Description
Describe the bug
In chapter 12, when working with the custom models, the Keras fails loading the model after saving. It is saved safely (not pure safely as it throws warning indicating the save format should be converted to Keras rather than h5) and after loading it throws an exceptions.
notebook name: 12_custom_models_and_training_with_tensorflow.ipynb
cell numbers: 91
To Reproduce
Please copy the code that fails here, using code blocks like this:
model = keras.models.load_model(
"my_model_with_many_custom_parts.h5",
custom_objects={
"my_l1_regularizer": my_l1_regularizer,
"my_positive_weights": my_positive_weights,
"my_glorot_initializer": my_glorot_initializer,
"my_softplus": my_softplus,
})
And if you got an exception, please copy the full stacktrace here:
[<ipython-input-43-137f634240f5>](https://m6gi76vt7u-496ff2e9c6d22116-0-colab.googleusercontent.com/outputframe.html?vrz=colab_20240405-060114_RC00_622131336#) in <cell line: 1>()
----> 1 model = keras.models.load_model(
2 "my_model_with_many_custom_parts.h5",
3 custom_objects={
4 "my_l1_regularizer": my_l1_regularizer,
5 "my_positive_weights": my_positive_weights,
2 frames
[/usr/local/lib/python3.10/dist-packages/keras/src/engine/base_layer.py](https://m6gi76vt7u-496ff2e9c6d22116-0-colab.googleusercontent.com/outputframe.html?vrz=colab_20240405-060114_RC00_622131336#) in from_config(cls, config)
868 return cls(**config)
869 except Exception as e:
--> 870 raise TypeError(
871 f"Error when deserializing class '{cls.__name__}' using "
872 f"config={config}.\n\nException encountered: {e}"
TypeError: Error when deserializing class 'Dense' using config={'name': 'dense_1', 'trainable': True, 'dtype': 'float32', 'units': 1, 'activation': {'module': 'builtins', 'class_name': 'function', 'config': 'my_softplus', 'registered_name': 'function'}, 'use_bias': True, 'kernel_initializer': {'module': 'builtins', 'class_name': 'function', 'config': 'my_glorot_initializer', 'registered_name': 'function'}, 'bias_initializer': {'module': 'keras.initializers', 'class_name': 'Zeros', 'config': {}, 'registered_name': None}, 'kernel_regularizer': {'module': 'builtins', 'class_name': 'function', 'config': 'my_l1_regularizer', 'registered_name': 'function'}, 'bias_regularizer': None, 'activity_regularizer': None, 'kernel_constraint': {'module': 'builtins', 'class_name': 'function', 'config': 'my_positive_weights', 'registered_name': 'function'}, 'bias_constraint': None}.
Exception encountered: Unknown activation function: 'function'. Please ensure you are using a `keras.utils.custom_object_scope` and that this object is included in the scope. See https://www.tensorflow.org/guide/keras/save_and_serialize#registering_the_custom_object for details.
Expected behavior
A clear and concise description of what you expected to happen.
I expect to load model without any error.
Versions (please complete the following information):
I have run it on google colab
- OS: Nil
- Python: 3.10.12
- TensorFlow: 2.15.0
- Scikit-Learn: 1.2.2