Basic-CNN

Well this is a basic Convolutional Neural network trained using Mnist data set

Requiremnets

Import Tensorflow and Mnist dataset keras to start building the model

load data into X and Y
for intial data visualiztaion use plt.imshow
use normalize function to scale the values between 0 and 1
create a model (defining type of model 'sequential')
flatten the model
create 2 Dense layer with 128 neurons (act=relu)
a final output layer (act=softmax)
compile the model (optimizer,loss,metrics)
fit the model with number of epochs
model.evaluate assign loss and acuuracy
save the model using the save function
save it with .h5 exeten (saves the model in hierarchial data format rather than .pb format as .pb is default extension of tf)