README.md

---
license: apache-2.0
tags:
- PEFT
- Mixture-of-Experts
- MoE-LoRA
- Multi-Task-Learning
- Large-Language-Models
- LLaMA
- LLaMA-2
- pytorch
---

<a id="top"></a>
<div align="center">
  <h1>🚀 D<sup>2</sup>MoRA: Diversity-Regulated Asymmetric MoE-LoRA Decomposition for Efficient Multi-Task Adaptation</h1>

  <p>
    <b>Jianhui Zuo</b><sup>1</sup>&nbsp;
    <b>Xuemeng Song</b><sup>2✉</sup>&nbsp;
    <b>Haokun Wen</b><sup>3,4</sup>&nbsp;
    <b>Meng Liu</b><sup>5</sup>&nbsp;
    <b>Yupeng Hu</b><sup>1</sup>&nbsp;
    <b>Jiuru Wang</b><sup>6</sup>&nbsp;
    <b>Liqiang Nie</b><sup>3✉</sup>
  </p>

  <p>
    <sup>1</sup>School of Software, Shandong University<br>
    <sup>2</sup>Department of Computer Science and Engineering, Southern University of Science and Technology<br>
    <sup>3</sup>School of Computer Science and Technology, Harbin Institute of Technology (Shenzhen)<br>
    <sup>4</sup>School of Data Science, City University of Hong Kong<br>
    <sup>5</sup>School of Computer and Artificial Intelligence, Shandong Jianzhu University<br>
    <sup>6</sup>School of Computer Science and Engineering, Linyi University
  </p>
</div>

These are the official pre-trained model weights and configuration files for **D<sup>2</sup>MoRA**, a novel **diversity-regulated asymmetric MoE-LoRA decomposition framework** for **parameter-efficient fine-tuning (PEFT)** of large language models in **multi-task adaptation** scenarios.

🔗 **Paper:** [Accepted by AAAI 2026]  
🔗 **GitHub Repository:** [AAAI26-D2MoRA](https://github.com/iLearn-Lab/AAAI26-D2MoRA)

---

## 📌 Model Information

### 1. Model Name
**D<sup>2</sup>MoRA** (**D**iversity-Regulated Asymmetric **MoE-LoRA** Decomposition) Checkpoints.

### 2. Task Type & Applicable Tasks
- **Task Type:** Parameter-Efficient Fine-Tuning (PEFT) / Low-Rank Adaptation (LoRA) / Mixture-of-Experts (MoE) / Multi-Task Learning
- **Applicable Tasks:** Efficient adaptation of large language models for heterogeneous downstream tasks, especially **multi-task commonsense reasoning** and related language understanding tasks.

### 3. Project Introduction
Low-Rank Adaptation (LoRA) has become a powerful parameter-efficient fine-tuning paradigm for adapting large language models. Recent studies further integrate LoRA with the Mixture-of-Experts (MoE) mechanism to improve multi-task adaptation. However, existing knowledge-sharing paradigms among LoRA experts still suffer from two major limitations:

1. **Constrained Functional Specialization**  
   Existing one-to-many sharing paradigms force all experts to operate in a single shared low-rank subspace, limiting the flexibility of expert-specific transformations.

2. **Induced Expert Homogenization**  
   Sharing a single down-projection matrix across experts may cause different experts to become overly similar, weakening expert diversity and reducing the benefit of MoE specialization.

To address these issues, **D<sup>2</sup>MoRA** introduces a **diversity-regulated asymmetric MoE-LoRA decomposition framework**. Instead of treating each LoRA expert as a fixed `(A, B)` pair, D<sup>2</sup>MoRA decomposes LoRA experts into two independent sets of base experts:

- **Down-projection experts:** A<sub>1</sub>, A<sub>1</sub>, ..., A<sub>M</sub>
- **Up-projection experts:** B<sub>1</sub>, B<sub>2</sub>, ..., B<sub>N</sub>

This design enables a novel **asymmetric many-to-many pairing** mechanism between down-projection and up-projection experts, allowing more flexible cross-expert knowledge sharing while preserving expert specialization. In addition, D<sup>2</sup>MoRA introduces:

- **Sample-Aware Down-Projection Expert Mixture**
- **Low-Rank Embedding-Aware Up-Projection Expert Mixture**
- **Dual Orthogonality Regularization**

to explicitly improve the diversity of both \(A\)-experts and \(B\)-experts and mitigate expert homogenization.

> 💡 **Note:** D<sup>2</sup>MoRA is evaluated in both **multi-task** and **single-task** settings, and consistently demonstrates strong effectiveness and generalization ability.

### 4. Training Data Source
The model was primarily trained and evaluated on the **Commonsense 170K** benchmark, which contains eight public commonsense reasoning datasets:
- **BoolQ**
- **PIQA**
- **SIQA**
- **HellaSwag**
- **WinoGrande**
- **ARC-c**
- **ARC-e**
- **OBQA**

---

## 🚀 Usage & Basic Inference

These weights are designed to be used directly with the official **D<sup>2</sup>MoRA** GitHub repository.

### Step 1: Prepare the Environment
Clone the GitHub repository and install dependencies following the official repository instructions:

```bash
git clone https://github.com/iLearn-Lab/AAAI26-D2MoRA.git
cd D2MoRA
```

Please refer to the official repository for the exact environment setup and dependency installation details.

### Step 2: Download Model Weights & Data

Download the checkpoint files (e.g., `best_model.pth`) from this Hugging Face repository and place them into your local checkpoint directory.

You should also prepare the **Commonsense 170K** benchmark and related processed data according to the official repository instructions.

### Step 3: Training / Evaluation

D<sup>2</sup>MoRA is built for PEFT-based adaptation of large language models such as **LLaMA-7B** and **LLaMA2-7B**.

In the paper, the method fine-tunes the **Query / Key / Value** projections of self-attention layers. Typical experimental settings include:

- **Backbones:** LLaMA-7B, LLaMA2-7B
- **Adapted modules:** Query / Key / Value projections
- **Orthogonality coefficient:** `λ = 1e-4`
- **Dropout:** `0.05`
- **Batch size:** `4` per A100 GPU (40GB)

Representative D<sup>2</sup>MoRA settings reported in the paper include:

- **LLaMA-7B**
  - `{M = 3, N = 8, r = 8}`
  - `{M = 3, N = 4, r = 16}`

- **LLaMA2-7B**
  - `{M = 3, N = 8, r = 8}`
  - `{M = 4, N = 3, r = 16}`



Please use the official repository scripts for training and evaluation.


## 📝⭐️ Citation

If you find our work or these model weights useful in your research, please consider leaving a **Star** ⭐️ on our GitHub repo and citing our paper:

```bibtex
@inproceedings{zuo2026d2mora,
  title={D2MoRA: Diversity-Regulated Asymmetric MoE-LoRA Decomposition for Efficient Multi-Task Adaptation},
  author={Zuo, Jianhui and Song, Xuemeng and Wen, Haokun and Liu, Meng and Hu, Yupeng and Wang, Jiuru and Nie, Liqiang},
  booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
  volume={40},
  number={34},
  pages={29286--29294},
  year={2026}
}
```