Invention Grant
- Patent Title: Using multiple functional blocks for training neural networks
-
Application No.: US16191359Application Date: 2018-11-14
-
Publication No.: US11880769B2Publication Date: 2024-01-23
- Inventor: Sudhanva Gurumurthi
- Applicant: Advanced Micro Devices, Inc.
- Applicant Address: US CA Santa Clara
- Assignee: Advanced Micro Devices, Inc.
- Current Assignee: Advanced Micro Devices, Inc.
- Current Assignee Address: US CA Santa Clara
- Agency: Park, Vaughan, Fleming & Dowler LLP
- Main IPC: G06N3/084
- IPC: G06N3/084 ; G06N3/04 ; G06N3/065

Abstract:
A system is described that performs training operations for a neural network, the system including an analog circuit element functional block with an array of analog circuit elements, and a controller. The controller monitors error values computed using an output from each of one or more initial iterations of a neural network training operation, the one or more initial iterations being performed using neural network data acquired from the memory. When one or more error values are less than a threshold, the controller uses the neural network data from the memory to configure the analog circuit element functional block to perform remaining iterations of the neural network training operation. The controller then causes the analog circuit element functional block to perform the remaining iterations.
Public/Granted literature
- US20200151572A1 Using Multiple Functional Blocks for Training Neural Networks Public/Granted day:2020-05-14
Information query