MLtuner: System Support for Automatic Machine Learning Tuning

TL;DR

MLtuner is an automated system that efficiently tunes critical training parameters in large-scale machine learning tasks, improving robustness and speed over existing methods.

Contribution

It introduces an online trial-and-error approach with snapshotting and optimization to automatically tune ML training settings during execution.

Findings

Successfully tunes various ML models and datasets.

Outperforms state-of-the-art auto-tuning methods in robustness and speed.

Effective for large-scale ML applications.

Abstract

MLtuner automatically tunes settings for training tunables (such as the learning rate, the momentum, the mini-batch size, and the data staleness bound) that have a significant impact on large-scale machine learning (ML) performance. Traditionally, these tunables are set manually, which is unsurprisingly error-prone and difficult to do without extensive domain knowledge. MLtuner uses efficient snapshotting, branching, and optimization-guided online trial-and-error to find good initial settings as well as to re-tune settings during execution. Experiments show that MLtuner can robustly find and re-tune tunable settings for a variety of ML applications, including image classification (for 3 models and 2 datasets), video classification, and matrix factorization. Compared to state-of-the-art ML auto-tuning approaches, MLtuner is more robust for large problems and over an order of magnitude faster.