1 min readNov 23, 2017
Hello Morgan, thank you for a great tutorial.
I have one question:
I trained my CNN on specified batch size = 75. So during training, the input tensor had size of {batch size, 244,244,3}
Everything worked great (testing). However, when I created the frozen model, batch size “bake in” to model and I am not able to test if for only single input tensor {1, 244, 244, 3} because it accept input of only {batch size, 244,244,3} size.
Do you know how can I create tensor of variable batch size so it won’t be assigned to frozen model?
Thank you for your time