[1610.02357] Deep Learning with Separable Convolutions

  • The observation leads us to propose a novel deep convolutional neural network architecture inspired by Inception, where Inception modules have been replaced with separable convolutions.
  • We show that the architecture, dubbed Xception, slightly outperforms Inception V3 on the ImageNet dataset (which Inception V3 was designed for), and significantly outperforms Inception V3 on a larger image classification dataset comprising 350 million images and 17,000 classes.
  • A separable convolution can be understood as an Inception module with a maximally large number of towers.
  • Abstract: We present an interpretation of Inception modules in convolutional neural networks as being an intermediate step in-between regular convolution and the recently introduced “separable convolution” operation.
  • Since the Xception architecture has the same number of parameter as Inception V3, the performance gains are not due to increased capacity but rather to a more efficient use of model parameters.


@fchollet: New paper: “Deep learning with separable convolutions”. – exploring what’s next in convnet design after Inception.

Abstract: We present an interpretation of Inception modules in convolutional neural networks as being an intermediate step in-between regular convolution and the recently introduced “separable convolution” operation. In this light, a separable convolution can be understood as an Inception module with a maximally large number of towers. This observation leads us to propose a novel deep convolutional neural network architecture inspired by Inception, where Inception modules have been replaced with separable convolutions. We show that this architecture, dubbed Xception, slightly outperforms Inception V3 on the ImageNet dataset (which Inception V3 was designed for), and significantly outperforms Inception V3 on a larger image classification dataset comprising 350 million images and 17,000 classes. Since the Xception architecture has the same number of parameter as Inception V3, the performance gains are not due to increased capacity but rather to a more efficient use of model parameters.

Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)

Link back to: arXiv, form interface, contact.

[1610.02357] Deep Learning with Separable Convolutions

You might also like More from author

Comments are closed, but trackbacks and pingbacks are open.