Corpus ID: 11164506

Deep Convolutional Neural Network Inference with Floating-point Weights and Fixed-point Activations

@article{Lai2017DeepCN,
  title={Deep Convolutional Neural Network Inference with Floating-point Weights and Fixed-point Activations},
  author={Liangzhen Lai and Naveen Suda and V. Chandra},
  journal={ArXiv},
  year={2017},
  volume={abs/1703.03073}
}
  • Liangzhen Lai, Naveen Suda, V. Chandra
  • Published 2017
  • Computer Science
  • ArXiv
  • Deep convolutional neural network (CNN) inference requires significant amount of memory and computation, which limits its deployment on embedded devices. [...] Key Method We show that using floating-point representation for weights is more efficient than fixed-point representation for the same bit-width and demonstrate it on popular large-scale CNNs such as AlexNet, SqueezeNet, GoogLeNet and VGG-16. We also show that such a representation scheme enables compact hardware multiply-and-accumulate (MAC) unit design…Expand Abstract
    52 Citations
    Phoenix: A Low-Precision Floating-Point Quantization Oriented Architecture for Convolutional Neural Networks
    Quantization of deep neural networks for accumulator-constrained processors
    • 1
    • PDF
    Deep Neural Network Approximation for Custom Hardware
    • 37
    • Highly Influenced
    • PDF
    Quantization of Constrained Processor Data Paths Applied to Convolutional Neural Networks
    • 2
    • PDF
    Low Precision Floating Point Arithmetic for High Performance FPGA-based CNN Acceleration
    An Energy-Efficient Sparse Deep-Neural-Network Learning Accelerator With Fine-Grained Mixed Precision of FP8–FP16
    • 1
    Effects of Approximate Multiplication on Convolutional Neural Networks
    A Variable Precision Approach for Deep Neural Networks

    References

    SHOWING 1-10 OF 29 REFERENCES
    Fixed Point Quantization of Deep Convolutional Networks
    • 434
    • Highly Influential
    • PDF
    Reduced-Precision Strategies for Bounded Memory in Deep Neural Nets
    • 75
    • Highly Influential
    • PDF
    Hardware-oriented Approximation of Convolutional Neural Networks
    • 227
    • Highly Influential
    • PDF
    Training deep neural networks with low precision multiplications
    • 351
    • Highly Influential
    • PDF
    Deep Learning with Limited Numerical Precision
    • 1,109
    • Highly Influential
    • PDF
    Accelerating Deep Convolutional Networks using low-precision and sparsity
    • 96
    • PDF
    Going Deeper with Embedded FPGA Platform for Convolutional Neural Network
    • 670
    • PDF
    Deep Compression: Compressing Deep Neural Network with Pruning, Trained Quantization and Huffman Coding
    • 3,844
    • Highly Influential
    • PDF