def start():
global train_dataset_len, train_dim, class_number, X, y, NN
test_size = simpledialog.askfloat(
"Set test size",
"Input the test size(0 ~ 0.9), which means the size of samples used to test model's accuracy: ",
initialvalue=0.2,
minvalue=0,
maxvalue=0.9)
test_samples = int(train_dataset_len) * test_size
train_samples = train_dataset_len - test_samples
info_label[
"text"] = "Train Data Size: %6d\nTest Data Size: %6d" % (
train_samples, test_samples)
epoch = simpledialog.askinteger(
"Set Epoch",
"Input the Epochs(0~INF), which mean the iteration of training: ",
initialvalue=8,
minvalue=1)
max_value = np.max(X)
print(max_value)
with open("max_value.txt", 'w') as f:
f.write(str(max_value))
X = X / max_value
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=test_size)
NN = NeuralNetework()
NN.create_model(train_dim, class_number)
y_train = to_categorical(y_train)
y_test = to_categorical(y_test)
NN.train(X_train, y_train, X_test, y_test, epochs=epoch, batch_size=1)
print(
"\nFinished Model Training! Press 'Save Model' Button to save this model.\n"
)
from Datasets import Datasets as dataset
from NN import NN
import os, glob, numpy as np
if __name__ == "__main__":
data = dataset()
data.read_data()
training, label = data.get_titanic_training_data()
testing, label_test = data.get_titanic_test_data()
print(training.shape)
print(testing.shape)
ai = NN()
# Training
ai.create_model()
ai.train_model(training, label, 500)
ai.model.summary()
# Testing
# ai.load_model()
ai.test_model(testing, label_test)
