UT High Performance Computing Center offers AI-related services

The University of Tartu High Performance Computing (HPC) Center provides modern and secure computing resources for training and running AI models. The service gives users access to high-performance GPU computing capacity through the university’s local high-performance computing cluster, Rocket, and LUMI (Large Unified Modern Infrastructure), one of Europe’s fastest supercomputers.

The centre’s computing resources support the development, fine-tuning and deployment of machine learning and deep learning models for research, development and application projects. The service can be used for the development and deployment of large language models (LLMs), image-processing and predictive models, and other machine learning systems.

For users who prefer not to deal with the technical complexities of deploying, configuring and managing language models, we offer access to pre-configured models through our LLM inference service. This enables them to focus on developing applications and creating new solutions without having to manage models, infrastructure or supporting systems. We provide access to state-of-the-art language models through an application programming interface (API) compatible with the standard OpenAI API specification. This makes it easy to integrate existing tools and applications, reducing development costs. To ensure performance and reliability, the service automatically scales to manage high-volume workloads.

The HPC Center is rapidly expanding its capabilities and supports AI projects from initial ideas through to fully functioning solutions by providing the necessary computing power and expertise. Users can quickly get started with machine learning frameworks, such as PyTorch, or with tools for running large language models, including vLLM and Ollama, for which the centre also offers consultancy and support. In addition, we provide a Kubernetes platform, which is well suited for managing long-term user services.