Back to results
US 12,632,303 B2Granted
Accelerator, method of operating the same, and electronic device including the same
AssigneesSamsung Electronics Co., Ltd., Seoul National University R&Db Foundation
InventorsWookeun Jung, Jaejin Lee, Seung Wook Lee
Life of this patent
Estimating the term…
Abstract
A processor-implemented accelerator method includes: reading, from a memory, an instruction to be executed in an accelerator; reading, from the memory, input data based on the instruction; and performing, on the input data and a parameter value included in the instruction, an inference task corresponding to the instruction.
Claims20
A processor-implemented method of operating an electronic device, the method comprising:
- reading an instruction to be executed in an accelerator included in the electronic device, the instruction embedding a parameter value of at least some portion of layers in a neural network for an inference task, wherein the reading of the instruction comprises reading a plurality of instructions each including a partial value of the parameter value to be used in the inference task, wherein a maximum length of the partial value of the parameter value included in each of the plurality of instructions is less than a preset threshold value;
- reading input data based on the instruction;
- reading, from the instruction, the parameter value for the inference task;
- and
- performing, on the input data and the parameter value embedded in the instruction, the inference task instructed by the instruction, and
- wherein the parameter value is a value fixed for the inference task.
3 more elements
An accelerator comprising:
- at least one processing element configured to:
- read an instruction to be executed in an accelerator, the instruction embedding a parameter value of at least some portion of layers in a neural network for an inference task, wherein the reading of the instruction comprises reading a plurality of instructions each including a partial value of the parameter value to be used in the inference task, wherein a maximum length of the partial value of the parameter value included in each of the plurality of instructions is less than a preset threshold value;
- read input data based on the instruction;
- reading, from the instruction, the parameter value for the inference task;
- and
- perform, on the input data and the parameter value embedded in the instruction, the inference task instructed by the instruction, and
- wherein the parameter value is a value fixed for the inference task.
4 more elements
An electronic device comprising:
- a memory configured to store an instruction to be executed in an accelerator;
- and
- the accelerator configured to:
- read the instruction from the memory, the instruction embedding a parameter value of at least some portion of layers in a neural network for an inference task,
- wherein the reading of the instruction comprises reading a plurality of instructions each including a partial value of the parameter value to be used in the inference task, wherein a maximum length of the partial value of the parameter value included in each of the plurality of instructions is less than a preset threshold value;
- read input data based on the instruction;
- read, from the instruction, the parameter value for the inference task;
- and
- perform, on the input data and the parameter value embedded in the instruction, the inference task instructed by the instruction, and
- wherein the parameter value is a value fixed for the inference task.
7 more elements
Dependencies are inferred from the claim text.