Home
:
Book details
:
Book description
Description of
Deep Learning From Scratch In Python
Published 3/2024 MP4 | Video: h264, 1920x1080 | Audio: AAC, 44.1 KHz Language: English | Size: 1.44 GB | Duration: 5h 16m Understand Convolutional Neural Networks and Implement your Object-Detection Framework From Scratch What you'll learn Understand how Deep Neural Networks work, practically and mathematically Understand Forward- and Backpropagation processes, mathematically and practically Design and implement a Deep Neural Network for multi-class classification Understand and implement the building blocks of Convolutional Neural Networks Understand and Implement cutting-edge Optimization, Regularization and Initialization techniques Train and validate a Convolutional Model on widely used datasets like MNIST and CIFAR-10 Understand and implement Transfer Learning Use a Convolutional Model to create a Real-Time, Multi-Object Detection System Requirements No prior knowledge is required Description This course is for anyone willing to really understand how Convolutional Neural Networks (CNNs) work. Every component of CNNs is first presented and explained mathematically, and the implemented in Python.Interactive programming exercises, executable within the course webpage, allow to gradually build a complete Object-Detection Framework based on an optimized Convolutional Neural Network model. No prior knowledge is required: the dedicated sections about Python Programming Basics and Calculus for Deep Learning provide the necessary knowledge to follow the course and implement Convolutional Neural Networks.In this course, students will be introduced to one of the latest and most successful algorithms for real-time multiple object detection. Throughout the course, they will gain a comprehensive understanding of the Backpropagation process, both from a mathematical and programming perspective, allowing them to build a strong foundation in this essential aspect of neural network training.By the course's conclusion, students will have hands-on experience implementing a sophisticated convolutional neural network framework. This framework will incorporate cutting-edge optimization and regularization techniques, enabling them to tackle complex real-world object detection tasks effectively and achieve impressive performance results. This practical knowledge will empower students to advance their capabilities in the exciting field of Computer Vision and Deep Learning. Overview Section 1: Neural Networks Basics Lecture 1 Introduction Lecture 2 Intuition about Fully-Connected Networks Lecture 3 Gradient Descent Algorithms Lecture 4 Training, Validation and Testing Section 2: Python Programming Basics Lecture 5 Python for CNNs Lecture 6 Working with Lists and Tuples Lecture 7 Working with NumPy Arrays Lecture 8 Object-Oriented Programming Section 3: Calculus for Deep Learning Lecture 9 The Derivative of a Function Lecture 10 The Product, Quotient and Power Rules Lecture 11 Derivatives by the Chain Rule Section 4: Cost Functions and Backpropagation Lecture 12 Backpropagation in Fully-Connected Networks Lecture 13 The Softmax Activation Function Lecture 14 The Cross-Entropy Cost Function Lecture 15 Backpropagation in the Output Layer Section 5: Building Blocks of Convolutional Neural Networks (CNNs) Lecture 16 Introduction to Convolutional Networks Lecture 17 Convolutions: Theory Lecture 18 Convolutions: Implementing an Edge Detector Lecture 19 Downsampling through Max Pooling Section 6: Backpropagation in Convolutional Neural Networks Lecture 20 Backpropagation in Convolutional Layers Lecture 21 Backpropagation in Pooling Layers Section 7: Integration of a Convolutional Model Lecture 22 Defining a Convolutional Model Lecture 23 The MNIST Dataset Lecture 24 Filter Visualization Section 8: Transfer Learning Lecture 25 What is Transfer Learning Section 9: Insights into Optimization and Regularization Lecture 26 Fully-Convolutional Implementation Lecture 27 The Vanishing Gradient and Dying ReLU Problems Lecture 28 Parameters Initialization Lecture 29 Learning Rate Decay Lecture 30 The Adam Optimizer Lecture 31 Testing the Optimized Model on MNIST Lecture 32 Testing the Optimized Model on CIFAR-10 Section 10: Multiple Object Detection in Real-Time Lecture 33 The YOLO Algorithm Lecture 34 Testing the YOLO Algorithm