Systems and methods for providing vector-wise sparsity in a neural network
AssigneeAlibaba Group Holding Limited
InventorsMaohua Zhu, Tao Zhang, Zhenyu Gu, Yuan Xie
Life of this patent
Estimating the term…
Abstract
The present disclosure relates to systems and methods for providing vector-wise sparsity in neural networks. In some embodiments, an exemplary method for providing vector-wise sparsity in a neural network, comprises: dividing a matrix associated with the neural network into a plurality of vectors; selecting a first subset of non-zero elements from the plurality of vectors to form a pruned matrix; and outputting the pruned matrix for executing the neural network using the pruned matrix.
Claims23
A method for providing vector-wise sparsity in a neural network, comprising:
- dividing a matrix associated with the neural network into a plurality of vectors;
- selecting a first subset of non-zero elements from the plurality of vectors to form a pruned matrix;
- and
- outputting the pruned matrix for executing the neural network using the pruned matrix.
A system for providing vector-wise sparsity in a neural network, comprising:
- at least one memory for storing instructions;
- and
- at least one processor to execute the instructions to cause the system to perform:
- dividing a matrix associated with the neural network into a plurality of vectors;
- selecting a first subset of non-zero elements from the plurality of vectors to form a pruned matrix;
- and
- outputting the pruned matrix for executing the neural network using the pruned matrix.
A non-transitory computer-readable storage medium storing a set of instructions that is executable by at least one processor to cause the computer to perform a method for providing vector-wise sparsity in a neural network, the method comprising:
- dividing a matrix associated with the neural network into a plurality of vectors;
- selecting a first subset of non-zero elements from the plurality of vectors to form a pruned matrix;
- and
- outputting the pruned matrix for executing the neural network using the pruned matrix.
A processor for executing a neural network, comprising:
- a memory for storing associated offsets of elements of compact vectors within original vectors;
- and
- at least one core configured to decode offsets of elements of a first plurality of compact vectors and perform a matrix multiplication of a first operand and a second operand based on the decoded offsets, the first operand comprising the first plurality of compact vectors.
Dependencies are inferred from the claim text.