The method of fusion of a unit neural network of a neural network fusion device according to the present invention comprises the steps of collecting meta information of a first neural network and a second neural network, and matching output data of the first neural network with input data of the second neural network. connecting a matching module between the first neural network and the second neural network using the collected meta information, and the first and second neural networks connected by the matching module from the input data of the first neural network. It may include executing a neural network interworking service to obtain output data of the second neural network.
展开▼