Is code in DenseNet dir used multi-scale attention?

[Zhang-O]

Is DenseNet used multi-scale attention? or just substitute encoder from vgg to densenet?
I found no multi-scale attention code in densenet dir mentioned in your paper "Multi-Scale Attention with Dense Encoder for Handwritten Mathematical Expression Recognition".
Should I add it myself according to your paper?

thanks very much.

best wishes!