Abstract: This application relates to the field of artificial intelligence, and provides a neural network training method and apparatus. The method includes: obtaining a to-be-trained neural network; grouping parameters of the to-be-trained neural network, to obtain M groups of parameters, where M is a positive integer greater than or equal to 1; obtaining sampling probability distribution and training iteration step arrangement, where the sampling probability distribution represents a probability that each of the M groups of parameters is sampled in each training iteration step, and the training iteration step arrangement includes interval arrangement and periodic arrangement; freezing or stopping updating a sampled parameter group based on the sampling probability distribution and the training iteration step arrangement; and training the to-be-trained neural network based on the parameter group that is frozen or stopped updating.

Abstract: An application programming interface (API) gateway cluster control method includes determining, by an API gateway cluster, that a quantity of API requests received per unit of time exceeds a traffic control parameter, obtaining a load parameter of the API gateway cluster and a load parameter of a service cluster, determining, based on the load parameter of the API gateway cluster, that the API gateway cluster is congested, determining, based on the load parameter of the service cluster, that the service cluster is not congested, and scaling out the API gateway cluster.

Abstract: An application programming interface (API) gateway cluster control method includes determining, by an API gateway cluster, that a quantity of API requests received per unit of time exceeds a traffic control parameter, obtaining a load parameter of the API gateway cluster and a load parameter of a service cluster, determining, based on the load parameter of the API gateway cluster, that the API gateway cluster is congested, determining, based on the load parameter of the service cluster, that the service cluster is not congested, and scaling out the API gateway cluster.

Abstract: An application programming interface (API) gateway cluster control method includes determining, by an API gateway cluster, that a quantity of API requests received per unit of time exceeds a traffic control parameter, obtaining a load parameter of the API gateway cluster and a load parameter of a service cluster, determining, based on the load parameter of the API gateway cluster, that the API gateway cluster is congested, determining, based on the load parameter of the service cluster, that the service cluster is not congested, and scaling out the API gateway cluster.

Abstract: The present invention provides an impeller assembly and cutting equipment provided with the impeller assembly. Compared with cutting equipment for an external dust cleaner in the prior art, the present invention has a dust suction function and a dust collection function itself, is more convenient to operate in no need of cooperating with an external vacuum cleaner, and thus is applied more widely and not limited by working conditions. By reasonably setting a curvature of each first vane, a first impeller which is in rotation can suck a radial movement of surrounding air along a motor. The first impeller is disposed in the cutting equipment for forming a negative pressure in the cutting equipment, thereby realizing a dust suction function of the cutting equipment. A heat dissipation area of a rotor can be increased by providing a plurality of second impellers on a second impeller, thereby achieving a heat dissipation function of the rotor.

Abstract: An application programming interface (API) gateway cluster control method includes determining, by an API gateway cluster, that a quantity of API requests received per unit of time exceeds a traffic control parameter, obtaining a load parameter of the API gateway cluster and a load parameter of a service cluster, determining, based on the load parameter of the API gateway cluster, that the API gateway cluster is congested, determining, based on the load parameter of the service cluster, that the service cluster is not congested, and scaling out the API gateway cluster.