Cambricon Technologies Corporation Limited

Chinese Academy of Sciences

Shanghai Cambricon Information Technology Co., Ltd

Cambricon (xi'an) Semiconductor Co., Ltd.

Yunji Chen

Tianshi Chen

Shaoli Liu

Qi Guo

Tian Zhi

Xiao Lin Zhang

Tao Luo

Shijin Zhang

Jinhua Tao

Dong Han

Zhen Li

Zidong Du

Xinkai Song

Cheng Qian

Yimin Zhuang

Huiying Lan

Yong Yu

Xunyu Chen

Daofu Liu

Yuzhe Luo

Yifan Hao

Xiaobing Chen

Xiao Zhang

Shangying Li

Yixuan Ren

Aspects for modifying data in a multi-layer neural network (MNN) acceleration processor for neural networks are described herein. As an example, the aspects may include receiving a predetermined weight value array and connection data. Further, the aspects may include modifying the weight values included in the predetermined weight value array based on the connection data. Further still, the aspects may include calculating one or more groups of output data based on the modified weight values.

Calculating device and method for a sparsely connected artificial neural network

Aspects for forward propagation of a convolutional artificial neural network are described herein. The aspects may include a direct memory access unit configured to receive input data from a storage device and a master computation module configured to select one or more portions of the input data based on a predetermined convolution window. Further, the aspects may include one or more slave computation modules respectively configured to convolute a convolution kernel with one of the one or more portions of the input data to generate a slave output value. Further still, the aspects may include an interconnection unit configured to combine the one or more slave output values into one or more intermediate result vectors, wherein the master computation module is further configured to merge the one or more intermediate result vectors into a merged intermediate vector.

Apparatus and methods for forward propagation in convolutional neural networks

Aspects for forward propagation of a multilayer neural network (MNN) in a neural network processor are described herein. As an example, the aspects may include a computation module that includes a master computation module and one or more slave computation modules. The master computation module may be configured to receive one or more groups of MNN data. The one or more groups of MNN data may include input data and one or more weight values and wherein at least a portion of the input data and the weight values are stored as discrete values. The one or more slave computation modules may be configured to calculate one or more groups of slave output values based on a data type of each of the one or more groups of MNN data.

Apparatus and methods for forward propagation in neural networks supporting discrete data

Aspects for neural network operations with fixed-point number of short bit length are described herein. The aspects may include a fixed-point number converter configured to convert one or more first floating-point numbers to one or more first fixed-point numbers in accordance with at least one format. Further, the aspects may include a neural network processor configured to process the first fixed-point numbers to generate one or more process results.

Apparatus and methods for neural network operations supporting fixed point numbers of short bit length

Aspects for backpropagation of a convolutional neural network are described herein. The aspects may include a direct memory access unit configured to receive input data from a storage device and a master computation module configured to select one or more portions of the input data based on a predetermined convolution window. Further, the aspects may include one or more slave computation modules respectively configured to convolute one of the one or more portions of the input data with one of one or more previously calculated first data gradients to generate a kernel gradient, wherein the master computation module is further configured to update a prestored convolution kernel based on the kernel gradient.

Apparatus and methods for training in convolutional neural networks

Aspects for backpropagation in a fully connect layer of a convolutional neural network are described herein. The aspects may include a direct memory access unit configured to receive input data and one or more first data gradients from a storage device. The aspects may further include a master computation module configured to transmit the input data and the one or more first data gradients to one or more slave computation modules. The slave computation modules are respectively configured to multiply one of the one or more first data gradients with the input data to generate a default weight gradient vector.

Apparatus and methods for training in fully connected layers of convolutional networks

Aspects for backpropagation of a multilayer neural network (MNN) in a neural network processor are described herein. The aspects may include a computation module configured to receive one or more groups of MNN data. The computation module may further include a master computation module configured to calculate an input gradient vector based on a first output gradient vector from an adjacent layer and based on a data type of each of the one or more groups of MNN data. Further still, the computation module may include one or more slave computation modules configured to parallelly calculate portions of a second output vector based on the input gradient vector calculated by the master computation module and based on the data type of each of the one or more groups of MNN data.

Apparatus and methods for backward propagation in neural networks supporting discrete data

Aspects for neural network operations with floating-point number of short bit length are described herein. The aspects may include a neural network processor configured to process one or more floating-point numbers to generate one or more process results. Further, the aspects may include a floating-point number converter configured to convert the one or more process results in accordance with at least one format of shortened floating-point numbers. The floating-point number converter may include a pruning processor configured to adjust a length of a mantissa field of the process results and an exponent modifier configured to adjust a length of an exponent field of the process results in accordance with the at least one format.

Apparatus and methods for neural network operations supporting floating point numbers of short bit length

Aspects of data modification for neural networks are described herein. The aspects may include a data modifier configured to receive input data and weight values of a neural network. The data modifier may include an input data configured to modify the received input data and a weight modifier configured to modify the received weight values. The aspects may further include a computing unit configured to calculate one or more groups of output data based on the modified input data and the modifier weight values.

Device and method for artificial neural network operation

A computing device and related products are provided. The computing device is configured to perform machine learning calculations. The computing device includes an operation unit, a controller unit, and a storage unit. The storage unit includes a data input/output (I/O) unit, a register, and a cache. Technical solution provided by the present disclosure has advantages of fast calculation speed and energy saving.

Vector and matrix computing device

Aspects for Long Short-Term Memory (LSTM) blocks in a recurrent neural network (RNN) are described herein. As an example, the aspects may include one or more slave computation modules, an interconnection unit, and a master computation module collectively configured to calculate an activated input gate value, an activated forget gate value, a current cell status of the current computation period, an activated output gate value, and a forward pass result.

Apparatus and method for executing recurrent neural network and LSTM computations

Aspects of a neural network operation device are described herein. The aspects may include a matrix element storage module configured to receive a first matrix that includes one or more first values, each of the first values being represented in a sequence that includes one or more bits. The matrix element storage module may be further configured to respectively store the one or more bits in one or more storage spaces in accordance with positions of the bits in the sequence. The aspects may further include a numeric operation module configured to calculate an intermediate result for each storage space based on one or more second values in a second matrix and an accumulation module configured to sum the intermediate results to generate an output value.

Multiplication and addition device for matrices, neural network computing device, and method

The present invention provides a fractal tree structure-based data transmit device and method, a control device, and an intelligent chip. The device comprises: a central node that is as a communication data center of a network-on-chip and used for broadcasting or multicasting communication data to a plurality of leaf nodes; the plurality of leaf nodes that are as communication data nodes of the network-on-chip and for transmitting the communication data to a central leaf node; and forwarder modules for connecting the central node with the plurality of leaf nodes and forwarding the communication data; the central node, the forwarder modules and the plurality of leaf nodes are connected in the fractal tree network structure, and the central node is directly connected to M the forwarder modules and/or leaf nodes, any the forwarder module is directly connected to M the next level forwarder modules and/or leaf nodes.

Fractal tree structure-based data transmit device and method, control device, and intelligent chip

The present disclosure provides a neural network processing system that comprises a multi-core processing module composed of a plurality of core processing modules and for executing vector multiplication and addition operations in a neural network operation, an on-chip storage medium, an on-chip address index module, and an ALU module for executing a non-linear operation not completable by the multi-core processing module according to input data acquired from the multi-core processing module or the on-chip storage medium, wherein the plurality of core processing modules share an on-chip storage medium and an ALU module, or the plurality of core processing modules have an independent on-chip storage medium and an ALU module. The present disclosure improves an operating speed of the neural network processing system, such that performance of the neural network processing system is higher and more efficient.

Method and system for processing neural network

Aspects of data modification for neural networks are described herein. The aspects may include a connection value generator configured to receive one or more groups of input data and one or more weight values and generate one or more connection values based on the one or more weight values. The aspects may further include a pruning module configured to modify the one or more groups of input data and the one or more weight values based on the connection values. Further still, the aspects may include a computing unit configured to update the one or more weight values and/or calculate one or more input gradients.

Growing community of inventors

Beijing, China

Yunji Chen