# Xiaomi-Robotics-0
**An Open-Sourced Vision-Language-Action Model with Real-Time Inference**
[](https://arxiv.org/abs/2602.12684)
[](https://robotics.xiaomi.com/xiaomi-robotics-0.html)
[](https://huggingface.co/collections/XiaomiRobotics/xiaomi-robotics-0)
[](xr0/)
[](LICENSE)
[](https://paperswithcode.co/api/v1/papers/2602.12684/leaderboard-badge-link?eval=13807)
[](https://paperswithcode.co/api/v1/papers/2602.12684/leaderboard-badge-link?eval=13803)
[](https://paperswithcode.co/api/v1/papers/2602.12684/leaderboard-badge-link?eval=13808)
[](https://paperswithcode.co/api/v1/papers/2602.12684/leaderboard-badge-link?eval=13810)
---
## ๐ก About Xiaomi-Robotics-0
**Xiaomi-Robotics-0** is a state-of-the-art **Vision-Language-Action (VLA) model** with 4.7B parameters, specifically engineered for high-performance and seamless real-time execution.
Try out [the post-training code](xr0/) on your data!
### Key Features:
* **๐ง Strong Generalization**: Pre-trained on diverse cross-embodiment trajectories and VL data to handle complex, unseen tasks.
* **๐ Real-Time Ready**: Optimized with asynchronous execution to minimize inference latency.
* **๐ ๏ธ Flexible Deployment**: Fully compatible with the Hugging Face `transformers` ecosystem and optimized for consumer GPUs.
## ๐
Updates
- **[Apr 27, 2026]** ๐ Open-sourced [**the post-training code**](xr0/) for **Xiaomi-Robotics-0**. See [**the Post-training section**](https://robotics.xiaomi.com/xiaomi-robotics-0.html) on the project page for details.
- **[Feb 2026]** ๐ Released the **Technical Report**.
- **[Feb 2026]** ๐ฅ Released **Pre-trained weights** and **Fine-tuned weights** for LIBERO, CALVIN, and SimplerEnv.
- **[Feb 2026]** ๐ป Inference code and evaluation scripts are now live!
---
## ๐ Benchmark
We evaluate **Xiaomi-Robotics-0** on three standard simulation benchmarks: **CALVIN**, **LIBERO**, and **SimplerEnv**. The table below summarizes the performance results across different embodiments and datasets. For each setting, we provide the corresponding fine-tuned checkpoint and a guide for running the evaluation.
| | ๐ค Name on Hugging Face | Description | Performance | Evaluation Guide |
| :------------- | :----------------------------------------------------------- | :-------------------------------- | :--------------------------------- | :------------------------------------------- |
| **LIBERO** | [**Xiaomi-Robotics-0-LIBERO**](https://huggingface.co/XiaomiRobotics/Xiaomi-Robotics-0-LIBERO) | Fine-tuned on four LIBERO suites. | **98.7%** (Avg Success) | [LIBERO Eval](eval_libero/README.md) |
| **CALVIN** | [**Xiaomi-Robotics-0-Calvin-ABCD_D**](https://huggingface.co/XiaomiRobotics/Xiaomi-Robotics-0-Calvin-ABCD_D) | Fine-tuned on ABCDโD Split. | **4.80** (Avg Length) | [CALVIN Eval](eval_calvin/README.md) |
| | [**Xiaomi-Robotics-0-Calvin-ABC_D**](https://huggingface.co/XiaomiRobotics/Xiaomi-Robotics-0-Calvin-ABC_D) | Fine-tuned on ABCโD Split. | **4.75** (Avg Length) | [CALVIN Eval](eval_calvin/README.md) |
| **SimplerEnv** | [**Xiaomi-Robotics-0-SimplerEnv-Google-Robot**](https://huggingface.co/XiaomiRobotics/Xiaomi-Robotics-0-SimplerEnv-Google-Robot) | Fine-tuned on Fractal dataset. | **85.5%** (VM)
**74.7%** (VA) | [SimplerEnv Eval](eval_simplerenv/README.md) |
| | [**Xiaomi-Robotics-0-SimplerEnv-WidowX**](https://huggingface.co/XiaomiRobotics/Xiaomi-Robotics-0-SimplerEnv-WidowX) | Fine-tuned on Bridge dataset. | **79.2%** | [SimplerEnv Eval](eval_simplerenv/README.md) |
| **Base** | [**Xiaomi-Robotics-0**](https://huggingface.co/XiaomiRobotics/Xiaomi-Robotics-0-Pretrain) | Pre-trained model. | - | - |
## ๐ Quick Start: Installation & Deployment
Our project relies primarily on HuggingFace Transformers ๐ค, making deployment extremely easy. If your environment supports transformers >= 4.57.1, you can use our project seamlesslyโwe recommend PyTorch 2.8.0 (paired with torchvision 0.23.0 and torchaudio 2.8.0), as this combination has been fully tested by our team and ensures optimal compatibility.
### 1๏ธโฃ Installation Guides
Hereโs a simple installation guide to get you started:
```bash
git clone https://github.com/XiaomiRobotics/Xiaomi-Robotics-0
cd Xiaomi-Robotics-0
# Create a Conda environment with Python 3.12
conda create -n mibot python=3.12 -y
conda activate mibot
# Install PyTorch
pip install torch==2.8.0 torchvision==0.23.0 torchaudio==2.8.0 --index-url https://download.pytorch.org/whl/cu128
# Install transformers
pip install transformers==4.57.1
# Install flash-attn
pip uninstall -y ninja && pip install ninja
pip install flash-attn==2.8.3 --no-build-isolation
# or pip install https://github.com/Dao-AILab/flash-attention/releases/download/v2.8.3/flash_attn-2.8.3+cu12torch2.8cxx11abiTRUE-cp312-cp312-linux_x86_64.whl
sudo apt-get install -y libegl1 libgl1 libgles2
```
### 2๏ธโฃ Deployment Guides
**Xiaomi-Robotics-0** is deployed on top of the HuggingFace Transformers ๐ค ecosystem, enabling straightforward deployment for robotic manipulation tasks. By leveraging Flash Attention 2 and bfloat16 precision, the model can be loaded and run efficiently on consumer-grade GPUs.
``` python
import torch
from transformers import AutoModel, AutoProcessor
# 1. Load model and processor
model_path = "XiaomiRobotics/Xiaomi-Robotics-0-LIBERO"
model = AutoModel.from_pretrained(
model_path,
trust_remote_code=True,
attn_implementation="flash_attention_2",
dtype=torch.bfloat16
).cuda().eval()
processor = AutoProcessor.from_pretrained(model_path, trust_remote_code=True, use_fast=False)
# 2. Construct the prompt with multi-view inputs
language_instruction = "Pick up the red block."
instruction = (
f"<|im_start|>user\nThe following observations are captured from multiple views.\n"
f"# Base View\n<|vision_start|><|image_pad|><|vision_end|>\n"
f"# Left-Wrist View\n<|vision_start|><|image_pad|><|vision_end|>\n"
f"Generate robot actions for the task:\n{language_instruction} /no_cot<|im_end|>\n"
f"<|im_start|>assistant\n<|im_end|>\n"
)
# 3. Prepare inputs
# Assuming `image_base`, `image_wrist`, and `proprio_state` are already loaded
inputs = processor(
text=[instruction],
images=[image_base, image_wrist], # [PIL.Image, PIL.Image]
videos=None,
padding=True,
return_tensors="pt",
).to(model.device)
# Add proprioceptive state and action mask
robot_type = "libero_all"
inputs["seed"] = 42
inputs["state"] = torch.from_numpy(proprio_state).to(model.device, model.dtype).view(1, 1, -1)
inputs["action_mask"] = processor.get_action_mask(robot_type).to(model.device, model.dtype)
# 4. Generate action
with torch.no_grad():
outputs = model(**inputs)
# Decode raw outputs into actionable control commands
action_chunk = processor.decode_action(outputs.actions, robot_type=robot_type)
print(f"Generated Action Chunk Shape: {action_chunk.shape}")
```
## ๐ ๏ธ Post-Training
Besides inference and evaluation, we also open-source the post-training pipeline for adapting **Xiaomi-Robotics-0** on real robots. The post-training codebase and sample data are available in [`xr0/`](xr0/). For detailed instructions on installation, data preparation, training, and deployment, please refer to [`xr0/README.md`](xr0/README.md).
## ๐ Citation
If you find this project useful, please consider citing:
```bibtex
@article{cai2026xiaomi,
title={Xiaomi-Robotics-0: An Open-Sourced Vision-Language-Action Model with Real-Time Execution},
author={Cai, Rui and Guo, Jun and He, Xinze and Jin, Piaopiao and Li, Jie and Lin, Bingxuan and Liu, Futeng and Liu, Wei and Ma, Fei and Ma, Kun and Qiu, Feng and Qu, Heng and Su, Yifei and Sun, Qiao and Wang, Dong and Wang, Donghao and Wang, Yunhong and Wu, Rujie and Xiang, Diyun and Yang, Yu and Ye, Hangjun and Zhang, Yuan and Zhou, Quanyun},
journal={arXiv preprint arXiv:2602.12684},
year={2026}
}
```
## ๐ License
This project is licensed under the [Apache License 2.0](LICENSE).