here is my implementation of the basic googLeNet described in the paper Going Deeper with Convolutions in keras
I wrapped the inception module described in the paper and put it in the file "inceptionModel.py"
the structure of the net is described as following: