Effective and High Computing Algorithms for Convolution Neural Networks

  • Authors

    • P Syamala Rao
    • Dr G.P.SaradhiVarma
    • Rajasekhar Mutukuri
    2018-08-24
    https://doi.org/10.14419/ijet.v7i3.31.18203
  • Deep convolution neural networks, fast convolution neural network Algorithm, CGEM Methodology
  • Training a large set of data takes GPU days using Deep convolution neural networks which are a time taking process. Self-driving cars require very low latency for pedestrian detection. Image recognition constrained by limited processing resources for mobile phones. The computation speed of the training set determines that in these situations convolution neural networks was a success. For large filters, Conventional Faster Fourier Transform based convolution is preferably fast, yet in case of small, 3 × 3 filters state of the art convolutional neural networks is used. By using Winograd's minimal filtering algorithms the new class of fast algorithms for convolutional neural networks was introduced by us. Instead of small tiles, minimal complexity convolution was computed by the algorithms, this increases the computing speed with small batch sizes and small filters.  With the VGG network, we benchmark a GPU implementation of our algorithm and at batch sizes from 1 to 64 state of the art throughput was shown.

     

  • References

    1. [1] Christian Szegedy and Sergey Ioffe, “Batch normalization: Accelerating deep network training by reducing internal covariate shiftâ€, 2015 In arXiv: 1502.03167.

      [2] R. K. Srivastava, K. Greff, and J. Schmidhuber. “Training very deep networksâ€. 2015. arXiv:1507.06228.

      [3] S.Winograd. “Arithmetic complexity of computationsâ€, volume 33. Siam, 1980:241-248

      [4] ann LeCun.†Fast training of convolutional networks through fftsâ€. 2014, Computer Research Repository, abs: 1312.5353.

      [5] Antonio J. Plaza, Chein-I Chang “High Performance Computing in Remote Sensing†CRC-2015 210-222.

      [6] cuDNN http://docs.nvidia.com/deeplearning/sdk/cudnn-install/index.html acessed on 2018 March .

      [7] Jason Cong and Bingjun Xiao. “Optimizing FPGA-based Accelerator Design for Deep Convolutional Neural Networks.†volume 24.Computer Vision and Pattern Recognition.2015 282-294.

  • Downloads

  • How to Cite

    Syamala Rao, P., G.P.SaradhiVarma, D., & Mutukuri, R. (2018). Effective and High Computing Algorithms for Convolution Neural Networks. International Journal of Engineering & Technology, 7(3.31), 66-71. https://doi.org/10.14419/ijet.v7i3.31.18203