---
title: Hy-Embodied-0.5-VLA-RoboTwin
canonical_url: "https://www.modelscope.cn/models/Tencent-Hunyuan/Hy-Embodied-0.5-VLA-RoboTwin"
md_url: "https://www.modelscope.cn/models/Tencent-Hunyuan/Hy-Embodied-0.5-VLA-RoboTwin.md"
repository: Tencent-Hunyuan/Hy-Embodied-0.5-VLA-RoboTwin
last_updated: 2026-07-01
license: apache-2.0
pipeline_tag: robotics
tasks:
  - robotics
architectures:
  - HyVLA
base_model:
  - tencent/Hy-Embodied-0.5-VLA-UMI
base_model_relation: finetune
parameters: 4.5B
tensor_type:
  - BF16
library_name:
  - pytorch
  - safetensors
frameworks:
  - pytorch
downloads: 234
stars: 2
tags:
  - robotics
  - VLA
  - bimanual
  - RoboTwin
  - imitation-learning
  - fine-tuned
---

# Hy-Embodied-0.5-VLA-RoboTwin

> Hy-Embodied-0.5-VLA-RoboTwin - Tencent-Hunyuan 在 ModelScope 开源的模型。Hy-Embodied-0.5-VLA From Vision-Language-Action Models to a Real-World Robot Learning Stack Tencent Robotics X × Tencent Hy Team

Tencent-Hunyuan/Hy-Embodied-0.5-VLA-RoboTwin 是 ModelScope 魔搭社区上的 4.5B 参数robotics模型，采用 apache-2.0 许可，基于 tencent/Hy-Embodied-0.5-VLA-UMI 构建。

- **Repository**: Tencent-Hunyuan/Hy-Embodied-0.5-VLA-RoboTwin
- **License**: apache-2.0
- **Tasks**: robotics
- **Parameters**: 4.5B
- **Base model**: tencent/Hy-Embodied-0.5-VLA-UMI
- **Tags**: robotics, VLA, bimanual, RoboTwin, imitation-learning, fine-tuned
- **Downloads**: 234
- **Stars**: 2
- **Last updated**: 2026-07-01

Source: https://www.modelscope.cn/models/Tencent-Hunyuan/Hy-Embodied-0.5-VLA-RoboTwin

---

<div align="center">
<h1>Hy-Embodied-0.5-VLA</h1>
<p><b>From Vision-Language-Action Models to a Real-World Robot Learning Stack</b></p>
<p><i>Tencent Robotics X × Tencent Hy Team</i></p>

<a href="https://tairos.tencent.com/openSourceModels/hy-embodied-0.5-vla"><img src="https://img.shields.io/badge/Website-Project_Page-blue?logo=internet-explorer" alt="Project Page"></a>
<a href="https://arxiv.org/abs/2606.14409"><img src="https://img.shields.io/badge/PDF-arXiv-red?logo=arxiv" alt="Tech Report"></a>
<a href="https://github.com/Tencent-Hunyuan/Hy-Embodied-0.5-VLA"><img src="https://img.shields.io/badge/Code-GitHub-black?logo=github" alt="Code"></a>
<br>
<a href="https://huggingface.co/tencent/Hy-Embodied-0.5-VLA-UMI"><img src="https://img.shields.io/badge/Model-HuggingFace-yellow?logo=huggingface" alt="Model"></a>
<a href="https://modelscope.cn/models/Tencent-Hunyuan/Hy-Embodied-0.5-VLA-RoboTwin"><img src="https://img.shields.io/badge/Model-ModelScope-purple?logo=modelscope" alt="ModelScope"></a>
<a href="https://huggingface.co/datasets/tencent/Hy-Embodied-0.5-VLA-Data"><img src="https://img.shields.io/badge/Data-HuggingFace-orange?logo=huggingface" alt="Data"></a>
<a href="https://modelscope.cn/datasets/Tencent-HunYuan/Hy-Embodied-0.5-VLA-Data"><img src="https://img.shields.io/badge/Data-ModelScope-purple?logo=modelscope" alt="Data"></a>
</div>

<div align="center">
<video src="https://github.com/user-attachments/assets/fdd1966c-8453-4f6a-9758-238076d08ac4" controls autoplay muted loop width="85%"></video>
</div>

## 📖 Abstract

We introduce **Hy-Embodied-0.5-VLA (Hy-VLA)** — 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](https://github.com/Tencent-Hunyuan/HY-Embodied) 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](https://wuyeyexvnainai.github.io/flowpro/) preference optimization and an asynchronous inference framework, Hy-VLA establishes a scalable paradigm for continuous dexterous manipulation.


## Overview

Hy-VLA-RoboTwin is the **supervised fine-tuned** (SFT) checkpoint of Hy-Embodied-0.5-VLA (Hy-VLA), an end-to-end Vision-Language-Action system built on the Hy-Embodied-0.5 MoT backbone. Fine-tuned on all 50 bimanual manipulation tasks in the RoboTwin 2.0 benchmark, it achieves **90.9% (Clean) / 90.1% (Randomized)** average success rate — state-of-the-art among published VLA methods.

## Architecture

Same as [Hy-Embodied-0.5-VLA-UMI](https://huggingface.co/tencent/Hy-Embodied-0.5-VLA-UMI), with the following SFT-specific settings:
- **Video Encoder**: K=6 frames (current + 5 historical), temporal-spatial attention enabled
- **Action Horizon**: H=20 at 3× downsampled rate (effective horizon covers a longer time span)
- All other architecture parameters identical to the pre-trained checkpoint

## Training

| Property | Value |
|---|---|
| Data | RoboTwin 2.0: 50 tasks × 550 episodes (50 clean + 500 randomized) |
| Initialization | `tencent/Hy-Embodied-0.5-VLA-UMI` |
| Objective | Conditional flow matching |
| Global batch size | 128 |
| Learning rate | 5e-5 (warmup 1K → cosine decay to 5e-6 over 150K) |
| Optimizer | AdamW, bfloat16 mixed precision |
| Hardware | 32 GPUs (4 nodes × 8) |

## Evaluation Performance (RoboTwin 2.0)

| Setting | Success Rate |
|---|---|
| Clean | **90.9%** |
| Randomized | **90.1%** |

## Contents

```
tencent/Hy-Embodied-0.5-VLA-RoboTwin/
├── model.safetensors         # Model weights
├── config.json               # HyVLA configuration
├── tokenizer.json            # Tokenizer for the VLM backbone
├── tokenizer_config.json
├── special_tokens_map.json
├── chat_template.jinja       # Chat template for instruction formatting
├── preprocessor_config.json  # Image preprocessing config
├── norm_stats.pkl            # Pre-computed normalization statistics
└── LICENSE
```

## Usage

### Basic Loading

```python
import torch
from huggingface_hub import snapshot_download
from hy_vla import HyVLA, HyVLAConfig

ckpt = snapshot_download("tencent/Hy-Embodied-0.5-VLA-RoboTwin")

config = HyVLAConfig.from_pretrained(ckpt)
policy = HyVLA.from_pretrained(ckpt, config=config)
policy.enable_video_encoder_if_needed()
policy = policy.to(device="cuda", dtype=torch.bfloat16).eval()

# (B, K, C, H, W); K=6 history slots
img = torch.zeros(1, 6, 3, 224, 224, device="cuda", dtype=torch.bfloat16)
# Normalized dual-arm EEF: [xyz(3) + rot6d(6) + gripper(1)] * 2
state = torch.zeros((1, config.max_state_dim), device="cuda", dtype=torch.bfloat16)
batch = {
    "observation.images.top_head":   img,
    "observation.images.hand_left":  img,
    "observation.images.hand_right": img,
    "observation.state": state,
    "task": ["pick up the bottle"],
}

with torch.no_grad():
    actions = policy.forward_evaluate(batch)["pred"]
    actions = actions[..., : config.action_feature.shape[0]]
print(actions.shape)
```

### RoboTwin Evaluation

To reproduce the benchmark results:

```bash
export ROBOTWIN_DIR=/path/to/RoboTwin
export CKPT_PATH=tencent/Hy-Embodied-0.5-VLA-RoboTwin

# Quick regression (6 tasks × 10 rollouts)
bash scripts/eval_robotwin_test.sh

# Full sweep (50 tasks × 100 rollouts, 8 GPUs)
bash scripts/eval_robotwin_full.sh
```

> **Note:** The eval scripts automatically symlink `Hy-VLA/robotwin_eval/` → `RoboTwin/policy/hy_vla`, so that RoboTwin's `eval_policy.py` can discover the Hy-VLA policy adapter without any manual configuration.

## Normalization Statistics

The checkpoint includes pre-computed `norm_stats.pkl` from the RoboTwin 2.0 training data. For fine-tuning on new robot platforms, regenerate using:

```bash
python scripts/compute_norm_hdf5.py \
    --csv /path/to/episodes.csv \
    --hdf5-dir /path/to/hdf5 \
    --output norm_stats.pkl
```

## 📚 Citation

If you find Hy-VLA useful for your research, please cite:

```bibtex
@article{zhang2026hy,
  title={Hy-Embodied-0.5-VLA: From Vision-Language-Action Models to a Real-World Robot Learning Stack},
  author={Zhang, He and Xiang, Lingzhu and Lin, Haitao and Huang, Zeyu and Wang, Minghui and Zhong, Dingyan and Dong, Yubo and Wu, Yihao and Rao, Yongming and Zhang, Dongsheng and others},
  journal={arXiv preprint arXiv:2606.14409},
  year={2026}
}
```

## License

This model is released under [Apache-2.0](https://www.apache.org/licenses/LICENSE-2.0).  
The base model [tencent/Hy-Embodied-0.5-VLA-UMI](https://huggingface.co/tencent/Hy-Embodied-0.5-VLA-UMI) is also released under Apache-2.0.