M1/M2 fix is hacky

[mgbckr]

To make things work on Apple M1/M2 machines, we are currently using a combination of installing tensorflow=2.10.0 by hand (see README.md) and some hack in __init.py__:

# TODO: hacky! 
# NOTE: Be aware of potential side-effects: but that may cause crashes or silently produce incorrect results.
# sources:
# - https://stackoverflow.com/questions/53014306/error-15-initializing-libiomp5-dylib-but-found-libiomp5-dylib-already-initial
# - https://stackoverflow.com/questions/55714135/how-can-i-fix-an-omp-error-15-initializing-libiomp5-dylib-but-found-libomp
import os
os.environ['KMP_DUPLICATE_LIB_OK']='True'

Used to fix the following issue:

OMP: Error #15: Initializing libiomp5.dylib, but found libomp.dylib already initialized.
OMP: Hint This means that multiple copies of the OpenMP runtime have been linked into the program. That is dangerous, since it can degrade performance or cause incorrect results. The best thing to do is to ensure that only a single OpenMP runtime is linked into the process, e.g. by avoiding static linking of the OpenMP runtime in any library. As an unsafe, unsupported, undocumented workaround you can set the environment variable KMP_DUPLICATE_LIB_OK=TRUE to allow the program to continue to execute, but that may cause crashes or silently produce incorrect results. For more information, please see http://www.intel.com/software/products/support/.

We should eventually fix this!

[mgbckr]

Note on closing this issue: I removed the "hack" from the __init__.py in the most recent merge. Its tested on M4 and I am assuming it will also work on M1/M2.