Accelerated Large Batch Optimization of BERT Pretraining in 54 minutes
arXiv:2006.13484
Abstract
BERT has recently attracted a lot of attention in natural language understanding (NLU) and achieved state-of-the-art results in various NLU tasks. However, its success requires large deep neural networks and huge amount of data, which result in long training time and impede development progress. Using stochastic gradient methods with large mini-batch has been advocated as an efficient tool to reduce the training time. Along this line of research, LAMB is a prominent example that reduces the training time of BERT from 3 days to 76 minutes on a TPUv3 Pod. In this paper, we propose an accelerated gradient method called LANS to improve the efficiency of using large mini-batches for training. As the learning rate is theoretically upper bounded by the inverse of the Lipschitz constant of the function, one cannot always reduce the number of optimization iterations by selecting a larger learning rate. In order to use larger mini-batch size without accuracy loss, we develop a new learning rate scheduler that overcomes the difficulty of using large learning rate. Using the proposed LANS method and the learning rate scheme, we scaled up the mini-batch sizes to 96K and 33K in phases 1 and 2 of BERT pretraining, respectively. It takes 54 minutes on 192 AWS EC2 P3dn.24xlarge instances to achieve a target F1 score of 90.5 or higher on SQuAD v1.1, achieving the fastest BERT training time in the cloud.
arXiv commentTechnical Report (not under reviewed in any venue)
References in corpus (7)
- ADADELTA: An Adaptive Learning Rate Method
- Accurate, Large Minibatch SGD: Training ImageNet in 1 Hour
- On the Convergence of Adam and Beyond
- Optimal Distributed Online Prediction using Mini-Batches
- Large Batch Training of Convolutional Networks
- AdaShift: Decorrelation and Convergence of Adaptive Learning Rate Methods
- Block-Normalized Gradient Method: An Empirical Study for Training Deep Neural Network