Possible Error in using Argsort

[lagidigu]

Hi,

There seems to be a mismatch between the description in your paper and the corresponding code:

First, instead of solving dynamical equations we use the currents as a proxy for ranking of the final activities. Given that ranking, the unit that responds the most to a given training example is pushed toward that example with activation g=1.

The corresponding code would be the following:

y=np.argsort(tot_input,axis=0)
yl=np.zeros((hid,Num))
yl[y[hid-1,:],np.arange(Num)]=1.0

Instead of taking the top ranked activation for a given sample, it rather takes the ranking of the last hidden unit, or did I get something wrong?