operation method and device of a convolutional neural network and a DPU
The invention discloses an operation method and device of a convolutional neural network and a DPU. The method comprises the steps that a first cache region obtains parameters of an nth layer of a neural network model and first data; The first operation unit calculates the first data according to th...
Gespeichert in:
Hauptverfasser: | , , , , , , , , , |
---|---|
Format: | Patent |
Sprache: | chi ; eng |
Schlagworte: | |
Online-Zugang: | Volltext bestellen |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | The invention discloses an operation method and device of a convolutional neural network and a DPU. The method comprises the steps that a first cache region obtains parameters of an nth layer of a neural network model and first data; The first operation unit calculates the first data according to the parameters of the nth layer of the neural network, and takes a calculation result as second data;The second cache region obtains the parameters of the (n + 1) th layer of the neural network model while the first operation unit calculates; After the calculation of the first operation unit is completed, the second cache region receives and stores second data; And the second operation unit calculates the second data according to the (n + 1) th layer of parameters of the neural network model stored in the second cache region, updates the first data according to a calculation result, stores the updated first data in the first cache region, and repeats the above steps until the calculation of the maximum layer number o |
---|