Embodiments of this application provide a communication method and a communication apparatus. The method includes: A central node obtains N indication parameters that are in a one-to-one correspondence with N distributed nodes, where a first indication parameter in the N indication parameters indicates a communication capability and a computing capability of a first distributed node, and the first distributed node is any node in the N distributed nodes; the central node determines M network parameters based on the N indication parameters, where the M network parameters are in a one-to-one correspondence with M distributed nodes in the N distributed nodes, and a first network parameter in the M network parameters indicates at least one of a communication configuration or a neural network architecture configuration of the first distributed node; and the central node sends the M network parameters. According to the method, in embodiments of this application, neural network architecture search may be performed in a variable communication environment, and a neural network architecture determined through the neural network architecture search can meet requirements of a node for a neural network with a low latency and high model accuracy.