Where is the code for MC-Dropout?

[liming-ai]

Hi @layumi , I noticed that in your paper, the uncertainty is obtained by computing both MC-Dropout and difference between two classifiers, but I cannot find the code used for MC-Dropout, could you please kindly tell me?

[layumi]

Hi @mitming MC-Dropout denotes that during inference stage, we still apply the dropout. In this way, we could estimate the uncertainty by the difference of the two predictions.

In our case, it is quite easy to implement. Since we estimate the uncertainty during training, dropout is already enabled.