modeltweak.py

# -*- coding: utf-8 -*-
"""modeltweak.ipynb

Automatically generated by Colaboratory.

Original file is located at
    https://colab.research.google.com/drive/1ae5IlVNbI9p-xi_z-WqC4hyS2VRiAxqg
"""

import keras
from keras.datasets import mnist
from keras.models import Sequential
from keras.utils.np_utils import to_categorical
from keras.layers import Dense
from keras.optimizers import Adam
from keras.backend import clear_session
import numpy

# Train model function

def train_model(neurons , model , epochs , test) : 
	print("\n" , " *** Summary *** " , "\n", "Iteration  : ", test , "\n" , "   Number of Neurons : ", neurons , "\n" , "   Number of Epochs : ",  epochs)
	model.add(Dense(units = neurons , input_dim = 28*28 , activation = 'relu'))
	model.add(Dense(units=200 , input_dim = 28*28 , activation = 'relu'))
	model.add(Dense(units=60 , input_dim = 28*28 , activation = 'relu'))
	model.add(Dense(units=10 , input_dim = 28*28 , activation = 'softmax'))
	model.compile( optimizer= "Adam" , loss='categorical_crossentropy', 
	             metrics=['accuracy'] )
	return model


def validate(fit_model, epochs):
	text = fit_model.history
	accuracy = text['accuracy'][epochs-1] * 100
	accuracy = int(accuracy)
	f= open("accuracy.txt","w+")
	f.write(str(accuracy))
	f.close()
	print("    Accuracy for this Iteration is : " , accuracy ,"%")
	return accuracy



# Load Model 
(train_X , train_y), (test_X , test_y) = mnist.load_data("mymnist.data")
# Reshape data and change type
test_X = test_X.reshape(-1 , 28*28)
train_X = train_X.reshape(-1 ,  28*28)
test_X = test_X.astype("float32")
train_X = train_X.astype("float32")
# One hot encoding 
test_y = to_categorical(test_y)
train_y = to_categorical(train_y)


#Initials
neurons = 10
accuracy = 0
epochs = 1
test = 1
flag = 0

while int(accuracy) < 95 :
	if flag == 1 :
		model = keras.backend.clear_session()
		neurons = neurons+10
		epochs = epochs+1 
		test = test + 1
	#model=reset_weights(model)
	model = Sequential()
	model = train_model(neurons , model , epochs , test)
	print("    calculating accuracy . . .")
	fit_model = model.fit(train_X ,  train_y , epochs = epochs , verbose =  False)
	accuracy=validate(fit_model , epochs)
	flag = 1

#sending mail

import smtplib
s = smtplib.SMTP('smtp.gmail.com', 587)
s.starttls()

s.login("messiisahot@gmail.com", "anuragp0010")



s.sendmail("messiisahot@gmail.com", "anuragp0010@gmail.com", "The Machine Learning model has achieved the accuracy of " + str(accuracy))
    

    # terminating the session 
quit()