Inference request processing method, device and equipment and readable storage medium

Abstract

The invention belongs to the technical field of computers, and discloses an inference request processing method, device and equipment and a readable storage medium, and the method comprises the steps: receiving a target inference request sent by a client, and then obtaining an inference request of which the processing sequence is before the target inference request; extracting a keyword of the reasoning request, and searching reasoning data matched with the keyword in a database; determining an approximate reasoning length based on the reasoning data; determining the reasoning time of the reasoning request by using the approximate reasoning length; and feeding back waiting time of the target reasoning request to the client based on the reasoning time. The reasoning time in the application fully considers the difference of different reasoning requests, and is combined with information such as historical reasoning request records and average reasoning lengths under different classifications. Therefore, the reasoning time is more close to the actual reasoning time, the waiting time fed back to the client is more accurate, and the user experience can be effectively improved.

Images