Support of Pipeline

[epigos]

Could you please let me know if this support sklearn.pipeline.Pipeline estimators. I tried it and had some errors.

Below is my pipeline

hog_svc_pipe = Pipeline(steps=[
    ('hog', HogTransformer()),
    ('svc', svm.SVC(random_state=42))
])
# define gridsearch parameters
param_grid = {
    'svc__kernel': ['linear', 'rbf'], 
    'svc__C': [1, 10, 100], 
    'svc__gamma': ['scale', 'auto'],
    'hog__block_norm': ['L2-Hys', 'L1', 'L1-sqrt'],
}

# est = svm.SVC(random_state=42)
hog_svc_grid = GridSearch(model=hog_svc_pipe, param_grid=param_grid)
# Grid-search all parameter combinations using a validation set.
hog_svc_grid.fit(train_data.images, train_data.target, val_data.images, val_data.target)

I get the following error

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-60-6ba446e48e56> in <module>()
----> 1 get_ipython().run_cell_magic('time', '', "# define gridsearch parameters\nparam_grid = {\n    'svc__kernel': ['linear', 'rbf'], \n    'svc__C': [1, 10, 100], \n    'svc__gamma': ['scale', 'auto'],\n    'hog__block_norm': ['L2-Hys', 'L1', 'L1-sqrt'],\n}\n\n# est = svm.SVC(random_state=42)\nhog_svc_grid = GridSearch(model=hog_svc_pipe, param_grid=param_grid)\n# Grid-search all parameter combinations using a validation set.\nhog_svc_grid.fit(train_data.images, train_data.target, val_data.images, val_data.target)")

3 frames
<decorator-gen-60> in time(self, line, cell, local_ns)

<timed exec> in <module>()

/usr/local/lib/python3.6/dist-packages/hypopt/model_selection.py in fit(self, X_train, y_train, X_val, y_val, scoring, scoring_params, verbose)
    361             else:
    362                 results = [_run_thread_job(job) for job in params]
--> 363             models, scores = list(zip(*results))
    364             self.model = models[np.argmax(scores)]
    365         else:

ValueError: not enough values to unpack (expected 2, got 0)