def main():
# Input parameters
INPUT_SHAPE = (123, 123, 3)
# Set working directory where files are located
workingDirectory = "C:\\Users\\Dillon\\Desktop\\Senior_Project"
trainingImgsPath = workingDirectory + "\\training_Images"
idDictPath = workingDirectory + "\\data\\data_points.geojson"
x_train, y_train = makeBatch(idDictPath, trainingImgsPath)
testIdDictPath = workingDirectory + "\\data\\test_data_points.geojson"
testingImgsPath = workingDirectory + "\\testing_Images"
x_test, y_test = makeBatch(testIdDictPath, testingImgsPath)
print("Batches made\nInitializing model...")
# model = initModel(0.5, INPUT_SHAPE)
# train(model,
# x_train,
# y_train,
# num_samples=20000,
# batch_size=128,
# learningRate = 5e-5,
# epochs=100)
runModel(workingDirectory, x_test, y_test, x_train, y_train)
def model(self):
try:
if (int(self.e2.get())) > self.df.shape[0] or (int(
self.e2.get())) < 3:
messagebox.showerror(
"K Mean Clustring",
"please enter a valid number of clusters")
else:
try:
if (int(self.e3.get())) > 50 or (int(self.e3.get())) < 1:
messagebox.showerror(
"K Mean Clustring",
"please enter a valid number of runs")
else:
try:
# self.master.button1['state'] = 'normal'
model.runModel(self.df, int(self.e2.get()),
int(self.e3.get()))
self.img = ImageTk.PhotoImage(
PIL.Image.open("countries.png"))
self.panel = Label(self.master, image=self.img)
self.panel.grid(row=8, column=0)
self.img1 = ImageTk.PhotoImage(
PIL.Image.open(
"Generosity_social_support.png"))
self.panel = Label(self.master, image=self.img1)
self.panel.grid(row=8, column=1)
messagebox.showinfo(
"K Mean Clustring",
"clustring completed successfully!")
except:
messagebox.showerror("K Mean Clustring",
"error while the model")
except:
messagebox.showerror(
"K Mean Clustring",
"please enter a valid number of runs")
except:
messagebox.showerror("K Mean Clustring",
"please enter a valid number of clusters")
def analysis():
print "Running Model...."
Alz = model.runModel()
print "Running Model"
if ('username' not in session):
session ['username'] = None
if (session.get('username') == None):
username = session ['username']
results = request.args.get("Results")
if (results == "Results"):
return redirect("/results")
return render_template("analysis.html")
def train(self, tfrecord_filename, begin=0, einde=1000):
print("Start training with data")
print("Image Height: %d, Width: %d, Channels: %d" % (self.height, self.width, self.channels))
T2 = imagesToTfRecord(tfrecord_filename, TRAIN_DIR, TRAIN_LABELS, self.width, self.height, self.channels)
if os.path.exists(tfrecord_filename):
""" Read images and labels (tensors) from TFRecord file """
parsed_image_dataset = T2.readRecord(tfrecord_filename)
print(parsed_image_dataset)
#for parsed_record in parsed_image_dataset.take(5):
#print("parsed_record: ", repr(parsed_record))
#labels = parsed_record[1][0]
parsed_image_dataset.batch(32)
iterator = parsed_image_dataset.make_one_shot_iterator()
image_ds, label_ds = iterator.get_next()
test_images, test_labels = image_ds, label_ds
# print("iterator: ", iterator)
runModel(image_ds, label_ds, test_images, test_labels)
else:
print("File does not exist: ", tfrecord_filename)
with open(modelFilename, 'rb') as inputFile, open(os.path.join(outputDirectory, MODEL_FILENAME), 'wb') as outputFile:
outputFile.write(inputFile.read())
with open(os.path.join(outputDirectory, SCORE_FILENAME), 'w')as outputFile:
outputFile.write("%f\n" % score)
if __name__ == '__main__':
experimentIndex = 0
for configuration in ConfigurationStore.getTestConfigurationFiles():
experimentIndex += 1
with open(CONFIG_FILENAME,'w') as configFile:
configFile.write(configuration)
import model as ModelTrainer
score = ModelTrainer.runModel(MODEL_FILENAME)
saveExperimentResults(experimentIndex, CONFIG_FILENAME, MODEL_FILENAME, score)
if os.path.isfile(CONFIG_FILENAME):
os.remove(CONFIG_FILENAME)
if os.path.isfile(MODEL_FILENAME):
os.remove(MODEL_FILENAME)
del ModelTrainer
selectedFeatures = df[[
"CashIn", "CashOut", "Weekday", "DayOfMonth", "yearlyDeviation",
"yearlyDeviationChange", "workingDayOrNot", "Season",
"isReligiousHoliday", "isNationalHoliday", "isSpecialDay"
]]
# "previousWeekCashInAvg",
# "previousWeekCashInStdDev"
categoricalFeatureNameList = list(
selectedFeatures.select_dtypes(exclude=["number", "datetime"]))
categoricalFeatureIndexList = [
selectedFeatures.columns.get_loc(c)
for c in categoricalFeatureNameList
]
scaledArray = model.preprocessModelDf(selectedFeatures,
categoricalFeatureIndexList)
model_, history, test_X, test_y, prediction_values = model.runModel(
scaledArray, n_features, n_days_to_feed_model, n_days_to_predict,
cash_in)
model.plotLossHistory(history)
rmse = model.makePrediction(test_X, test_y, prediction_values, model_)
print("rmse : {}".format(rmse))
except Exception as e:
errorLog = str(traceback.format_exc())
print(errorLog)