Distributed training micro-batch data determination method and device, equipment and medium

The invention provides a distributed training micro-batch data determination method and device, equipment and a medium, and relates to the technical field of emerging information, for example, the method comprises the steps that a training data set is divided into a plurality of training data subsets based on the size of pre-determined micro-batch data, so that corresponding gradient values are calculated; receiving gradient values returned by the preset number of training nodes, and screening out a target training node from the training nodes; judging whether a gradient value which is returned by the target training node and meets a preset condition is received within a preset duration or not; inputting all the received gradient values, the total duration consumed for receiving all the gradient values and the model performance parameters of all the training nodes into a pre-trained neural network model, outputting the updated model performance parameters of all the training nodes, and finally determining the