Bank-balanced-sparse activation feature maps for neural network models
AssigneeMoffett Technologies Co., Limited
InventorsENXU YAN, DONGKUAN XU, JIACHAO LIU
Life of this patent
Estimating the term…
Abstract
Embodiments disclose bank-balanced-sparse activation neural network models and methods to generate the bank-balanced-sparse activation neural network models. According to one embodiment, a neural network sparsification engine determines a first deep neural network (DNN) model having two or more hidden layers. The engine determines a bank size, a bank layout, and a target sparsity. The engine segments the activation feature maps into a plurality of banks based on the bank size and the bank layout. The engine generates a second DNN model by increasing a sparsity for each bank of activation feature map based on the target sparsity, wherein the second DNN model is used for inferencing.
Claims21
A method to generate a deep neural network (DNN) model, comprising:
- determining a first deep neural network (DNN) model having one or more hidden layers;
- determining a bank size, a bank layout, and a target sparsity;
- grouping a plurality of activation feature maps into a plurality of banks based on the bank size and the bank layout;
- and
- generating a second DNN model by increasing a sparsity for each bank of an activation feature map based on the target sparsity, wherein the second DNN model is used for inferencing.
A non-transitory machine-readable medium having instructions stored therein, which when executed by a processor, cause the processor to perform operations, the operations comprising:
- determining a first deep neural network (DNN) model having one or more hidden layers;
- determining a bank size, a bank layout, and a target sparsity;
- grouping a plurality of activation feature maps into a plurality of banks based on the bank size and the bank layout;
- and
- generating a second DNN model by increasing a sparsity for each bank of an activation feature map based on the target sparsity, wherein the second DNN model is used for inferencing.
A bank-balanced-sparse activation deep neural network (DNN) model, comprising:
- an input layer;
- one or more hidden layers;
- and
- an output layer, wherein the one or more hidden layers are disposed between the input layer and the output layer, and each hidden layer is to output a plurality of bank-balanced-sparse activation feature maps,
- wherein the bank-balanced-sparse activation DNN model is generated by
- determining a first DNN model having one or more hidden layers,
- determining a bank size, a bank layout, and a target sparsity,
- grouping a plurality of activation feature maps into a plurality of banks based on the bank size and the bank layout, and
- generating a second DNN model as the bank-balanced-sparse activation DNN model by increasing a sparsity for each bank of an activation feature map based on the target sparsity, wherein the second DNN model is used for inferencing.
Dependencies are inferred from the claim text.