公开(公告)号：US20250077839A1

公开(公告)日：2025-03-06

申请号：US18239764

申请日：2023-08-30

Applicant: MEDIATEK INC.

Inventor： Chia-Hsiang Liu ,  Cheng-Sheng Chan ,  Min-Fong Horng ,  Sheng-Je Hung ,  Hung-Jen Chen ,  Jui-Yang Hsu

IPC: G06N3/045 ,  G06N3/082

Abstract: A method for generating a dynamic neural network includes: utilizing a neural architecture search (NAS) method to obtain a searched result, wherein the searched result comprises a plurality of sub-networks; combining the plurality of sub-networks to generate a combined neural network; and fine-tuning the combined neural network to generate the dynamic neural network.

公开(公告)号：US20240119283A1

公开(公告)日：2024-04-11

申请号：US18377315

申请日：2023-10-06

Applicant: MEDIATEK INC.

Inventor： Jui-Yang Hsu ,  Cheng-Sheng Chan ,  Jen-Chieh Tsai ,  Huai-Ting Li ,  Bo-Yu Kuo ,  Yen-Hao Chen ,  Kai-Ling Huang ,  Ping-Yuan Tseng ,  Tao Tu ,  Sheng-Je Hung

IPC: G06N3/08

CPC classification number: G06N3/08

Abstract: A method of performing automatic tuning on a deep learning model includes: utilizing an instruction-based learned cost model to estimate a first type of operational performance metrics based on a tuned configuration of layer fusion and tensor tiling; utilizing statistical data gathered during a compilation process of the deep learning model to determine a second type of operational performance metrics based on the tuned configuration of layer fusion and tensor tiling; performing an auto-tuning process to obtain a plurality of optimal configurations based on the first type of operational performance metrics and the second type of operational performance metrics; and configure the deep learning model according to one of the plurality of optimal configurations.

公开(公告)号：US20250156708A1

公开(公告)日：2025-05-15

申请号：US18939492

申请日：2024-11-06

Applicant: MEDIATEK INC.

Inventor： Tzu-Yun Chien ,  Bo-Yu Kuo ,  Jui-Yang Hsu ,  Kai-Ling Huang ,  Sheng-Je Hung

IPC: G06N3/08

Abstract: A method for optimizing deep learning models includes: initializing a plurality of pools, each including a plurality of candidate solutions; concurrently performing a plurality of tuning algorithms respectively within the plurality of pools during a single tuning run, thereby obtaining a plurality of selected candidate solutions; and generating an optimized model configuration based on the plurality of selected candidate solutions.

公开(公告)号：US20250077892A1

公开(公告)日：2025-03-06

申请号：US18239759

申请日：2023-08-30

Applicant: MEDIATEK INC.

Inventor： Chun-Wei Yang ,  Cheng-Sheng Chan ,  Min-Fong Horng ,  Jui-Yang Hsu ,  Sheng-Je Hung

IPC: G06N3/0985

Abstract: A neural network optimization method with a preview mechanism includes: in a preview stage, building an optimization space and obtaining multiple previewed results from the optimization space; generating an updating signal according to a reference value corresponding to the multiple previewed results, and processing the optimization space received in the preview stage according to the reference value; and in a view stage, receiving the optimization space and the updating signal, and processing the optimization space received in the view stage according to the updating signal to generate an optimization result.

公开(公告)号：US20250077838A1

公开(公告)日：2025-03-06

申请号：US18239753

申请日：2023-08-29

Applicant: MEDIATEK INC.

Inventor： Chia-Hsiang Liu ,  Cheng-Sheng Chan ,  Min-Fong Horng ,  Jui-Yang Hsu ,  Sheng-Je Hung ,  Hung-Jen Chen

IPC: G06N3/045

Abstract: A system for dynamically adjusting neural network efficiency of a dynamic neural network running on a device includes a detector and a signal generator. The detector is arranged to detect a change of a status of the device, to generate a trigger signal. The signal generator is arranged to generate a control signal according to the trigger signal, to dynamically adjust the neural network efficiency of the dynamic neural network.