High-Performance Data Loading for Machine Learning Workloads with Squirrel
In this talk, our Machine Learning Researcher, Alexander Koenig, presents our open-sourced data-loading library called Squirrel. Alex introduced Squirrel in the AI Colloquium of the safe.trAIn research consortium spearheaded by Siemens to provide insights into efficient data loading for machine learning workloads. His presentation focuses on the main features of Squirrel, its high-level software architecture, Squirrel’s integration into the data-loading stack, and common usage patterns. To find out more about the research consortium, please visit https://safetrain-projekt.de/en/.