[Bug] after I saved rnet to saved model. when I try to load, get errors.

[iysheng]

Describe the bug
I saved rnet mode to tensorflow saved model in red_rnet dir success.
But when I try to load this saved model ,I get errors as

red_rnet has saved before
Traceback (most recent call last):
  File "/home/yjoy/gProjects/mtcnn/./test00.py", line 22, in <module>
    detector = MTCNN(device="CPU:0")
  File "/home/yjoy/gProjects/mtcnn/mtcnn/mtcnn.py", line 61, in __init__
    self._stages = [stage() if isinstance(stage, type) else stage for stage in stages]
  File "/home/yjoy/gProjects/mtcnn/mtcnn/mtcnn.py", line 61, in <listcomp>
    self._stages = [stage() if isinstance(stage, type) else stage for stage in stages]
  File "/home/yjoy/gProjects/mtcnn/mtcnn/stages/stage_rnet.py", line 65, in __init__
    model = tf.keras.models.load_model('red_rnet/', compile=False)
  File "/home/yjoy/.conda/envs/mtcnn_tf12/lib/python3.10/site-packages/keras/saving/saving_api.py", line 212, in load_model
    return legacy_sm_saving_lib.load_model(
  File "/home/yjoy/.conda/envs/mtcnn_tf12/lib/python3.10/site-packages/keras/utils/traceback_utils.py", line 70, in error_handler
    raise e.with_traceback(filtered_tb) from None
  File "/home/yjoy/.conda/envs/mtcnn_tf12/lib/python3.10/site-packages/keras/initializers/initializers.py", line 171, in __call__
    return tf.zeros(shape, dtype)
ValueError: Attempt to convert a value (None) with an unsupported type (<class 'NoneType'>) to a Tensor.

Besides， I use the same way to save pnet saved mode, and load the saved mode success.

the code with releate as

+            model = RNet()
+            model.build()  # Building the RNet model
+            model.set_weights(load_weights(weights))  # Load pre-trained weights
+            print(model, model.summary(), "\nsummary done of rnet")
+            # for save mode， use dummy_input here
+            dummy_input=np.random.rand(1,24,24,3)
+            model(dummy_input)
+            #  model.save("red_rnet/", include_optimizer=True)
+            model.save("red_rnet/")

the load code as

model = tf.keras.models.load_model('red_rnet/', compile=False)

To Reproduce
save rnet mode, and then load the mode.

Expected behavior
load rnet model success.

Screenshots

Environment (please complete the following information):

• OS: [e.g., Windows, macOS, Linux]
• Python version: [3.10]
• Version of the project: [master version]

Additional context
Add any other context about the problem here.

[ladyshen]

您的邮件已接收！