Algorithm

Layer (electronics)

Theoretical computer science

Set (abstract data type)

Computer science

Scaling

Code (set theory)

Organic chemistry

Parameterized complexity

Mathematics

Inference

Artificial intelligence

Geometry

Chemistry

Programming language

Artificial neural network

Convolutional neural network

Machine learning

Convolution (computer science)

arXiv (Cornell University)

Convolutional layers are one of the basic building blocks of modern deep neural networks. One fundamental assumption is that convolutional kernels should be shared for all examples in a dataset. We propose conditionally parameterized convolutions (CondConv), which learn specialized convolutional kernels for each example. Replacing normal convolutions with CondConv enables us to increase the size and capacity of a network, while maintaining efficient inference. We demonstrate that scaling networks with CondConv improves the performance and inference cost trade-off of several existing convolutional neural network architectures on both classification and detection tasks. On ImageNet classification, our CondConv approach applied to EfficientNet-B0 achieves state-of-the-art performance of 78.3% accuracy with only 413M multiply-adds. Code and checkpoints for the CondConv Tensorflow layer and CondConv-EfficientNet models are available at: https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet/condconv.

CondConv: Conditionally Parameterized Convolutions for Efficient Inference