I am trying to reproduce the Image Captioning model shared by tensorflow in this blog: Image captioning with visual attention | Text | TensorFlow
However, when I am trying to save/load the model I am getting the following error:
Kindly help on how to resolve it
Welcome to the Tensorflow Forum,
I see that you are adding model.build while loading the weights which is not a right approach. The model is built while saving the weights and therefore when it is built on a new model instance with random weights, we face the error.
Please find the working gist for reference.
Thank you!
