performance

[aoyanl]

Why is the result on the val set more than three points different from that published by the author? If I want to use it in my WSSS job，will it have a big impact？ Thanks！！

[isht7]

What do you mean by WSSS job? I am not sure why there is the difference in performance, it is most possibly due several reasons including small changes in caffe compared to pytorch, and some implementation details not mentioned in the paper.

If performance is an issue for you, you can consider using deeplab v3+. This repository has an older version(deeplab v2).