This repository contains a
Qwen2.5-1.5B student model distilled from
Qwen2.5-7B-Thinking using
Hybrid Policy Distillation (HPD), as presented in the paper
Hybrid Policy Distillation for LLMs.
Knowledge distillation (KD) is a powerful paradigm for compressing large language models (LLMs). Hybrid Policy Distillation (HPD) is a framework designed to make policy distillation more stable and efficient for reasoning-oriented models. It integrates the complementary advantages of forward and reverse KL to balance mode coverage and mode-seeking, and combines off-policy data with lightweight, approximate on-policy sampling.
The following table shows the performance of the distilled student model compared to the teacher model across various reasoning benchmarks:
If you find this model or the HPD framework useful in your research, please cite the following work:
1@article{hong2024hybrid,
2 title={Hybrid Policy Distillation for LLMs},
3 author={Hong, Zhang-Wei and others},
4 journal={arXiv preprint arXiv:2604.20244},
5 year={2024}
6}
