Invention Grant
- Patent Title: Architecture optimized training of neural networks
-
Application No.: US15677311Application Date: 2017-08-15
-
Publication No.: US11676004B2Publication Date: 2023-06-13
- Inventor: Kristof Denolf , Kornelis A. Vissers
- Applicant: Xilinx, Inc.
- Applicant Address: US CA San Jose
- Assignee: XILINX, INC.
- Current Assignee: XILINX, INC.
- Current Assignee Address: US CA San Jose
- Agency: Patterson + Sheridan, LLP
- Main IPC: G06N3/08
- IPC: G06N3/08 ; G06N3/084 ; G06N3/063 ; G06N3/04

Abstract:
An example a method of optimizing a neural network having a plurality of layers includes: obtaining an architecture constraint for circuitry of an inference platform that implements the neural network; training the neural network on a training platform to generate network parameters and feature maps for the plurality of layers; and constraining the network parameters, the feature maps, or both based on the architecture constraint.
Public/Granted literature
- US20190057305A1 ARCHITECTURE OPTIMIZED TRAINING OF NEURAL NETWORKS Public/Granted day:2019-02-21
Information query