Computer programs that use decoders, encoders, methods, and sorting to provide decoding parameters for neural networks.

The decoder and encoder system addresses inefficiencies in tensor-based neural network parameter processing by rearranging dimensions with a single scalar shift, enhancing efficiency and flexibility while maintaining compatibility with existing frameworks.

JP7868173B2Active Publication Date: 2026-06-01FRAUNHOFER GESELLSCHAFT ZUR FORDERUNG DER ANGEWANDTEN FORSCHUNG EV
View PDF 3 Cites 0 Cited by

Patent Information

Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
FRAUNHOFER GESELLSCHAFT ZUR FORDERUNG DER ANGEWANDTEN FORSCHUNG EV
Filing Date
2023-04-13
Publication Date
2026-06-01

Smart Images

  • Figure 0007868173000364
    Figure 0007868173000364
  • Figure 0007868173000365
    Figure 0007868173000365
  • Figure 0007868173000366
    Figure 0007868173000366
Patent Text Reader

Abstract

An embodiment according to the invention relates to a decoder for providing decoded parameters of a neural network based on an encoded representation, the decoder configured to use decoding of neural network parameters to obtain a first multidimensional array comprising a plurality of neural network parameter values, the decoder configured to use reordering to obtain a reordered multidimensional array, wherein a first dimension of the first multidimensional array is rearranged to a different dimension within the reordered multidimensional array. Further disclosed is an encoder, a method and a computer program using reordering.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Neural network model compression and optimization

    WO2020190772A1

  • Method and apparatus for three-dimensional (3D)-tree coding for neural network model compression

    WO2021102123A1

  • Neural network model compression

    WO2021211522A1