Retrieving code of a saved model

Course Q&A Deep Learning Specialization Convolutional Neural Networks
1

This question is being formulated at Week 3 of Course 4, but could equally apply to Week 2 assignments.
I saved a trained model with tf.keras.models.save_model(MyM), it produced a folder with the contents expected. When I retrieve it, using tf.keras.models.load_model(MyM), everything works, except I can’t see the source code. (I am using Jupyter.notebooks). What to do if I want to continue working with the code of a “save_model” model?

2

If you want to save your notebook, use the “File → Download As” tool from the menu bar.

3

And you shouldn’t expect to have the notebook code saved as part of the model, since per the doc

The SavedModel and HDF5 file contains:

- the model’s configuration (topology)
- the model’s weights
- the model’s optimizer’s state (if any)

Thus models can be reinstantiated in the exact same state, without any of the code used for model definition or training.

4

Thanks ai_curious. Yes, I learnt the hard way that saving the model with save_model does not retain the code.

5

There is nothing that says you can’t write your own Python Class that incorporates Keras Model functionality, either by composition (has-a) or inheritance (is-a), and implementing a custom save model to capture meta information about how the model was configured and trained.

6

Thanks Curious. But your answer is way above my “pay level”…:frowning: