Yet Another Accelerated SGD: ResNet-50 Training on ImageNet in 74.7 seconds
This addresses the demand for faster machine learning execution, particularly in distributed settings, but appears incremental as it builds on existing accelerated SGD methods.
The paper tackled the challenge of accelerating distributed deep learning with large mini-batches without compromising accuracy, achieving a training time of 74.7 seconds for ResNet-50 on ImageNet using 2,048 GPUs, with a throughput of over 1.73 million images/sec and 75.08% top-1 validation accuracy.
There has been a strong demand for algorithms that can execute machine learning as faster as possible and the speed of deep learning has accelerated by 30 times only in the past two years. Distributed deep learning using the large mini-batch is a key technology to address the demand and is a great challenge as it is difficult to achieve high scalability on large clusters without compromising accuracy. In this paper, we introduce optimization methods which we applied to this challenge. We achieved the training time of 74.7 seconds using 2,048 GPUs on ABCI cluster applying these methods. The training throughput is over 1.73 million images/sec and the top-1 validation accuracy is 75.08%.