make_trainable() does not freeze weights

[embanner]

You define a function make_trainable() which sets every layer's trainable attribute to either True or False and call this repeatedly during training. However, setting keras.layers.Layer.trainable doesn't have any effect unless you follow it up with recompiling the model. So I'm pretty sure that your layers are unfrozen during the entire training process since you only compile once.

I'll take a stab at verifying this shortly.