AIにアミバとトキを学習させよう④

from keras.models import Sequential
from keras.layers import Conv2D, MaxPooling2D
from keras.layers import Activation, Dropout, Flatten, Dense
from keras.optimizers import RMSprop # TensorFlow1系
from keras.utils import np_utils
import keras
import numpy as np

classes = ["amiba", "toki"]
num_classes = len(classes)
image_size = 64

def load_data():
    X_train, X_test, y_train, y_test = np.load("./amiba_toki.npy", allow_pickle=True)
    # 入力データの各画素値を0-1の範囲で正規化(学習コストを下げるため)
    X_train = X_train.astype("float") / 255
    X_test  = X_test.astype("float") / 255

    y_train = np_utils.to_categorical(y_train, num_classes)
    y_test  = np_utils.to_categorical(y_test, num_classes)

    return X_train, y_train, X_test, y_test

def train(X, y, X_test, y_test):
    model = Sequential()


    model.add(Conv2D(32,(3,3), padding='same',input_shape=X.shape[1:]))
    model.add(Activation('relu'))
    model.add(Conv2D(32,(3,3)))
    model.add(Activation('relu'))
    model.add(MaxPooling2D(pool_size=(2,2)))
    model.add(Dropout(0.1))

    model.add(Conv2D(64,(3,3), padding='same'))
    model.add(Activation('relu'))
    model.add(Conv2D(64,(3,3)))
    model.add(Activation('relu'))
    model.add(MaxPooling2D(pool_size=(2,2)))
    model.add(Dropout(0.25))

    model.add(Flatten())
    model.add(Dense(512))
    model.add(Activation('relu'))
    model.add(Dropout(0.45))
    model.add(Dense(2)) #2クラス分類指定
    model.add(Activation('softmax'))

    
    opt = RMSprop(lr=0.00005, decay=1e-6)
    model.compile(loss='categorical_crossentropy',optimizer=opt,metrics=['accuracy'])
    model.fit(X, y, batch_size=28, epochs=40)
    model.save('./cnn.h5')#h5ファイル形式でcnnモデルを保存

    return model

def main():
    # データの読み込み
    X_train, y_train, X_test, y_test = load_data()

    # モデルの学習
    model = train(X_train, y_train, X_test, y_test)

main()