I was looking for a solution for Profiling in tensorflow. See Tensorflow profiling in TF2.0.
Since I could not find a solution for tensorflow2.0, I am using https://www.tensorflow.org/versions/r2.0/api_docs/python/tf/summary/trace_on
As indicated in the documentation, it incurrs high memory overload, but between two given points in code, I am able to trace the runtime and memory.
What are the other differences between these methods(i.e, Profiler and Trace)?