def predict(network, x):
w1, w2, w3 = network['W1'], network['W2'], network['W3']
b1, b2, b3 = network['b1'], network['b2'], network['b3']
a1 = np.dot(x, w1) + b1
z1 = sigmoid(a1)
a2 = np.dot(z1, w2) + b2
z2 = sigmoid(a2)
a3 = np.dot(z2, w3) + b3
y = softmax(a3)
return y
def loss(self, x, t):
z = self.predict(x)
y = softmax(z)
loss = cross_entropy_error(y, t)
return loss
def loss(self, x, t): z = self.predict(x) y = softmax(z) loss = cross_entropy_error(y, t) return loss
def forward(self, x):
self.out = softmax(x)
return self.out
def loss(self, x, t):
z = self.predict(x)
y = softmax(z)
loss = cross_entropy_error(y, t) # t is one-hot encoded answer label
return loss
def loss(self, x, t):
z = self.predict(x) # 演算
y = softmax(z) # 活性化関数としてソフトマックス関数を用いて確率を求める
loss = cross_entropy_error(y, t) # 損失関数の平均値
return loss
def forward(self, x, t):
self.t = t
self.y = softmax(x)
self.loss = cross_entropy_error(self.y, self.t)
return self.loss
def loss(self, x, t): #t:正解ラベル #損失関数の値を求めるメソッド
z = self.predict(x)
y = softmax(z)
loss = cross_entropy_error(y, t)
return loss
def forward(self, x, d):
self.d = d
self.y = functions.softmax(x)
self.loss = functions.cross_entropy_error(self.d, self.y)
return self.loss
input_silhouette = util.resizeImage(input_silhouette, data_size, rect,
True)
if flip:
input_silhouette = cv2.flip(input_silhouette, 1)
input_silhouette_path = os.path.join(data_path, 'input_silhouette')
input_silhouette_path = os.path.join(
input_silhouette_path, file_name + '_v' + extractor.version + '.png')
cv2.imwrite(input_silhouette_path, input_silhouette)
input_silhouette = np.where(input_silhouette > 0, 1, 0)
input_silhouette = input_silhouette.flatten()
result = estimator.predict(input_silhouette)
result = functions.softmax(result[0])
index = np.argmax(result)
_label = label.copy()
final_result = []
for i in range(3):
for j in range(i + 1, len(result)):
if result[i] < result[j]:
temp = result[i]
temp_label = _label[i]
result[i] = result[j]
result[j] = temp
_label[i] = _label[j]
_label[j] = temp_label
res = {'keyword': _label[i], 'probability': round(result[i], 2)}
def forward(self, x, t):
self.t = t
self.y = functions.softmax(x)
self.loss = loss_functions.cross_entropy_error(self.y, self.t)
return self.loss
def predict(self, x):
a = np.dot(x, self.W) # 신경망 각 층의 계산은 행렬의 내젹으로 처리할 수 있다. a = xW
y = functions.softmax(a) # 출력층 활성화함수로 소프트맥스함수 사용
return y
