You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
As the fig.2 show that you get the CxC matrix by the fc layer, why did you say the CNNs so many times? And the CNNs in Fig2 is just to compress as you said in session 5.1.
The text was updated successfully, but these errors were encountered:
I means the CNNs you mentioned in the paper's section 3.2 is for transform T, but actually is done by Fc layer, right? The CNNs is just for reducing dimensions.
As the fig.2 show that you get the CxC matrix by the fc layer, why did you say the CNNs so many times? And the CNNs in Fig2 is just to compress as you said in session 5.1.
The text was updated successfully, but these errors were encountered: