Bitpredict predicament while building model_5_lstm (TF 2.7.0)

[EimantasZaranka]

TensorFlow version: 2.7.0

Model:

inputs = layers.Input(shape=(WINDOW))
x = layers.Lambda(lambda x: tf.expand_dims(x, axis=1))(inputs)
x = layers.LSTM(128, activation='relu')(x)
outputs = layers.Dense(HORIZON, activation='linear')(x)
model_5 = tf.keras.Model(inputs, outputs, name='model_5')

While saving a model checkpoint I get an error, which results in inability to load a saved checkpoint (evaluation returns 0s).

WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7f75f5880890> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.

Tried using a custom_objects in loading, but it's a confusing part for me and I hardly get a grasp of how it works.

model_5_loaded = tf.keras.models.load_model('/content/model_experiments/model_5_lstm',
                                            custom_objects={'CustomModel':model_5})

Tried googling the error but with no success also tried to give a name for LSTM Layer. Any suggestions or literature on how to solve the error?

[arrshsh]

Hey, did you get the thing solved? I am doing the course now I am facing the same error.

[EimantasZaranka]

For me it was the issue with TF 2.7, after upgrading it the error disappeared.