AntBatchInfer: Elastic Batch Inference in the Kubernetes Cluster

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Li, Siyuan, Xiao, Youshao, Meng, Fanzhuang, Ju, Lin, Liang, Lei, Wang, Lin, Zhou, Jun
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866914755157950464
author Li, Siyuan
Xiao, Youshao
Meng, Fanzhuang
Ju, Lin
Liang, Lei
Wang, Lin
Zhou, Jun
author_facet Li, Siyuan
Xiao, Youshao
Meng, Fanzhuang
Ju, Lin
Liang, Lei
Wang, Lin
Zhou, Jun
contents Offline batch inference is a common task in the industry for deep learning applications, but it can be challenging to ensure stability and performance when dealing with large amounts of data and complicated inference pipelines. This paper demonstrated AntBatchInfer, an elastic batch inference framework, which is specially optimized for the non-dedicated cluster. AntBatchInfer addresses these challenges by providing multi-level fault-tolerant capabilities, enabling the stable execution of versatile and long-running inference tasks. It also improves inference efficiency by pipelining, intra-node, and inter-node scaling. It further optimizes the performance in complicated multiple-model batch inference scenarios. Through extensive experiments and real-world statistics, we demonstrate the superiority of our framework in terms of stability and efficiency. In the experiment, it outperforms the baseline by at least $2\times$ and $6\times$ in the single-model or multiple-model batch inference. Also, it is widely used at Ant Group, with thousands of daily jobs from various scenarios, including DLRM, CV, and NLP, which proves its practicability in the industry.
format Preprint
id arxiv_https___arxiv_org_abs_2404_09686
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle AntBatchInfer: Elastic Batch Inference in the Kubernetes Cluster
Li, Siyuan
Xiao, Youshao
Meng, Fanzhuang
Ju, Lin
Liang, Lei
Wang, Lin
Zhou, Jun
Machine Learning
Distributed, Parallel, and Cluster Computing
Offline batch inference is a common task in the industry for deep learning applications, but it can be challenging to ensure stability and performance when dealing with large amounts of data and complicated inference pipelines. This paper demonstrated AntBatchInfer, an elastic batch inference framework, which is specially optimized for the non-dedicated cluster. AntBatchInfer addresses these challenges by providing multi-level fault-tolerant capabilities, enabling the stable execution of versatile and long-running inference tasks. It also improves inference efficiency by pipelining, intra-node, and inter-node scaling. It further optimizes the performance in complicated multiple-model batch inference scenarios. Through extensive experiments and real-world statistics, we demonstrate the superiority of our framework in terms of stability and efficiency. In the experiment, it outperforms the baseline by at least $2\times$ and $6\times$ in the single-model or multiple-model batch inference. Also, it is widely used at Ant Group, with thousands of daily jobs from various scenarios, including DLRM, CV, and NLP, which proves its practicability in the industry.
title AntBatchInfer: Elastic Batch Inference in the Kubernetes Cluster
topic Machine Learning
Distributed, Parallel, and Cluster Computing
url https://arxiv.org/abs/2404.09686