TensorFlow is an open source library for high-performance numerical computation. Developed by Google Brain team for machine and deep learning. Supports computation across various platforms including CPUs, GPUs, TPUs. Licensed under Apache 2.0
Keras model consists of architecture, weights, optimizer state, and metrics. Models can be saved in TensorFlow SavedModel (tf) or H5 format. tf format is default and supports deployment with TensorFlow.js
Course divided into two volumes: Supervised and Unsupervised Deep Learning. Each volume focuses on three distinct algorithms. Tutorials emphasize intuitive understanding of Deep Learning concepts
Latest stable version requires Python 3.9-3.12 and pip 19.0+. GPU support available for CUDA architectures 3.5-8.0. Ubuntu 16.04+ and macOS 12.0+ (64-bit) needed. Windows 7+ (64-bit) required for Windows Native. Windows WSL2 needs Windows 10 19044+
Tutorial demonstrates loading flowers dataset containing 3,670 images. Dataset includes five subdirectories per flower class. Images downloaded as 218MB file
3-month, 10-hour weekly program with flexible learning pace. 893,607 students have completed the program. Includes 5 course series covering various deep learning topics. Offers shareable certificate and LinkedIn profile enhancement