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. |
Year | Venue | Field |
---|---|---|
2018 | arXiv: Learning | Domain knowledge,Matrix decomposition,Momentum,Artificial intelligence,Contextual image classification,Order of magnitude,Mathematics,Machine learning,Branching (version control) |
DocType | Volume | Citations |
Journal | abs/1803.07445 | 0 |
PageRank | References | Authors |
0.34 | 33 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Henggang Cui | 1 | 307 | 11.66 |
Gregory R. Ganger | 2 | 4560 | 383.16 |
Phillip B. Gibbons | 3 | 6863 | 624.14 |