SavedModel format in TensorFlow - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to save a TensorFlow model in SavedModel format.

TensorFlow

model.save('[1]')

Drag options to blanks, or click blank then click option'

Asaved_model/my_model

Bmodel.h5

Ccheckpoint.ckpt

Dmodel.json

Attempts:

3 left

💡 Hint

Common Mistakes

Using a filename with .h5 extension which saves in HDF5 format, not SavedModel.

Using checkpoint or JSON filenames which are not for SavedModel format.

To save a TensorFlow model in SavedModel format, you provide a directory path. The path 'saved_model/my_model' is a folder where the model will be saved.

2fill in blank

medium

Complete the code to load a TensorFlow model saved in SavedModel format.

TensorFlow

loaded_model = tf.keras.models.[1]('saved_model/my_model')

Drag options to blanks, or click blank then click option'

Asave_weights

Bload_weights

Csave_model

Dload_model

Attempts:

3 left

💡 Hint

Common Mistakes

Using load_weights which only loads weights, not the full model.

Using save_model or save_weights which are for saving, not loading.

To load a model saved in SavedModel format, use tf.keras.models.load_model() with the folder path.

3fill in blank

hard

Fix the error in the code to save the model in SavedModel format.

TensorFlow

model.save('my_model[1]')

Drag options to blanks, or click blank then click option'

A.h5

B/

C.json

D.ckpt

Attempts:

3 left

💡 Hint

Common Mistakes

Adding file extensions which save in other formats.

Not specifying a folder path causes errors or saves in wrong format.

To save in SavedModel format, the path should be a directory. Adding a slash '/' at the end ensures it's treated as a folder.

4fill in blank

hard

Fill both blanks to save a model and then load it back correctly.

TensorFlow

model.save('[1]')
loaded_model = tf.keras.models.[2]('[1]')

Drag options to blanks, or click blank then click option'

Asaved_model/my_model

Bload_model

Cload_weights

Dmodel.h5

Attempts:

3 left

💡 Hint

Common Mistakes

Using different paths for saving and loading.

Using load_weights instead of load_model for loading.

Use the same folder path to save and load the model. Use load_model to load the full model from SavedModel format.

5fill in blank

hard

Fill all three blanks to save a model, load it, and evaluate it on test data.

TensorFlow

model.save('[1]')
loaded_model = tf.keras.models.[2]('[1]')
loss, accuracy = loaded_model.evaluate([3])

Drag options to blanks, or click blank then click option'

Asaved_model/my_model

Bload_model

Ctest_data

Dload_weights

Attempts:

3 left

💡 Hint

Common Mistakes

Using load_weights instead of load_model before evaluation.

Passing training data instead of test data to evaluate.

Save and load the model using the same folder path and load_model. Then evaluate the loaded model on the test dataset.