TensorFlow

Back to Module 3 | Back to Table of Contents

3.6 TensorFlow Overview

TensorFlow is an open-source machine learning framework developed by Google. It is widely used for building, training, and deploying deep learning models in areas such as image classification, object detection, NLP, and time-series forecasting.

On Jetson platforms, TensorFlow can leverage GPU acceleration to improve training and inference performance, making it suitable for edge AI development.

Key Features

  • High-level API support through Keras (tf.keras) for fast model development.
  • Flexible model export and deployment options (SavedModel, TensorFlow Lite, ONNX conversion workflows).
  • Rich ecosystem tools including TensorBoard for visualization and debugging.
  • Scalable execution from local devices to cloud infrastructure.

Typical Workflow

  1. Prepare and preprocess data.
  2. Build and train a model with tf.keras.
  3. Evaluate model accuracy and performance.
  4. Export and deploy the model for inference.

Quick Example (Keras)

python
import tensorflow as tf

model = tf.keras.Sequential([
    tf.keras.layers.Dense(64, activation='relu', input_shape=(32,)),
    tf.keras.layers.Dense(10, activation='softmax')
])

model.compile(optimizer='adam',
              loss='sparse_categorical_crossentropy',
              metrics=['accuracy'])

print("TensorFlow version:", tf.__version__)

Useful TensorFlow Links

Back to Module 3