ICML 2020poster20 citations

On Efficient Constructions of Checkpoints

Yu Chen, Zhenming Liu, Bin Ren, Xin Jin

Abstract

Efficient construction of checkpoints/snapshots is a critical tool for training and diagnosing deep learning models. In this paper, we propose a lossy compression scheme for checkpoint constructions (called LC-Checkpoint). LC-Checkpoint simultaneously maximizes the compression rate and optimizes the recovery speed, under the assumption that SGD is used to train the model. LC-Checkpoint uses quantization and priority promotion to store the most crucial information for SGD to recover, and then uses a Huffman coding to leverage the non-uniform distribution of the gradient scales. Our extensive experiments show that LC-Checkpoint achieves a compression rate up to 28{\texttimes} and recovery speedup up to 5.77{\texttimes} over a state-of-the-art algorithm (SCAR).

BibTeX
@InProceedings{pmlr-v119-chen20m,
  title = 	 {On Efficient Constructions of Checkpoints},
  author =       {Chen, Yu and Liu, Zhenming and Ren, Bin and Jin, Xin},
  booktitle = 	 {Proceedings of the 37th International Conference on Machine Learning},
  pages = 	 {1627--1636},
  year = 	 {2020},
  editor = 	 {III, Hal Daumé and Singh, Aarti},
  volume = 	 {119},
  series = 	 {Proceedings of Machine Learning Research},
  month = 	 {13--18 Jul},
  publisher =    {PMLR},
  pdf = 	 {http://proceedings.mlr.press/v119/chen20m/chen20m.pdf},
  url = 	 {https://proceedings.mlr.press/v119/chen20m.html},
  abstract = 	 {Efficient construction of checkpoints/snapshots is a critical tool for training and diagnosing deep learning models. In this paper, we propose a lossy compression scheme for checkpoint constructions (called LC-Checkpoint). LC-Checkpoint simultaneously maximizes the compression rate and optimizes the recovery speed, under the assumption that SGD is used to train the model. LC-Checkpoint uses quantization and priority promotion to store the most crucial information for SGD to recover, and then uses a Huffman coding to leverage the non-uniform distribution of the gradient scales. Our extensive experiments show that LC-Checkpoint achieves a compression rate up to 28{\texttimes} and recovery speedup up to 5.77{\texttimes} over a state-of-the-art algorithm (SCAR).}
}
On Efficient Constructions of Checkpoints · ICML 2020