Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,58 @@
|
|
1 |
---
|
2 |
license: mit
|
|
|
|
|
|
|
|
|
|
|
3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
license: mit
|
3 |
+
language:
|
4 |
+
- en
|
5 |
+
metrics:
|
6 |
+
- accuracy
|
7 |
+
pipeline_tag: text-generation
|
8 |
---
|
9 |
+
# MathCoder: Seamless Code Integration in LLMs for Enhanced Mathematical Reasoning
|
10 |
+
|
11 |
+
Paper: [https://arxiv.org/pdf/2310.03731.pdf](https://arxiv.org/pdf/2310.03731.pdf)
|
12 |
+
|
13 |
+
Repo: [https://github.com/mathllm/MathCoder](https://github.com/mathllm/MathCoder)
|
14 |
+
|
15 |
+
|
16 |
+
## Introduction
|
17 |
+
We introduce MathCoder, a series of open-source large language models (LLMs) specifically tailored for general math problem-solving.
|
18 |
+
|
19 |
+
| Base Model: Llama-2 | Base Model: Code Llama |
|
20 |
+
|-------------------------------------------------------------------|-----------------------------------------------------------------------|
|
21 |
+
| [MathCoder-L-7B](https://huggingface.co/MathLLM/MathCoder-L-7B) | [MathCoder-CL-7B](https://huggingface.co/MathLLM/MathCoder-CL-7B) |
|
22 |
+
| [MathCoder-L-13B](https://huggingface.co/MathLLM/MathCoder-L-13B) | [MathCoder-CL-34B](https://huggingface.co/MathLLM/MathCoder-CL-34B) |
|
23 |
+
|
24 |
+
|
25 |
+
## Training Data
|
26 |
+
The models are trained on the [MathCodeInstruct](https://huggingface.co/datasets/MathLLM/MathCodeInstruct) Dataset.
|
27 |
+
|
28 |
+
|
29 |
+
## Training Procedure
|
30 |
+
The models are fine-tuned with the MathCodeInstruct dataset using the original Llama-2 and CodeLlama models as base models. Check out our paper and repo for more details.
|
31 |
+
|
32 |
+
## Evaluation
|
33 |
+
|
34 |
+
<br>
|
35 |
+
<div align="center">
|
36 |
+
<img src="result.png" width="100%" title="Result Figure">
|
37 |
+
</div>
|
38 |
+
|
39 |
+
|
40 |
+
|
41 |
+
## Usage
|
42 |
+
You can use the models through Huggingface's Transformers library. Use the pipeline function to create a text-generation pipeline with the model of your choice, then feed in a math problem to get the solution.
|
43 |
+
Check our Github repo for datails.
|
44 |
+
|
45 |
+
|
46 |
+
## Citation
|
47 |
+
Please cite the paper if you use our data, model or code.
|
48 |
+
|
49 |
+
```
|
50 |
+
@misc{wang2023mathcoder,
|
51 |
+
title={MathCoder: Seamless Code Integration in LLMs for Enhanced Mathematical Reasoning},
|
52 |
+
author={Ke Wang and Houxing Ren and Aojun Zhou and Zimu Lu and Sichun Luo and Weikang Shi and Renrui Zhang and Linqi Song and Mingjie Zhan and Hongsheng Li},
|
53 |
+
year={2023},
|
54 |
+
eprint={2310.03731},
|
55 |
+
archivePrefix={arXiv},
|
56 |
+
primaryClass={cs.CL}
|
57 |
+
}
|
58 |
+
```
|