Accurate, Large Minibatch SGD: Training ImageNet in 1 Hour

by   Priya Goyal, et al.

Deep learning thrives with large neural networks and large datasets. However, larger networks and larger datasets result in longer training times that impede research and development progress. Distributed synchronous SGD offers a potential solution to this problem by dividing SGD minibatches over a pool of parallel workers. Yet to make this scheme efficient, the per-worker workload must be large, which implies nontrivial growth in the SGD minibatch size. In this paper, we empirically show that on the ImageNet dataset large minibatches cause optimization difficulties, but when these are addressed the trained networks exhibit good generalization. Specifically, we show no loss of accuracy when training with large minibatch sizes up to 8192 images. To achieve this result, we adopt a linear scaling rule for adjusting learning rates as a function of minibatch size and develop a new warmup scheme that overcomes optimization challenges early in training. With these simple techniques, our Caffe2-based system trains ResNet-50 with a minibatch size of 8192 on 256 GPUs in one hour, while matching small minibatch accuracy. Using commodity hardware, our implementation achieves 90 GPUs. This system enables us to train visual recognition models on internet-scale data with high efficiency.


page 1

page 2

page 3

page 4


Layered SGD: A Decentralized and Synchronous SGD Algorithm for Scalable Deep Neural Network Training

Stochastic Gradient Descent (SGD) is the most popular algorithm for trai...

An argument in favor of strong scaling for deep neural networks with small datasets

In recent years, with the popularization of deep learning frameworks and...

Extremely Large Minibatch SGD: Training ResNet-50 on ImageNet in 15 Minutes

We demonstrate that training ResNet-50 on ImageNet for 90 epochs can be ...

AMPNet: Asynchronous Model-Parallel Training for Dynamic Neural Networks

New types of machine learning hardware in development and entering the m...

Hidden Progress in Deep Learning: SGD Learns Parities Near the Computational Limit

There is mounting empirical evidence of emergent phenomena in the capabi...

Yet Another Accelerated SGD: ResNet-50 Training on ImageNet in 74.7 seconds

There has been a strong demand for algorithms that can execute machine l...

Please sign up or login with your details

Forgot password? Click here to reset