Documentation Request: How to predict with "Separate preprocessing from the same preset"

[hatdropper1977]

Is your feature request related to a problem? Please describe. Documentation Request Please update: https://github.com/keras-team/keras-io/blob/master/guides/keras_nlp/getting_started.py

Describe the solution you'd like I would like to call my custom preprocessed model.

In the example given, I execute the following code:

preprocessor = keras_nlp.models.BertPreprocessor.from_preset(
    "bert_tiny_en_uncased",
    sequence_length=512,
)
# Apply the preprocessor to every sample of train and test data using `map()`.
# `tf.data.AUTOTUNE` and `prefetch()` are options to tune performance, see
# https://www.tensorflow.org/guide/data_performance for details.
imdb_train_cached = (
    imdb_train.map(preprocessor, tf.data.AUTOTUNE).cache().prefetch(tf.data.AUTOTUNE)
)
imdb_test_cached = (
    imdb_test.map(preprocessor, tf.data.AUTOTUNE).cache().prefetch(tf.data.AUTOTUNE)
)

classifier = keras_nlp.models.BertClassifier.from_preset(
    "bert_tiny_en_uncased", preprocessor=None, num_classes=2
)
classifier.fit(
    imdb_train_cached,
    validation_data=imdb_test_cached,
    epochs=3,
)

I get a trained model:

Downloading data from https://storage.googleapis.com/keras-nlp/models/bert_tiny_en_uncased/v1/vocab.txt
231508/231508 [==============================] - 0s 2us/step
Downloading data from https://storage.googleapis.com/keras-nlp/models/bert_tiny_en_uncased/v1/model.h5
17602216/17602216 [==============================] - 2s 0us/step
Epoch 1/3
1563/1563 [==============================] - 330s 198ms/step - loss: 0.4165 - sparse_categorical_accuracy: 0.8064 - val_loss: 0.3525 - val_sparse_categorical_accuracy: 0.8452
Epoch 2/3
1563/1563 [==============================] - 274s 175ms/step - loss: 0.2653 - sparse_categorical_accuracy: 0.8927 - val_loss: 0.3167 - val_sparse_categorical_accuracy: 0.8683
Epoch 3/3
1563/1563 [==============================] - 278s 178ms/step - loss: 0.1976 - sparse_categorical_accuracy: 0.9257 - val_loss: 0.3445 - val_sparse_categorical_accuracy: 0.8663
<keras.src.callbacks.History at 0x78a5f5b4c130>

I would like to predict with my new model:

classifier.predict(["I love modular workflows in keras-nlp!"])

and I get an error: "*Layer "bert_classifier" expects 3 input(s), but it received 1 input tensors. *"

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
[<ipython-input-6-3ebc2f7f2482>](https://localhost:8080/#) in <cell line: 1>()
----> 1 classifier.predict(["I love modular workflows in keras-nlp!"])

3 frames
[/usr/local/lib/python3.10/dist-packages/keras/src/engine/training.py](https://localhost:8080/#) in tf__run_step(data)
      9                 do_return = False
     10                 retval_ = ag__.UndefinedReturnValue()
---> 11                 outputs = ag__.converted_call(ag__.ld(model).predict_step, (ag__.ld(data),), None, fscope)
     12                 with ag__.ld(tf).control_dependencies(ag__.ld(_minimum_control_deps)(ag__.ld(outputs))):
     13                     ag__.converted_call(ag__.ld(model)._predict_counter.assign_add, (1,), None, fscope)

ValueError: in user code:

    File "/usr/local/lib/python3.10/dist-packages/keras/src/engine/training.py", line 2341, in predict_function  *
        return step_function(self, iterator)
    File "/usr/local/lib/python3.10/dist-packages/keras/src/engine/training.py", line 2315, in run_step  *
        outputs = model.predict_step(data)
    File "/usr/local/lib/python3.10/dist-packages/keras/src/engine/training.py", line 2283, in predict_step  **
        return self(x, training=False)
    File "/usr/local/lib/python3.10/dist-packages/keras/src/utils/traceback_utils.py", line 70, in error_handler
        raise e.with_traceback(filtered_tb) from None
    File "/usr/local/lib/python3.10/dist-packages/keras/src/engine/input_spec.py", line 219, in assert_input_compatibility
        raise ValueError(

    ValueError: Layer "bert_classifier" expects 3 input(s), but it received 1 input tensors. Inputs received: [<tf.Tensor 'data:0' shape=(None,) dtype=string>]

I would appreciate a quick code block or documentation block (could be a "prologue") to demonstrate how to call this model.

Thanks!