AttributeError: 'DenseBlock' object has no attribute 'get_adapter'

[yurimarimo]

This is my code.

import autokeras as ak

input_node = ak.Input()
gru_node = ak.RNNBlock(return_sequences=True, layer_type="gru")(input_node)
output_node = ak.DenseBlock()(gru_node)

auto_model = ak.AutoModel(
    inputs=input_node, 
    outputs=output_node,
    overwrite=True,
    max_trials=2)

auto_model.fit(X_train, y_train, batch_size=100, epochs=1, validation_split=0.1)

When I tried, I got the following error.

AttributeError: 'DenseBlock' object has no attribute 'get_adapter'

I don't know how to solve this problem, so I hope you can tell me something.

Thank you.

[KiLJ4EdeN]

This is my code.

import autokeras as ak

input_node = ak.Input()
gru_node = ak.RNNBlock(return_sequences=True, layer_type="gru")(input_node)
output_node = ak.DenseBlock()(gru_node)

auto_model = ak.AutoModel(
    inputs=input_node, 
    outputs=output_node,
    overwrite=True,
    max_trials=2)

auto_model.fit(X_train, y_train, batch_size=100, epochs=1, validation_split=0.1)

When I tried, I got the following error.

AttributeError: 'DenseBlock' object has no attribute 'get_adapter'

I don't know how to solve this problem, so I hope you can tell me something.

Thank you.

Hello I reproduced your error with random vectors,

The error can be solved by adding a head to the end of the network, this also can be a RegressionHead.
A full example would look like:

import autokeras as ak

input_node = ak.Input()
gru_node = ak.RNNBlock(return_sequences=True, layer_type="lstm")(input_node)
dense_node = ak.DenseBlock()(gru_node)
output_node = ak.ClassificationHead()(dense_node)

auto_model = ak.AutoModel(
inputs=input_node,
outputs=output_node,
overwrite=True,
max_trials=2)

auto_model.fit(X_train, y_train, batch_size=100, epochs=1, validation_split=0.1)

I think the problem is that you are using a non final layer as an output layer.

[yurimarimo]

Thank you for your answer.
I've solved the problem.