论文标题
repbnn：通过重复进行具有增强功能图的精确二进制神经网络
RepBNN: towards a precise Binary Neural Network with Enhanced Feature Map via Repeating
论文作者
论文摘要
二进制神经网络（BNN）是卷积神经网络（CNN）的极端量化版本，其所有功能和权重映射到仅1位。尽管BNN节省了大量的内存和计算需求，以使CNN适用于边缘或移动设备，但由于二进制后的表示能力降低，BNN遭受了网络性能的下降。在本文中，我们提出了一个新的可更换且易于使用的卷积模块reponv，该模块通过复制输入或沿信道维度的输出来增强特征地图，而无需额外的参数和卷积计算费用。我们还定义了一组Reptran规则，可以在BNN模块中使用Repconv，例如二进制卷积，完全连接的层和批处理标准化。实验表明，在Reptran转换之后，一组高度引用的BNN与原始BNN版本相比，实现了普遍的性能。例如，Rep-Recu-Resnet-20的前1位准确性，即REPBCONV增强的RECU-RESNET-20，在CIFAR-10上达到了88.97％，比原始网络高1.47％。 Rep-Adambnn-Reactnet-A在Imagenet上获得了71.342％的TOP-1精度，这是BNN的最新结果。代码和型号可在以下网址提供：https：//github.com/imfinethanks/rep_adambnn。
Binary neural network (BNN) is an extreme quantization version of convolutional neural networks (CNNs) with all features and weights mapped to just 1-bit. Although BNN saves a lot of memory and computation demand to make CNN applicable on edge or mobile devices, BNN suffers the drop of network performance due to the reduced representation capability after binarization. In this paper, we propose a new replaceable and easy-to-use convolution module RepConv, which enhances feature maps through replicating input or output along channel dimension by $β$ times without extra cost on the number of parameters and convolutional computation. We also define a set of RepTran rules to use RepConv throughout BNN modules like binary convolution, fully connected layer and batch normalization. Experiments demonstrate that after the RepTran transformation, a set of highly cited BNNs have achieved universally better performance than the original BNN versions. For example, the Top-1 accuracy of Rep-ReCU-ResNet-20, i.e., a RepBconv enhanced ReCU-ResNet-20, reaches 88.97% on CIFAR-10, which is 1.47% higher than that of the original network. And Rep-AdamBNN-ReActNet-A achieves 71.342% Top-1 accuracy on ImageNet, a fresh state-of-the-art result of BNNs. Code and models are available at:https://github.com/imfinethanks/Rep_AdamBNN.
