Deep Generative Models

Deep Generative Models

Published 3/2023

Created by Digital tech academy

MP4 | Video: h264, 1280x720 | Audio: AAC, 44.1 KHz, 2 Ch

Genre: eLearning | Language: English | Duration: 52 Lectures ( 4h 20m ) | Size: 1.93 GB



Learn generative AI with Stable diffusion, Generative adversarial networks, VAEs and much more using Colab and Python



What you'll learn

Explore every aspect of generative models and gain a comprehensive understanding of their mechanics and capabilities.

Gain a comprehensive understanding of the principles and intricacies of autoregressive (AR) models, and learn how to write and train them in Python from scratch

Learn principles and details of variational autoencoder (VAE) models and how to code and train them in Python

Acquire a deep understanding of the principles and specifics of Generative Adversarial Networks (GANs), and master the art of coding and training them in Python

Familiarize yourself with the fundamental principles of Normalizing Flows models.

Discover the principles and intricacies of diffusion models, including Dall-E2, Imagen, and Stable diffusion, and learn how to code and train them in Python

Acquire the skills to become a master of this disruptive technology that is poised to revolutionize the future.



Requirements

No hard prerequisites

To fully understand all the details, a basic knowledge of mathematics, statistics and neural networks is suggested

To fully understand the coding parts, a basic understanding of Python is suggested

No software installation is required on your PC



Description

Generative AI is an emerging field in Artificial Intelligence with incredible potential. Its ability to generate synthetic data and support human creativity has made it increasingly popular in diverse fields such as multimedia, art, design, and product development.According to Gartner, which ranks it in the Top Strategic Technology Trends, Generative AI, or Generative Artificial Intelligence, is a «disruptive technology capable of generating artifacts that were based on human creativity, guaranteeing innovative results free from those prejudices typical of human experience and its thought processes».With this course, you can dive into the theory and practice of generative AI. You'll explore the different types of models currently used in generative AI, master them, and write your own code to generate unique images.Each section of the course starts with an intuitive explanation of how the model works before diving into the specifics of implementation. We use Python and Pytorch in a Colab notebook to walk you through every line of code, ensuring you have a complete understanding of the implementation.By the end of the course, you'll have a clear understanding of both the main concepts and the details of the coding process. We focus on valuable and essential topics to provide 100% useful information without any unnecessary chatter.Join us to unlock the potential of Generative AI and prepare for a future of innovation and creativity!Some keywords:Autoencoder, autoregressive, backward diffusion, backpropagation, Cifar 10, CLIP, classifier-free, Colab, CycleGAN, Dall-E2, DC-GAN, denoiser, Diagonal BiLSTM, Diffusion, dynamic thresholding, eDiff-I, encoder, embeddings, fashion-MNIST, forward diffusion, Generative adversarial networks, Glow, Imagen, Imagenet, inpainting, knowledge distillation, Kullbak-Leibler divergence, LSTM, log-likelihood, MNIST, neural networks, Normalizing Flow, openCLIP, paint-with-words, PixelCNN, Pixel RNN, python, pytorch, Row LSTM, stable diffusion, StyleGAN, superresolution, text encoder, text-to-image, unCLIP, universal approximation theory, WGAN, WGAN-GPCREDIT: Presentation template: slidescarnival



Who this course is for

Curious people who want to learn more about generative AI

Tech enthusiasts who want to stay up to date with state of the art AI

People interested in having a detailed description of how generative models work (autoregressive, variational autoencoders, generative adversarial networks, flow, diffusion)

Developers who want to learn how to generate images by writing Python code from scratch

Machine learning experts who want to expand their knowledge