datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
jepa-qwen3-32b-pure-baselines-2026-05-25
JEPA-Align: Qwen3-32B Safety Defense Matrix
The complete 11-condition Qwen3-32B experiment for Predictive Representation
Alignment (PRA), the paired-view objective introduced in Predictive
Representation Alignment Improves Generalization in LLM Safety.
PRA aligns adversarially rewritten prompts with clean prompts expressing the
same intent. This release contains trained adapters, attack traces, benign
capability evaluations, machine-readable results, and paper-ready tables for… See the full description on the dataset page: https://huggingface.co/datasets/memo-ozdincer/jepa-qwen3-32b-pure-baselines-2026-05-25.jepa-wms
🌍 JEPA-WMs Datasets
Robotics trajectories for world model training 🤖
Meta AI Research, FAIR
This 🤗 HuggingFace repository hosts datasets for training JEPA-WM world models.
👉 See the main repository for training code and pretrained models.
👁️ Preview Images: To view example images in the Dataset Viewer above, select a dataset configuration (e.g., metaworld, pusht) and click "Run query".
📦 Downloading Data
Use the download script… See the full description on the dataset page: https://huggingface.co/datasets/facebook/jepa-wms.repro-causal-jepa-learning-world-models-through-object-level-latent-masking-traces
Agent traces
Agent sessions published from a Trackio Logbook.
CGM-JEPA-Downstream
CGM-JEPA Downstream Evaluation Splits
Paper | Code
Labeled cohort splits used to evaluate CGM encoders on two binary metabolic outcomes — insulin resistance and β-cell dysfunction — in the paper CGM-JEPA: Learning Consistent Continuous Glucose Monitor Representations via Predictive Self-Supervised Pretraining.
Downstream-only. For the unlabeled pretraining corpus (Stanford + Colas), see CRUISEResearchGroup/CGM-JEPA-Pretraining. For pretrained encoder weights, see… See the full description on the dataset page: https://huggingface.co/datasets/CRUISEResearchGroup/CGM-JEPA-Downstream.fin-jepa-study0
Fin-JEPA Study 0: XBRL Financial Distress Dataset
A tabular dataset of 35,709 company-year observations for
4,520 unique SEC filers (2012--2023), linking XBRL financial
statement features extracted from 10-K filings to five binary distress
outcomes. Built for the first gate of the
Financial JEPA project.
Quick Start
from datasets import load_dataset
ds = load_dataset("elroyg/fin-jepa-study0")
print(ds)
# DatasetDict({
# train: Dataset(num_rows=14748, ...),
#… See the full description on the dataset page: https://huggingface.co/datasets/elroyg/fin-jepa-study0.rollout_vla_jepa_so101_d6231_step25000_large_green_cube_sync_0810_1125_20260810_112659This 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/geonmin-kim/rollout_vla_jepa_so101_d6231_step25000_large_green_cube_sync_0810_1125_20260810_112659.Bar-JEPA
Bar Chart Annotation Dataset
This dataset accompanies the paper Bar-JEPA: Extracting Values from Bar Chart with Joint-Embedding Predictive Architecture. Code is available at https://github.com/dralois/Bar-JEPA.
A collection of synthetic bar chart images with detailed structural annotations. Split into three subsets for different stages of a chart understanding pipeline.
Subsets
Subset
Folder
Train
Test
encoder_training
hf_ready/encoder_training/
100… See the full description on the dataset page: https://huggingface.co/datasets/dralois/Bar-JEPA.rollout_vla_jepa_so101_pick_toy_v9_2_plate_20260701_183313This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 15,
"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/kamada-riku/rollout_vla_jepa_so101_pick_toy_v9_2_plate_20260701_183313.rollout_vla_jepa_so101_pick_toy_v9_plate_20260630_210212This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 15,
"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/kamada-riku/rollout_vla_jepa_so101_pick_toy_v9_plate_20260630_210212.rollout_vla_jepa_so101_pick_toy_v9_plate_20260630_205933This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 15,
"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/kamada-riku/rollout_vla_jepa_so101_pick_toy_v9_plate_20260630_205933.rollout_vla_jepa_so101_cup_toy_to_plate_v1.1_20260703_200718This 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/kamada-riku/rollout_vla_jepa_so101_cup_toy_to_plate_v1.1_20260703_200718.rollout_vla_jepa_so101_d6231_mdrift_step20000_large_green_cube_sync_0810_1131_20260810_113240This 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/geonmin-kim/rollout_vla_jepa_so101_d6231_mdrift_step20000_large_green_cube_sync_0810_1131_20260810_113240.mimic-iv-echo-jepa-embeddings
MIMIC-IV-Echo V-JEPA & EchoJEPA Embeddings
Pre-computed video embeddings for all ~525K MIMIC-IV-Echo echocardiography clips using V-JEPA2 / V-JEPA2.1 and EchoJEPA fine-tuned variants.
Access requirement: You must have PhysioNet credentialed access and a signed DUA for MIMIC-IV-Echo before requesting access.
Models
Folder names map directly to source checkpoints. V-JEPA 2 and V-JEPA 2.1 are distinct — see the Description column. EchoJEPA models are all V-JEPA 2.1… See the full description on the dataset page: https://huggingface.co/datasets/MITCriticalData/mimic-iv-echo-jepa-embeddings.rollout_vla_jepa_so101_pick_toy_v9_plate_20260630_211916This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 15,
"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/kamada-riku/rollout_vla_jepa_so101_pick_toy_v9_plate_20260630_211916.rollout_vla_jepa_so101_pick_toy_v9_2_plate_20260701_182635This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 15,
"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/kamada-riku/rollout_vla_jepa_so101_pick_toy_v9_2_plate_20260701_182635.rollout_eval_vla_jepa_20260616_182230This 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/witsense-ai/rollout_eval_vla_jepa_20260616_182230.rollout_eval_vla_jepa_20260616_182611This 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/witsense-ai/rollout_eval_vla_jepa_20260616_182611.rollout_vla_jepa_so101_pick_toy_v9_plate_20260630_190853This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 15,
"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/kamada-riku/rollout_vla_jepa_so101_pick_toy_v9_plate_20260630_190853.rollout_vla_jepa_so101_pick_toy_v9_plate_20260630_210711This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 15,
"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/kamada-riku/rollout_vla_jepa_so101_pick_toy_v9_plate_20260630_210711.rollout_jepa_20260727_183646This 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/Shirish24/rollout_jepa_20260727_183646.rollout_eval_vla_jepa_20260616_182849This 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/witsense-ai/rollout_eval_vla_jepa_20260616_182849.rollout_vla_jepa_so101_pick_toy_v9_2_plate_20260701_182010This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 15,
"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/kamada-riku/rollout_vla_jepa_so101_pick_toy_v9_2_plate_20260701_182010.rollout_omx_pickandplace_vla_jepa_20260519_110909This 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/maximellerbach/rollout_omx_pickandplace_vla_jepa_20260519_110909.rollout_vla_jepa_so101_pick_toy_v9_plate_20260624_204228This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 15,
"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/kamada-riku/rollout_vla_jepa_so101_pick_toy_v9_plate_20260624_204228.rollout_vla_jepa_20260615_085547_117_20260615_085605This 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/yuto-urushima/rollout_vla_jepa_20260615_085547_117_20260615_085605.rollout_omx_pickandplace_vla_jepa_20260519_112600This 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/maximellerbach/rollout_omx_pickandplace_vla_jepa_20260519_112600.jepa-spillover-datachan-jepa-daily