Computing Resources, Software & Datasets
Computing Resources
Lotty Bruzelius - DGX cluster
https://www.ltu.se/en/research/research-subjects/machine-learning/gpu---computation-lab
Aggregate Summary of Computing Resources:
Server Infrastructure:
- 1 x Nvidia DGX Gold
- 2 x Inspur Servers
- 2 x Dell 9680 Servers
CPU Power:
- 80 VCPUs (2 x Xeon(R) E5-2698)
- 48 VCPUs (2 x AMD EPYC 7352) per server (2 servers)
- 192 VCPUs (2 x Xeon(R) Platinum 8468) per server (2 servers)
Nvidia GPU Acceleration:
- 8 x Tesla V100 32 GB GPUs
- 16 x A100 40 GB GPUs
- 8 X A100 80GB GPUs
- 8 x H100 80 GB GPUs
Collective System Memory/RAM: 3.5TB
Collective disk storage: 70 TB
Networking: High-speed Ethernet connectivity with Dell S6000 switch supporting 40Gbit to 100GbE QSFP Adapter
Software Environment: Ubuntu 22.04.3 LTS, Kernel 5.15.0-92-generic, CUDA Version 12.3, Supporting Conda and Docker Environment
GPU Computation Lab
https://www.ltu.se/en/research/research-subjects/machine-learning/kubernetes-cluster
Sweden's Berzelius Supercomputer
https://wasp-sweden.org/the-most-powerful-supercomputer-in-sweden-installed/ External link.
Unlock the Power of AI with Luleå Tekniska Universitet's Machine Learning Cluster
Are you ready to push the boundaries of AI research and innovation? Look no further than the cutting-edge computing resources offered by the Machine Learning group at Luleå Tekniska Universitet. Our state-of-the-art cluster, titled Lotty Bruzelius cluster, is equipped with top-tier hardware tailored to meet the demands of modern AI and related tasks.
Why Collaborate with Us?
- Unmatched Performance: Our cluster boasts a remarkable combination of CPUs and GPUs, offering unparalleled computing power for AI research and development.
- Versatile Hardware: Whether you're delving into deep learning, natural language processing, computer vision, or other AI disciplines, our diverse range of hardware ensures you have the tools you need.
- Exceptional Storage and Networking: With extensive storage capacity and high-speed networking capabilities, you can seamlessly manage and analyze vast datasets for your research projects.
- Reliable Support: Our team is dedicated to providing assistance and support, ensuring a smooth and productive research experience.
- Collaborative Opportunities: Partnering with our Machine Learning group opens doors to collaborative research projects, knowledge exchange, and access to cutting-edge resources, fostering innovation and advancement in the field of AI.
Join Us Today!
Embark on a journey of innovation and discovery with Luleå Tekniska Universitet's Machine Learning group. Whether you're a researcher or academic institution, we invite you to harness the power of our computing resources and embark on groundbreaking AI projects together. Contact us now to explore collaboration opportunities and unlock the full potential of AI technology!
Code and Datasets
Imaging dataset for inner-speech recognition
Dataset for condition monitoring vibration data annotated with technical language, from paper machine industries in northern Sweden (see Löwenmark, Karl DIVA)
Code: WORDSTYLIST: this https URL. https://github.com/koninik/WordStylist
Nordic Vehicle Dataset and Code
Dataset collected to evaluate the performance of vehicle detectors from UAV in different snowy weather conditions.
Publications based on NVD
- Nordic Vehicle Dataset (NVD): Performance of vehicle detectors using newly captured NVD from UAV in different snowy weather conditions.
- Fractional B-Spline Wavelets and U-Net Architecture for Robust and Reliable Vehicle Detection in Snowy Conditions
- Vehicle Detection Performance in Nordic Region
- Challenging YOLO and Faster RCNN in Snowy Conditions: UAV Nordic Vehicle Dataset (NVD) as an Example
Other Resources
Hardware
EEG/fMRI compatible 64 channels headset for synchronous recordings, Brain Products
Bipol: https://huggingface.co/datasets/tosin/mab_english External link. and https://github.com/LTU-Machine-Learning/bipolmulti
External link.
Papers based on bipol: (see above)
iDocVQA: Instruction Document Visual Question Answering dataset https://github.com/LTU-Machine-Learning/iDocVQA External link.
Paper based on iDocVQA: https://arxiv.org/abs/2402.00453 External link.
Updated: