Neural network computing device

The invention discloses a neural networking computing device, and belongs to the field of electronic chips. The device comprises a first arithmetic processor and a second arithmetic processor. The first arithmetic processor is used for carrying out the first operation or second operation in a training operation of a neural network, wherein the first operation comprises vector dot product operation, and the second operation comprises vector dot product operation and nonlinear transformation operation. The second arithmetic processor is used for correcting at least one weight vector of the neural network according to at least one correction value, and obtaining at least one corrected weight vector, wherein each correction value is corresponding to one vector element in the weight vector of the neural network, and at least one correction value is obtained according to the operation result of the first arithmetic processor. The device provided by the invention can improve the operation efficiency of the neural ne