https://github.com/hpssjellis/easy-tensorflow-on-cloud9/blob/master/aymericdamien-Examples/examples/alexnet.py
When I read the original paper, I think these are the differences.
- only 3 convolutional layers are coded, where Alexnet Have 5 convolutional layer
- Normalization follows only first and second convolution
- Normalization should be before pooling, not after.