PPTAgent/Zenodo10K
This is the dataset used in PPTAgent, crawled from zenodo.
To the best of our knowledge, it is the largest presentation dataset currently available, comprising over 10,000 PowerPoint (.pptx) files, all distributed under a clear and compliant license.
For more information, please visit our github repo.
dirname = f"zenodo-pptx/pptx/{task['license']}/{task['created'][:4]}/"
basename = f"{task['checksum'][4:]}-{task['filename']}"
filepath = dirname + basename
try:… See the full description on the dataset page: https://huggingface.co/datasets/Forceless/Zenodo10K.UltraPresent
UltraPresent
Project: https://github.com/icip-cas/PPTAgent
Paper: https://arxiv.org/abs/2602.22839
UltraPresent is a dataset for automatic presentation generation, designed to support research on large language model (LLM) and agent-based systems that generate structured slide presentations from natural language instructions.
The dataset contains a collection of diverse presentation generation tasks, where each example specifies a user instruction describing the topic, structure… See the full description on the dataset page: https://huggingface.co/datasets/Forceless/UltraPresent.ForceVLA-real-dataorange_cube_pnp_with_forceThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "franka_fer_xhand",
"total_episodes": 100,
"total_frames": 19340,
"total_tasks": 1,
"total_videos": 303,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 30,
"splits": {
"train": "0:100"
},
"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/wengmister/orange_cube_pnp_with_force.bimanual_assembly_insert_force_lerobot_mix854This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 30,
"features": {
"observation.images.ego": {
"dtype": "video",
"shape": [
640,
640,
3
],
"names": [
"height",
"width",
"channel"
],
"info": {
"video.height": 640… See the full description on the dataset page: https://huggingface.co/datasets/rpwang/bimanual_assembly_insert_force_lerobot_mix854.catheter_force_depth_eef
catheter_force_depth_eef
This dataset was created using LeRobot and is stored in the episode-split LeRobot v2.1 layout used by OpenPI.
Dataset Structure
The dataset uses parquet files for per-frame robot observations/actions and episode-split videos for camera streams.
meta/
info.json
tasks.jsonl
episodes.jsonl
stats.json
episodes_stats.jsonl
data/
chunk-000/
episode_000000.parquet
videos/
chunk-000/
<video_key>/
episode_000000.mp4… See the full description on the dataset page: https://huggingface.co/datasets/giakhuyendihoc/catheter_force_depth_eef.ForceFlow
ForceFlow Dataset
ForceFlow: Learning to Feel and Act via Contact-Driven Flow Matching
[Project Page] | [Code]
Contact-rich manipulation remains one of the hardest problems in robot learning: vision alone cannot capture the high-frequency contact dynamics that determine whether a plug seats correctly, a stamp triggers cleanly, or a wipe exerts consistent pressure. This dataset is collected to support ForceFlow, a force-aware reactive framework built on flow matching that… See the full description on the dataset page: https://huggingface.co/datasets/JokerESC/ForceFlow.omx_bilateral_force_v2This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 30,
"features": {
"action": {
"dtype": "float32",
"names": [
"shoulder_pan.pos",
"shoulder_lift.pos",
"elbow_flex.pos",
"wrist_flex.pos",
"wrist_roll.pos",
"gripper.pos"
],
"shape": [
6… See the full description on the dataset page: https://huggingface.co/datasets/masato-ka/omx_bilateral_force_v2.forcesight-demos
ForceSight Demos (6_3)
Teleoperated demonstration datasets in LeRobot
format (v2.1) for the ForceSight manipulation project. Recorded on a Franka Panda
with a parallel-jaw hand and paired vision + tactile sensing.
Companion fine-tuned pi0.5 policies: mlshehab/forcesight.
Tasks
Four manipulation tasks, teleoperated:
balance — balance a white box on top of blue beams.
gear — insert a gear onto a peg.
medicine — pick up the bottle and place it in the orange bin if… See the full description on the dataset page: https://huggingface.co/datasets/mlshehab/forcesight-demos.delta-force-bc-sft-shuffled
Delta Force BC SFT Dataset (Shuffled)
三角洲行动（Delta Force）PC 端键鼠操作的行为克隆（Behavior Cloning）SFT 数据集，专为 VLM（视觉语言模型）的多图多轮训练设计。
数据规模
指标
数值
Parquet 文件数
19
总轨迹数（数据集 size）
3,797
每条轨迹包含的图像数
20
每条轨迹的对话轮数
1 system + 20 human + 20 gpt
每张图分辨率
1280 × 720 (JPEG)
每张图大小
~260 KB
总数据大小
~18.3 GB
总图像数（= 训练时实际预测目标数）
75,940
时间覆盖
每条轨迹 = 4 秒连续操作（20 帧 × 200ms）
一条 trajectory 对应玩家连续 4 秒的操作。 训练时 1 行 parquet = 1 个 multi-turn 训练样本，但每个样本里 20 个 gpt 回复都参与 loss 计算。… See the full description on the dataset page: https://huggingface.co/datasets/zhiyuanhucs/delta-force-bc-sft-shuffled.libero_goal_force_openvla_predicted_action_fixedtest_no_forcebimanual_assembly_insert_force_lerobotForceASR
owner: RK 2423 FRASCAL (https://zenodo.org/records/7445749)
license: cc-by-4.0
data_production:
type: simulation
physics: phase-field fracture models for brittle fracture
script: Subset 'res-SENS' of the initial dataset, 1/5th time steps, converted to
PLAID format for standardized access; no changes to data content.
num_samples:
res_SENS: 28
storage_backend: hf_datasets
This dataset was generated with plaid, we refer to this documentation for additional details on how to… See the full description on the dataset page: https://huggingface.co/datasets/PhysArena/ForceASR.lerobot_test_force_sensorThis dataset was created using LeRobot.
Dataset Description
A quick test to see force measurements on the pressure sensor installed in the gripper's lower jaw of LeRobot SO-101 arm. The bot is being teleoperated and squeezing a large marshmallow.
Note the 7th dimension of the observations (the gripping force); default so-100/so-101 arms only have 6 observations, one for each motor.
See more about the project:
https://github.com/ayerone/lerobot_robot_so_sensor_arm… See the full description on the dataset page: https://huggingface.co/datasets/primordial-spork/lerobot_test_force_sensor.libero_spatial_force_fixedpptagent-code26komx_bilateral_forceThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 30,
"features": {
"action": {
"dtype": "float32",
"names": [
"shoulder_pan.pos",
"shoulder_lift.pos",
"elbow_flex.pos",
"wrist_flex.pos",
"wrist_roll.pos",
"gripper.pos"
],
"shape": [
6… See the full description on the dataset page: https://huggingface.co/datasets/masato-ka/omx_bilateral_force.ur5_ur5_force_sensor_testThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "ur5",
"total_episodes": 4,
"total_frames": 2401,
"total_tasks": 2,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 200,
"fps": 30,
"splits": {
"train": "0:4"
},
"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/DORLR/ur5_ur5_force_sensor_test.lerobot-record-force-test_20260822_163839This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 30,
"features": {
"action": {
"dtype": "float32",
"names": [
"shoulder_pan.pos",
"shoulder_lift.pos",
"elbow_flex.pos",
"wrist_flex.pos",
"wrist_roll.pos",
"gripper.pos"
],
"shape": [
6… See the full description on the dataset page: https://huggingface.co/datasets/masato-ka/lerobot-record-force-test_20260822_163839.lerobot_test_force_sensor_limitingThis dataset was created using LeRobot.
Dataset Description
A short recording of teleoperating the so_sensor_arm with a modified record script. The modification limits the teleoperator action based on the robot's gripper sensor reading; if there already force on the gripper, the action is clipped at the current gripper position (based on the robot's observed state).
The force threshold was set to 20 (which corresponds to about 90 grams) but you can see that in reality the robot… See the full description on the dataset page: https://huggingface.co/datasets/primordial-spork/lerobot_test_force_sensor_limiting.Forces_are_not_enough
Cite this dataset Fu, X., Wu, Z., Wang, W., Xie, T., Keten, S., Gomez-Bombarelli, R., and Jaakkola, T. Forces are not enough. ColabFit, 2023. https://doi.org/10.60732/62c08514
This dataset has been curated and formatted for the ColabFit Exchange
This dataset is also available on the ColabFit Exchange:
https://materials.colabfit.org/id/DS_tku3ae1rtxiy_0
Visit the ColabFit Exchange to search additional datasets by author, description, element content… See the full description on the dataset page: https://huggingface.co/datasets/colabfit/Forces_are_not_enough.ablation_force_doubt_logic_DeepSeek_R1_Distill_Qwen_7Bablation_force_doubt_logic_Phi_4_reasoning_plusj_ablation_force_doubt_logic_EXAONE_Deep_32Bsample-force-alignment-datasetplug_baseline_high_stiff_3_no_forceafrica-egypt-capmas-labor-force-survey-annual-7240deb5
Labor Force Survey - Annual | Africa (CAPMAS Egypt Open Data)
8,213 rows - 1 Africa country/area - 2010-2024 - 55 indicators - Engineered by Electric Sheep Africa
TL;DR
This dataset contains 8,213 rows from CAPMAS Egypt Open Data, covering Labor Force Survey - Annual. It is published as ML-ready Parquet with consistent Hugging Face metadata, source provenance, and analysis-friendly loading examples.
What This Dataset Measures
Labour and… See the full description on the dataset page: https://huggingface.co/datasets/electricsheepafrica/africa-egypt-capmas-labor-force-survey-annual-7240deb5.give_forcep_40This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "so101_follower",
"total_episodes": 100,
"total_frames": 44415,
"total_tasks": 1,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 500,
"fps": 30,
"splits": {
"train": "0:100"
},
"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/yamilama/give_forcep_40.