tensorflow: tf.keras.models.load_model fails on Sequential model

System information

• Have I written custom code (as opposed to using a stock example script provided in TensorFlow): Yes
• OS Platform and Distribution (e.g., Linux Ubuntu 16.04): Linux Ubuntu 18.04
• TensorFlow installed from (source or binary): Binary
• TensorFlow version (use command below): tf-nightly-gpu-2.0-preview==2.0.0-dev20190718
• Python version: 3.7

Describe the current behavior Attempt to run tf.keras.models.load_model on a tf.keras.Sequential with lazily generated inputs produces an error.

Describe the expected behavior No error is thrown when running the script below. model2 is a clone of model.

Code to reproduce the issue

import tempfile

import tensorflow as tf


def main():
    batch_size = 3

    image_shape = (32, 32, 3)
    inputs = tf.random.uniform((batch_size, *image_shape))

    model = tf.keras.Sequential((
        tf.keras.layers.Conv2D(
            filters=16,
            kernel_size=3,
            strides=2,
            padding='SAME',
            activation='linear'),
    ))

    _ = model(inputs)

    with tempfile.NamedTemporaryFile(suffix='.hdf5', delete=True) as fd:
        tf.keras.models.save_model(model, fd.name, overwrite=True)
        model2 = tf.keras.models.load_model(fd.name, compile=False)

    print(model2.summary())


if __name__ == '__main__':
    main()

Other info / logs Running the script above produces the following error:

Traceback (most recent call last):
  File "/home/kristian/conda/envs/softlearning-2/lib/python3.7/runpy.py", line 193, in _run_module_as_main
    "__main__", mod_spec)
  File "/home/kristian/conda/envs/softlearning-2/lib/python3.7/runpy.py", line 85, in _run_code
    exec(code, run_globals)
  File "/home/kristian/github/hartikainen/softlearning-2/tests/test_sequential_serialize.py", line 31, in <module>
    main()
  File "/home/kristian/github/hartikainen/softlearning-2/tests/test_sequential_serialize.py", line 25, in main
    model2 = tf.keras.models.load_model(fd.name, compile=False)
  File "/home/kristian/conda/envs/softlearning-2/lib/python3.7/site-packages/tensorflow_core/python/keras/saving/save.py", line 138, in load_model
    return hdf5_format.load_model_from_hdf5(filepath, custom_objects, compile)
  File "/home/kristian/conda/envs/softlearning-2/lib/python3.7/site-packages/tensorflow_core/python/keras/saving/hdf5_format.py", line 165, in load_model_from_hdf5
    load_weights_from_hdf5_group(f['model_weights'], model.layers)
  File "/home/kristian/conda/envs/softlearning-2/lib/python3.7/site-packages/tensorflow_core/python/keras/saving/hdf5_format.py", line 671, in load_weights_from_hdf5_group
    ' layers.')
ValueError: You are trying to load a weight file containing 1 layers into a model with 0 layers.
Exception ignored in: <function _TensorCacheDeleter.__del__ at 0x7f9fabdde158>
Traceback (most recent call last):
  File "/home/kristian/conda/envs/softlearning-2/lib/python3.7/site-packages/tensorflow_core/python/eager/context.py", line 316, in __del__
TypeError: argument of type 'NoneType' is not iterable