Research on Cloud Side Collaboration Architecture and Lightweight Model of Distribution Network

With the continuous development of the deep learning model, the prediction accuracy of the model has reached a high enough level. The high memory consumption in the prediction of the deep learning model has become a problem that needs to be solved. Moreover, the large size of the model not only brings about the memory capacity problem, but also requires sufficient memory bandwidth. At the same time, when the model is running, the battery will also be rapidly consumed, which is also a test of the endurance of the mobile terminal. In order to solve the above problems, this paper proposes a cloud-side collaboration method based on federated learning, carries out research on deep neural network model compression algorithm, reduces the size of the model under the condition of ensuring the accuracy of the model, so that the prediction effect will not have too much loss, while reducing the requirements for the performance of the edge and end devices, realizes the cloud-side collaboration of the distribution network, and accelerates the digital transformation of the distribution network.