The issue I'm running into is that VGG16 is defined as a Model (not a Sequential) so there is no `.add` method. #deeplearning
-
to have a defined input shape, pass an `input_tensor` argument (with defined shape) to VGG16. See docs.
else you could use a GlobalAveragePooling layer, too, which is dynamic.
-
-
basically when you use `include_top=False` then the model can accept any input size, and shape inference propagates "None".
-
It was the `input_tensor` that I was missing. The `None` in the error message should have clued me in. Thanks again!
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.