# Set the parameters by cross-validation tuned_parameters = [{'kernel': ['rbf'], 'gamma': [1e-3, 1e-4], 'C': [1, 10, 100, 1000]}, {'kernel': ['linear'], 'C': [1, 10, 100, 1000]}] scores = [ ('precision', ), ('recall', recall_score), ] # for score_name, score_func in scores: # print "# Tuning hyper-parameters for %s" % score_name print clf = IPythonGridSearchCV(SVC(C=1), tuned_parameters, score_func=precision_score, view=v) clf.fit(X_train, y_train, cv=5) print "fit submitted" while v.outstanding: v.wait(timeout=0.1) grid_scores = clf.collect_results() # import IPython # IPython.embed() # # print "Best parameters set found on development set:" # print # print clf.best_estimator_ print print "Grid scores on development set:" print
# turn the data in a (samples, feature) matrix: n_samples = len(digits.images) X = digits.images.reshape((n_samples, -1)) y = digits.target # Split the dataset in two equal parts X_train, X_test, y_train, y_test = train_test_split( X, y, test_fraction=0.5, random_state=0) # Set the parameters by cross-validation tuned_parameters = [{'kernel': ['rbf'], 'gamma': [1e-3, 1e-4], 'C': [1, 10, 100, 1000]}, {'kernel': ['linear'], 'C': [1, 10, 100, 1000]}] search = IPythonGridSearchCV(SVC(C=1), tuned_parameters, score_func=precision_score, view=v, cv=5, local_store='/tmp') search.fit_async(X_train, y_train) print "Launched asynchronous fit on a cluster." def print_scores(scores): for params, mean_score, scores, mean_duration, durations in scores: print "%0.3f (+/-%0.03f) [%i] for %r (%0.3fs)" % ( mean_score, scores.std() / 2, len(scores), params, mean_duration) while v.outstanding: v.wait(timeout=0.5) completed_scores, n_remaining = search.collect_results() top_scores = completed_scores[:3]