Document Type

Conference Proceeding

Publication Title

Advances in Neural Information Processing Systems

Abstract

Scaling up model sizes can lead to fundamentally new capabilities in many machine learning (ML) tasks. However, training big models requires strong distributed system expertise to carefully design model-parallel execution strategies that suit the model architectures and cluster setups. In this paper, we develop AMP, a framework that automatically derives such strategies. AMP identifies a valid space of model parallelism strategies and efficiently searches the space for high-performed strategies, by leveraging a cost model designed to capture the heterogeneity of the model and cluster specifications. Unlike existing methods, AMP is specifically tailored to support complex models composed of uneven layers and cluster setups with more heterogeneous accelerators and bandwidth. We evaluate AMP on popular models and cluster setups from public clouds and show that AMP returns parallel strategies that match the expert-tuned strategies on typical cluster setups. On heterogeneous clusters or models with heterogeneous architectures, AMP finds strategies with 1.54× and 1.77× higher throughput than state-of-the-art model-parallel systems, respectively.

Publication Date

12-2022

Keywords

Design models, Distributed systems, Execution strategies, Learning tasks, Machine-learning, Model size, Modeling architecture, Parallel executions, Parallel strategies, Scaling-up

Comments

IR conditions: non-described

Open Access version available on NeurIPS Proceedings

Download

Share

COinS