-
公开(公告)号：US20240346317A1
公开(公告)日：2024-10-17
申请号：US18752163
申请日：2024-06-24
Applicant: Samsung Electronics Co., Ltd.
Inventor： Minkyoung CHO , Wonjo LEE , Seungwon LEE
Abstract: A method and apparatus for the pruning of a neural network is provided. The method sets a weight threshold value based on a weight distribution of layers included in a neural network, predicts a change of inference accuracy of a neural network by pruning of each layer based on the weight threshold value, determines a current subject layer to be pruned with a weight threshold value among the layers included in the neural network, and prunes a determined current subject layer.
-
公开(公告)号：US20240231944A1
公开(公告)日：2024-07-11
申请号：US18351737
申请日：2023-07-13
Applicant: Samsung Electronics Co., Ltd.
Inventor： Myeong Woo KIM , Yongdeok KIM , Changin CHOI , Seungwon LEE
CPC classification number: G06F9/5055 , G06F16/1724
Abstract: A processor-implemented method with data loading includes: based on sizes of a plurality of data files in a training dataset, dividing the training dataset into a plurality of sub-sets; loading some data files in each sub-set into a plurality of processors; determining a packing combination of one or more data files loaded to processors in a same group among the plurality of processors, based on a ratio of a number of data files between the plurality of sub-sets and a batch size of distributed training; determining packed data files by packing the one or more data files according to the packing combination; and reallocating the packed data files to the processors in the same group.
-
公开(公告)号：US20240069866A1
公开(公告)日：2024-02-29
申请号：US18115032
申请日：2023-02-28
Applicant: SAMSUNG ELECTRONICS CO., LTD.
Inventor： Jinseong KIM , Jinpil LEE , Seungwon LEE
CPC classification number: G06F7/4876 , G06F9/5027 , G06F17/16
Abstract: A method of performing a floating-point operation using a memory processor (the floating-point operation being a multiplication of a first matrix and a second matrix that are double-precision floating-point matrices) includes: determining whether an emulation is to be used to perform the floating-point operation, based on a result of the determining whether the emulation is to be used, determining whether to use the memory processor for the emulation, the emulation comprising stages, based on a result of the determining whether to use the memory processor for the emulation, individually determining whether to use the memory processor for each stage of the emulation, and multiplying the first matrix and the second matrix based on a result of the individually determining whether to use the memory processor.
-
公开(公告)号：US20230206031A1
公开(公告)日：2023-06-29
申请号：US18116553
申请日：2023-03-02
Applicant: Samsung Electronics Co., Ltd.
Inventor： Wonjo LEE , Seungwon LEE , Junhaeng LEE
Abstract: According to a method and apparatus for neural network quantization, a quantized neural network is generated by performing learning of a neural network, obtaining weight differences between an initial weight and an updated weight determined by the learning of each cycle for each of layers in the first neural network, analyzing a statistic of the weight differences for each of the layers, determining one or more layers, from among the layers, to be quantized with a lower-bit precision based on the analyzed statistic, and generating a second neural network by quantizing the determined one or more layers with the lower-bit precision.
-
公开(公告)号：US20220253248A1
公开(公告)日：2022-08-11
申请号：US17483083
申请日：2021-09-23
Applicant: Samsung Electronics Co., Ltd.
Inventor： Hosang YOON , Seungwon LEE
IPC: G06F3/06
Abstract: A memory controller includes: a first buffer configured to receive a memory request from a host and store therein the received memory request; a command generator configured to generate a first command corresponding to the memory request, and set a type of the first command indicating whether an address comprised in the memory request corresponds to a processing in memory (PIM) memory; a second buffer configured to store therein a plurality of commands comprising the first command; and a command scheduler configured to determine whether to change an order of the first command stored in the second buffer based on the type of the first command.
-
Patent Agency Ranking
公开(公告)号：US20220253247A1
公开(公告)日：2022-08-11
申请号：US17472990
申请日：2021-09-13
Applicant: SAMSUNG ELECTRONICS CO., LTD.
Inventor： Hyunsoo KIM , Seungwon LEE , Seungwoo SEO , Hosang YOON
IPC: G06F3/06
Abstract: A memory controller, a method of operating the memory controller, and an electronic device including the memory controller are disclosed. The method of operating a memory controller, comprising receiving, from a host core, a plurality of commands for a memory, identifying, from among the plurality of commands, processing in memory (PIM) commands to execute one or more operations in the memory, verifying ordering information from a data field in each of the PIM commands, and reordering the PIM commands based on the ordering information and transmitting the reordered PIM commands to the memory.
-
公开(公告)号：US20220253246A1
公开(公告)日：2022-08-11
申请号：US17408584
申请日：2021-08-23
Applicant: Samsung Electronics Co., Ltd.
Inventor： Seungwoo SEO , Seungwon LEE
IPC: G06F3/06
Abstract: A memory controller and a memory control method are disclosed. The memory controller includes a first buffer configured to receive a first memory request from a host and to store the first memory request, a request scheduler configured to determine an order in which the first memory request is transferred to a command generator, a request generator configured to generate one or more second memory requests based on a generation parameter of the first memory request, in response to an address of the first memory request corresponding to a processing in memory (PIM) memory, the command generator being configured to generate a first command corresponding to the first memory request and one or more second commands corresponding to the one or more second memory requests, and to store the generated first command and the one or more second commands in a second buffer, and a command scheduler configured to schedule the first command and the one or more second commands.
-
公开(公告)号：US20220188070A1
公开(公告)日：2022-06-16
申请号：US17689454
申请日：2022-03-08
Inventor： Seungwon LEE , Namhyung KIM , Hanmin PARK , Kiyoung CHOI
Abstract: A processor-implemented method of processing neural network data includes: setting first limit data by performing a first operation based on first input data and weight data generated from weights included in a filter; comparing the first limit data with an intermediate result of a second operation performed based on second input data and the weight data; and determining whether to perform a subsequent second operation based on a result of the comparing.
-
公开(公告)号：US20220156081A1
公开(公告)日：2022-05-19
申请号：US17323171
申请日：2021-05-18
Applicant: SAMSUNG ELECTRONICS CO., LTD.
Inventor： Seungwoo SEO , Hyunsoo KIM , Seungwon LEE
IPC: G06F9/38
Abstract: A processing-in-memory (PIM) and a method of outputting an instruction using a PIM. The PIM includes an internal processor, a memory, and a register configured to store instruction meta information. The memory is configured to store a lookup table generated by predicting a future instruction. The external processor is configured to retrieve an instruction corresponding to the instruction meta information from the lookup table in response to a PIM application programming interface (API) call instruction including the instruction meta information. The instruction corresponding to the instruction meta information is output based on a retrieval result.
-
公开(公告)号：US20220108180A1
公开(公告)日：2022-04-07
申请号：US17191954
申请日：2021-03-04
Applicant: Samsung Electronics Co., Ltd.
Inventor： Minkyoung CHO , Saerom CHOI , Seungwon LEE
IPC: G06N3/08
Abstract: A method and apparatus for compressing an artificial neural network may acquire weights corresponding to an artificial neural network trained in advance, wherein the artificial neural network includes a plurality of layers, and a processor configured to generate data for acquiring a change of behavior of the artificial neural network due to pruning of the artificial neural network based on the weights, determine a pruning threshold for pruning of the artificial neural network based on the change of the behavior of the artificial neural network, and compress the neural network based on the pruning threshold.
-
-
-
-
-
-
-
-
-
Search Results
56
records
(took 0.031 seconds)
