Data processing method and apparatus, storage medium, and electronic device

A data processing apparatus, comprising:

• at least one memory configured to store program code;
• and
• at least one processor configured to read the program code and operate as instructed by the program code, the program code comprising:
• first processing code configured to cause at least one of the at least one processor to perform first processing on data by using a first sub-model corresponding to an artificial intelligence (AI) processing model, to obtain an intermediate processing result, the AI processing model being split into M neural network layers and K neural network layers, the first sub-model being generated according to the split M neural network layers in the AI processing model, a second sub-model being generated according to the split K neural network layers in the AI processing model, M and K being positive integers greater than or equal to 1;
• transmitting code configured to cause at least one of the at least one processor to transmit the intermediate processing result to a first server;
• and
• first receiving code configured to cause at least one of the at least one processor to receive a target processing result from the first server, the target processing result being based on a result of second processing on the intermediate processing result by using the second sub-model,
• wherein the program code further comprises:
• second processing code configured to cause at least one of the at least one processor to obtain a first processing result obtained by performing the first processing on a training object by using the first sub-model;
• and
• training code configured to cause at least one of the at least one processor to train the K neural network layers that are split from the AI processing model, to obtain the second sub-model meeting a training constraint condition, an input of the K neural network layers being the first processing result, the training constraint condition being:
• a processing accuracy difference between an output result of the second sub-model and a second processing result is less than or equal to a target threshold, the second processing result being a processing result obtained by processing the training object by using N network layers of the AI processing model, N being greater than or equal to a sum of M and K.

A non-transitory computer readable storage medium, storing a computer program, the computer program being executable by at least one processor to perform:

• performing first processing on data by using a first sub-model corresponding to an artificial intelligence (AI) processing model, to obtain an intermediate processing result, the AI processing model being split into M neural network layers and K neural network layers, the first sub-model being generated according to the split M neural network layers in the AI processing model, a second sub-model being generated according to the split K neural network layers in the AI processing model, M and K being positive integers greater than or equal to 1;
• transmitting the intermediate processing result to a first server;
• and
• receiving a target processing result from the first server, the target processing result being based on a result of second processing on the intermediate processing result by using the second sub-model,
• wherein the computer program is executable by the at least one processor to further perform:
• obtaining a first processing result obtained by performing the first processing on a training object by using the first sub-model;
• and
• training the K neural network layers that are split from the AI processing model, to obtain the second sub-model meeting a training constraint condition, an input of the K neural network layers being the first processing result, the training constraint condition being:
• a processing accuracy difference between an output result of the second sub-model and a second processing result is less than or equal to a target threshold, the second processing result being a processing result obtained by processing the training object by using N network layers of the AI processing model, N being greater than or equal to a sum of M and K.