ValueError when reloading EfficientNet Model in tf.keras

[lyndonteng]

While training, tf returned a (non fatal) error indicating that there were multiple overlapping variable names.

Input:

import efficientnet.tfkeras
model = tf.keras.models.load_model(load_model_path, custom_objects={'custom_loss':custom_loss})

Ouptut:

ValueError                                Traceback (most recent call last)
<ipython-input-17-55e7f4d0b56f> in <module>()
      8   model = transfer_model_7x7_14x14(backbone_model, input_shape=input_shape, dims_list=dims_list, num_aspect_ratios=len(aspect_ratios), wt_decay=wt_decay, model_name=model_context+'-res50')
      9 else:
---> 10   model = tf.keras.models.load_model(load_model_path, custom_objects={'custom_loss':custom_loss})
     11 model.summary()

5 frames
/usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/utils/generic_utils.py in class_and_config_for_serialized_keras_object(config, module_objects, custom_objects, printable_module_name)
    319   cls = get_registered_object(class_name, custom_objects, module_objects)
    320   if cls is None:
--> 321     raise ValueError('Unknown ' + printable_module_name + ': ' + class_name)
    322 
    323   cls_config = config['config']

ValueError: Unknown layer: Functional