Datasheet

Chapter3 Functional Description 15
both cores
• Support for software interrupts, and each core can trigger cross-core inter-
rupts
• Built-in CPU timer interrupt, both cores are freely configurable
• Advanced external interrupt management, supporting 64 external interrupt
sources, each interrupt source can be configured with 7 priority levels
3.1.4 Debugging Support
• Support performance monitoring instructions to count instruction execution
cycles
• High-speed UART and JTAG interface for debugging
• Support DEBUG mode and hardware breakpoints
3.2 Neural Network Processor (KPU)
KPU is a general-purpose neural network processor with built-in convolution,
batch normalization, activation, and pooling operations. It can detect faces or
objects in real time. The specific characteristics are as follows:
• Supports the fixed-point model that the mainstream training framework trains
according to specific restriction rules
• There is no direct limit on the number of network layers, and each layer of
convolutional neural network parameters can be configured separately, includ-
ing the number of input and output channels, and the input and output line
width and column height
• Support for 1x1 and 3x3 convolution kernels
• Support for any form of activation function
• The maximum supported neural network parameter size for real-time work is 5MiB
to 5.9MiB
• The maximum supported network parameter size when working in non-real time is
(flash size - software size)
Mode
Maximum fixed point
model size (MiB)
Maximum pre-quantisation
floating point model size
（MiB）
Realtime（≥ 30fps） 5.9 11.8