Error: model must be a callable, a Keras Model instance or None

[tserversbfs]

I am way over my head... I keep getting this error and I have no idea what it means. I am hoping someone can help me fix this error. I have searched for "must be a callable, a Keras Model instance or None" and have found nothing.

def build_model(self):
    model = Sequential()
    model.add(Dense(64, input_dim=7, activation='relu'))
    model.add(Dropout(0.2))
    model.add(Dense(32, activation='relu'))
    model.add(Dense(1, activation='sigmoid'))
    model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])
    return model

def train_model(self, X, y):
    # Wrap the Keras model for use in scikit-learn
    keras_model = KerasClassifier(model=self.build_model, epochs=10, batch_size=32, verbose=0)

    # Perform hyperparameter tuning
    param_grid = {
        'model__epochs': [10],
        'model__batch_size': [32]
    }
    grid_search = GridSearchCV(estimator=keras_model, param_grid=param_grid, cv=3, scoring='accuracy', error_score='raise')
    grid_search.fit(X, y)

    # Set the best model
    self.model = grid_search.best_estimator_

thank you for you patience.

[Surya2k1]

@tserversbfs , There was an issue with wrapping a compiled model with SKLearn wrappers which was addressed in this PR #20888.
I think this issue also might probably relevant to the same. Can you please check with latest nightly and if still issue exists providing replicable code snippet might help to look into the issue.