Large-Batch Training for LSTM and Beyond
This work addresses the problem of accelerating training for deep learning models like LSTMs and CNNs, offering a practical solution for researchers and practitioners, though it is incremental as it builds on existing large-batch techniques.
The paper tackles the challenge of large-batch training for deep neural networks, proposing a linear-epoch gradual-warmup (LEGW) approach that enables scaling batch sizes up to 64x for LSTMs without accuracy loss and achieves a 5.3x average speedup over baselines for LSTM applications.
Large-batch training approaches have enabled researchers to utilize large-scale distributed processing and greatly accelerate deep-neural net (DNN) training. For example, by scaling the batch size from 256 to 32K, researchers have been able to reduce the training time of ResNet50 on ImageNet from 29 hours to 2.2 minutes (Ying et al., 2018). In this paper, we propose a new approach called linear-epoch gradual-warmup (LEGW) for better large-batch training. With LEGW, we are able to conduct large-batch training for both CNNs and RNNs with the Sqrt Scaling scheme. LEGW enables Sqrt Scaling scheme to be useful in practice and as a result we achieve much better results than the Linear Scaling learning rate scheme. For LSTM applications, we are able to scale the batch size by a factor of 64 without losing accuracy and without tuning the hyper-parameters. For CNN applications, LEGW is able to achieve the same accuracy even as we scale the batch size to 32K. LEGW works better than previous large-batch auto-tuning techniques. LEGW achieves a 5.3X average speedup over the baselines for four LSTM-based applications on the same hardware. We also provide some theoretical explanations for LEGW.