Add model card for DeepSeekMath-7B-Caco with metadata and links

README.md

@@ -0,0 +1,33 @@
+---
+license: apache-2.0
+library_name: transformers
+pipeline_tag: text-generation
+---
+
+# DeepSeekMath-7B-Caco
+
+This repository hosts the **DeepSeekMath-7B-Caco** model, which is fine-tuned using the Caco framework. Caco (Code-Assisted Chain-of-ThOught) is a novel framework that automates the synthesis of high-quality, verifiable, and diverse instruction-Chain-of-Thought (CoT) reasoning data through code-driven augmentation. It leverages executable code steps to generate comprehensive and logically correct reasoning paths, enhancing the reasoning capabilities of Large Language Models (LLMs).
+
+The **DeepSeekMath-7B-Caco** model is built upon the `deepseek-math-7b-base` and has been trained on the Caco-1.3M dataset, achieving strong competitive performance on mathematical reasoning benchmarks, as detailed in the research paper.
+
+For more details on the Caco framework, its methodology, and experimental results, please refer to the official paper:
+[**Scaling Code-Assisted Chain-of-Thoughts and Instructions for Model Reasoning**](https://huggingface.co/papers/2510.04081) (accepted by NeurIPS 2025).
+
+## Code and Further Information
+The official implementation, detailed instructions for data generation, training, and evaluation, can be found on the GitHub repository:
+[https://github.com/LHL3341/Caco](https://github.com/LHL3341/Caco)
+
+You can also find the Caco dataset and other Caco-trained models within the Hugging Face collection:
+[https://huggingface.co/collections/LHL3341/caco-68e0cb7b8a5f0071fac1f611](https://huggingface.co/collections/LHL3341/caco-68e0cb7b8a5f0071fac1f611)
+
+## Citation
+If you find this work useful, please cite the paper:
+
+```bibtex
+@article{caco,
+ title={Scaling Code-Assisted Chain-of-Thoughts and Instructions for Model Reasoning},
+ author={Honglin Lin and Qizhi Pei and Xin Gao and Zhuoshi Pan and Yu Li and Juntao Li and Conghui He and Lijun Wu},
+ journal={arXiv preprint arXiv:2510.04081},
+ year={2025}
+}
+```