Distributed deep learning data recovery method and system and readable storage medium

The invention provides a distributed deep learning data recovery method and system and a readable storage medium, and the method comprises the following steps: S1, collecting parameter information for training a neural network model, and obtaining global information of to-be-recovered data distribution by using an erasure code; s2, calculating the recovery priority of each type of training data by using artificial intelligence, and sorting the strips where the to-be-recovered data are located; and S3, combining the calculated data recovery priority and the global information of the to-be-recovered data distribution, and constructing a parallel recovery scheme. According to the method, the weighted priority is established for the data to be recovered through artificial intelligence, the parallel recovery scheme is constructed in combination with the global information of data block distribution obtained by the erasure codes, and the capability of obtaining the global information of the data by the erasure code