Transformers: The Heart of modern AI

Large language models (LLMs) are the latest major breakthrough in technology. Companies like Google, OpenAI, or Microsoft are racing to train increasingly powerful models capable of generating text, reasoning, and solving complex tasks. But how do they actually work? Do they really “understand” language? Well, to the surprise of many, it all comes down to… matrix multiplications! We will learn how these models can: - Turn words into numbers - Represent semantic relationships - Generate coherent responses To do this, we will use open-source models from HuggingFace for real tasks such as image classification or text generation. Download workshop resources Download the slides