Abstract: A method of compressing a set of weight values is provided in which an uncompressed set of weight values is obtained, which uncompressed set of weight values includes a plurality of weight values associated with a neural network. A frequently occurring value is identified among the plurality of weight values within the set of weight values and each occurrence of the frequently occurring weight value is replaced within the set of weight values with an index value. The frequently occurring weight value and the index value are associated with the set of weight values. The index value is selected to be less storage intensive than the frequently occurring weight value that it replaces.

Abstract: A method for compressing activation data of a neural network to be written to a storage is provided. The activation data is formed into a plurality of groups and a state indicator indicates whether there are any data elements within each group that have a non-zero value. A second state indicator indicates, for groups having a non-zero value, whether sub-groups within the group contain a data element having a non-zero value. A sub-group state indicator indicates, for each sub-group having a non-zero value, which data elements within that sub-group have a non-zero value. Non-zero values of data elements in the activation data are encoded and a compressed data set is formed comprising the first state indicators, any second state indicators, any sub-group state indicators and the encoded non-zero values.