[FIXED] Missing Positional Argument X

Issue

I'm currently working to set up a randomized search to help find the best parameters for a model. But I've hit an error that I can't find anywhere on google: Missing positional argument 'x' in the .fit function for the randomized search model.

This is what the code for my model and randomized search model look like.

# Build model
def build_model(il_neurons=100, hl_neurons=50, num_hl=10):
    model = Sequential()
    model.add(Dense(il_neurons, input_dim=num_inputs, activation='relu'))
    for i in range(num_hl):
        model.add(Dense(hl_neurons, activation='relu'))
    model.add(Dense(1))
    
    model.compile(optimizer=optimizer, loss="mean_squared_error", metrics=[RootMeanSquaredError()])
    return model

'''Random Search Model'''
# Parameters
il_neurons = range(75, 151, 5)
hl_neurons = range(10, 61, 3)
num_hl = range(1, 16, 1)

parameters = {"model__il_neurons": il_neurons,
              "model__hl_neurons": hl_neurons,
              "model__num_hl": num_hl}


# Model
model = KerasRegressor(build_fn=build_model)

rscv_model = RandomizedSearchCV(estimator=model,
                                param_distributions=parameters,
                                n_iter=20,
                                scoring='neg_root_mean_squared_error',
                                cv=kfold,
                                n_jobs=-1,
                                verbose=0)

history = rscv_model.fit(x=X_train_pca,
                         y=y_train,
                         validation_split=0.1,
                         epochs=epochs,
                         batch_size=batch_size,
                         callbacks=callbacks,
                         shuffle=True,
                         verbose=0)

Running that same model on it's own doesn't throw any errors. Works totally fine. But running the random search model code gives this error:

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
/tmp/ipykernel_41/3876581477.py in <module>
     28                          callbacks=callbacks,
     29                          shuffle=True,
---> 30                          verbose=0)
     31 
     32 print("Best Parameters:")

/opt/conda/lib/python3.7/site-packages/sklearn/utils/validation.py in inner_f(*args, **kwargs)
     70                           FutureWarning)
     71         kwargs.update({k: arg for k, arg in zip(sig.parameters, args)})
---> 72         return f(**kwargs)
     73     return inner_f
     74 

TypeError: fit() missing 1 required positional argument: 'X'

I've tried cutting out everything except my X and y arguments. That doesn't help. I've double checked to make sure that my X data does actually exist; it does. I've double checked that my X data is the right type; it is. I've googled the issue in every way that I can think to word it, but I can't find anyone who's gotten this same problem.

Basically, I have no idea what to do. Help.

Solution

rscv_model.fit requires an X (uppercase X) parameter, not lowercase. Have a look at the docs.

Answered By - mcsoini