Ionut Cosmin Duta comments

Results 5 comments of


                                            Ionut Cosmin Duta

Yes, the SE block can bring additional improvement.

You have there the links to the models: https://github.com/iduta/pyconv/blob/master/README.md

Hi, thanks for the interest in our work. Yes, in this case you need to train the modified network from scratch.

You can just change the layers information from 50 to 152 in the yaml file to obtain the ade20k_pyvonvsegnet152_pyvonvsegnet.yaml

yes, and also to potentially update the directory names for saving the results there.