TensorFlow Saving and Loading Models

Question 1

Which method saves a Keras model in HDF5 format?

Accepted Answer

model.save('model.h5')

Answer

model.export('model.h5')

Answer

model.serialize('model.h5')

Answer

model.write('model.h5')

Question 2

Which function is used to reload a previously saved Keras model?

Accepted Answer

tf.keras.models.load_model()

Answer

tf.keras.models.restore()

Answer

tf.keras.models.import_model()

Answer

tf.keras.models.deserialize()

Question 3

What does the SavedModel format store that HDF5 does not?

Accepted Answer

The full TensorFlow computation graph and serving signatures

Answer

Only the optimizer state

Answer

Only the model weights

Answer

Only the model architecture as JSON

Question 4

Which TensorFlow class is the modern API for saving and restoring model variables as checkpoints?

Accepted Answer

tf.train.Checkpoint

Answer

tf.train.Saver

Answer

tf.train.ModelSaver

Answer

tf.train.VariableStore

Question 5

Which Keras callback automatically saves the model during training at configurable intervals?

Accepted Answer

tf.keras.callbacks.ModelCheckpoint

Answer

tf.keras.callbacks.SaveModel

Answer

tf.keras.callbacks.CheckpointSaver

Answer

tf.keras.callbacks.TrainingSnapshot

Question 6

Which method saves only a model's weight values without the architecture or optimizer?

Accepted Answer

model.save_weights()

Answer

model.export_weights()

Answer

model.write_weights()

Answer

model.checkpoint_weights()

Question 7

What file extensions does TensorFlow create for checkpoint index and data files respectively?

Accepted Answer

.ckpt.index and .ckpt.data-00000-of-00001

Answer

.checkpoint and .checkpointdata

Answer

.tfstate.index and .tfstate.data

Answer

.weights.index and .weights.data