For profiling of models, almost everything needed is already there. You only need to pass in a StepStatsCollector through the Session::Run() method (I called i

For profiling of models, almost everything needed is already there. You only need to pass in a StepStatsCollector through the Session::Run() method (I called it RunWithStats() ) and hook it up to the Executor Args by filling in this variable: https://github.com/tensorflow/tensorflow/blob/master/tensorflow/core/common_runtime/executor.h#L79 https://github.com/tensorflow/tensorflow/blob/master/tensorf... You then get a very usable set of profiling statistics out by aggregating the StepStats object. For profiling individual ops on the CPU, perf is very useful.