Model training method and device and server

The invention provides a model training method and device and a server. In some embodiments, in the process of cooperatively and jointly training a target model, when data processing of a current batch of a current round is carried out, whether the data processing of the current batch of the current round meets a preset triggering condition or not can be detected firstly; when it is determined that a preset triggering condition is met, according to a preset protocol, a third data party does not need to generate a new random matrix, but utilizes an existing matrix in a locally stored random matrix set to generate a first random vector, a second random vector and a third random vector, and the first random vector, the second random vector and the third random vector are distributed; a second data party can utilize a locally stored intermediate matrix set to calculate a second fragment with a secret processing result; and a first data party does not need to additionally generate and send an intermediate matrix t