grid_search.py

from sklearn.model_selection import StratifiedKFold, GridSearchCV

# Declare grid of hyper-parameters
dict_hyperparams = dict(svm__gamma=[0.001,0.01,0.1],
                        svm__C = [1,10,100,1_000])
# Cross Validation for GridSearchCV
crossVal = StratifiedKFold(n_splits=5, shuffle=True, random_state=0)

model_02 = Pipeline(steps= [
    ('hogdescriptor', DescriptorHOG()),
    ('pca', PCA(n_components=0.9, svd_solver='full')),
    ('scaler', StandardScaler()),
    ('svm', SVC(kernel='rbf'))
    ])

grid = GridSearchCV(
    estimator= model_02, # Model
    param_grid= dict_hyperparams, # hyper-parameters
    cv= crossVal, # Cross Validation
    n_jobs=-1, # Use all CPU's
    verbose= 5 # Info
)

# Train the model
grid.fit(X_train,y_train)

# Best hyper-parameters
gamma_02 = grid.best_params_['svm__gamma']
C_02 = grid.best_params_['svm__C']