Deep Learning Foundations – TensorFlow Basics

TensorFlow Basics

TensorFlow is an open‑source library for numerical computation and large‑scale machine learning. Keras, now integrated as tf.keras, provides a high‑level API.

Installing TensorFlow

pip install tensorflow

Hello World: a simple dense network

import tensorflow as tf
from tensorflow import keras
from tensorflow.keras import layers

# Define a simple sequential model
model = keras.Sequential([
    layers.Dense(64, activation='relu', input_shape=(10,)),
    layers.Dense(1, activation='linear')
])

model.compile(optimizer='adam', loss='mse', metrics=['mae'])
print(model.summary())

Training workflow

# Dummy data
import numpy as np
X = np.random.rand(1000, 10)
y = X @ np.random.rand(10,1) + np.random.randn(1000,1)*0.1

history = model.fit(X, y, epochs=30, batch_size=32, validation_split=0.2)

Visualizing training curves

import matplotlib.pyplot as plt
plt.plot(history.history['loss'], label='train loss')
plt.plot(history.history['val_loss'], label='val loss')
plt.legend()
plt.title('Training & Validation Loss')
plt.show()

Saving and loading models

model.save('my_model.h5')
# later
loaded = keras.models.load_model('my_model.h5')

GPU acceleration (optional)

• Install tensorflow-gpu or use the default TensorFlow (GPU support is built‑in for recent builds).
• Verify with tf.config.list_physical_devices('GPU').

Key concepts

• Tensors: N‑dimensional arrays (like NumPy, but GPU‑compatible).
• Eager execution (default) vs. graph execution.
• Layers: building blocks (Dense, Conv2D, LSTM, …).
• Model subclassing for custom architectures.