STEAM (ArXiv)

STEAM (Squeeze and Transform Enhanced Attention Module) is under submission at a leading conference. However, I have decided to share some code files which can help people replicate results of SOTA modules easily but haven't provided STEAM's entire code for obvious reasons. We will upload all files once our work is accepted.
Specifically, one can use resnet_se.py, resnet_eca.py to see results of SENet and ECANet. resnet_gat.py provides some idea about our module however, the main graph transformer file is not uploaded yet.