aymericdamien/tensorflow-examples

TensorFlow-Examples is a learning collection that helps beginners get started with TensorFlow, Google's popular machine learning framework. The problem it solves is simple: TensorFlow can feel overwhelming at first, and this project provides a step-by-step series of clear, runnable examples that teach concepts gradually without assuming prior expertise. The collection is organized as a curriculum. It starts with the very basics, printing "hello world" using TensorFlow, then works up through basic mathematical operations, simple statistical models like linear and logistic regression, and then into neural networks. These include convolutional networks (good at recognizing images), recurrent networks with LSTM cells (good at sequences and time-series data), and generative adversarial networks (which can create new images from noise). There are also sections on practical topics like saving and loading trained models, visualizing training progress with TensorBoard, and loading different types of data efficiently. Each example exists as a Jupyter Notebook, which is an interactive document where you can read explanations, run code cells one at a time, and immediately see the output. The notebooks support both TensorFlow version 1 and the newer version 2, and most examples use the MNIST dataset, a standard collection of handwritten digit images used as a beginner benchmark in machine learning. Someone would use this project when they are new to machine learning or TensorFlow and want concrete, working code alongside explanations rather than pure theory. It is also useful for developers familiar with other frameworks who want to understand how TensorFlow's specific APIs like layers, models, and datasets work in practice. The tech stack is Python, TensorFlow (both v1 and v2), and Jupyter Notebooks.