python-kerasHow can I use Python Keras to perform a grid search?

You can use Python Keras to perform a grid search by using the KerasClassifier class from the keras.wrappers.scikit_learn module. This class allows you to use the scikit-learn library's GridSearchCV method to perform a grid search.

Example code

# Import necessary modules
from keras.wrappers.scikit_learn import KerasClassifier
from sklearn.model_selection import GridSearchCV

# Create function to build model
def create_model(optimizer='rmsprop', init='glorot_uniform'):
    # Create model
    model = Sequential()
    # Add layers
    # Compile model
    model.compile(optimizer=optimizer, loss='categorical_crossentropy', metrics=['accuracy'])
    return model

# Create model
model = KerasClassifier(build_fn=create_model, verbose=0)

# Define the grid search parameters
optimizer = ['SGD', 'RMSprop', 'Adagrad', 'Adadelta', 'Adam', 'Adamax', 'Nadam']
init = ['uniform', 'lecun_uniform', 'normal', 'zero', 'glorot_normal', 'glorot_uniform', 'he_normal', 'he_uniform']

# Create the grid
param_grid = dict(optimizer=optimizer, init=init)

# Perform the grid search
grid = GridSearchCV(estimator=model, param_grid=param_grid, n_jobs=-1, cv=3)
grid_result = grid.fit(X, Y)

# Print the best parameters
print("Best: %f using %s" % (grid_result.best_score_, grid_result.best_params_))

The code above will search through a set of optimizers and initialization parameters to find the best combination for the model. The GridSearchCV method will then output the best score and parameters.

Code explanation

• KerasClassifier: This class allows you to use the scikit-learn library's GridSearchCV method to perform a grid search.
• GridSearchCV: This method will search through a set of optimizers and initialization parameters to find the best combination for the model.
• fit: This method will perform the grid search.
• best_score_: This will output the best score from the grid search.
• best_params_: This will output the best parameters from the grid search.

Helpful links

• KerasClassifier
• GridSearchCV

Edit this code on GitHub