US 12,406,175 B2
Method and apparatus with model optimization, and accelerator system
Jae-Ki Hong, Seongnam-si (KR)
Assigned to Samsung Electronics Co., Ltd., Suwon-si (KR)
Filed by Samsung Electronics Co., Ltd., Suwon-si (KR)
Filed on Jul. 8, 2020, as Appl. No. 16/923,382.
Claims priority of application No. 10-2019-0163853 (KR), filed on Dec. 10, 2019.
Prior Publication US 2021/0174202 A1, Jun. 10, 2021
Int. Cl. G06N 3/063 (2023.01); G06F 9/50 (2006.01); G06F 18/20 (2023.01); G06F 18/21 (2023.01); G06N 3/08 (2023.01); G06N 20/00 (2019.01)
CPC G06N 3/063 (2013.01) [G06F 9/5027 (2013.01); G06F 18/217 (2023.01); G06F 18/29 (2023.01); G06N 3/08 (2013.01); G06N 20/00 (2019.01)] 24 Claims
OG exemplary drawing
 
1. A processor-implemented method with model optimization, comprising:
determining a graph representing operations to be performed in a target model related to input data input into the target model;
determining, by using a first machine learning (ML) model, separate from the target model, the first ML model receiving the input data, an attribute of the input data based on the inputted input data;
determining a predicted performance of the target model based on a behavior pattern of hardware that executes the target model;
optimizing the operations performed in the target model based on the graph, the attribute of the input data, and the predicted performance of the target model; and
executing the target model, including inputting the input data to the target model, by the hardware according to the optimized operations.