US 11,948,086 B2
Accelerated embedding layer computations
Rahul Nagarajan, Sunnyvale, CA (US); Lifeng Nai, Sunnyvale, CA (US); George Kurian, Sunnyvale, CA (US); and Hema Hariharan, Mountain View, CA (US)
Assigned to Google LLC, Mountain View, CA (US)
Filed by Google LLC, Mountain View, CA (US)
Filed on Apr. 21, 2023, as Appl. No. 18/305,297.
Application 18/305,297 is a continuation of application No. 16/659,527, filed on Oct. 21, 2019, granted, now 11,651,209.
Claims priority of provisional application 62/909,667, filed on Oct. 2, 2019.
Prior Publication US 2023/0376759 A1, Nov. 23, 2023
Int. Cl. G06N 3/08 (2023.01); G06F 1/03 (2006.01); G06N 3/063 (2023.01); G06N 20/10 (2019.01)
CPC G06N 3/08 (2013.01) [G06F 1/03 (2013.01); G06N 3/063 (2013.01); G06N 20/10 (2019.01)] 20 Claims
OG exemplary drawing
 
1. A method comprising:
identifying, in a request from a first processor, a set of addresses for locations in memory;
performing, by the first processor, a deduplication operation on the set of addresses by:
identifying duplicate addresses in the set of addresses;
generating a unique identifier mapped to the duplicate addresses; and
filtering the duplicate addresses from the set of addresses based on the unique identifier;
retrieving, by a second processor and using the unique identifier, data from memory locations identified by addresses in the set of addresses after the duplicate addresses are filtered from the set of addresses; and
generating an output in response to processing the data using the second processor.