Decomposition of machine learning operations

A system comprising:

• a processor;
• a memory device containing instructions, which when executed by the processor cause the processor to:
• receive a representation of a neural network (NN) model to be executed on an electronic device, the representation of the NN model including nodes corresponding to layers of the NN model, wherein at least one of the nodes corresponds to an operation of a corresponding layer of the NN model to be performed by the electronic device;
• determine a set of operations that is mathematically equivalent to the operation such that an aggregation of outputs of the set of operations is equivalent to an output of the operation, wherein the determining the set of operations comprises:
• determining a first plurality of operation for the operation;
• determining a second plurality of operations for the operation;
• and
• selecting one of the first or second plurality of operations based at least in part on an analysis of the first and second plurality of operations, wherein the analysis is based at least in part on statistics indicating computational overhead and memory accesses and the analysis indicated which of the first or second plurality of operations utilize fewer memory resources;
• generate a graph based on the set of operations, wherein the graph includes a set of branches, each branch includes a plurality of operations including at least one operation from the set of operations;
• determine a respective order for executing each branch of the graph;
• and
• store the graph and the respective order for executing each branch of the graph for compilation of the NN model.

A non-transitory computer-readable medium comprising instructions, which when executed by a computing device, cause the computing device to perform operations comprising:

• receiving a representation of a neural network (NN) model to be executed on an electronic device, the representation of the NN model including nodes corresponding to layers of the NN model, wherein at least some of the nodes each corresponds to a respective operation of a respective layer of the NN model to be performed by the electronic device;
• determining, for the respective operation corresponding to at least one node in layer of the NN model, a respective set of operations that is mathematically equivalent to the respective operation such that an aggregation of outputs of the respective set of operations is equivalent to an output of the respective operation, wherein the determining, for the respective operation corresponding to at least one node in at least one layer of the NN model, the respective set of operations comprises:
• determining a first plurality of operations for the respective operation;
• determining a second plurality of operations for the respective operation;
• and
• selecting one of the first or second plurality of operations based at least in part on an analysis of the first and second plurality of operations, wherein the analysis is based at least in part on statistics indicating computational overhead and memory accessed and the analysis indicates which of the first or second plurality of operations utilize fewer memory resources;
• generating a graph based on each respective set of operations, wherein the graph includes a set of branches, each branch includes a plurality of operations, the plurality of operations including a particular operation from each respective set of operations;
• and
• determining a respective order for executing each branch of the graph.