def restoreNeuralNetwork(self):
fname, _ = QFileDialog.getOpenFileName(self, 'Open Network File', '.', 'Network File(*.nf)')
if not fname:
# self.logStatus('打开神经网络文件{}失败!'.format(fname), 'red', 'E')
return
# clear previous plots
self.clearPlots()
training_x, training_y = DataHandler.split_xy(self.training_data)
self.network = NeuralNetwork(training_x, training_y)
try:
self.network.load(fname)
except ShapeError as e:
self.logStatus('加载神经网络文件{}失败!'.format(fname), 'red', 'E')
QMessageBox.warning(self, '警告', '加载神经网络文件失败, 请检查文件格式是否正确!')
return
self.logStatus('神经网络文件{}加载成功'.format(fname))
self.predictDatakAct.setEnabled(True)
self.logStatus('请执行数据预测', '#FF8C00', 'T')
def update_figure(self):
if GlobalMap().get('userdefined') == 'Null_': #检测从自定义对话框传递过来的数据
return
if not self.begin: #防止重绘
return
self.begin = 0
GlobalMap().set(time_point=self.times, )
self.times += 1
if self.times == 2:
self.begin = 1
self.temp0 = NN.main([], [],
GlobalMap().get('userdefined')) #传入BP网络预测
return
else:
temp01 = self.temp0
self.temp0 = NN.main([], [], temp01)
imgP = imread('core/map.png', )
self.axes.imshow(imgP,
extent=[112.1495, 112.288, 30.308, 30.391],
aspect='auto')
x = np.array(self.temp0[0])
y = np.array(self.temp0[1])
z = gaussian_kde(self.temp0)(self.temp0) #生成密度颜色
self.axes.hold(True)
self.axes.scatter(x, y, marker='.', s=100, c=z, edgecolors='')
self.axes.grid(True)
self.axes.set_xlim(112.150, 112.284)
self.axes.set_ylim(30.309, 30.385)
self.axes.set(xlabel="longitude", ylabel="latitude")
self.axes.set_title('The ' + str(self.times) +
':00 ’s distribution prediction')
self.draw()
self.begin = 1
self.axes.hold(False)
if self.times >= 2:
self.pause()
def run(self):
if self.working == True:
times = GlobalMap().get('time_point')
GlobalMap().set(progress_num=0)
lists = []
for i in range(24):
lists += [i]
GlobalMap().set(progress_num=2)
if exists('db/temp/20180515.csv') and exists(
'db/temp/20180516.csv'):
temp1 = DataFrame(read_csv('db/temp/20180515.csv', header=0))
temp2 = DataFrame(read_csv('db/temp/20180516.csv', header=0))
else:
temp1 = dataclear(20180515).run()
temp2 = dataclear(20180516).run()
GlobalMap().set(progress_num=5)
temp11 = temp1[temp1['time'] == lists[times - 1]]
temp12 = temp1[temp1['time'] == lists[times]]
temp1112 = merge(temp11, temp12, how='outer', on='id').dropna()
temp21 = temp2[temp2['time'] == lists[times - 1]]
temp22 = temp2[temp2['time'] == lists[times]]
temp2122 = merge(temp21, temp22, how='outer', on='id').dropna()
temp000 = np.vstack((temp1112['x_x'], temp1112['y_x']))
temp001 = np.vstack((temp1112['x_y'], temp1112['y_y']))
temp020 = np.vstack((temp2122['x_x'], temp2122['y_x']))
#temp021 = np.vstack((temp2122['x_y'],temp2122['y_y']))
cells_array = np.array(NN.main(temp000, temp001, temp020))
np.save(GlobalMap().get('temp_path') + '/test' + str(times),
cells_array)
GlobalMap().set(progress_num=100, running_global=0)
self.changed2_signal.emit(1)
gpu_options=tf.GPUOptions(
allocator_type='BFC',
# allow_growth=True, # it will cause fragmentation.
per_process_gpu_memory_fraction=0.1)))
graph_def = graph.as_graph_def()
with sess.graph.as_default():
sess.run(tf.global_variables_initializer())
tf.keras.backend.set_session(session=sess)
# with tf.gfile.GFile(COMPILE_MODEL_PATH.replace('.pb', '_{}.pb'.format(int(0.95 * 10000))), "rb") as f:
# graph_def_file = f.read()
# graph_def.ParseFromString(graph_def_file)
# print('{}.meta'.format(tf_checkpoint))
model = NeuralNetwork(model_conf, RunMode.Predict, model_conf.neu_cnn,
model_conf.neu_recurrent)
model.build_graph()
saver = tf.train.Saver(var_list=tf.global_variables())
"""从项目中加载最后一次训练的网络参数"""
saver.restore(sess,
tf.train.latest_checkpoint(model_conf.model_root_path))
# _ = tf.import_graph_def(graph_def, name="")
"""定义操作符"""
dense_decoded_op = sess.graph.get_tensor_by_name("dense_decoded:0")
x_op = sess.graph.get_tensor_by_name('input:0')
"""固定网络"""
sess.graph.finalize()
true_count = 0
def testing(self, image_dir, limit=None):
graph = tf.Graph()
sess = tf.Session(
graph=graph,
config=tf.ConfigProto(
# allow_soft_placement=True,
# log_device_placement=True,
gpu_options=tf.GPUOptions(
allocator_type='BFC',
# allow_growth=True, # it will cause fragmentation.
per_process_gpu_memory_fraction=0.1)))
with sess.graph.as_default():
sess.run(tf.global_variables_initializer())
tf.keras.backend.set_session(session=sess)
model = NeuralNetwork(self.model_conf, RunMode.Predict,
self.model_conf.neu_cnn,
self.model_conf.neu_recurrent)
model.build_graph()
saver = tf.train.Saver(var_list=tf.global_variables())
"""从项目中加载最后一次训练的网络参数"""
saver.restore(
sess,
tf.train.latest_checkpoint(self.model_conf.model_root_path))
# _ = tf.import_graph_def(graph_def, name="")
"""定义操作符"""
dense_decoded_op = sess.graph.get_tensor_by_name("dense_decoded:0")
x_op = sess.graph.get_tensor_by_name('input:0')
"""固定网络"""
sess.graph.finalize()
true_count = 0
false_count = 0
"""
以下为根据路径调用预测函数输出结果的demo
"""
# Fill in your own sample path
dir_list = os.listdir(image_dir)
random.shuffle(dir_list)
lines = []
for i, p in enumerate(dir_list):
n = os.path.join(image_dir, p)
if limit and i > limit:
break
with open(n, "rb") as f:
b = f.read()
batch = self.get_image_batch(b)
st = time.time()
predict_text = self.predict_func(
batch,
sess,
dense_decoded_op,
x_op,
)
et = time.time()
# t = p.split(".")[0].lower() == predict_text.lower()
# csv_output = "{},{}".format(p.split(".")[0], predict_text)
# lines.append(csv_output)
# print(csv_output)
is_mark = '_' in p
if is_mark:
if 'LOWER' in self.model_conf.category_param:
label = p.split("_")[0].lower()
t = label == predict_text.lower()
elif 'UPPER' in self.model_conf.category_param:
label = p.split("_")[0].upper()
t = label == predict_text.upper()
else:
label = p.split("_")[0]
t = label == predict_text
# Used to verify test sets
if t:
true_count += 1
else:
false_count += 1
print(i, p, label, predict_text, t,
true_count / (true_count + false_count),
(et - st) * 1000)
else:
print(i, p, predict_text,
true_count / (true_count + false_count),
(et - st) * 1000)
X = np.transpose(np.asarray(X))
Y = np.transpose(np.asarray(Y))
Y = np.expand_dims(Y, axis=0)
X = X / 255 #normalization
lr = 0.001
epochs = 5000
amount_data = X.shape[0]
hidden_neurons = 64
output_neurons = 1
nn = NeuralNetwork(lr=lr,
epochs=epochs,
amount_data=amount_data,
hidden_neurons=hidden_neurons,
output_neurons=output_neurons)
X_train, X_validate, Y_train, Y_validate = nn.split_data(X, Y, 0.3)
loss_train, acc_train, loss_validate, acc_validate =\
nn.train(
X_train = X_train,
Y_train = Y_train,
X_validate = X_validate,
Y_validate = Y_validate)
graph = Graph(file="nn1",
amount_data=amount_data,
hidden_neurons=hidden_neurons,
))
)
graph_def = graph.as_graph_def()
with sess.graph.as_default():
sess.run(tf.global_variables_initializer())
tf.keras.backend.set_session(session=sess)
# with tf.gfile.GFile(COMPILE_MODEL_PATH.replace('.pb', '_{}.pb'.format(int(0.95 * 10000))), "rb") as f:
# graph_def_file = f.read()
# graph_def.ParseFromString(graph_def_file)
# print('{}.meta'.format(tf_checkpoint))
model = NeuralNetwork(
model_conf,
RunMode.Predict,
model_conf.neu_cnn,
NETWORK_MAP[model_conf.neu_recurrent] if model_conf.neu_recurrent else None
)
model.build_graph()
saver = tf.train.Saver(var_list=tf.global_variables())
"""从项目中加载最后一次训练的网络参数"""
saver.restore(sess, tf.train.latest_checkpoint(model_conf.model_root_path))
# _ = tf.import_graph_def(graph_def, name="")
"""定义操作符"""
dense_decoded_op = sess.graph.get_tensor_by_name("dense_decoded:0")
x_op = sess.graph.get_tensor_by_name('input:0')
"""固定网络"""
def run(self):
if self.working == True:
times = GlobalMap().get('time_point')
GlobalMap().set(progress_num=0)
lists = []
for i in range(24):
lists += [i]
GlobalMap().set(progress_num=2)
if exists('db/temp/20180515.csv') and exists(
'db/temp/20180516.csv'):
temp1 = DataFrame(read_csv('db/temp/20180515.csv', header=0))
temp2 = DataFrame(read_csv('db/temp/20180516.csv', header=0))
else:
temp1 = dataclear(20180515).run()
temp2 = dataclear(20180516).run()
GlobalMap().set(progress_num=5)
temp11 = temp1[temp1['time'] == lists[times - 1]]
temp12 = temp1[temp1['time'] == lists[times]]
temp1112 = merge(temp11, temp12, how='outer', on='id').dropna()
'''begin 另一种数据处理方式'''
# temp1_x = DataFrame(read_csv('db/temp/20180515-x.csv' ,header=0))
# temp1_y = DataFrame(read_csv('db/temp/20180515-y.csv' ,header=0))
# temp2_x = DataFrame(read_csv('db/temp/20180516-x.csv' ,header=0))
# temp2_y = DataFrame(read_csv('db/temp/20180516-y.csv' ,header=0))
#
# temp11_x = temp1_x[str(lists[times])]
# temp11_y = temp1_y[str(lists[times])]
# temp12_x = temp2_x[str(lists[times])]
# temp12_y = temp2_y[str(lists[times])]
#
# temp21_x = temp1_x[str(lists[times+1])]
# temp21_y = temp1_y[str(lists[times+1])]
# temp22_x = temp2_x[str(lists[times+1])]
# temp22_y = temp2_y[str(lists[times+1])]
#
# cc01 = np.vstack((temp11_x, temp11_y))
# cc02 = np.vstack((temp21_x, temp21_y))
# cc03 = np.vstack((temp12_x, temp12_y))
# cc04 = np.vstack((temp22_x, temp22_y))
'''end'''
temp21 = temp2[temp2['time'] == lists[times - 1]]
temp22 = temp2[temp2['time'] == lists[times]]
temp2122 = merge(temp21, temp22, how='outer', on='id').dropna()
temp000 = np.vstack((temp1112['x_x'], temp1112['y_x']))
temp001 = np.vstack((temp1112['x_y'], temp1112['y_y']))
temp020 = np.vstack((temp2122['x_x'], temp2122['y_x']))
temp021 = np.vstack((temp2122['x_y'], temp2122['y_y']))
#用20180515 T 和 T+1 时刻的训练 来预测 20180516 T 时刻 在 T+1 时刻的分布情况
cells_array = np.array(NN.main(temp000, temp001, temp020))
#cells_array = np.array(NN.main(cc01, cc02, cc03 ))
np.save(GlobalMap().get('temp_path') + '/test' + str(times),
cells_array)
GlobalMap().set(progress_num=100)
self.changed_signal.emit(cells_array, temp021)
def runNetworkTraining(self):
if self.network is not None:
ans = QMessageBox.question(self, '警告',
'系统中已存在训练好的神经网络，请问您需要重新训练神经网络吗?')
if ans == QMessageBox.No:
return
# clear previous plots
self.clearPlots()
self.logStatus("正在初始化神经网络结构...", 'blue', 'I')
training_x, training_y = DataHandler.split_xy(self.training_data)
# retrieve settings
epoch0 = 2000
tol0 = 0.1
retry_num = 3
h0size = 4
h1size = 4
if self.networkSettings:
epoch0 = self.networkSettings['epoch']
tol0 = self.networkSettings['tol']
retry_num = self.networkSettings['retry']
h0size = self.networkSettings['h0size']
h1size = self.networkSettings['h1size']
self.logStatus("神经网络信息：layer1={}, layer2={}, epoch0={}, retry={}, tol0={}"
.format(h0size, h1size, epoch0, retry_num, tol0), 'blue', 'I')
net = [(training_x.shape[0], ''), (h0size, 'sigmoid'), (h1size, 'sigmoid'), (1, 'feedthrough')]
try:
self.network = NeuralNetwork(training_x, training_y, sizes=net)
except ShapeError as e:
self.logStatus('初始化神经网络结构失败!')
QMessageBox.warning(self, '警告', '初始化神经网络结构失败, 请重试!')
return
# training
mu0 = 0.1
beta = 10
retry = 0
self.logStatus("使用LM算法开始训练神经网络...", 'blue', 'I')
while retry < retry_num:
residual, mu, citer, msg = \
self.network.train(retry=retry, epoch=epoch0, mu0=mu0,
beta=beta, tol=tol0, cb=self.logStatus)
if residual is None:
if retry == (retry_num - 1):
self.logStatus("训练失败!".format(msg), 'red', 'E')
return
else:
self.logStatus("训练失败:{}, 重试中...".format(msg), '#FFA07A', 'I')
self.network.randomize_wb()
# continue
elif residual > tol0:
if retry == (retry_num - 1):
self.logStatus("训练失败!".format(msg), 'red', 'E')
return
else:
self.logStatus("训练失败: 运算未能收敛, 重试中...", '#FFA07A', 'I')
self.network.randomize_wb()
# continue
else:
self.logStatus("神经网络训练完成, 迭代次数={1}, 最终残差={0}"
.format(residual, citer+retry*epoch0), 'blue', 'I')
break
retry += 1
self.predictDatakAct.setEnabled(True)
self.saveNetworkAct.setEnabled(True)
self.logStatus('请执行数据预测', '#FF8C00', 'T')
class AppWindow(QMainWindow):
def __init__(self):
QMainWindow.__init__(self)
self.setAttribute(Qt.WA_DeleteOnClose)
self.setWindowTitle(progname)
self.init_ui()
self.network = None
self.isDataLoaded = False
self.training_data = None
self.test_data = None
self.predict_y = None
self.networkSettings = None
def init_ui(self):
self.init_menus()
self.init_toolbar(QSize(40, 40))
mainWidget = QWidget(self)
mainbox = QHBoxLayout(mainWidget)
# self.setLayout(mainbox)
topFrame = QFrame(mainWidget)
topFrame.setFrameShape(QFrame.StyledPanel)
btmFrame = QFrame(mainWidget)
btmFrame.setFrameShape(QFrame.StyledPanel)
splitter = QSplitter(Qt.Vertical)
splitter.addWidget(topFrame)
splitter.addWidget(btmFrame)
# logText 30%, Plot 70%
splitter.setStretchFactor(0, 4)
splitter.setStretchFactor(1, 1)
mainbox.addWidget(splitter)
self.init_plot_area(topFrame)
vboxLog = QVBoxLayout(btmFrame)
self.logTextEdit = QPlainTextEdit("")
self.logTextEdit.appendHtml("""<font size='4'>欢迎使用{}</font><p>""".format(progname))
self.logTextEdit.setReadOnly(True)
vboxLog.addWidget(self.logTextEdit)
mainWidget.setFocus()
self.setCentralWidget(mainWidget)
self.statusBar().showMessage("Ready")
self.setWindowIcon(QIcon('res/load_network.png'))
self.show()
def init_plot_area(self, parent):
hboxPlot = QHBoxLayout(parent)
# errplot_labels = {'t': u'供水温度预测误差', 'x': u'时间', 'y': u'误差百分比(%)'}
# predplot_labels = {'t': u'供水温度预测值', 'x': u'时间', 'y': u'供水温度(℃)'}
errplot_labels = {'t': 'Prediction Errors', 'x': 'Time', 'y': 'Error Percent(%)'}
predplot_labels = {'t': 'Predicted Temperature', 'x': 'Time', 'y': 'Temperature(℃)'}
self.errPlot = StaticMplotCanvas(parent, labels=errplot_labels)
self.predPlot = StaticMplotCanvas(parent, labels=predplot_labels)
hboxPlot.addWidget(self.errPlot)
hboxPlot.addWidget(self.predPlot)
def init_toolbar(self, iconSize):
# data file
self.loadDataAct = QAction(QIcon('res/load_data.png'), 'Import Training Data', self)
self.loadDataAct.setShortcut('Ctrl+L')
self.loadDataAct.triggered.connect(self.loadTrainingDataFile)
self.saveDataAct = QAction(QIcon('res/save_data.png'), 'Export Predicted Data', self)
self.saveDataAct.setShortcut('Ctrl+E')
self.saveDataAct.triggered.connect(self.savePredictDataToFile)
self.saveDataAct.setEnabled(False)
# network
self.loadNetworkAct = QAction(QIcon('res/load_network.png'), 'Load Trained Network', self)
self.loadNetworkAct.setShortcut('Ctrl+N')
self.loadNetworkAct.triggered.connect(self.restoreNeuralNetwork)
self.loadNetworkAct.setEnabled(False)
self.saveNetworkAct = QAction(QIcon('res/save_network.png'), 'Save Trained Network', self)
self.saveNetworkAct.setShortcut('Ctrl+S')
self.saveNetworkAct.triggered.connect(self.saveNeuralNetwork)
self.saveNetworkAct.setEnabled(False)
# run & predict
self.runTrainingAct = QAction(QIcon('res/train_network.png'), 'Train Network', self)
self.runTrainingAct.setShortcut('Ctrl+R')
self.runTrainingAct.triggered.connect(self.runNetworkTraining)
self.runTrainingAct.setEnabled(False)
self.predictDatakAct = QAction(QIcon('res/predict.png'), 'Predict Data', self)
self.predictDatakAct.setShortcut('Ctrl+P')
self.predictDatakAct.triggered.connect(self.predictData)
self.predictDatakAct.setEnabled(False)
# clear
self.resetAct = QAction(QIcon('res/clear.png'), 'Clear data and network', self)
self.resetAct.setEnabled(False)
self.resetAct.triggered.connect(self.clearDataAndNetwork)
dataToolbar = self.addToolBar('Data ToolBar')
dataToolbar.addAction(self.loadDataAct)
dataToolbar.addAction(self.saveDataAct)
dataToolbar.setIconSize(iconSize)
networkToolbar = self.addToolBar('Network ToolBar')
networkToolbar.addAction(self.loadNetworkAct)
networkToolbar.addAction(self.runTrainingAct)
networkToolbar.addAction(self.predictDatakAct)
networkToolbar.addAction(self.saveNetworkAct)
networkToolbar.setIconSize(iconSize)
resetToolbar = self.addToolBar('Reset ToolBar')
resetToolbar.addAction(self.resetAct)
resetToolbar.setIconSize(iconSize)
def init_menus(self):
# File
settingAct = QAction(QIcon('res/settings.png'), '设置', self)
settingAct.triggered.connect(self.showSettingDialog)
exitAct = QAction(QIcon('exit.png'), '退出', self)
exitAct.setShortcut('Ctrl+Q')
exitAct.triggered.connect(self.fileQuit)
menubar = self.menuBar()
fileMenu = menubar.addMenu('&File')
fileMenu.addAction(settingAct)
fileMenu.addSeparator()
fileMenu.addAction(exitAct)
# Help
helpMenu = QMenu('&Help', self)
self.menuBar().addMenu(helpMenu)
helpMenu.addAction('使用说明', self.usage)
helpMenu.addAction('关于', self.about)
@pyqtSlot()
def loadTrainingDataFile(self):
fname, _ = QFileDialog.getOpenFileName(self, 'Open Training Data', '.', 'Data File(*.csv)')
if not fname:
# self.logStatus("加载数据文件{}失败!".format(fname), 'red', 'E')
return
dl = DataHandler(fname)
self.training_data, self.test_data = dl.load(cb=self.logStatus)
self.isDataLoaded = True
self.runTrainingAct.setEnabled(True)
self.loadNetworkAct.setEnabled(True)
self.resetAct.setEnabled(True)
self.logStatus("加载数据文件{}成功".format(fname))
self.logStatus('请训练神经网络或者加载已经训练的神经网络模型', '#FF8C00', 'T')
@pyqtSlot()
def savePredictDataToFile(self):
if self.predict_y is None:
# self.logStatus('没有未保存的预测数据, 请先进行数据预测!', 'red', 'E')
return
fname, _ = QFileDialog.getSaveFileName(self, 'Save Predicted Data', '.', 'Data File(*.csv)')
if not fname:
self.logStatus('保存预测数据文件{}失败!'.format(fname), 'red', 'E')
return
test_x, _ = DataHandler.split_xy(self.test_data)
status = DataHandler.save(np.concatenate((test_x, self.predict_y), axis=0), fname)
if status:
self.logStatus('保存预测数据文件{}成功'.format(fname))
@pyqtSlot()
def restoreNeuralNetwork(self):
fname, _ = QFileDialog.getOpenFileName(self, 'Open Network File', '.', 'Network File(*.nf)')
if not fname:
# self.logStatus('打开神经网络文件{}失败!'.format(fname), 'red', 'E')
return
# clear previous plots
self.clearPlots()
training_x, training_y = DataHandler.split_xy(self.training_data)
self.network = NeuralNetwork(training_x, training_y)
try:
self.network.load(fname)
except ShapeError as e:
self.logStatus('加载神经网络文件{}失败!'.format(fname), 'red', 'E')
QMessageBox.warning(self, '警告', '加载神经网络文件失败, 请检查文件格式是否正确!')
return
self.logStatus('神经网络文件{}加载成功'.format(fname))
self.predictDatakAct.setEnabled(True)
self.logStatus('请执行数据预测', '#FF8C00', 'T')
@pyqtSlot()
def saveNeuralNetwork(self):
fname, _ = QFileDialog.getSaveFileName(self, 'Save Network File', '.', 'Network File(*.nf)')
if not fname:
# self.logStatus('保存神经网络文件{}失败!'.format(fname), 'red', 'E')
return
self.network.dump(fname)
self.logStatus('保存神经网络文件{}成功'.format(fname))
@pyqtSlot()
def runNetworkTraining(self):
if self.network is not None:
ans = QMessageBox.question(self, '警告',
'系统中已存在训练好的神经网络，请问您需要重新训练神经网络吗?')
if ans == QMessageBox.No:
return
# clear previous plots
self.clearPlots()
self.logStatus("正在初始化神经网络结构...", 'blue', 'I')
training_x, training_y = DataHandler.split_xy(self.training_data)
# retrieve settings
epoch0 = 2000
tol0 = 0.1
retry_num = 3
h0size = 4
h1size = 4
if self.networkSettings:
epoch0 = self.networkSettings['epoch']
tol0 = self.networkSettings['tol']
retry_num = self.networkSettings['retry']
h0size = self.networkSettings['h0size']
h1size = self.networkSettings['h1size']
self.logStatus("神经网络信息：layer1={}, layer2={}, epoch0={}, retry={}, tol0={}"
.format(h0size, h1size, epoch0, retry_num, tol0), 'blue', 'I')
net = [(training_x.shape[0], ''), (h0size, 'sigmoid'), (h1size, 'sigmoid'), (1, 'feedthrough')]
try:
self.network = NeuralNetwork(training_x, training_y, sizes=net)
except ShapeError as e:
self.logStatus('初始化神经网络结构失败!')
QMessageBox.warning(self, '警告', '初始化神经网络结构失败, 请重试!')
return
# training
mu0 = 0.1
beta = 10
retry = 0
self.logStatus("使用LM算法开始训练神经网络...", 'blue', 'I')
while retry < retry_num:
residual, mu, citer, msg = \
self.network.train(retry=retry, epoch=epoch0, mu0=mu0,
beta=beta, tol=tol0, cb=self.logStatus)
if residual is None:
if retry == (retry_num - 1):
self.logStatus("训练失败!".format(msg), 'red', 'E')
return
else:
self.logStatus("训练失败:{}, 重试中...".format(msg), '#FFA07A', 'I')
self.network.randomize_wb()
# continue
elif residual > tol0:
if retry == (retry_num - 1):
self.logStatus("训练失败!".format(msg), 'red', 'E')
return
else:
self.logStatus("训练失败: 运算未能收敛, 重试中...", '#FFA07A', 'I')
self.network.randomize_wb()
# continue
else:
self.logStatus("神经网络训练完成, 迭代次数={1}, 最终残差={0}"
.format(residual, citer+retry*epoch0), 'blue', 'I')
break
retry += 1
self.predictDatakAct.setEnabled(True)
self.saveNetworkAct.setEnabled(True)
self.logStatus('请执行数据预测', '#FF8C00', 'T')
@pyqtSlot()
def predictData(self):
# don't forget to clear previous plots
self.clearPlots()
self.logStatus("开始进行数据预测...", 'blue', 'I')
test_x, test_y = DataHandler.split_xy(self.test_data, False)
self.predict_y = self.network.predict(test_x)
self.logStatus("开始计算误差...", 'blue', 'I')
self.predPlot.update_plot(x=np.arange(len(self.predict_y[0])),
y=self.predict_y[0])
# error plot
err_percent = (self.predict_y - test_y) * 100.0 / test_y
self.errPlot.error_plot(x=np.arange(len(err_percent[0])),
y=err_percent[0])
abs_err = np.abs(err_percent)
self.logStatus("数据预测完成, 最大绝对值误差={}%, 平均绝对值误差={}%"
.format(abs_err.max(), abs_err.mean()), 'blue', 'I')
self.saveDataAct.setEnabled(True)
@pyqtSlot()
def clearDataAndNetwork(self):
ans = QMessageBox.question(self, '警告',
'您希望删除所有的数据和已经训练好的神经网络吗?')
if ans == QMessageBox.No:
return
# reset
self.network = None
self.isDataLoaded = False
self.training_data = None
self.test_data = None
# update UI
self.loadNetworkAct.setEnabled(False)
self.runTrainingAct.setEnabled(False)
self.saveDataAct.setEnabled(False)
self.saveNetworkAct.setEnabled(False)
self.predictDatakAct.setEnabled(False)
self.resetAct.setEnabled(False)
self.logTextEdit.clear()
# clear plots
self.clearPlots()
@pyqtSlot(dict)
def updateSettings(self, settings):
self.networkSettings = settings
@pyqtSlot()
def showSettingDialog(self):
dlg = SettingDialog(self, self.networkSettings)
dlg.show()
@pyqtSlot()
def fileQuit(self):
self.close()
@pyqtSlot()
def about(self):
QMessageBox.about(self, "关于",
"""<b>{}</b><p>版本号: {}""".format(progname, progversion)
)
@pyqtSlot()
def usage(self):
dlg = UsageDialog(self)
dlg.show()
def closeEvent(self, ce):
self.fileQuit()
def logStatus(self, text, color='green', tag='S'):
self.logTextEdit.appendHtml("<p><font color='{0}'><b>[{1}]:</b> {2}"
"</font></p>".format(color, tag, text))
# force UI update. An alternative is to use QThread
QApplication.processEvents()
def clearPlots(self):
self.predPlot.clear()
self.errPlot.clear()
errplot_labels = {'t': 'Prediction Errors', 'x': 'Time', 'y': 'Error Percent(%)'}
predplot_labels = {'t': 'Predicted Temperature', 'x': 'Time', 'y': 'Temperature(℃)'}
self.errPlot.initial_figure(errplot_labels)
self.predPlot.initial_figure(predplot_labels)