Enregistré dans:
| Auteurs principaux: | , , , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | https://arxiv.org/abs/2603.04135 |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866917314173075456 |
|---|---|
| author | Zhu, Haodong Ren, Yangyang Li, Yanjing Lin, Mingbao Yang, Linlin Liu, Xuhui Zhen, Xiantong Liu, Haiguang Zhang, Baochang |
| author_facet | Zhu, Haodong Ren, Yangyang Li, Yanjing Lin, Mingbao Yang, Linlin Liu, Xuhui Zhen, Xiantong Liu, Haiguang Zhang, Baochang |
| contents | Group Relative Policy Optimization (GRPO) effectively scales LLM reasoning but incurs prohibitive computational costs due to its extensive group-based sampling requirement. While recent selective data utilization methods can mitigate this overhead, they could induce estimation bias by altering the underlying sampling distribution, compromising theoretical rigor and convergence behavior. To address this limitation, we propose Dynamic Pruning Policy Optimization (DPPO), a framework that enables dynamic pruning while preserving unbiased gradient estimation through importance sampling-based correction. By incorporating mathematically derived rescaling factors, DPPO significantly accelerates GRPO training without altering the optimization objective of the full-batch baseline. Furthermore, to mitigate the data sparsity induced by pruning, we introduce Dense Prompt Packing, a window-based greedy strategy that maximizes valid token density and hardware utilization. Extensive experiments demonstrate that DPPO consistently accelerates training across diverse models and benchmarks. For instance, on Qwen3-4B trained on MATH, DPPO achieves 2.37$\times$ training speedup and outperforms GRPO by 3.36% in average accuracy across six mathematical reasoning benchmarks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_04135 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Unbiased Dynamic Pruning for Efficient Group-Based Policy Optimization Zhu, Haodong Ren, Yangyang Li, Yanjing Lin, Mingbao Yang, Linlin Liu, Xuhui Zhen, Xiantong Liu, Haiguang Zhang, Baochang Machine Learning Artificial Intelligence Group Relative Policy Optimization (GRPO) effectively scales LLM reasoning but incurs prohibitive computational costs due to its extensive group-based sampling requirement. While recent selective data utilization methods can mitigate this overhead, they could induce estimation bias by altering the underlying sampling distribution, compromising theoretical rigor and convergence behavior. To address this limitation, we propose Dynamic Pruning Policy Optimization (DPPO), a framework that enables dynamic pruning while preserving unbiased gradient estimation through importance sampling-based correction. By incorporating mathematically derived rescaling factors, DPPO significantly accelerates GRPO training without altering the optimization objective of the full-batch baseline. Furthermore, to mitigate the data sparsity induced by pruning, we introduce Dense Prompt Packing, a window-based greedy strategy that maximizes valid token density and hardware utilization. Extensive experiments demonstrate that DPPO consistently accelerates training across diverse models and benchmarks. For instance, on Qwen3-4B trained on MATH, DPPO achieves 2.37$\times$ training speedup and outperforms GRPO by 3.36% in average accuracy across six mathematical reasoning benchmarks. |
| title | Unbiased Dynamic Pruning for Efficient Group-Based Policy Optimization |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2603.04135 |