Sören Brunk - Storch - GPU Accelerated Deep Learning for Scala 3

Подробнее о видео

Deep learning allows computers to automatically learn and improve from data, enabling them to solve complex problems that were previously difficult or impossible with traditional programming approaches. It is the technology behind almost all recent groundbreaking AI applications, fundamentally changing the way we interact with computers.

Scala is a natural fit for deep learning due to its powerful type system, functional programming support, collections library, and rich ecosystem of data-processing libraries. But until now, building deep learning models in Scala has been a challenge.

In this talk, we are going to learn about Storch, a new library based on PyTorch that makes it easy and fast to develop and train deep learning models in Scala.

We’ll have a look at Storch’s key features: GPU accelerated tensor operations, automatic differentiation, and the neural network API. We’ll see how it seamlessly integrates with existing Scala code, making it well-suited for building enterprise-grade machine learning applications. We’ll also dive a bit into the details of how Storch leverages Scala 3 features like match types to provide a type-safe yet easy to use API. Finally, we’ll see a hands-on demonstration of how it can be used to build and deploy a state-of-the-art deep learning model.

Join me on this journey to learn about the future of deep learning in Scala!