gridsearch_tensorflow.py

classifier = skflow.TensorFlowEstimator(model_fn=conv_model, n_classes=outer_name+1,
                                        batch_size=10)


# use a full grid over all parameters
param_grid = {"steps": [1000, 1500, 2000, 2500, 3000],
              "learning_rate": [0.01, 0.03, 0.05, 0.08],
              "batch_size": [8, 10, 12]}

# run grid search
grid_search = GridSearchCV(classifier, param_grid=param_grid, scoring = 'accuracy', verbose=10, n_jobs=-1,cv=2)
grid_search.fit(X_train, y_train)

print(grid_search)
# summarize the results of the grid search
print(grid_search.best_score_)
print(grid_search.best_params_)