Title:
ニューラルネットワークプロセッサにおけるベクトル計算ユニット
Document Type and Number:
Japanese Patent JP7346510
Kind Code:
B2
Abstract:
A circuit for performing neural network computations for a neural network comprising a plurality of layers, the circuit comprising: activation circuitry configured to receive a vector of accumulated values and configured to apply a function to each accumulated value to generate a vector of activation values; and normalization circuitry coupled to the activation circuitry and configured to generate a respective normalized value from each activation value.
More Like This:
| JPH05204405 | DEVICE FOR ESTIMATING INPUT/OUTPUT RELATION |
| JP2019046282 | SEMICONDUCTOR DEVICE |
| JPH0622125 | PICTURE PROCESSING UNIT |
Inventors:
Gregory Michael, Thorson
Clark, Christopher Aaron
Lou, Dan
Clark, Christopher Aaron
Lou, Dan
Application Number:
JP2021148010A
Publication Date:
September 19, 2023
Filing Date:
September 10, 2021
Export Citation:
Assignee:
Google LLC
International Classes:
G06G7/60; G06N3/063
Domestic Patent References:
| JP2015036939A | ||||
| JP3253966A | ||||
| JP63206828A |
Foreign References:
| US20140180989 | ||||
| US20140280989 |
Other References:
石井 智大 他,ConvolutionalNeural Networkを用いた一般物体認識手法の解析,情報処理学会 研究報告 コンピュータビジョンとイメージメディア(CVIM),日本,情報処理学会,2014年05月08日,Vol.2014-CVIM-192 No.14,1-8,インターネット
Attorney, Agent or Firm:
Patent Attorney Fukami Patent Office
Previous Patent: Friction element for friction type shift element for vehicle transmission
Next Patent: molded nonwoven fabric
Next Patent: molded nonwoven fabric