I am trying to fine-tune an AE-LSTM, using keras tuner, to use the output of the embedding layer in the rest of my project.
To this end, I have defined two classes, one for train and test the fine-tuned model, which I called "LstmAe", and the other to fine-tune it, which I called "ApplyLstmAe."
The problem is when I try execute my fine_tune_the_model() method in the latter class I face the following error:
TypeError: Inputs to a layer should be tensors. Got: <keras_tuner.src.engine.hyperparameters.hyperparameters.HyperParameters object at 0x7fdc7c3438e0>
My failed attempted:
Changing build method as staticmethod, define "hps" inside the build function, changing InputLayer to Input etc., unfortunately I failed.
Googling also did not help me to find a proper answer (although one unanswered post on SO and one issue on GitHub are available).
Describe the bug
I am trying to fine-tune an AE-LSTM, using keras tuner, to use the output of the embedding layer in the rest of my project.
To this end, I have defined two classes, one for train and test the fine-tuned model, which I called "LstmAe", and the other to fine-tune it, which I called "ApplyLstmAe."
The problem is when I try execute my fine_tune_the_model() method in the latter class I face the following error:
TypeError: Inputs to a layer should be tensors. Got: <keras_tuner.src.engine.hyperparameters.hyperparameters.HyperParameters object at 0x7fdc7c3438e0>
My failed attempted:
Changing build method as staticmethod, define "hps" inside the build function, changing InputLayer to Input etc., unfortunately I failed.
Googling also did not help me to find a proper answer (although one unanswered post on SO and one issue on GitHub are available).
To Reproduce
In the link below you can find the google-colab file, however, I cannot share the data https://colab.research.google.com/drive/1rJMqaeAjsKvWvI3EcdiHY2ZPxCJCKu-y?usp=sharing
Expected behavior
Method fine_tune_the_model does not accept any inputs, I expect it to initiate the tuner; however, instead it produces such an error.
Would you like to help us fix it? Unofortunately, I cannot help.