Processing for multiple input data sets

A method of processing a plurality of data sets in a multi-layer neural network, the multi-layer neural network comprising a first neural network layer, a second neural network layer that processes outputs generated by the first neural network layer, and a third neural network layer that processes outputs generated by the second neural network layer, the method comprising:

• storing first data sets in a memory;
• processing, using the first neural network layer, first data set from a memory to generate the first layer output of the first data sets;
• storing the first layer output of the first data sets in the memory;
• removing the first data set from the memory;
• adding second data sets to the first layer output of the first data sets stored in the memory;
• processing, using the first neural network layer, the second data sets from the memory to generate a first layer output of the second data set;
• adding the first layer output of the second data set to the first layer output of the first data sets stored in the memory;
• removing the second data sets from the memory;
• processing, using the second neural network layer, the first layer output of the first data sets and the first layer output of the second data set from the memory to generate, respectively, a second layer output of the first data sets and a second layer output of the second data sets;
• storing the second layer output of the first data sets and the second layer output of the second data sets in the memory;
• removing the first layer output of the first data sets and the first layer output of the second data sets from the memory;
• adding third data sets to the second layer output of the first data sets and the second layer output of the second data sets stored in the memory;
• processing, using the first neural network layer, the third data sets from the memory to generate a first layer output of the third data sets after the second layer output of the first data sets and the second layer output of the second data sets are generated;
• storing the first layer output of the third data sets in the memory;
• removing the third data sets from the memory;
• adding fourth data sets to the second layer output of the first data sets, the second layer output of the second data sets, and the first layer output of the third data sets stored in the memory;
• processing, using the first neural network layer, the fourth data sets from the memory to generate a first layer output of the fourth data sets;
• adding the first layer output of the fourth data sets to the first layer output of the third data sets, the second layer output of the first data sets, and the second layer output of the second data sets stored in the memory;
• removing the fourth data sets from the memory;
• processing, using the second neural network layer, the first layer output of the third data sets and the first layer output of the fourth data sets to generate, respectively, a second layer output of the third data sets and a second layer output of the fourth data sets;
• adding the second layer output of the third data sets and the second layer output of the fourth data sets to the second layer output of the first data sets and the second layer output of the second data sets stored in the memory;
• and
• processing, using the third neural network layer, the second layer output of the first data sets, the second layer output of the second data sets, the second layer output of the third data sets, and the second layer output of the fourth data sets from the memory to generate, respectively, a third layer output of the first data sets, a third layer output of the second data sets, a third layer output of the third data sets, and a third layer output of the fourth data sets;
• and
• storing the third layer output of the first data sets, the third layer output of the second data sets, the third layer output of the third data sets, and the third layer output of the fourth data sets in the memory.