# **LinearRAG: Linear Graph Retrieval-Augmented Generation on Large-scale Corpora**  

> A relation-free graph construction method for efficient GraphRAG. It eliminates LLM token costs during graph construction, making GraphRAG faster and more efficient than ever.

<p align="center">
  <a href="https://arxiv.org/abs/2510.10114" target="_blank">
    <img src="https://img.shields.io/badge/Paper-Arxiv-red?logo=arxiv&style=flat-square" alt="arXiv:2506.08938">
  </a>
  <a href="https://huggingface.co/datasets/Zly0523/linear-rag/tree/main" target="_blank">
    <img src="https://img.shields.io/badge/HuggingFace-Model-yellow?logo=huggingface&style=flat-square" alt="HuggingFace">
  </a>
  <a href="https://github.com/LuyaoZhuang/linear-rag" target="_blank">
    <img src="https://img.shields.io/badge/GitHub-Project-181717?logo=github&style=flat-square" alt="GitHub">
  </a>
</p>

---

## 🚀 **Highlights**
- ✅ **Context-Preserving**: Relation-free graph construction, relying on lightweight entity recognition and semantic linking to achieve comprehensive contextual comprehension. 
- ✅ **Complex Reasoning**: Enables deep retrieval via semantic bridging, achieving multi-hop reasoning in a single retrieval pass without requiring explicit relational graphs.
- ✅ **High Scalability**: Zero LLM token consumption, faster processing speed, and linear time/space complexity.
  
<p align="center">
  <img src="figure/main_figure.png" width="95%" alt="Framework Overview">
</p>

---
## 🎉 **News**
- **[2026-01-26]** Our **[LinearRAG](https://github.com/DEEP-PolyU/LinearRAG)** is accepted by ICLR’26. 
- **[2025-10-27]** We release **[LinearRAG](https://github.com/DEEP-PolyU/LinearRAG)**, a relation-free graph construction method for efficient GraphRAG.
- **[2025-06-06]** We release **[GraphRAG-Bench](https://github.com/GraphRAG-Bench/GraphRAG-Benchmark.git)**, the benchmark for evaluating GraphRAG models.
- **[2025-01-21]** We release the **[GraphRAG survey](https://github.com/DEEP-PolyU/Awesome-GraphRAG)**.

---

## 🛠️ **Usage**

### 1️⃣ Install Dependencies  

**Step 1: Install Python packages**

```bash
pip install -r requirements.txt
```

**Step 2: Download Spacy language model**

```bash
python -m spacy download en_core_web_trf
```

> **Note:** For the `medical` dataset, you need to install the scientific/biomedical Spacy model:
```bash
pip install https://s3-us-west-2.amazonaws.com/ai2-s2-scispacy/releases/v0.5.3/en_core_sci_scibert-0.5.3.tar.gz
```

**Step 3: Set up your OpenAI API key**

```bash
export OPENAI_API_KEY="your-api-key-here"
export OPENAI_BASE_URL="your-base-url-here"
```

**Step 4: Download Datasets**

Download the datasets from HuggingFace and place them in the `dataset/` folder:

```bash
git clone https://huggingface.co/datasets/Zly0523/linear-rag
cp -r linear-rag/* dataset/
```

**Step 5: Prepare Embedding Model**

Make sure the embedding model is available at:

```
model/all-mpnet-base-v2/
```


### 2️⃣ Quick Start Example

```bash
SPACY_MODEL="en_core_web_trf"
EMBEDDING_MODEL="model/all-mpnet-base-v2"
DATASET_NAME="2wikimultihop"
LLM_MODEL="gpt-4o-mini"
MAX_WORKERS=16

python run.py \
    --spacy_model ${SPACY_MODEL} \
    --embedding_model ${EMBEDDING_MODEL} \
    --dataset_name ${DATASET_NAME} \
    --llm_model ${LLM_MODEL} \
    --max_workers ${MAX_WORKERS} 
    # --use_vectorized_retrieval # optional, use vectorized matrix-based retrieval for GPU acceleration if Strong GPU is available, otherwise use BFS iteration.
```

## 🎯 **Performance**

<div align="center">
<img src="figure/generation_results.png" alt="framework" width="1000">

**Main results of end-to-end performance**
</div>
<div align="center">
<img src="figure/efficiency_result.png" alt="framework" width="1000">

**Efficiency and performance comparison.**
</div>


## 📬 Citation

If you find this work helpful, please consider citing us:
```bibtex
@article{zhuang2025linearrag,
  title={LinearRAG: Linear Graph Retrieval Augmented Generation on Large-scale Corpora},
  author={Zhuang, Luyao and Chen, Shengyuan and Xiao, Yilin and Zhou, Huachi and Zhang, Yujing and Chen, Hao and Zhang, Qinggang and Huang, Xiao},
  journal={arXiv preprint arXiv:2510.10114},
  year={2025}
}
```