Optimize fitfuns.py

[CloudyDory]

The following changes are made in bmtk/simulator/filternet/lgnmodel/fitfuns.py to significantly speed up network preparation:

• Optimize function ff() to use vectorized computation instead of looping over an NumPy array.
• Replace the use of np.tile in function normalizecols() and makeBasis_StimKernel() by automatic broadcasting in NumPy.
• In function makeBasis_StimKernel(), compute variable kbasis0 in a single step instead of using a for-loop in original code.