arXivDaily arXiv每日学术速递 周一至周五更新
北美退伍军人节(Veterans Day),12 日无更新
arXiv 2302.05045cs.LGcs.AIcs.DCcs.PF
收藏

利用剪枝神经网络中的稀疏性优化大规模模型训练

Exploiting Sparsity in Pruned Neural Networks to Optimize Large Model Training

University of Maryland(马里兰大学)

Siddharth Singh, Abhinav Bhatele

AI总结:

提出利用剪枝神经网络的稀疏性优化数据并行和层间并行中的内存与通信,集成于AxoNN框架,在512块V100 GPU上实现27亿参数模型内存减少74%、通信时间减少40%,整体加速34%。

摘要:

大规模神经网络的并行训练因通信带来的显著开销而颇具挑战性。近期,深度学习研究者开发了多种剪枝算法,能够将神经网络中80-90%的参数置零,从而生成与未剪枝的父网络精度相等的稀疏子网络。在本工作中,我们提出了一种新颖的方法,利用这些稀疏子网络来优化两种流行的并行深度学习算法——即数据并行和层间并行——中的内存利用和通信。我们将该方法集成到AxoNN中,这是一个高度可扩展的并行深度学习框架,依赖于数据和层间并行,并展示了通信时间和内存利用的减少。在512块NVIDIA V100 GPU上,我们的优化将27亿参数模型的内存消耗减少了74%,总通信时间减少了40%,从而相对于AxoNN实现了34%的整体加速,相对于DeepSpeed-3D实现了32%的加速,相对于稀疏矩阵计算基线Sputnik实现了46%的加速。

英文摘要:

Parallel training of neural networks at scale is challenging due to significant overheads arising from communication. Recently, deep learning researchers have developed a variety of pruning algorithms that are capable of pruning (i.e. setting to zero) 80-90% of the parameters in a neural network to yield sparse subnetworks that equal the accuracy of the unpruned parent network. In this work, we propose a novel approach that exploits these sparse subnetworks to optimize the memory utilization and communication in two popular algorithms for parallel deep learning namely -- data and inter-layer parallelism. We integrate our approach into AxoNN, a highly scalable framework for parallel deep learning that relies on data and inter-layer parallelism, and demonstrate the reduction in communication time and memory utilization. On 512 NVIDIA V100 GPUs, our optimizations reduce the memory consumption of a 2.7 billion parameter model by 74%, and the total communication time by 40%, thus providing an overall speedup of 34% over AxoNN, 32% over DeepSpeed-3D and 46% over Sputnik, a sparse matrix computation baseline.

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。

↑