Invention Grant
- Patent Title: Efficient convolutional neural networks
-
Application No.: US16676757Application Date: 2019-11-07
-
Publication No.: US11449729B2Publication Date: 2022-09-20
- Inventor: Lingchuan Meng , Danny Daysang Loh , Ian Rudolf Bratt , Alexander Eugene Chalfin , Tianmu Li
- Applicant: Arm Limited
- Applicant Address: GB Cambridge
- Assignee: Arm Limited
- Current Assignee: Arm Limited
- Current Assignee Address: GB Cambridge
- Agency: Leveque Intellectual Property Law, P.C.
- Main IPC: G06N3/04
- IPC: G06N3/04 ; G06N3/06 ; G06N3/08 ; G06N3/10 ; G06F17/15 ; G06F17/16 ; G06F17/18 ; G06F30/18 ; G06F30/20 ; G06F30/27 ; G06F30/33 ; G06F30/367

Abstract:
The present disclosure advantageously provides a system and a method for convolving data in a quantized convolutional neural network (CNN). The method includes selecting a set of complex interpolation points, generating a set of complex transform matrices based, at least in part, on the set of complex interpolation points, receiving an input volume from a preceding layer of the quantized CNN, performing a complex Winograd convolution on the input volume and at least one filter, using the set of complex transform matrices, to generate an output volume, and sending the output volume to a subsequent layer of the quantized CNN.
Public/Granted literature
- US20200151541A1 Efficient Convolutional Neural Networks Public/Granted day:2020-05-14
Information query