def nn(x, t, batch_size, epochs, feature=None, validation=None):
"""
簡単なニューラルネットワークのモデルを作成する関数
Parameters
----------
x : ndarray
学習用データ
t : ndarray
教師データ
batch_size : int
バッチサイズ
eopchs : int
エポック数
feature : int
Feature Scalingの選択
"""
data = {}
data['x'] = x
data['t'] = t
# データの仕分け
#data['x'], data['t'], data['val'] = __sorting__(x)
# データの前処理
for i in data.keys():
data[i] = __feature__(data[i], feature) # 標準化と正規化
data[i] = __shuffle__(data[i], 0) # データシャッフル
# データのシャッフル
#x, t = __shuffle__(x, t, 0)
#data['x'], data['t'] = __shuffle__(data['x'], data['t'], 0)
model = Sequential()
model.add(Input(input_shape=x.shape[1]))
#model.add(Dense(50, activation='relu', weight_initializer='relu'))
#model.add(Dense(50, activation='relu', weight_initializer='relu'))
model.add(Dense(50, activation='sigmoid', weight_initializer='sigmoid'))
model.add(Dense(50, activation='sigmoid', weight_initializer='sigmoid'))
#model.add(Dense(t.shape[1], activation='softmax'))
#model.compile(loss='cross_entropy_error')
model.add(Dense(t.shape[1], activation='liner'))
model.compile(loss='mean_squared_error')
#history=model.fit(x, t, batch_size=batch_size, epochs=epochs, validation=validation)
history = model.fit(data['x'],
data['t'],
batch_size=batch_size,
epochs=epochs,
validation=validation)
# lossグラフ
loss = history['loss_ave']
val_loss = history['val_loss']
nb_epoch = len(loss)
plt.plot(range(nb_epoch), loss, marker='.', label='loss')
plt.plot(range(nb_epoch), val_loss, marker='.', label='val_loss')
plt.legend(loc='best', fontsize=10)
plt.grid()
plt.xlabel('epoch')
plt.ylabel('loss')
plt.show()
def nn(x, t, batch_size, epochs, feature=None, validation=None):
"""
簡単なニューラルネットワークのモデルを作成する関数
Parameters
----------
x : ndarray
学習用データ
t : ndarray
教師データ
batch_size : int
バッチサイズ
eopchs : int
エポック数
feature : int
Feature Scalingの選択
"""
model = Sequential()
model.add(Input(input_shape=x.shape[1]))
model.add(Dense(50, activation='relu', weight_initializer='relu'))
model.add(Dense(50, activation='relu', weight_initializer='relu'))
#model.add(Dense(50, activation='sigmoid', weight_initializer='sigmoid'))
#model.add(Dense(50, activation='sigmoid', weight_initializer='sigmoid'))
#model.add(Dense(t.shape[1], activation='softmax'))
#model.compile(loss='cross_entropy_error')
model.add(Dense(t.shape[1], activation='liner'))
model.compile(loss='mean_squared_error')
#history=model.fit(x, t, batch_size=batch_size, epochs=epochs, validation=validation)
history = model.fit(x,
t,
batch_size=batch_size,
epochs=epochs,
validation=validation)
# lossグラフ
loss = history['loss_ave']
val_loss = history['val_loss']
train_acc = history['train_acc']
nb_epoch = len(loss)
plt.plot(range(nb_epoch), loss, marker='.', label='loss')
plt.plot(range(nb_epoch), val_loss, marker='.', label='val_loss')
plt.plot(range(nb_epoch), train_acc, marker='.', label='train_acc')
plt.legend(loc='best', fontsize=10)
plt.grid()
plt.xlabel('epoch')
plt.ylabel('loss')
plt.show()
def nn(x, t, batch_size, epochs, feature):
"""
簡単なニューラルネットワークのモデルを作成する関数
Parameters
----------
x : ndarray
学習用データ
t : ndarray
教師データ
batch_size : int
バッチサイズ
eopchs : int
エポック数
feature : int
Feature Scalingの選択
"""
# 標準化
if (feature == 0 or feature == 2):
x = data_std(x)
t = data_std(t)
if (feature == 1 or feature == 2):
x = data_nom(x)
t = data_nom(t)
model = Sequential()
model.add(Input(input_shape=x.shape[1]))
#model.add(Dense(50, activation='relu', weight_initializer='relu'))
#model.add(Dense(50, activation='relu', weight_initializer='relu'))
model.add(Dense(50, activation='sigmoid', weight_initializer='sigmoid'))
model.add(Dense(50, activation='sigmoid', weight_initializer='sigmoid'))
model.add(Dense(t.shape[1], activation='liner'))
model.compile(loss='mean_squared_error')
history = model.fit(x, t, batch_size=batch_size, epochs=epochs)
# lossグラフ
loss = history['loss_ave']
nb_epoch = len(loss)
plt.plot(range(nb_epoch), loss, marker='.', label='loss')
plt.legend(loc='best', fontsize=10)
plt.grid()
plt.xlabel('epoch')
plt.ylabel('loss')
plt.show()
def LayerAdd_click(layerName, Node, weight=None, bias=None, activation=None):
""" NetWorkにレイヤを追加していく関数 """
Node = int(Node)
if layerName == 'input':
model.add(Input(input_shape=Node))
elif layerName == 'Dense':
model.add(Dense(Node, activation, weight, bias))
model.printParams()
def LayerAdd_click(self, layerName, Node, weight=None, bias=None, activation=None):
""" NetWorkにレイヤを追加していく関数 """
Node = int(Node)
if layerName == 'input':
self.model.add(Input(input_shape=Node))
elif layerName == 'Dense':
self.model.add(Dense(Node, activation, weight, bias))
print(layerName + ':' + str(Node))
return layerName
def nn(x, t, batch_size, epochs, feature=None, validation=None):
"""
簡単なニューラルネットワークのモデルを作成する関数
Parameters
----------
x : ndarray
学習用データ
t : ndarray
教師データ
batch_size : int
バッチサイズ
eopchs : int
エポック数
feature : int
Feature Scalingの選択
"""
#-------------------------------
# DataFeature
#-------------------------------
if feature != None:
x = Datafeature(x, feature)
t = Datafeature(t, feature)
#-------------------------------
# Validation
#-------------------------------
if validation != None: # バリデーションが最初からセットされているとき
x_val = validation[0]
t_val = validation[1]
else:
x = __shuffle__(x)
t = __shuffle__(t)
x_val, x = __sorting__(x, 100)
t_val, t = __sorting__(t, 100)
# 学習曲線を可視化するコールバックを用意する
higher_better_metrics = ['r2']
visualize_cb = LearningVisualizationCallback(higher_better_metrics)
callbacks = [
visualize_cb,
]
model = Sequential()
model.add(Input(input_shape=x.shape[1]))
model.add(Dense(50, activation='relu', weight_initializer='relu'))
model.add(Dense(50, activation='relu', weight_initializer='relu'))
#model.add(Dense(50, activation='sigmoid', weight_initializer='sigmoid'))
#model.add(Dense(50, activation='sigmoid', weight_initializer='sigmoid'))
#model.add(Dense(t.shape[1], activation='softmax'))
#model.compile(loss='cross_entropy_error')
model.add(Dense(t.shape[1], activation = 'liner'))
model.compile(loss='mean_squared_error', metrics = ['r2', 'rsme'])
#history=model.fit(x, t, batch_size=batch_size, epochs=epochs, validation=validation)
history = model.fit(x, t, batch_size=batch_size,
epochs=epochs, validation=(x_val, t_val), callbacks=callbacks)
# lossグラフ
loss = history['loss_ave']
val_loss = history['val_loss']
nb_epoch = len(loss)
plt.plot(range(nb_epoch), loss, marker = '.', label = 'loss')
plt.plot(range(nb_epoch), val_loss, marker='.', label='val_loss')
plt.legend(loc = 'best', fontsize = 10)
plt.grid()
plt.xlabel('epoch')
plt.ylabel('loss')
plt.show()
#正解データの加工
t_train = keras.utils.to_categorical(t_train, 10) # one_hot_labelに変換
t_test = keras.utils.to_categorical(t_test, 10)
"""
# 学習曲線を可視化するコールバックを用意する
higher_better_metrics = ['r2']
visualize_cb = LearningVisualizationCallback(higher_better_metrics)
callbacks = [
visualize_cb,
]
model = Sequential()
model.add(Input(input_shape=x_train.shape[1]))
model.add(Dense(50, activation='relu', weight_initializer='relu'))
model.add(Dense(50, activation='relu', weight_initializer='relu'))
#model.add(Dense(50, activation='sigmoid', weight_initializer='sigmoid'))
#model.add(Dense(50, activation='sigmoid', weight_initializer='sigmoid'))
#model.add(Dense(t.shape[1], activation='softmax'))
#model.compile(loss='cross_entropy_error')
model.add(Dense(t_train.shape[1], activation='liner'))
model.compile(loss='mean_squared_error', optimizer='sgd', metrics=['r2'])
epochs = 10
batch_size = 128
history = model.fit(x_train,
t_train,
batch_size=batch_size,
epochs=epochs,
#正解データの加工
t_train = keras.utils.to_categorical(t_train, 10) # one_hot_labelに変換
t_test = keras.utils.to_categorical(t_test, 10)
"""
# 学習曲線を可視化するコールバックを用意する
higher_better_metrics = ["r2"]
visualize_cb = LearningVisualizationCallback(higher_better_metrics)
callbacks = [
visualize_cb,
]
model = Sequential()
model.add(Input(input_shape=x_train.shape[1]))
model.add(Dense(50, activation="relu", weight_initializer="relu"))
model.add(Dense(50, activation="relu", weight_initializer="relu"))
# model.add(Dense(50, activation='sigmoid', weight_initializer='sigmoid'))
# model.add(Dense(50, activation='sigmoid', weight_initializer='sigmoid'))
# model.add(Dense(t.shape[1], activation='softmax'))
# model.compile(loss='cross_entropy_error')
model.add(Dense(t_train.shape[1], activation="liner"))
model.compile(loss="mean_squared_error", optimizer="sgd", metrics=["r2"])
epochs = 50
batch_size = 128
history = model.fit(
x_train,
t_train,
batch_size=batch_size,
