Parle: parallelizing stochastic gradient descent
This addresses the need for efficient parallel training in deep learning, offering a practical solution for both single-machine and distributed settings without extra hyper-parameters.
The paper tackles the problem of slow convergence in parallel deep network training by proposing Parle, which converges 2-4x faster than data-parallel SGD and achieves nearly state-of-the-art error rates on benchmarks like CIFAR-10 and CIFAR-100.
We propose a new algorithm called Parle for parallel training of deep networks that converges 2-4x faster than a data-parallel implementation of SGD, while achieving significantly improved error rates that are nearly state-of-the-art on several benchmarks including CIFAR-10 and CIFAR-100, without introducing any additional hyper-parameters. We exploit the phenomenon of flat minima that has been shown to lead to improved generalization error for deep networks. Parle requires very infrequent communication with the parameter server and instead performs more computation on each client, which makes it well-suited to both single-machine, multi-GPU settings and distributed implementations.