class simple_cnn_model(object):
def __init__(self, epochs, batch_size, lr):
self.epochs = epochs
self.batch_size = batch_size
self.lr = lr
def load_data(self):
# load data from cifar100 folder
(x_train, y_train), (x_test, y_test) = cifar100(1211506319)
return x_train, y_train, x_test, y_test
def train_model(self, layers, loss_metrics, x_train, y_train):
# build model
self.model = Sequential(layers, loss_metrics)
# train the model
loss = self.model.fit(x_train,
y_train,
self.epochs,
self.lr,
self.batch_size,
print_output=True)
avg_loss = np.mean(np.reshape(loss, (self.epochs, -1)), axis=1)
return avg_loss
def test_model(self, x_test, y_test):
# make a prediction
pred_result = self.model.predict(x_test)
accuracy = np.mean(pred_result == y_test)
return accuracy
# Please make sure that cifar-100-python is present in the same folder as dataset.py
(x_train, y_train), (x_test, y_test) = cifar100(1212356299)
from layers import (FullLayer, ReluLayer, SoftMaxLayer, CrossEntropyLayer,
Sequential)
model = Sequential(layers=(FullLayer(3072,
500), ReluLayer(), FullLayer(500, 4),
SoftMaxLayer()),
loss=CrossEntropyLayer())
lr_accuracies = np.zeros((3, ))
loss1 = model.fit(x_train, y_train, lr=0.01, epochs=15)
y_predict = model.predict(x_test)
count = 0
for i in range(np.size(y_test)):
if y_predict[i] == y_test[i]:
count += 1
lr_accuracies[0] = (100.0 * count) / np.shape(y_predict)[0]
loss2 = model.fit(x_train, y_train, lr=0.1, epochs=15)
y_predict = model.predict(x_test)
count = 0
for i in range(np.size(y_test)):
if y_predict[i] == y_test[i]:
for j in range(len(lr_vals)):
train_loss, test_loss = model.fit(x_train,
y_train,
x_test,
y_test,
epochs=8,
lr=lr_vals[j],
batch_size=128)
losses_train.append(train_loss)
losses_test.append(test_loss)
print("--- RUN TIME %s seconds --- \n" % (time.clock() - start_time))
# print("Losses are: ",losses)
plt.plot(range(1, 9), losses_train[j], label='Train loss')
plt.plot(range(1, 9), losses_test[j], label='Test Loss')
predictions_train = model.predict(x_train)
y_train_class = np.array(
[np.argmax(y_train[i]) for i in range(y_train.shape[0])])
print("Train accuracy is: ", np.mean(predictions_train == y_train_class))
predictions_test = model.predict(x_test)
y_test_class = np.array(
[np.argmax(y_test[i]) for i in range(y_test.shape[0])])
test_acc[j] = np.mean(predictions_test == y_test_class)
print("Test accuracy is: ", np.mean(predictions_test == y_test_class))
plt.xlabel('Epoch Number --->')
plt.ylabel('Loss value --->')
plt.title('Loss v/s epochs curve')
plt.legend()
plt.show()
finalloss = []
testloss = []
model = Sequential(layers=(ConvLayer(3, 16, 3), ReluLayer(), MaxPoolLayer(2),
ConvLayer(16, 32, 3), ReluLayer(), MaxPoolLayer(2),
FlattenLayer(), FullLayer(8 * 8 * 32,
4), SoftMaxLayer()),
loss=CrossEntropyLayer())
train_loss, valid_loss = model.fit(x_train,
y_train,
x_test,
y_test,
epochs=15,
lr=0.1,
batch_size=128)
y_pred = model.predict(x_test)
accuracy = (np.mean(y_test == onehot(y_pred)))
print('Accuracy: %.2f' % accuracy)
#np.append(test_accuracy, accuracy)
plt.plot(range(len(train_loss)), train_loss, label='Training loss')
plt.plot(range(len(valid_loss)), valid_loss, label='Testing loss')
plt.xlabel('epochs')
plt.ylabel('Training & Testing loss')
plt.title('Training & Testing loss Vs Epochs for the CIFAR100 data set')
plt.legend()
plt.show()
# plt.plot(lr, test_accuracy)
# plt.title('Test Accuracy Vs Learning Rate for Modified NN')
# plt.xlabel('Learning rate')
# plt.ylabel('Testing Accuracy')
# plt.show()