Quantization made by Richard Erkhov.

OpenMath-Mistral-7B-v0.1-hf - GGUF

Model creator: https://huggingface.co/nvidia/
Original model: https://huggingface.co/nvidia/OpenMath-Mistral-7B-v0.1-hf/

Name	Quant method	Size
OpenMath-Mistral-7B-v0.1-hf.Q2_K.gguf	Q2_K	2.53GB
OpenMath-Mistral-7B-v0.1-hf.IQ3_XS.gguf	IQ3_XS	2.81GB
OpenMath-Mistral-7B-v0.1-hf.IQ3_S.gguf	IQ3_S	2.96GB
OpenMath-Mistral-7B-v0.1-hf.Q3_K_S.gguf	Q3_K_S	2.95GB
OpenMath-Mistral-7B-v0.1-hf.IQ3_M.gguf	IQ3_M	3.06GB
OpenMath-Mistral-7B-v0.1-hf.Q3_K.gguf	Q3_K	3.28GB
OpenMath-Mistral-7B-v0.1-hf.Q3_K_M.gguf	Q3_K_M	3.28GB
OpenMath-Mistral-7B-v0.1-hf.Q3_K_L.gguf	Q3_K_L	3.56GB
OpenMath-Mistral-7B-v0.1-hf.IQ4_XS.gguf	IQ4_XS	3.67GB
OpenMath-Mistral-7B-v0.1-hf.Q4_0.gguf	Q4_0	3.83GB
OpenMath-Mistral-7B-v0.1-hf.IQ4_NL.gguf	IQ4_NL	3.87GB
OpenMath-Mistral-7B-v0.1-hf.Q4_K_S.gguf	Q4_K_S	3.86GB
OpenMath-Mistral-7B-v0.1-hf.Q4_K.gguf	Q4_K	4.07GB
OpenMath-Mistral-7B-v0.1-hf.Q4_K_M.gguf	Q4_K_M	4.07GB
OpenMath-Mistral-7B-v0.1-hf.Q4_1.gguf	Q4_1	4.24GB
OpenMath-Mistral-7B-v0.1-hf.Q5_0.gguf	Q5_0	4.65GB
OpenMath-Mistral-7B-v0.1-hf.Q5_K_S.gguf	Q5_K_S	4.65GB
OpenMath-Mistral-7B-v0.1-hf.Q5_K.gguf	Q5_K	4.78GB
OpenMath-Mistral-7B-v0.1-hf.Q5_K_M.gguf	Q5_K_M	4.78GB
OpenMath-Mistral-7B-v0.1-hf.Q5_1.gguf	Q5_1	5.07GB
OpenMath-Mistral-7B-v0.1-hf.Q6_K.gguf	Q6_K	5.53GB
OpenMath-Mistral-7B-v0.1-hf.Q8_0.gguf	Q8_0	7.17GB

Original model description:

license: apache-2.0 base_model: - mistralai/Mistral-7B-v0.1 datasets: - nvidia/OpenMathInstruct-1 language: - en tags: - nvidia - code - math

OpenMath-Mistral-7B-v0.1-hf

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

	greedy		majority@50
model	GSM8K	MATH	GMS8K	MATH
OpenMath-CodeLlama-7B (nemo \| HF)	75.9	43.6	84.8	55.6
OpenMath-Mistral-7B (nemo \| HF)	80.2	44.5	86.9	57.2
OpenMath-CodeLlama-13B (nemo \| HF)	78.8	45.5	86.8	57.6
OpenMath-CodeLlama-34B (nemo \| HF)	80.7	48.3	88.0	60.2
OpenMath-Llama2-70B (nemo \| HF)	84.7	46.3	90.1	58.3
OpenMath-CodeLlama-70B (nemo \| HF)	84.6	50.7	90.8	60.4

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

See our paper for more details!

How to use the models?

Try to run inference with our models with just a few commands!

Reproducing our results

We provide all instructions to fully reproduce our results.

Improving other models

To improve other models or to learn more about our code, read through the docs below.

NeMo-Skills Pipeline

In our pipeline we use NVIDIA NeMo, an end-to-end, cloud-native framework to build, customize, and deploy generative AI models anywhere. It includes training and inferencing frameworks, guardrailing toolkits, data curation tools, and pretrained models, offering enterprises an easy, cost-effective, and fast way to adopt generative AI.

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}
}