README.md

---
license: apache-2.0
datasets:
- internlm/Spatial-SSRL-81k
language:
- en
base_model:
- Qwen/Qwen3-VL-4B-Instruct
pipeline_tag: image-text-to-text
tags:
- multimodal
- spatial
- sptial understanding
- self-supervised learning
library_name: transformers
---


# Spatial-SSRL-Qwen3VL-4B

📖<a href="https://arxiv.org/abs/2510.27606">Paper</a>| 🏠<a href="https://github.com/InternLM/Spatial-SSRL">Github</a> |🤗<a href="https://huggingface.co/internlm/Spatial-SSRL-7B">Spatial-SSRL-7B Model</a> | 
🤗<a href="https://huggingface.co/internlm/Spatial-SSRL-Qwen3VL-4B">Spatial-SSRL-Qwen3VL-4B Model</a> |
  🤗<a href="https://huggingface.co/datasets/internlm/Spatial-SSRL-81k">Spatial-SSRL-81k Dataset</a> | 📰<a href="https://huggingface.co/papers/2510.27606">Daily Paper</a> 

Spatial-SSRL-Qwen3VL-4B is a large vision-language model targeting spatial understanding, built on the base of Qwen3-VL-4B-Instruct. It's optimized by applying Spatial-SSRL, a lightweight self-supervised reinforcement learning
paradigm which can scale RLVR efficiently. The model demonstrates strong spatial intelligence while preserving the original general visual capabilities of the base model. 

## 📢 News
- 🚀 [2025/11/24] We have released the [🤗Spatial-SSRL-Qwen3VL-4B Model](https://huggingface.co/internlm/Spatial-SSRL-Qwen3VL-4B), initialized from Qwen3-VL-4B-Instruct.
- 🚀 [2025/11/03] Now you can try out Spatial-SSRL-7B on [🤗Spatial-SSRL Space](https://huggingface.co/spaces/yuhangzang/Spatial-SSRL).
- 🚀 [2025/11/03] We have released the [🤗Spatial-SSRL-7B Model](https://huggingface.co/internlm/Spatial-SSRL-7B), and [🤗Spatial-SSRL-81k Dataset](https://huggingface.co/datasets/internlm/Spatial-SSRL-81k).
- 🚀 [2025/11/02] We have released the [🏠Spatial-SSRL Repository](https://github.com/InternLM/Spatial-SSRL).

## 🌈 Overview
We are thrilled to introduce <strong>Spatial-SSRL</strong>, a novel self-supervised RL paradigm aimed at enhancing LVLM spatial understanding. 
By optimizing Qwen2.5-VL-7B with Spatial-SSRL, the model exhibits stronger spatial intelligence across seven spatial understanding benchmarks in both image and video settings.
</p>
<p style="text-align: center;"> 
  <img src="assets/teaser_1029final.png" alt="Teaser" width="100%"> 
</p>
Spatial-SSRL is a <strong>lightweight</strong> tool-free framework that is natually compatible with the RLVR training paradigm and easy to extend to a multitude of pretext tasks.
Five tasks are currently formulated in the framework, requiring only ordinary RGB and RGB-D images. <strong>And we welcome you to join Spatial-SSRL with effective pretext tasks to further strengthen the capabilities of LVLMs!</strong>

<p style="text-align: center;"> 
  <img src="assets/pipeline_1029final.png" alt="Pipeline" width="100%"> 
</p>

## 💡 Highlights
- 🔥 **Highly Scalable:** Spatial-SSRL uses ordinary raw RGB and RGB-D images instead of richly-annotated public datasets or manual labels for data curation, making it highly scalable.
- 🔥 **Cost-effective:** Avoiding the need for human labels or API calls for general LVLMs throughout the entire pipeline endows Spatial-SSRL with cost-effectiveness.
- 🔥 **Lightweight:** Prior approaches for spatial understanding heavily rely on annotation of external tools, incurring inherent errors in training data and additional cost. In constrast, Spatial-SSRL is completely tool-free and can easily be extended to more self-supervised tasks. 
- 🔥 **Naturally Verifiable:** Intrinsic supervisory signals determined by pretext objectives are naturally verifiable, aligning Spatial-SSRL well with the RLVR paradigm.
<p style="text-align: center;"> 
  <img src="assets/comparison_v2.png" alt="Teaser" width="100%"> 
</p>

## 📊 Results
We train Qwen3-VL-4B-Instruct with our Spatial-SSRL paradigm and the average experimental results on spatial understanding and general VQA benchmarks are shown below.
<p style="text-align: center;"> 
  <img src="assets/exp_result_new3.png" alt="Pipeline" width="100%"> 
</p>

## 🛠️ Usage
Here we provide a code snippet for you to start a simple trial of <strong>Spatial-SSRL-Qwen3VL-4B</strong> on your own device. You can download the model from 🤗<a href="https://huggingface.co/internlm/Spatial-SSRL-Qwen3VL-4B">Spatial-SSRL-Qwen3VL-4B Model</a > before your trial!
</p>

```python
from transformers import AutoProcessor, AutoModelForImageTextToText #transformers==4.57.1
from qwen_vl_utils import process_vision_info #0.0.14
import torch

model_path = "internlm/Spatial-SSRL-Qwen3VL-4B" #You can change it to your own local path if deployed already

#Change the path of the input image
img_path = "assets/eg1.jpg"

#Change your question here
question = "Question: Consider the real-world 3D locations and orientations of the objects. If I stand at the man's position facing where it is facing, is the menu on the left or right of me?\nOptions:\nA. on the left\nB. on the right\n"

question += "Please select the correct answer from the options above. \n"
#We recommend using the format prompt to make the inference consistent with training
format_prompt = "You FIRST think about the reasoning process as an internal monologue and then provide the final answer. The reasoning process MUST BE enclosed within <think> </think> tags. The final answer MUST BE put in \\boxed{}."

model = AutoModelForImageTextToText.from_pretrained(
    model_path, torch_dtype=torch.float16, device_map='auto', attn_implementation='flash_attention_2'
    )
processor = AutoProcessor.from_pretrained(model_path)

messages = [
    {
        "role": "user",
        "content": [
            {
                "type": "image",
                "image": img_path,
            },
            {"type": "text", "text": question + format_prompt},
        ],
    }
]

text = processor.apply_chat_template(
    messages, tokenize=False, add_generation_prompt=True
)
image_inputs, video_inputs = process_vision_info(messages)
inputs = processor(
    text=[text],
    images=image_inputs,
    videos=video_inputs,
    padding=True,
    return_tensors="pt",
)
inputs = inputs.to("cuda")

generated_ids = model.generate(**inputs, max_new_tokens=4096, do_sample=False)
generated_ids_trimmed = [
    out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids)
]
output_text = processor.batch_decode(
    generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False
)
print("Model Response:", output_text[0])
```

## Cases
<p style="text-align: center;"> 
  <img src="assets/case-qwen3vl.jpg" alt="Teaser" width="100%"> 
</p>


## ✒️Citation
If you find our model useful, please kindly cite:
```
@article{liu2025spatial,
  title={Spatial-SSRL: Enhancing Spatial Understanding via Self-Supervised Reinforcement Learning},
  author={Liu, Yuhong and Zhang, Beichen and Zang, Yuhang and Cao, Yuhang and Xing, Long and Dong, Xiaoyi and Duan, Haodong and Lin, Dahua and Wang, Jiaqi},
  journal={arXiv preprint arXiv:2510.27606},
  year={2025}
}
```

## 📄 License
![Code License](https://img.shields.io/badge/Code%20License-Apache_2.0-green.svg) ![Data License](https://img.shields.io/badge/Data%20License-CC%20By%20NC%204.0-red.svg) 

**Usage and License Notices**: The data and code are intended and licensed for research use only.