Custoizable Neural Network

About the dataset:

The MNIST dataset is an acronym that stands for the Modified National Institute of Standards and Technology dataset.

It is a dataset of 60,000 small square 28×28 pixel grayscale images of handwritten single digits between 0 and 9.

About the task:

The task is to classify a given image of a handwritten digit into one of 10 classes representing integer values from 0 to 9, inclusively.

used Sequential model with bach of layers built deployed using streamlit.

And the app is Customizable for the number of neurons, number of epochs, and activation function.

How it works

To use streamlit you have to create an env to work in:

1. Go to your directory.

2. open the cmd and type: - py -m venv .env - .env\Scripts\activate

3. 3- To run the streamlite app, type: - streamlit run yourscript.py

See what inside the dataset:

import matplotlib.pyplot as pltplt.imshow(X_train[0])

Preprocessing the image:

def preprocess_image(images):images = images / 255return imagesX_train = preprocess_image(X_train)X_test = preprocess_image(X_test)

Create the model:

model = Sequential()model.add(InputLayer((28,28)))model.add(Flatten())model.add(Dense(32, 'relu'))model.add(Dense(10))model.add(Softmax())model.compile(loss='sparse_categorical_crossentropy', metrics=['accuracy'])model.summary()

Model: "sequential_7" _________________________________________________________________ Layer (type)                 Output Shape              Param #    ================================================================= flatten_7 (Flatten)          (None, 784)               0          _________________________________________________________________ dense_13 (Dense)             (None, 32)                25120      _________________________________________________________________ dense_14 (Dense)             (None, 10)                330        _________________________________________________________________ softmax_6 (Softmax)          (None, 10)                0          ================================================================= Total params: 25,450 Trainable params: 25,450 Non-trainable params: 0

Plot model accuracy

import pandas as pdimport matplotlib.pyplot as plthistory = pd.read_csv('history.csv')fig = plt.figure()plt.plot(history['epoch'], history['accuracy'])plt.plot(history['epoch'], history['val_accuracy'])plt.title('Model accuracy vs epochs')plt.ylabel('Accuracy')plt.xlabel('Epoch')plt.legend(['Train', 'Val'])plt.show()

Check code: https://github.com/geehaad/Custoizable-Neural-Network