Find out if the version for $F(\lambda)$ is really much faster than a naiive implementation of the matrix multiplications.
If performance is similar, don't use the complicated formulation and make it more legible. See if we can then re-use elements of $F$ in computing_mode.
if __name__ == "__main__":
cProfile.run('foo()', sort=1)
Find out if the version for$F(\lambda)$ is really much faster than a naiive implementation of the matrix multiplications.$F$ in
If performance is similar, don't use the complicated formulation and make it more legible. See if we can then re-use elements of
computing_mode.Profiling in Python