The paper mentions that dropblock is present in the last two layers however, does not give the probability used. In the code also the probability is always 0 as far as I can see.
Was it used in the best results (especially for segmentation)? If so can you tell me what was the probability used?
gfotedar
commented
3 years ago
The paper mentions that dropblock is present in the last two layers however, does not give the probability used. In the code also the probability is always 0 as far as I can see.
Was it used in the best results (especially for segmentation)? If so can you tell me what was the probability used?