def test_KNN_train(sample_train, sample_test):
#this test is designed to verify that input shapes are correct
Xtrain, ytrain = sample_train(count=40)
Xtest, ytest = sample_test(count=10)
with pytest.raises(ValueError):
knn = KNearestNeighbor()
knn.train(Xtrain, ytrain)
def test_KNN_train_reshape_input(sample_train, sample_test):
Xtrain, ytrain = sample_train(count=40)
Xtest, ytest = sample_test(count=10)
Xtrain = np.reshape(Xtrain, (Xtrain.shape[0], -1))
Xtest = np.reshape(Xtest, (Xtest.shape[0], -1))
knn = KNearestNeighbor()
knn.train(Xtrain,ytrain)
def test_KNN_dists_noloop_shape(sample_train, sample_test, in_count):
Xtrain, ytrain = sample_train(count=in_count)
Xtest, ytest = sample_test(count=in_count-30)
Xtrain = np.reshape(Xtrain, (Xtrain.shape[0], -1))
Xtest = np.reshape(Xtest, (Xtest.shape[0], -1))
knn = KNearestNeighbor()
knn.train(Xtrain,ytrain)
assert knn.compute_distances_no_loops(Xtest).shape == (Xtest.shape[0], Xtrain.shape[0])
def test_KNN_predict_loop_parameter(sample_train, sample_test, k, num_loops):
Xtrain, ytrain = sample_train(count=40)
Xtest, ytest = sample_test(count=10)
Xtrain = np.reshape(Xtrain, (Xtrain.shape[0], -1))
Xtest = np.reshape(Xtest, (Xtest.shape[0], -1))
knn = KNearestNeighbor()
knn.train(Xtrain,ytrain)
assert knn.predict(Xtest,k,num_loops).shape == ytest.shape
def test_KNN_predict_labels_shape(sample_train, sample_test):
Xtrain, ytrain = sample_train(count=40)
Xtest, ytest = sample_test(count=10)
Xtrain = np.reshape(Xtrain, (Xtrain.shape[0], -1))
Xtest = np.reshape(Xtest, (Xtest.shape[0], -1))
knn = KNearestNeighbor()
knn.train(Xtrain,ytrain)
dist_no = knn.compute_distances_no_loops(Xtest)
assert knn.predict_labels(dist_no, k=1).shape == ytest.shape
assert knn.predict_labels(dist_no, k=2).shape == ytest.shape
assert knn.predict_labels(dist_no, k=3).shape == ytest.shape
assert knn.predict_labels(dist_no, k=4).shape == ytest.shape
def test_KNN_predict_incorrect_shape(sample_train, sample_test):
Xtrain, ytrain = sample_train(count=500)
Xtest, ytest = sample_test(count=125)
Xtrain = np.reshape(Xtrain, (Xtrain.shape[0], -1))
Xtest = np.reshape(Xtest, (Xtest.shape[0], -1))
knn = KNearestNeighbor()
knn.train(Xtrain,ytrain)
with pytest.raises(ValueError):
knn.predict(ytrain)#using ytrain, shich has incorrect dimensions;
def test_KNN_predict_num_loop_parameter(sample_train, sample_test, num_loops):
Xtrain, ytrain = sample_train(count=40)
Xtest, ytest = sample_test(count=10)
Xtrain = np.reshape(Xtrain, (Xtrain.shape[0], -1))
Xtest = np.reshape(Xtest, (Xtest.shape[0], -1))
knn = KNearestNeighbor()
knn.train(Xtrain,ytrain)
with pytest.raises(ValueError):
knn.predict(Xtest,0,num_loops).shape
def test_KNN_dists_one_to_none(sample_train, sample_test):
Xtrain, ytrain = sample_train(count=40)
Xtest, ytest = sample_test(count=10)
Xtrain = np.reshape(Xtrain, (Xtrain.shape[0], -1))
Xtest = np.reshape(Xtest, (Xtest.shape[0], -1))
knn = KNearestNeighbor()
knn.train(Xtrain,ytrain)
dist_one = knn.compute_distances_one_loop(Xtest)
dist_no = knn.compute_distances_no_loops(Xtest)
assert np.linalg.norm(dist_one - dist_no, ord='fro') < 0.001
num_test = 500
mask = list(range(num_test))
X_test = X_test[mask]
y_test = y_test[mask]
# Reshape the image data into rows
X_train = np.reshape(X_train, (X_train.shape[0], -1))
X_test = np.reshape(X_test, (X_test.shape[0], -1))
print(X_train.shape, X_test.shape)
from cs231n.classifiers.k_nearest_neighbor import KNearestNeighbor
# Create a kNN classifier instance.
# Remember that training a kNN classifier is a noop:
# the Classifier simply remembers the data and does no further processing
classifier = KNearestNeighbor()
classifier.train(X_train, y_train)
"""
dists = classifier.compute_distances_two_loops(X_test)
pickle.dump(dists,open(r"D:\python\CS231n\assignment1\tmp.txt","wb"))
print(dists.shape)
print(dists)
plt.imshow(dists, interpolation='none')
plt.show()
"""
with open(r"D:\python\CS231n\assignment1\tmp.txt", "rb") as file:
dists = pickle.load(file)
y_test_pred = classifier.predict_labels(dists, k=1)
# Compute and print the fraction of correctly predicted examples
if i == 0:
plt.title(cls)
plt.show()
# Subsample the data for more efficient code execution in this exercise
num_training = 5000
mask = range(num_training)
X_train = X_train[mask]
y_train = y_train[mask]
num_test = 500
mask = range(num_test)
X_test = X_test[mask]
y_test = y_test[mask]
# Reshape the image data into rows
# CV: each row is a data vector of 3072 items.
# 5k for training and 500 for testing
X_train = np.reshape(X_train, (X_train.shape[0], -1))
X_test = np.reshape(X_test, (X_test.shape[0], -1))
print X_train.shape, X_test.shape
# CV: Just saves the data
# Create a kNN classifier instance.
# Remember that training a kNN classifier is a noop:
# the Classifier simply remembers the data and does no further processing
classifier = KNearestNeighbor()
classifier.train(X_train, y_train)
num_test = 500 mask = list(range(num_test)) X_test = X_test[mask] y_test = y_test[mask] # Reshape the image data into rows X_train = np.reshape(X_train, (X_train.shape[0], -1)) X_test = np.reshape(X_test, (X_test.shape[0], -1)) print(X_train.shape, X_test.shape) from cs231n.classifiers.k_nearest_neighbor import KNearestNeighbor # Create a kNN classifier instance. # Remember that training a kNN classifier is a noop: # the Classifier simply remembers the data and does no further processing classifier = KNearestNeighbor() classifier.train(X_train, y_train) # Open cs231n/classifiers/k_nearest_neighbor.py and implement # compute_distances_two_loops. # Test your implementation: dists = classifier.compute_distances_two_loops(X_test) # We can visualize the distance matrix: each row is a single test example and # its distances to training examples plt.imshow(dists, interpolation='none') plt.show() # Now implement the function predict_labels and run the code below: # We use k = 1 (which is Nearest Neighbor).
X_train = np.reshape(X_train, (X_train.shape[0], -1)) X_test = np.reshape(X_test, (X_test.shape[0], -1)) print(X_train.shape, X_test.shape) #%% #import sys #new = cwd+'\\cs231n' #sys.path.intert(0, new+'\\classifier') # where the classifier is stored from cs231n.classifiers.k_nearest_neighbor import KNearestNeighbor # Create a kNN classifier instance. # Remember that training a kNN classifier is a noop: # the Classifier simply remembers the data and does no further processing classifier = KNearestNeighbor() classifier.train(X_train, y_train) #% calcualte the distance # Open cs231n/classifiers/k_nearest_neighbor.py and implement # compute_distances_two_loops. # Test your implementation: dists = classifier.compute_distances_no_loops(X_test) #dists = classifier.compute_distances_two_loops(X_test) #dists = classifier.compute_distances_one_loop(X_test) #%% # We can visualize the distance matrix: each row is a single test example and # its distances to training examples plt.figure()
num_training=50000
mask=range(num_training)
x_train=x_train[mask]
y_train=y_train[mask]
num_test=10000
mask=range(num_test)
x_test=x_test[mask]
y_test=y_test[mask]
x_train=np.reshape(x_train,(x_train.shape[0],-1))
x_test=np.reshape(x_test,(x_test.shape[0],-1))
print(x_train.shape,x_test.shape)
classifier=KNearestNeighbor()
classifier.train(x_train,y_train)
ks = range(1 , 10)
pre = []
num_correct = []
accuracy = []
for k in ks:
pr = classifier.predict(x_test , k)
num = np.sum(pr == y_test)
pre.append(pr)
num_correct.append(num)
accuracy.append(float(num) / num_test)
plt.plot(ks , accuracy)
# Split the arrays into individual folds
X_train_folds = np.split(X_train, num_folds)
y_train_folds = np.split(y_train, num_folds)
# Dictionary holding the accuracies (list) for different values of k
k_to_accuracies = {}
# k-fold cross validation using fold i as validation, and all others as training
for choice in k_choices:
for i in range(num_folds):
# Partition training and test arrays
X_tr = np.vstack([X_train_folds[x] for x in range(num_folds) if x!=i])
y_tr = np.hstack([y_train_folds[x] for x in range(num_folds) if x!=i])
X_te = X_train_folds[i]
y_te = y_train_folds[i]
# Create kNN classifier instance
clf = KNearestNeighbor()
clf.train(X_tr, y_tr)
# Predict
pred = clf.predict(X_te, k=choice)
acc = float(np.sum(pred == y_te)) / y_te.shape[0]
print(f"k = {choice}, accuracy = {acc}")
if i == 0:
k_to_accuracies[choice] = [acc]
else:
k_to_accuracies[choice].append(acc)
# Plot results
for k in k_choices:
accs = k_to_accuracies[k]
plt.scatter([k] * len(accs), accs)
# Plot trend line with error bars corresponding to standard deviation
[np.mean(v) for k, v in sorted(k_to_accuracies.items())])
accuracies_std = np.array(
[np.std(v) for k, v in sorted(k_to_accuracies.items())])
plt.errorbar(k_choices, accuracies_mean, yerr=accuracies_std)
plt.title('Cross-validation on k')
plt.xlabel('k')
plt.ylabel('Cross-validation accuracy')
plt.show()
best_k = 1
classifier.train(X_train, y_train)
y_test_pred = classifier.predict(X_test, k=best_k)
determine_accuracy(y_test_pred, y_test, num_test)
X_train, y_train, X_test, y_test, num_test = load_data()
raw_input('Any key to continue...')
classifier = KNearestNeighbor()
compute_distances(X_train, y_train, X_test, y_test, num_test)
print '# 1'
print classifier.y_train
print classifier.y_train.shape
raw_input('Any key to continue...')
no_loop_time = time_function(classifier.compute_distances_no_loops, X_test)
print 'No loop version took %f seconds' % no_loop_time
raw_input('Any key to continue...')
cross_validation(X_train, y_train)
