课程: Introduction to Generative Adversarial Networks (GANs)

今天就学习课程吧！

今天就开通帐号，24,100 门业界名师课程任您挑！

How were they developed?

How were they developed?

课程: Introduction to Generative Adversarial Networks (GANs)

开始一个月试用 购买团队方案

How were they developed?

“

- [Instructor] So how exactly are these generative models developed and how are they different to traditional machine learning models? It all comes down to the types of tools that were being used. So over time, Python libraries, like PyTorch and TensorFlow, had more and more investment and they made it easy and easier and more accessible to be able to train neural networks like this, especially neural networks that work together. So it's not a standard A to B type of flow. As you'll see in the architecture, it's kind of A to B, A to C, B to C connection between the different training and data sources. A lot of helper libraries, like Torchvision and Keras, also came about, which gave heaps of support for data loaders, for simple code management, and other tools to make it really easy to start training models. Additional cloud resources as well, like Azure and Amazon Web Services, they provide a number of code bases and…

内容