参考：https://www.kaggle.com/toregil/welcome-to-deep-learning-cnn-99
import pandas as pd
import numpy as np
from sklearn.model_selection import train_test_split
from sklearn.metrics import confusion_matrix
from keras.utils.np_utils import to_categorical # convert to one-hot-encoding
from keras.models import Sequential
from keras.layers import Dense, Dropout, Flatten, Conv2D, MaxPool2D, BatchNormalization
from keras.optimizers import Adam
from keras.preprocessing.image import ImageDataGenerator
from keras.callbacks import LearningRateScheduler
train_data=pd.read_csv('train.csv')
test_data=pd.read_csv('test.csv')
train_data.shape,test_data.shape
train_data.head()
labels=train_data['label'].values
images=train_data.drop(['label'],axis=1).values
labels.shape,images.shape
x_train, x_val, y_train, y_val = train_test_split(
images, labels, test_size=0.04)
x_train.shape,x_val.shape
x_train = x_train.reshape(-1, 28, 28, 1)
x_val = x_val.reshape(-1, 28, 28, 1)
x_train = x_train.astype("float32")/255.
x_val = x_val.astype("float32")/255.
y_train = to_categorical(y_train)
y_val = to_categorical(y_val)
#example:
print(y_train[0])
model = Sequential()
model.add(Conv2D(filters = 16, kernel_size = (3, 3), activation='relu',
input_shape = (28, 28, 1)))
model.add(BatchNormalization())
model.add(Conv2D(filters = 16, kernel_size = (3, 3), activation='relu'))
model.add(BatchNormalization())
model.add(MaxPool2D(strides=(2,2)))
model.add(Dropout(0.25))
model.add(Conv2D(filters = 32, kernel_size = (3, 3), activation='relu'))
model.add(BatchNormalization())
model.add(Conv2D(filters = 32, kernel_size = (3, 3), activation='relu'))
model.add(BatchNormalization())
model.add(MaxPool2D(strides=(2,2)))
model.add(Dropout(0.25))
model.add(Flatten())
model.add(Dense(512, activation='relu'))
model.add(Dropout(0.25))
model.add(Dense(1024, activation='relu'))
model.add(Dropout(0.5))
model.add(Dense(10, activation='softmax'))
datagen = ImageDataGenerator(zoom_range = 0.1,
height_shift_range = 0.1,
width_shift_range = 0.1,
rotation_range = 10)
model.compile(loss='categorical_crossentropy', optimizer = Adam(lr=1e-4), metrics=["accuracy"])
annealer = LearningRateScheduler(lambda x: 1e-3 * 0.9 ** x)
hist = model.fit_generator(datagen.flow(x_train, y_train, batch_size=16),
steps_per_epoch=500,
epochs=100, #Increase this when not on Kaggle kernel
verbose=2, #1 for ETA, 0 for silent
validation_data=(x_val[:400,:], y_val[:400,:]), #For speed
callbacks=[annealer])
final_loss, final_acc = model.evaluate(x_val, y_val, verbose=0)
print("Final loss: {0:.4f}, final accuracy: {1:.4f}".format(final_loss, final_acc))
y_hat = model.predict(x_val)
y_pred = np.argmax(y_hat, axis=1)
y_true = np.argmax(y_val, axis=1)
cm = confusion_matrix(y_true, y_pred)
print(cm)
test_data=test_data.values.astype('float32')
x_test = test_data.reshape(-1, 28, 28, 1)/255.
y_hat = model.predict(x_test, batch_size=64)
y_pred = np.argmax(y_hat,axis=1)
with open('submission_5.csv', 'w') as f :
f.write('ImageId,Label\n')
for i in range(len(y_pred)) :
f.write("".join([str(i+1),',',str(y_pred[i]),'\n']))