No "tanh" activation in "convert_weights.py"

[xushanthu-2014]

Thanks for sharing this package! I run your example in my local machine by changing your network to my own network. And I got following warning:

../convert_weights.py:234: UserWarning: Unsupported activation, tanh, found! Replacing with Linear.
  warnings.warn(warning_str)

So I checked the convert_weights.py and found the supported activation methods are ['relu', 'linear', 'leakyrelu', 'sigmoid']. But I found that in the fortran code FKB/src/lib/mod_activation.F90, there are indeed functions for tanh. So do I just need to change ACTIVATIONS = ['relu', 'linear', 'leakyrelu', 'sigmoid'] in convert_weights.py to run my model? Thanks a lot!