Reuse Model for text classification

Research & Models
,
#1

Hello,
I’m a complete newbie and just getting my hands dirty.
Regarding this example: Google Colab

I now find out how to save the generated model:

model_for_inference.compile(
    optimizer=keras.optimizers.RMSprop(),  # Optimizer
    # Minimize loss:
    loss=keras.losses.SparseCategoricalCrossentropy(),
    # Monitor metrics:
    metrics=[keras.metrics.SparseCategoricalAccuracy()],
)

# Save the entire model as a SavedModel.
model_for_inference.save('saved_model/model')

And I think I also found out how to load this model with another Python file like:

model = keras.models.load_model('saved_model/model')

But I don’t found out how to re-use this model.

I plan to send a String of text to it, and then I’ll get the top 3 predicted labels.
How to do this?

cheers
Charlie

#3

Hi @Charlie_Schaubmair Are you stuck on a specific point? I mean Tensorflow provides with documentation dedicated to this, here: Save and load Keras models and it is sraightforward in practice.