table_name stringlengths 9 9 | total_frames int64 7.03M 10.8M | total_episodes int64 7.59k 11.7k | duration_hours float32 65.1 99.9 | fps int32 30 30 |
|---|---|---|---|---|
table_000 | 10,790,284 | 11,604 | 99.900002 | 30 |
table_001 | 10,788,879 | 11,594 | 99.900002 | 30 |
table_002 | 10,788,912 | 11,484 | 99.900002 | 30 |
table_003 | 10,789,312 | 11,463 | 99.900002 | 30 |
table_004 | 10,790,641 | 11,603 | 99.900002 | 30 |
table_005 | 10,788,952 | 11,559 | 99.900002 | 30 |
table_006 | 10,789,030 | 11,565 | 99.900002 | 30 |
table_007 | 10,789,113 | 11,655 | 99.900002 | 30 |
table_008 | 10,789,076 | 11,531 | 99.900002 | 30 |
table_009 | 10,789,252 | 11,451 | 99.900002 | 30 |
table_010 | 10,789,779 | 11,535 | 99.900002 | 30 |
table_011 | 10,789,767 | 11,471 | 99.900002 | 30 |
table_012 | 10,790,134 | 11,539 | 99.900002 | 30 |
table_013 | 10,789,433 | 11,528 | 99.900002 | 30 |
table_014 | 10,787,184 | 11,537 | 99.900002 | 30 |
table_015 | 10,790,311 | 11,522 | 99.900002 | 30 |
table_016 | 10,789,326 | 11,663 | 99.900002 | 30 |
table_017 | 10,789,019 | 11,586 | 99.900002 | 30 |
table_018 | 10,789,577 | 11,576 | 99.900002 | 30 |
table_019 | 10,783,947 | 11,618 | 99.900002 | 30 |
table_020 | 10,789,298 | 11,633 | 99.900002 | 30 |
table_021 | 7,029,088 | 7,587 | 65.099998 | 30 |
Hy-Embodied-0.5-VLA
From Vision-Language-Action Models to a Real-World Robot Learning Stack
Tencent Robotics X Γ Tencent Hy Team
π Abstract
We introduce Hy-Embodied-0.5-VLA (Hy-VLA) β an end-to-end Vision-Language-Action system that spans the full robot learning stack: data collection, model design, pre-training, supervised fine-tuning, RL post-training, and real-world deployment. Built on the Hy-Embodied-0.5 MoT backbone, Hy-VLA integrates a flow-matching action expert, a compact memory encoder for multi-frame history, and a delta-chunk action representation decoupled from embodiment-specific kinematics.
Powered by 10,000+ hours of high-fidelity UMI demonstrations collected via a custom fingertip interface with optical motion-capture, Hy-VLA achieves state-of-the-art results on the RoboTwin 2.0 benchmark (90.9% / 90.1% on Clean / Randomized) and demonstrates robust cross-embodiment transfer across four real-world robot platforms. Paired with FlowPRO preference optimization and an asynchronous inference framework, Hy-VLA establishes a scalable paradigm for continuous dexterous manipulation.
Overview
Hy-Embodied-0.5-VLA-Data is a large-scale bimanual manipulation dataset for training Vision-Language-Action (VLA) foundation models. Powered by 2000+ hours of high-fidelity demonstrations collected via a custom fingertip UMI device with optical motion-capture, it spans 70+ manipulation tasks. The dataset is released in Lance format compatible with LeRobot v3.0.
Note: The open-source release contains approximately 20% of the full corpus.
Dataset Statistics
| Property | Value |
|---|---|
| Total Episodes | 250,304 |
| Total Frames | 233,600,314 |
| Total Duration | 2,163 hours |
| Total Size | ~18.8 TB (22 tables) |
| Frequency | 30 Hz |
| Cameras | 3 views (head + left wrist + right wrist) |
| Resolution | 240 Γ 424 px per camera |
| Format | Lance (LeRobot v3.0 schema) |
| Tables | 22 (table_000 ~ table_021, ~100h each) |
Directory Structure
Each table is a self-contained LeRobot v3.0 dataset root:
table_000/
βββ table_000.lance/ # Lance columnar data (5GB shards)
β βββ _versions/
β βββ data/
β βββ ...-data-0.lance
βββ meta/ # LeRobot v3.0 metadata
βββ info.json # Table-level summary
βββ stats.json # Per-feature statistics
βββ tasks.parquet # Task β index mapping
βββ episodes/ # Episode boundary parquet files
Data Schema
Each row (frame) contains:
Observations
| Column | Type | Shape | Description |
|---|---|---|---|
observation.state |
float32 |
[16] |
Dual-arm EEF state: left [x,y,z,qx,qy,qz,qw,gripper], right [x,y,z,qx,qy,qz,qw,gripper] |
observation.images.cam_high |
image |
[240,424,3] |
Overhead camera RGB image |
observation.images.cam_left_wrist |
image |
[240,424,3] |
Left wrist-mounted camera RGB image |
observation.images.cam_right_wrist |
image |
[240,424,3] |
Right wrist-mounted camera RGB image |
Actions
| Column | Type | Shape | Description |
|---|---|---|---|
action |
float32 |
[2] |
Gripper openness [left, right] derived from state |
Metadata
| Column | Type | Shape | Description |
|---|---|---|---|
task_index |
int32 |
[1] |
Task ID mapping to meta/tasks.parquet |
task |
string |
[1] |
Task description (Chinese, e.g., "ζεηΊ’θ²ζΉεεΉΆζΎε ₯ηε") |
episode_index |
int32 |
[1] |
Global unique episode index |
frame_index |
int32 |
[1] |
Frame index within episode (starts at 0) |
timestamp |
float32 |
[1] |
Seconds from episode start |
Usage
LanceTableReader reads a single Lance table (local or HF Hub):
from hy_vla.data.lance_dataset import LanceTableReader
# Local directory
reader = LanceTableReader(root="./table_000")
# HF Hub
reader = LanceTableReader(
repo_id="tencent/Hy-Embodied-0.5-VLA-Data",
table_name="table_000",
)
# Access
frame = reader[42] # single frame dict
episode = reader.get_episode(3) # all frames of episode 3
Also compatible with raw
lance,lancedb, andlerobot-lancedb(LeRobotLanceDataset).
Episode Visualization
# Use the HF Hub dataset, pick table_000 episode 666
python scripts/vis_umi_episode.py -t table_000 -e 666
# Local Lance root
python scripts/vis_umi_episode.py /path/to/Hy-Embodied-0.5-Data -e 0 --no-3d
Downloading Specific Tables
Due to the large total size (~18.8 TB), you may prefer to download individual tables:
from huggingface_hub import snapshot_download
# Download only table_000 (~890 GB)
snapshot_download(
"tencent/Hy-Embodied-0.5-VLA-Data",
allow_patterns="table_000/**",
repo_type="dataset"
)
π Citation
If you find Hy-VLA useful for your research, please cite:
@article{tencent2026hyembodied05vla,
title={Hy-Embodied-0.5-VLA: From Vision-Language-Action Models to a Real-World Robot Learning Stack},
author={Tencent Robotics X and Tencent Hy Team},
journal={arXiv preprint arXiv:2606.14409},
year={2026}
}
License
This dataset is released under CC-BY-4.0.
- Downloads last month
- 14,648