Openmath Mistral 7B V0.1 Hf

🚀 OpenMath-Mistral-7B-v0.1-hf

OpenMath models are designed to solve mathematical problems by integrating text-based reasoning with code blocks executed by a Python interpreter. These models were trained on OpenMathInstruct-1, a math instruction tuning dataset featuring 1.8M problem-solution pairs generated using the permissively licensed Mixtral-8x7B model.

✨ Features

The OpenMath models have the following features:

• Integrate text-based reasoning with Python code execution to solve math problems.
• Trained on a large math instruction tuning dataset.
• The pipeline for producing these models is fully open-sourced.

📦 Installation

No installation steps are provided in the original README.

💻 Usage Examples

No code examples are provided in the original README.

📚 Documentation

Model Performance Table

How to Use the Models

You can run inference with our models with just a few commands.

Reproducing Results

We provide all instructions to fully reproduce our results.

Improving Other Models

To improve other models or learn more about our code, refer to the following docs:

• NeMo-Skills Pipeline
  • Generating synthetic data
  • Finetuning models
  • Evaluating models

In our pipeline, we use NVIDIA NeMo, an end-to-end, cloud-native framework for building, customizing, and deploying generative AI models anywhere. It includes training and inferencing frameworks, guardrailing toolkits, data curation tools, and pretrained models, providing enterprises with an easy, cost-effective, and fast way to adopt generative AI.

🔧 Technical Details

The pipeline used to produce these models is fully open-sourced:

• Code
• Models
• Dataset

See our paper for more details.

📄 License

The models are licensed under the Apache-2.0 license.

📚 Additional Information

📖 Citation

If you find our work useful, please consider citing us:

@article{toshniwal2024openmath,
  title   = {OpenMathInstruct-1: A 1.8 Million Math Instruction Tuning Dataset},
  author  = {Shubham Toshniwal and Ivan Moshkov and Sean Narenthiran and Daria Gitman and Fei Jia and Igor Gitman},
  year    = {2024},
  journal = {arXiv preprint arXiv: Arxiv-2402.10176}
}