少即是多：8位量化改善大型语言模型的持续学习
When Less is More: 8-bit Quantization Improves Continual Learning in Large Language Models
AI总结：
本文系统研究量化精度（FP16、INT8、INT4）与重放缓冲区策略在大型语言模型持续学习中的交互，发现量化模型在后续任务上优于FP16，INT8在学习可塑性与知识保留间取得最佳平衡，量化噪声充当隐式正则化防止过拟合。
摘要：
灾难性遗忘是持续学习中的根本挑战，尤其当模型为部署效率而量化时。我们系统研究了大型语言模型中量化精度（FP16、INT8、INT4）与重放缓冲区策略之间的相互作用，揭示了意想不到的动态。虽然FP16在初始任务上表现优异（NLU上74.44%），但我们在后续任务上观察到惊人的反转：量化模型在最终任务前向准确率上超过FP16 8-15%，其中INT4在代码生成任务上的性能几乎是FP16的两倍（40% vs 20%）。关键的是，即使是最小的重放缓冲区（0.1%）也能显著提高保留率——在所有精度水平下，经过数学训练后NLU保留率从45%提升至65%——而INT8始终在学习可塑性与知识保留之间取得最佳平衡。我们假设量化引入的噪声充当隐式正则化，防止了困扰高精度模型的对新任务梯度的过拟合。这些发现挑战了更高精度总是更好的传统观念，反而表明INT8量化既提供了计算效率又提供了更优的持续学习动态。我们的结果为在持续学习场景中部署压缩模型提供了实用指南：小重放缓冲区（1-2%）足以应对NLU任务，而数学和代码任务受益于中等缓冲区（5-10%），且量化模型相比FP16需要更少的重放即可达到相当的保留率。代码见此 https URL。
英文摘要：
Catastrophic forgetting poses a fundamental challenge in continual learning, particularly when models are quantized for deployment efficiency. We systematically investigate the interplay between quantization precision (FP16, INT8, INT4) and replay buffer strategies in large language models, revealing unexpected dynamics. While FP16 achieves superior initial task performance (74.44% on NLU), we observe a striking inversion on subsequent tasks: quantized models outperform FP16 by 8-15% on final task forward accuracy, with INT4 achieving nearly double FP16's performance on Code generation (40% vs 20%). Critically, even minimal replay buffers (0.1%) dramatically improve retention - increasing NLU retention after Math training from 45% to 65% across all precision levels - with INT8 consistently achieving the optimal balance between learning plasticity and knowledge retention. We hypothesize that quantization-induced noise acts as implicit regularization, preventing the overfitting to new task gradients that plagues high-precision models. These findings challenge the conventional wisdom that higher precision is always preferable, suggesting instead that INT8 quantization offers both computational efficiency and superior continual learning dynamics. Our results provide practical guidelines for deploying compressed models in continual learning scenarios: small replay buffers (1-2%) suffice for NLU tasks, while Math and Code benefit from moderate buffers (5-10%), with quantized models requiring less replay than FP16 to achieve comparable retention. Code is available at https://github.com/Festyve/LessIsMore.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
