fashion_mnist.py

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

Automatically generated by Colab.

Original file is located at
    https://colab.research.google.com/drive/1gncgD_q5H7Hb7D4jD04jGcrIsK-8JrRK
"""

import tensorflow as tf
import numpy as np
import matplotlib.pyplot as plt
from sklearn.metrics import classification_report

(xtrain,ytrain),(xtest,ytest) = tf.keras.datasets.fashion_mnist.load_data()

plt.imshow(xtrain[0])
plt.title(f"object = {ytrain[0]}")
plt.colorbar()

xtrain[0].shape

xtrain.shape

label_to_remove = 8
def remove_label(X, y, label):
    mask = y != label  # Create a mask where the label is not equal to the target label
    return X[mask], y[mask]

# Remove class 8 from both training and testing datasets
xtrain, ytrain = remove_label(xtrain, ytrain, label_to_remove)
xtest, ytest = remove_label(xtest, ytest, label_to_remove)

label_to_remove = 6
def remove_label(X, y, label):
    mask = y != label  # Create a mask where the label is not equal to the target label
    return X[mask], y[mask]

# Remove class 8 from both training and testing datasets
xtrain, ytrain = remove_label(xtrain, ytrain, label_to_remove)
xtest, ytest = remove_label(xtest, ytest, label_to_remove)

class_images = {}
for img, label in zip(xtrain, ytrain):
    if label not in class_images:
        class_images[label] = img
    if len(class_images) == 10:  # Stop when we have all 10 classes
        break

plt.figure(figsize=(10, 2))
for i in range(10):
    if i in class_images:
      plt.imshow(class_images[i])
      plt.title(f"Class {i}")
      plt.show()

"""NORMALIZATION"""

xtrain,xtest = xtrain/255,xtest/255

"""NEURAL NETWORKS USING TENSORFLOW ANS KERAS."""

model = tf.keras.Sequential([
    tf.keras.Input(shape = (28,28)),
    tf.keras.layers.Flatten(),
    tf.keras.layers.Dense(128,activation = 'relu'),
    tf.keras.layers.Dense(10,activation = 'softmax')
])
model.compile(optimizer = 'adam', metrics = ['accuracy'],loss = "sparse_categorical_crossentropy")
model.fit(xtrain,ytrain,epochs = 20,verbose = 1)

loss , accuracy = model.evaluate(xtest,ytest)

"""CLASSIFICATION REPORT ON TEST DATA"""

y_pred = model.predict(xtest)
y_pred_classes = y_pred.argmax(axis=1)
print(classification_report(ytest, y_pred_classes))

np.argmax(model.predict(xtest)[0])

"""TEST WITH ANY NEW IMAGE"""

test_image = tf.keras.preprocessing.image.load_img("NEW_IMAGE_PATH")
test_image = tf.image.rgb_to_grayscale(test_image)
test_image = tf.keras.preprocessing.image.img_to_array(test_image)
test_image = tf.image.resize(test_image,[28,28])
test_image = test_image.numpy().squeeze()
test_image = test_image.astype("float64")
plt.imshow(test_image)
test_image = test_image.reshape(1,28,28)
test_image = test_image/255
plt.title(f"ID = {np.argmax(model.predict(test_image))}")