-
RepBNN: towards a precise Binary Neural Network with Enhanced Feature Map via...
Binary neural network (BNN) is an extreme quantization version of convolutional neural networks (CNNs) with all features and weights mapped to just 1-bit. -
The CIFAR-10 dataset is used to evaluate the performance of FracBNN.
The CIFAR-10 dataset is used to evaluate the performance of FracBNN. -
The ImageNet dataset is used to evaluate the performance of FracBNN.
The ImageNet dataset is used to evaluate the performance of FracBNN. -
The first convolutional layer is not binarized.
The first convolutional layer is not binarized. -
Binary neural networks (BNNs) have 1-bit weights and activations.
Binary neural networks (BNNs) have 1-bit weights and activations. -
FracBNN: Accurate and FPGA-Efficient Binary Neural Networks with Fractional A...
FracBNN: Accurate and FPGA-Efficient Binary Neural Networks with Fractional Activations