MMOS

| ArXiv | Models | Data | Code |

🔥 News

• [2024/6/22] Revised the article and added attempts on automatic theorem proving tasks. Codes are in MMOS-F2F.
• [2024/3/30] Update result on MMOS-Code 34B and MMOS-LLEMMA 34B Notice the vllm and transformers version.
• [2024/3/8] 🔥🔥🔥Models MMOS-DeepSeekMath 7B show nice performence with self-consistency and k=50 !!
• [2024/2/28] 🔥 Models MMOS-DeepSeekMath 7B show nice performence and released at MMOS-DeepSeekMath 7B !!
• [2024/2/27] 🔥 Models MMOS-LLEMMA 7B show nice performence and released at MMOS-LLEMMA 7B !!
• [2024/2/27] 🔥 Models MMOS-CODE 13B and MMOS-CODE 34B released at MMOS-CODE 13B and MMOS-CODE 34B !!
• [2024/2/27] 🔥 Models MMOS-CODE 7B released at MMOS-CODE 7B !!
• [2024/2/26] 🔥🔥🔥 Dataset MMOS released at 😊 HuggingFace !!
• [2024/2/23] 🔥🔥🔥Arxiv released at An Empirical Study of Data Ability Boundary in LLMs' Math Reasoning ~

💡 Introductions & Performances

Mix of Minimal Optimal Sets (MMOS) of dataset has two advantages for two aspects, higher performance and lower construction costs on math reasoning.

| Model | Size | GSM8K | SVAMP | ASDiv | MATH | Size | GSM8K | SVAMP | ASDiv | MATH |Size | GSM8K | SVAMP | ASDiv | MATH | |------------------|------|-------|-------|-------|------|------|-------|-------|-------|------|------|-------|-------|-------|------| | WizardMath | 7B | 54.9 | 57.3 | 59.1 | 10.7 | 13B | 63.9 | 64.3 | 65.8 | 14.0 | 34B | - | - | - | - | | MAMMOTH | 7B | 53.6 | 67.7 | 31.5 | - | 13B | 62.0 | 72.4 | - | 34.2 | 34B | - | - | - | - | | MetaMath | 7B | 66.5 | - | - | 19.8 | 13B | 72.3 | - | - | 22.4 | 34B | - | - | - | - | | MathCoder-L | 7B | 64.2 | 71.5 | - | 23.3 | 13B | 72.6 | 76.9 | - | 29.9 | 34B | - | - | - | - | | MathCoder-CL | 7B | 67.8 | 70.7 | - | 30.2 | 13B | 74.1 | 78.0 | - | 35.9 | 34B | - | - | - | - | | TORA | 7B | 68.8 | 68.2 | 73.9 | 40.1 | 13B | 72.7 | 72.9 | 77.2 | 43.0 | 34B | - | - | - | - | | TORA-CODE | 7B | 72.6 | 70.4 | 78.7 | 44.6 | 13B | 75.8 | 75.7 | 81.4 | 48.1 | 34B | 80.7 | 80.5 | 84.2 | 50.8 | | MMOS | 7B | 69.9 | 73.4 | 76.8 | 40.2 | 13B | 74.8 | 77.0 | 80.0 | 43.2 | 34B | - | - | - | - | | MMOS-CODE | 7B | 73.9 | 76.4 | 78.6 | 44.3 | 13B | 77.1 | 77.5 | 81.9 | 48.1 | 34B |81.7|81.9|82.8|48.8| | MMOS-MinCODE | 7B | 70.3 | 72.5 | 76.7 | 44.6 | 13B | - | - | - | - | 34B | - | - | - | - | | MMOS-LLEMMA | 7B | 76.5 | 77.7 | 81.4 | 48.8 | 13B | - | - | - | - | 34B |82.8|81.8|84.8|51.3| | MMOS-DeepSeekMath | 7B | 80.5 | 79.3 | 87.6 | 55.0 | 13B | - | - | - | - | 34B | - | - | - | - | | MMOS-DeepSeekMath(SC,k=50) | 7B | 87.2 | - | - | 63.7 | 13B | - | - | - | - | 34B | - | - | - | - |

💾 Install

git clone https://github.com/cyzhh/MMOS.git
cd MMOS
conda create -n MMOS python=3.10 
conda activate MMOS
pip install -r requirements.txt

📚 Dataset

To identify the minimal optimal set, we follow these steps: 1) Sample a sufficient number of correct reasoning paths to form initial set. 2) Implement a deduplication algorithm to obtain its deduplicated subset. 3) Conduct a statistical analysis on the upper limit of reasoning paths per question k with the subset data amount N. 4) Perform SFT on several subsets to analyze the impact of removing duplicates and keeping varied reasoning paths.

We use ToRA series to generate QA-pairs from open source dataset GSM8K, MATH, TAL-SCQ. The QA-pairs are processed by our deduplication algorithm, resulting in the dataset MMOS. The total number of QA-pairs is 135K.

The DATA, which we publish at 😊 HuggingFace, need to be placed under the relative path, ./train_data/MMOS/.

If you are interested in our work, we will publish details about the data processing aspects after the paper is published.

Create your own MMOS dataset

Following scripts/generate.sh:

1. Prepare your sampling results.
2. Combine the results.
3. Extract the true cases.
4. Dedup the cases.
5. (Filter) and rerank.

⚙️ Auto Problem Generator

You can generate a data set for testing the numerical robustness of model performance by executing the following script command：

bash scripts/generate.sh
bash scripts/attack.sh
bash scripts/rerank.sh

🚀 Training

Due to resource constraints, we performed supervised fine-tuning on CodeLLaMA 7B, CodeLLaMA 13B and CodeLLaMA 34B using our dataset on A100 40G GPUs. To reproduce our work from CodeLLaMA 7B/13B, you can train according to the following instruction. You can also train the 34B model through DDP script instructions.

bash scripts/train_single.sh codellama 7b
bash scripts/train.sh codellama 34b

💻 Inference

bash scripts/infer.sh

📜 Citations

If you find this repository helpful, please consider citing our paper:

@misc{chen2024empirical,
      title={An Empirical Study of Data Ability Boundary in LLMs' Math Reasoning}, 
      author={Zui Chen and Yezeng Chen and Jiaqi Han and Zhijie Huang and Ji Qi and Yi Zhou},
      year={2024},
      eprint={2403.00799},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}

😇 Acknowledgements

• ToRA

🌟Star History