README.md

---
license: apache-2.0
library_name: lerobot
pipeline_tag: robotics
---

# WALL-OSS: Wall-OSS-0.5

<div align="left">

<p align="center">
    <img src="https://huggingface.co/X-Square-Robot/wall-oss-flow/resolve/main/assets/logo.png" width="600"/>
<p>

<div align="center">

[![Paper](https://img.shields.io/badge/📄%20Paper-PDF-EA1B22?style=for-the-badge&logo=adobeacrobatreader&logoColor=fff)](https://huggingface.co/papers/2605.30877)
&nbsp;&nbsp;
[![Hugging Face](https://img.shields.io/badge/Hugging%20Face-x--square--robot-FFB000?style=for-the-badge&logo=huggingface&logoColor=000)](https://huggingface.co/x-square-robot)
&nbsp;&nbsp;
[![GitHub](https://img.shields.io/badge/GitHub-181717?style=for-the-badge&logo=github&logoColor=fff)](https://github.com/X-Square-Robot/wall-x)
&nbsp;&nbsp;
[![Project Page](https://img.shields.io/badge/Project-1E90FF?style=for-the-badge&logo=google-chrome&logoColor=fff)](https://x2robot.com/en/oss)

</div>

</div>

Wall-OSS-0.5 is an open-source 4B Vision-Language-Action (VLA) model built upon a 3B VLM backbone augmented with action-generation components. It is designed to provide executable robot behavior directly from pretraining. The model was introduced in the [Wall-OSS-0.5 Technical Report](https://huggingface.co/papers/2605.30877).

Before task-specific fine-tuning, Wall-OSS-0.5 achieves non-trivial zero-shot real-robot behavior. After fine-tuning, it serves as a strong adaptation prior, significantly outperforming previous baselines in manipulation tasks.

## 🎬 Video Demos

<div align="center">
    <video width="80%" controls>
        <source src="https://x2robot.com/api/videos/file/wall-oss_top_720p-1.mp4" type="video/mp4">
        Your browser does not support the video tag.
    </video>
    <p><strong>WALL-OSS in Action: Demonstrating advanced manipulation capabilities and embodied AI performance</strong></p>
</div>

## 🚀 Quick Start

### Installation

```bash
# Create conda environment
conda create --name wallx python=3.10
conda activate wallx

# Install base requirements
pip install torch torchvision transformers
pip install huggingface_hub

# Install Wall-X from GitHub
git clone https://github.com/X-Square-Robot/wall-x.git
cd wall-x
pip install -e .
```

### Basic Usage

```python
import torch
from wall_x.model.qwen2_5_based.modeling_qwen2_5_vl_act import Qwen2_5_VLMoEForAction

# Load the model
model_path = "X-Square-Robot/wall-oss-flow"  # or your local path
model = Qwen2_5_VLMoEForAction.from_pretrained(model_path)
model.eval()

# Configuration
device = "cuda" if torch.cuda.is_available() else "cpu"
model = model.to(device).bfloat16()

# Basic inference example
batch_size = 1
seq_length = 50

# Prepare inputs (example with synthetic data)
input_ids = torch.randint(0, len(model.processor.tokenizer), (batch_size, seq_length), dtype=torch.long).to(device)
attention_mask = torch.ones((batch_size, seq_length), dtype=torch.long).to(device)
proprioception = torch.randn((batch_size, 1, 20), dtype=torch.float32).to(device).bfloat16()

with torch.no_grad():
    outputs = model(input_ids=input_ids, attention_mask=attention_mask, proprioception=proprioception, mode="validate")
    print(f"Output logits shape: {outputs.logits.shape}")
```

## 🎯 Supervised Fine-Tuning (SFT)

For training Wall-X on your robotics datasets using the [LeRobot](https://github.com/huggingface/lerobot) format, please refer to the [Training Documentation](https://github.com/X-Square-Robot/wall-x/blob/main/workspace/README.md).

## 📄 Citation

If you find WALL-OSS models useful, please cite:

```bibtex
@misc{walloss_paper_2025,
  title        = {Wall-OSS-0.5 Technical Report},
  author       = {X Square Robot},
  year         = {2026},
  howpublished = {\url{https://huggingface.co/papers/2605.30877}},
  note         = {Technical Report}
}
```