arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1902.00465cs.LGcs.AIcs.DCstat.ML
收藏

TF-Replicator:面向研究人员的分布式机器学习

TF-Replicator: Distributed Machine Learning for Researchers

DeepMind(深度思维)

Peter Buchlovsky, David Budden, Dominik Grewe, Chris Jones, John Aslanides, Frederic Besse, Andy Brock, Aidan Clark, Sergio Gómez Colmenarejo, Aedan Pope, Fabio Viola, Dan Belov

摘要:

我们描述了TF-Replicator,这是一个为DeepMind研究人员设计的分布式机器学习框架,并作为TensorFlow的抽象实现。TF-Replicator简化了数据并行和模型并行研究代码的编写。使用同步或异步训练方案,可以毫不费力地将相同的模型部署到不同的集群架构(即一台或多台包含CPU、GPU或TPU加速器的机器)。为了证明TF-Replicator的通用性和可扩展性,我们实现并测试了三个非常不同的模型:(1)用于ImageNet分类的ResNet-50,(2)用于类条件ImageNet图像生成的SN-GAN,以及(3)用于连续控制的D4 PG强化学习代理。我们的研究结果表明,强大的可扩展性的性能,而不需要任何分布式系统的专业知识的用户。TF-Replicator编程模型将作为TensorFlow 2.0的一部分开源(参见https://github.com/tensorflow/community/pull/25)。

英文摘要:

We describe TF-Replicator, a framework for distributed machine learning designed for DeepMind researchers and implemented as an abstraction over TensorFlow. TF-Replicator simplifies writing data-parallel and model-parallel research code. The same models can be effortlessly deployed to different cluster architectures (i.e. one or many machines containing CPUs, GPUs or TPU accelerators) using synchronous or asynchronous training regimes. To demonstrate the generality and scalability of TF-Replicator, we implement and benchmark three very different models: (1) A ResNet-50 for ImageNet classification, (2) a SN-GAN for class-conditional ImageNet image generation, and (3) a D4PG reinforcement learning agent for continuous control. Our results show strong scalability performance without demanding any distributed systems expertise of the user. The TF-Replicator programming model will be open-sourced as part of TensorFlow 2.0 (see https://github.com/tensorflow/community/pull/25).

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。

↑