In the third part of our blog series, we cover an essential topic that has gained significant traction in the ML-community in the past years: Explainability. Explainable AI is essential to establish trust in the models we develop. We discuss various approaches for CNN networks, with a particular focus on the Grad-CAM method.
Car Model Classification II: Deploying TensorFlow Models in Docker using TensorFlow Serving
How do we interact with machine learning models in practice? In the second part of our 4-part blog series on car model classification, you will learn how models can be deployed using TensorFlow Serving, and how we can run model queries.
Car Model Classification I: Transfer Learning with ResNet
In this 4-part blog series on car model classification, we want to illustrate how an end-to-end deep learning project can be implemented. In the first part, we will show how you can use transfer learning to tackle car image classification. In particular, you will learn how a pre-trained ResNet model can be fine-tuned to tackle a downstream task.
Generative Adversarial Networks: How Data Can Be Generated With Neural Networks
In this article, our colleague Stephan is first going to explain how GANs work in general. Afterward, he will discuss several use cases that can be implemented with the help of GANs, and to sum up, he will present current trends that are emerging in the area of generative networks.