Cambricon Technologies Corporation Limited

Shanghai Cambricon Information Technology Co., Ltd

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

Chinese Academy of Sciences

Sanghai Cambricon Information Technology Co., Ltd.

Tianshi Chen

Shaoli Liu

Yunji Chen

Qi Guo

Shuai Hu

Zai Wang

Tian Zhi

Zidong Du

Xiao Lin Zhang

Yao Zhang

Bingrui Wang

Shengyuan Zhou

Jinhua Tao

Dong Han

Zhen Li

Yifan Hao

Tao Luo

Yimin Zhuang

Xiaobing Chen

Shijin Zhang

Xuda Zhou

Yuzhe Luo

Lei Zhang

Xinkai Song

Cheng Qian

Daofu Liu

Wei Li

Xunyu Chen

Huiying Lan

Jie Wei

Xiaobin Chen

Yufeng Gao

Yong Yu

Zhou Fang

Xishan Zhang

Yixuan Ren

Shangying Li

Haoyuan He

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

Disclosed are an information processing method and a terminal device. The method comprises: acquiring first information, wherein the first information is information to be processed by a terminal device, calling an operation instruction in a calculation apparatus to calculate the first information so as to obtain second information, and outputting the second information. By means of the examples in the present disclosure, a calculation apparatus of a terminal device can be used to call an operation instruction to process first information, so as to output second information of a target desired by a user, thereby improving the information processing efficiency. The present technical solution has advantages of a fast computation speed and high efficiency.

Information processing method and system

Disclosed are an information processing method and a terminal device. The method comprises: acquiring first information, wherein the first information is information to be processed by a terminal device; calling an operation instruction in a calculation apparatus to calculate the first information so as to obtain second information; and outputting the second information. By means of the examples in the present disclosure, a calculation apparatus of a terminal device can be used to call an operation instruction to process first information, so as to output second information of a target desired by a user, thereby improving the information processing efficiency. The present technical solution has advantages of a fast computation speed and high efficiency.

Information processing method and terminal device

An information processing method applied to a computation circuit is disclosed. The computation circuit includes a communication circuit and an operation circuit. The method includes controlling, by the computation circuit, the communication circuit to obtain a voice to be identified input by a user; controlling, by the computation circuit, the operation circuit to obtain and call an operation instruction to perform voice identification processing on the voice to be identified to obtain target text information corresponding to the voice to be identified. The operation instruction is a preset instruction for voice identification.

Voice recognition based on neural 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

The present disclosure relates to a video retrieval method, a method, system and device for generating a video retrieval mapping relationship, and a storage medium. The video retrieval method comprises: acquiring a retrieval instruction, wherein the retrieval instruction carries retrieval information for retrieving a target frame picture; and obtaining the target frame picture according to the retrieval information and a preset mapping relationship. The method for generating a video retrieval mapping relationship comprises: performing a feature extraction operation on each frame picture in a video stream by using a feature extraction model so as to obtain a key feature sequence corresponding to each frame picture; inputting the key feature sequence corresponding to each frame picture into a text sequence extraction model for processing so as to obtain a text description sequence corresponding to each frame picture; and constructing a mapping relationship according to the text description sequence corresponding to each frame picture. By means of the video retrieval method and the method for generating a video retrieval mapping relationship provided in the present application, the efficiency of video retrieval can be improved, and human-machine interaction is made more intelligent.

Video retrieval method, and method and apparatus for generating video retrieval mapping relationship

Image retouching method and terminal device

The present disclosure provides a computing method that is applied to a computing device. The computing device includes: a memory, a register unit, and a matrix computing unit. The method includes the following steps: controlling, by the computing device, the matrix computing unit to obtain a first operation instruction, where the first operation instruction includes a matrix reading instruction for a matrix required for executing the instruction; controlling, by the computing device, an operating unit to send a reading command to the memory according to the matrix reading instruction; and controlling, by the computing device, the operating unit to read a matrix corresponding to the matrix reading instruction in a batch reading manner, and executing the first operation instruction on the matrix. The technical solutions in the present disclosure have the advantages of fast computing speed and high efficiency.

Calculation method and related product

The present application provides an operation device and related products. The operation device is configured to execute operations of a network model, wherein the network model includes a neural network model and/or non-neural network model; the operation device comprises an operation unit, a controller unit and a storage unit, wherein the storage unit includes a data input unit, a storage medium and a scalar data storage unit. The technical solution provided by this application has advantages of a fast calculation speed and energy-saving.

Sparse processing in neural network processors

The present disclosure provides a processing device including: a coarse-grained pruning unit configured to perform coarse-grained pruning on a weight of a neural network to obtain a pruned weight, an operation unit configured to train the neural network according to the pruned weight. The coarse-grained pruning unit is specifically configured to select M weights from the weights of the neural network through a sliding window, and when the M weights meet a preset condition, all or part of the M weights may be set to 0. The processing device can reduce the memory access while reducing the amount of computation, thereby obtaining an acceleration ratio and reducing energy consumption.

Processing method and accelerating device

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

Growing community of inventors

Beijing, China

Tianshi Chen