Understanding Neural Networks with Keras and TensorFlow

The session was about understanding how neural networks are built and how you can use TensorFlow and Keras to build one.

Transcript

1. Understanding Neural Networks with Keras and TensorFlow Cloud Mbarara
2. What is TensorFlow? It's an open-source deep learning framework that offers a vast array of tools, libraries, and resources to create and implement machine learning and deep learning models. TensorFlow is and maintained developed by the Google Brain team import tensorflow as tf
3. What is Keras? It's a Python-based, open-source neural network API designed to simplify the process of building, training, and deploying deep learning models. Keras was developed by François Chollet in 2015 import keras // from tensorflow import keras
4. What are Neural Networks? They are computational models inspired by the human brain's interconnected neurons, utilized in machine learning to process and learn from data, making them capable of complex pattern recognition and decision-making. Neural network layers can have a state (i.e have weights) or be stateless.
5. Input Layer The input layer is where data is fed into the neural network. Each node (or neuron) in this layer represents a feature of the input data. Hidden Layers Between the input and output layers, we have one or more hidden layers. Each layer consists of neurons that apply weighted sums and activation functions to their inputs. A typical neural network in practice can have hundreds of hidden layers. Neurons Each neuron takes the weighted sum of its inputs (from the previous layer) plus a bias term.
8. Activation Functions - Rectified Linear Unit (ReLU) - Tanh - Sigmoid - Softmax Optimizers - RMSProp - Adam
9. Let me explain it in code… Cloud Mbarara  The Innovation Village @TheVillageUG Following  Institute for Nonprofit News @INN  INNA @inna_ro  Mo Rocca @MoRocca  Innersloth  @InnerslothDevs  Inna Afinogenova @inafinogenova  Inner City Press @innercitypress
10. 
 from tensorflow.keras.datasets import mnist
 from tensorflow import keras
 from tensorflow.keras import layers
 
 (train_images,train_labels),(test_images,test_labels) = mnist.load_data()

11. Then we visualize the data Cloud Mbarara
12. from matplotlib import pyplot as plt
 from matplotlib import image as mpimg
 
 plt.title("Sample Image")
 plt.xlabel("X pixel scaling")
 plt.ylabel("Y pixels scaling")
 
 sample_image = test_images[0]
 plt.imshow(sample_image,cmap=plt.cm.binary)
 plt.show()

14. Build the neural network with 2 layers… Cloud Mbarara
15. model = keras.Sequential([
 layers.Dense(512, activation ="relu"), layers.Dense(10, activation="softmax")
 ])
 
 
 

16. Compile the model… Cloud Mbarara
17. model.compile(
 optimizer ="rmsprop",
 loss = "sparse_categorical_crossentropy",
 metrics=["accuracy"]
 )
 
 

18. Reshape the data and fit the model… Cloud Mbarara
19. train_images = train_images.reshape((60000,28*28))
 train_images = train_images.astype("float32")/255
 test_images = test_images.reshape((10000,28*28))
 test_images = test_images.astype("float32")/255
 
 model.fit(
 train_images,train_labels, 
 epochs=10, 
 batch_size=128
 )

20. Evaluate the model… Cloud Mbarara
21. test_loss, test_acc = model.evaluate(
 test_images, test_labels
 )
 
 print(f"Test Accuracy is: {test_acc}")

22. Perform an inference to make predictions… Cloud Mbarara
23. image_test = mpimg.imread("test_image.png")
 plt.imshow(image_test)
 
 from mnst_image_format_converter import imageprepare
 image_test = imageprepare("test_image.png")
 plt.imshow(image_test, cmap=plt.cm.binary )

25. Make predictions… Cloud Mbarara
26. import numpy as np
 imgNumpyData = np.array(image_test)
 
 type(imgNumpyData)
 imgNumpyData.shape
 imgNumpyData = imgNumpyData.reshape((1,784))
 imgNumpyData = imgNumpyData.astype("float32")/255
 
 prediction_test = model.predict(imgNumpyData)

28. Let’s see this in play… Cloud Mbarara
29. Cloud Mbarara Thank you! Any questions? Wesley Kambale @weskambale kambale.dev