2nd Dropout Layer missing in VGG Models #1127
Answered
by
michaelschleiss
michaelschleiss
asked this question in
Q&A
-
|
Usually there are two dropout layers in the classification layers of a VGG model. However, I noticed that the second dropout is missing (see L79). This second dropout is mentioned in the original paper and can also be found in the torchvision implementation. Is this a purposeful design decision or should this be modified? |
Beta Was this translation helpful? Give feedback.
Answered by
michaelschleiss
Feb 8, 2022
Replies: 1 comment
-
|
Found the answer: The second dropout is hidden in the ClassifierHead. It is implemented invoking the functional version of the Dropout layer. I guess that is the reason why isn't listed when printing the model summary. |
Beta Was this translation helpful? Give feedback.
0 replies
Answer selected by
michaelschleiss
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Found the answer: The second dropout is hidden in the ClassifierHead. It is implemented invoking the functional version of the Dropout layer. I guess that is the reason why isn't listed when printing the model summary.