[CS231n@Stanford] Assignment1-Q5 (python) features实现

features.ipynb

from linear_classifier import LinearSVMlearning_rates = [1e-9, 1e-8, 1e-7]regularization_strengths = [1e5, 1e6, 1e7]results = {}best_val = -1best_svm = Nonepass################################################################################# TODO:                                                                        ## Use the validation set to set the learning rate and regularization strength. ## This should be identical to the validation that you did for the SVM; save    ## the best trained classifer in best_svm. You might also want to play          ## with different numbers of bins in the color histogram. If you are careful    ## you should be able to get accuracy of near 0.44 on the validation set.       #################################################################################iters = 2000  for lr in learning_rates:      for reg in regularization_strengths:          svm = LinearSVM()          svm.train(X_train_feats, y_train, learning_rate=lr, reg=reg, num_iters=iters)                    y_train_pred = svm.predict(X_train_feats)          acc_train = np.mean(y_train == y_train_pred)                    y_val_pred = svm.predict(X_val_feats)          acc_val = np.mean(y_val == y_val_pred)            results[(lr, reg)] = (acc_train, acc_val)                    if best_val < acc_val:              best_val = acc_val              best_svm = svm              pass#################################################################################                              END OF YOUR CODE                                ################################################################################## Print out results.for lr, reg in sorted(results):    train_accuracy, val_accuracy = results[(lr, reg)]    print 'lr %e reg %e train accuracy: %f val accuracy: %f' % (                lr, reg, train_accuracy, val_accuracy)    print 'best validation accuracy achieved during cross-validation: %f' % best_val# Evaluate your trained SVM on the test sety_test_pred = best_svm.predict(X_test_feats)test_accuracy = np.mean(y_test == y_test_pred)print 'test_accuracy',test_accuracy

lr 1.000000e-09 reg 1.000000e+05 train accuracy: 0.108265 val accuracy: 0.123000
lr 1.000000e-09 reg 1.000000e+06 train accuracy: 0.111939 val accuracy: 0.133000
lr 1.000000e-09 reg 1.000000e+07 train accuracy: 0.414204 val accuracy: 0.413000
lr 1.000000e-08 reg 1.000000e+05 train accuracy: 0.110429 val accuracy: 0.112000
lr 1.000000e-08 reg 1.000000e+06 train accuracy: 0.413490 val accuracy: 0.406000
lr 1.000000e-08 reg 1.000000e+07 train accuracy: 0.419245 val accuracy: 0.422000
lr 1.000000e-07 reg 1.000000e+05 train accuracy: 0.414082 val accuracy: 0.418000
lr 1.000000e-07 reg 1.000000e+06 train accuracy: 0.414673 val accuracy: 0.414000
lr 1.000000e-07 reg 1.000000e+07 train accuracy: 0.325633 val accuracy: 0.347000
best validation accuracy achieved during cross-validation: 0.422000
test_accuracy 0.417

from neural_net import TwoLayerNetinput_dim = X_train_feats.shape[1]hidden_dim = 500num_classes = 10#net = TwoLayerNet(input_dim, hidden_dim, num_classes)best_net = None################################################################################# TODO: Train a two-layer neural network on image features. You may want to    ## cross-validate various parameters as in previous sections. Store your best   ## model in the best_net variable.                                              #################################################################################results = {} best_val = -1  learning_rate = [1e-1, 5e-1, 1]regularization_strengths = [1e-3, 5e-3, 1e-2] iters = 2000    for lr in learning_rate:        for reg in regularization_strengths:                   print lr,reg        net = TwoLayerNet(input_dim, hidden_dim, num_classes)        net.train(X_train_feats, y_train, X_val_feats, y_val,                   learning_rate=lr, reg=reg, learning_rate_decay=0.95,                   num_iters=iters, verbose=False)                        y_train_pred = net.predict(X_train_feats)            acc_train = np.mean(y_train == y_train_pred)                        y_val_pred = net.predict(X_val_feats)            acc_val = np.mean(y_val == y_val_pred)                results[(lr, reg)] = (acc_train, acc_val)                        if best_val < acc_val:                best_val = acc_val                best_net = net   pass#################################################################################                              END OF YOUR CODE                                #################################################################################for lr, reg in sorted(results):    train_accuracy, val_accuracy = results[(lr, reg)]    print 'lr %e reg %e train accuracy: %f val accuracy: %f' % (                lr, reg, train_accuracy, val_accuracy)    print 'best validation accuracy achieved during cross-validation: %f' % best_val # Run your neural net classifier on the test set. You should be able to# get more than 55% accuracy.test_acc = (net.predict(X_test_feats) == y_test).mean()print 'test_acc:',test_acc

lr 1.000000e-01 reg 1.000000e-03 train accuracy: 0.554041 val accuracy: 0.539000
lr 1.000000e-01 reg 5.000000e-03 train accuracy: 0.544286 val accuracy: 0.529000
lr 1.000000e-01 reg 1.000000e-02 train accuracy: 0.533510 val accuracy: 0.526000
lr 5.000000e-01 reg 1.000000e-03 train accuracy: 0.704449 val accuracy: 0.596000
lr 5.000000e-01 reg 5.000000e-03 train accuracy: 0.625367 val accuracy: 0.569000
lr 5.000000e-01 reg 1.000000e-02 train accuracy: 0.569918 val accuracy: 0.535000
lr 1.000000e+00 reg 1.000000e-03 train accuracy: 0.731633 val accuracy: 0.586000
lr 1.000000e+00 reg 5.000000e-03 train accuracy: 0.616347 val accuracy: 0.575000
lr 1.000000e+00 reg 1.000000e-02 train accuracy: 0.548939 val accuracy: 0.533000
best validation accuracy achieved during cross-validation: 0.596000
test_acc: 0.536