arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2401.09149cs.DC
收藏

InternEvo:通过混合并行与冗余分片实现高效长序列大语言模型训练

InternEvo: Efficient Long-sequence Large Language Model Training via Hybrid Parallelism and Redundant Sharding

Qiaoling Chen, Diandian Gu, Guoteng Wang, Xun Chen, YingTong Xiong, Ting Huang, Qinghao Hu, Xin Jin, Yonggang Wen, Tianwei Zhang, Peng Sun

AI总结:

针对长序列 LLM 训练效率低且与 FlashAttention 不兼容的问题,本文提出 InternEvo,通过分层分片解耦、混合并行、选择性通信重叠和内存管理提升模型 FLOPs 利用率。

摘要:

支持长序列的大语言模型(LLM)正开始为我们日常使用的越来越多根本性新应用提供动力。现有的长序列 LLM 训练方法既不高效,也无法与 FlashAttention 等常用训练算法兼容。我们设计了 InternEvo 来解决这些问题。InternEvo 将所有分片维度解耦到一个新的分层空间中,并系统分析 LLM 训练的内存与通信开销。随后,它生成有效的混合并行策略。我们设计了一种新的选择性重叠机制,以缓解混合并行带来的通信开销。我们还实现了内存管理技术,以减少 GPU 内存碎片。评估结果表明,InternEvo 生成的并行化策略在模型 FLOPs 利用率上与现有方法相当或更优。

英文摘要:

Large language models (LLMs) with long sequences begin to power more and more fundamentally new applications we use every day. Existing methods for long-sequence LLM training are neither efficient nor compatible with commonly-used training algorithms such as FlashAttention. We design InternEvo to address these issues. InternEvo decouples all of the sharding dimensions into a new hierarchical space, and systematically analyzes the memory and communication cost of LLM training. Then, it generates an effective hybrid parallelism strategy. We design a new selective overlap mechanism to mitigate the communication overhead introduced by the hybrid parallelism. We also implement memory management techniques to reduce GPU memory fragmentation. Evaluation results show that InternEvo generates parallelization strategies that match or outperform existing methods in model FLOPs utilization.

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。

↑