Model looks unoptimised

[mrgloom]

1. BN can be fused
2. batchnorm and scale are separated layers here this is because of bad Caffe implementation.

BTW is there any app where we just can benchmarks some other pertained classification models on iPhone?

[hollance]

Yes, you're correct. But this is also an old model so it's not really worth fixing. I'd recommend getting a more up-to-date checkpoint from the TensorFlow models repo instead, and converting that.

(P.S. I believe Core ML automatically does the BN fusing behind the scenes.)