When loaded a saved model from disk, predict is missing

[Howard-ll]

Hi

• Issue symptom When loaded a saved model from disk, predict is missing

• How to reproduce 1) Create a model 2) Save it to disk 3) load it from disk 4) call model.predict (If I do like model(X, training=False), it works well)

  # model creation
  model = tfdf.keras.RandomForestModel(num_trees=n_trees, max_depth=depth, min_examples=1)
  model.fit(x=x_selected, y=y)
  
  # model save
  path = './random_forest'
  os.makedirs(path, exist_ok=True)
  file_name = tempfile.TemporaryDirectory(dir=path).name
  model.save(file_name)
  # model load
  loaded_model = tf.saved_model.load(file_name)
  # inference
  score = loaded_model.predict(x_selected)

AttributeError: '_UserObject' object has no attribute 'predict'

Thanks!

[achoum]

TLDR

Try replacing

loaded_model = tf.saved_model.load(file_name)

by

loaded_model = tf.keras.models.load_model(file_name)

Details

The predict method is not part of the TensorFlow SavedModel object itself. Instead, it is created and restored by Keras (which is build on top of TensorFlow). If you work work with Keras models, it is best to load them with tf.keras.models.load_model.

[Howard-ll]

tf.keras.models.load_model

Thanks! You are right.