In this video we will guide you through the complete pipeline of fine-tuning large language models (LLMs) for specialised tasks such as medical question-answering using NeMo Framework and Triton Inference Server.
Prepare and preprocess open-source datasets for fine-tuning.
Apply Parameter-Efficient Fine-Tuning (PEFT) using LoRA with NVIDIA NeMo Framework.
Deploy optimised LLMs using NVIDIATriton Inference Server and TensorRT-LLM.
Generate a synthetic Q&A dataset using Label Studio connected to a live inference backend.
Fine-tune and evaluate your customised LLM for domain-specific applications.
All workflows will be executed inside a UCloud project environment with access to GPU resources.
Target audience: Machine learning practitioners, researchers, and engineers interested in LLM customisation, domain adaptation, or scalable model deployment.
In this video we will go through the process of developing a course on UCloud, using UCloud Courses – a tool for hosting and managing university courses on UCloud. Want to know more about this new feature, check out our webinar recording introducing UCloud Courses.
Introduction
00:00 – Outline of the workshop agenda 03:10 – Introduction to the UCloud Courses concept 04:23 – Advantages of using UCloud Courses 05:56 – Outline of the steps involved in developing a UCloud Course 09:54 – Requesting and planning a UCloud Course 10:20 – Example of the planning and structuring of an existing UCloud Course
Developing a UCloud Course step by step
Setting up the environment 15:53 – Preliminary remarks 17:15 – Showing the existing UCloud Course that will be re-developed in the workshop 18:02 – Showing the UCloud-Courses GitHub repository 18:40 – Software requirements and -recommendations 20:22 – Short introduction to Git and GitHub 24:27 – Cloning the repository 25:15 – Creating a working branch 26:12 – What should go in the UCloud-Courses repository and what shouldn’t 27:24 – Installing the required dependencies
Creating a new UCloud Course and modifying the templates 28:00 – Creating a UCloud Course using a prepared script 30:31 – Walk-through of the different auto-generated files/folders in the course folder 33:35 – Short introduction to Docker 38:15 – Modifying the Dockerfile 42:32 – What should go in the UCloud-Courses repository and what shouldn’t 45:28 – Building the Docker image locally using a prepared script 47:54 – Modifying the starting script 01:00:50 – What can, shouldn’t, and mustn’t be changed while the course is running
Building and testing the course locally 01:05:00 – Re-building the Docker image locally using a prepared script 01:05:25 – Running a Docker container locally using a prepared script 01:10:58 – Opening the JupyterLab interface on localhost
Finalising the course app 01:12:48 – Opening a pull request and requesting code review 01:19:44 – Testing the course app on UCloud before it’s deployed
Closing remarks
01:25:44 – Reusing/updating an existing UCloud course 01:28:19 – The financial model in brief 01:29:38 – Useful links and resources
AlphaFold 3, the new AI model developed by Google DeepMind and Isomorphic Labs, can predict the structures of complexes formed by different biomolecules, including protein-DNA, protein-RNA, and protein-ligand interactions.
AlphaFold 3 deployment on Interactive HPC – UCloud is optimized for use with GPUs, and includes the necessary multiple genetic (sequence) protein and RNA databases to run.
Beside the Batch Mode, additional app flavors (Lab Mode and Visualization Mode) allow respectively to use the software in a JupyterLab environment and to visualize protein structures obtained by running AlphaFold 3.
AlphaFold 3 has been implemented on Interactive HPC – UCloud to support specific type of work, e.g., accurate structure prediction and modelling, and is ready for use by all Interactive HPC – UCloud users.
The position of Chair of the DeiC Board is currently vacant, leading to delays in decision-making processes and affecting upcoming national HPC calls.
John Renner Hansen has stepped down as Chair of the DeiC Board. As a result, the DeiC Board is currently unable to make financial decisions, which impacts several ongoing and upcoming HPC activities.
A new Chair is expected to be appointed by the Rectors’ Conference of Danish Universities in early August 2025.
Fewer Resources in Upcoming National HPC Call
The next national HPC call will open on 15 July 2025 and covers the compute period from 1 January to 31 December 2026. DeiC’s current contracts with DeiC Interactive HPC will expire on 31 December 2025, and work is underway to establish a new agreement. However, as the Board is not currently quorate, there is no guarantee that new agreements will be in place before the call opens. Therefore, the upcoming national call will only include national compute time on LUMI.
For more information, please contact your local Front Office.
ONLYOFFICE is a browser-based office suite for creating and editing text documents, spreadsheets, and presentations. Fully compatible with Microsoft Office formats, the interface supports a seamless user experience for those familiar with traditional desktop editors.
Designed to support document work directly within the UCloud environment, ONLYOFFICE allows users to:
Edit and create documents stored on UCloud
Work with word documents, presentations, and spreadsheets
Upload and download files without leaving the platform
Maintain full compatibility with Microsoft Office file types
The app provides core functionality through a lightweight, browser-accessible interface. No additional software installation is required.
ONLYOFFICE is available to all UCloud users and is particularly relevant for:
In this video we will introduce the new UCloud Courses app – a tool for hosting and managing university courses on UCloud.
Dr. Federica Lo Verso will walk you through the concept and background of the app. You’ll get a hands-on demonstration of how to access and use the tool, explore its integration with GitHub, and hear directly from Dr. Himanshu Khandelia, who shares his experience using UCloud Courses in real teaching scenarios.
We’ll also guide you through the application process, show you the technical and financial requirements, and give you a preview of an upcoming hands-on course.
00:00 – Introduction by Dr. Federica Lo Verso 01:23 – Background: Launch of the new UCloud Courses app 04:12 – Walkthrough: Where to find the Courses app and how to use it 07:00 – Recap of the UCloud Courses app demonstration 07:56 – Walkthrough: How the GitHub repository works 10:07 – Advantages of using UCloud Courses 11:12 – Use case introduction: Dr. Himanshu Khandelia 11:58 – Testimonial: Dr. Khandelia shares his experience using UCloud Courses 14:00 – Live demo: Dr. Khandelia shows his use of Courses and GitHub integration 23:00 – Application procedure 24:21 – Necessary resources: Compute and storage 25:28 – Re-use and update existing UCloud Courses 26:20 – Financial model for support 27:12 – Outro and teaser for the upcoming hands-on course
Indico is a robust open-source platform for event organization, archiving, and collaboration, ideal for managing conferences, workshops, and scientific meetings.
Key features include:
Event Management: Indico allows you to organize a wide range of events, from small seminars to large conferences. It supports the entire event lifecycle, including registration, abstract and final paper submission.
Hierarchical Protection: Indico uses a hierarchical protection scheme that allows for different roles and protection levels within events.
Integration with Collaborative Tools: Indico integrates seamlessly with tools like Zoom, enhancing the productivity of web-driven meetings.
Document Management: Indico facilitates easy upload and retrieval of presentations, papers, and other documents.
Indico has been implemented on UCloud and is ready for use by all UCloud users.
For more details, visit UCloud docs for full documentation.
Developing your own UCloud course app using a newly developed template-based approach
Join us for a hands-on workshop where we guide you through all the steps of developing a UCloud course app using our newly developed template-based approach. The concept involves having a dedicated app on UCloud for your university course that students can use in e.g., exercise/lab session and/or at home. An introduction to the approach can be found in this webinar recording.
In this workshop, you’ll learn how to:
Translate your course structure to a structure that is compatible with a UCloud course app.
Set up the course development environment which involves cloning the GitHub repository and running a course setup script.
Customize the provided templates to build a UCloud course app that includes all required components – software dependencies, scripts, datasets, and more.
Test the course on your computer during development to ensure everything functions correctly. We will also see how the final course app looks once it has been put on UCloud.
Git/GitHub and Docker are essential tools in the UCloud course development process. In the workshop, we will give short introductions to both tools primarily aimed at those participants who do not have experience with them. Participants can benefit from following introductory tutorials to Git and Docker in advance, although this is not a requirement.
Date: 11 June 2025
Time: 12:30-14:30 (CET)
Location: Online, via Zoom (link TBA)
Target audience: Researchers and teachers across all departments at all the Danish universities
Users of DeiC Interactive HPC – UCloud can now access Apache Kafka, a high-performance event streaming platform designed for managing real-time data at scale.
Kafka deployments on UCloud support multi-node configurations, enabling production-grade distributed logging, stream processing, data integration, and publish/subscribe messaging systems.
Students and researchers can use Apache Kafka to collect, process, and analyse large volumes of real-time data – for example, streaming data from lab sensors, social media, or web clickstreams. It’s especially useful for building data pipelines in projects involving machine learning, IoT, or real-time analytics, enabling reproducible and scalable research workflows.
For more details on the Apache Kafka application, visit UCloud docs for the full documentation.
With growing demand for transcription/AI services on Interactive HPC – UCloud, Speech Analyzer has been added to the platform.
Speech Analyzer is a specialised application built on top of Label Studio, optimised for dialogue analysis. It enables users to label, edit, and annotate speech transcriptions generated by the integrated Transcriber engine. It also allows you to share tasks with collaborators.
