Inception-v4, Inception-ResNet and the Impact of Residual Connections on Learning

Christian Szegedy, Sergey Ioffe, Vincent Vanhoucke

Feb 18, 2016 (modified: Feb 18, 2016) ICLR 2016 workshop submission readers: everyone
  • Abstract: Very deep convolutional networks have been central to the largest advances in image recognition performance in recent years. One example is the Inception architecture that has been shown to achieve good performance at relatively low computational cost. Recently, the introduction of residual connections in conjunction with a more traditional architecture has yielded state-of-the-art performance in the 2015 ILSVRC challenge; its performance was similar to the latest generation Inception-v3 network. This raises the question of whether there are any benefit in combining the Inception architecture with residual connections. Here we give clear empirical evidence that training with residual connections accelerates the training of Inception networks significantly, however, when fully trained, the final quality of the non-residual Inception variants seem to be close to those of residual versions. We present several new streamlined architectures for both residual and non-residual Inception networks. With an ensemble of three residual and one pure Inception-v4, we achieve 3.08\% top-5 error on the test set of the ImageNet classification (CLS) challenge
  • Conflicts: