The Machine Learning eResearch Platform (MLeRP) is now available to Australian researchers to efficiently allocate compute resources for their machine learning (ML) needs.
The demand for machine learning in research is growing rapidly in Australia. To cater for this, allocation of graphics processing unit (GPU) resources for machine learning computation on research clouds and high-performance computing (HPC) systems has to be flexible. This is so the GPUs can be appropriately utilised and more machine learning users can be accommodated simultaneously on the systems.
The new MLeRP – developed by Monash University with ARDC co-investment through the National Machine Learning Service project – does just that. The platform combines the interactivity of a notebook and the power of an HPC environment, allowing users to develop and debug their algorithms while processing their dataset. Users can change compute requirements like RAM and GPU sizes and number of processes without leaving the notebook environment, and when their code is not being executed, compute resources can be released for other users.
Dr Slava Kitaeff, Associate Director of the Monash e-Research Centre and leader of the National Machine Learning Service project, said, “High-end GPUs are expensive and in high demand by machine learning and AI researchers. We want to deliver the technology to as many researchers as possible without escalating the infrastructure cost.
“From what we’ve identified, the development phase seems to be the gap where GPUs are required but underutilised. We’ve opted for the Dask framework in the Jupyter environment to increase the seats available in national machine learning services, and we’ve designed and delivered the MLeRP system with various options for the researchers to access powerful GPUs.”
MLeRP will benefit researchers working in neuroscience, clinical science, molecular imaging, robotics, economics, design and architecture, and many more areas by efficiently using compute resources in HPC clusters and on the national ARDC Nectar Research Cloud.
Ben Chiu, Director of Services, ARDC, said, “The ARDC is delighted to have partnered with Monash University and helped steer the development of this crucial element of machine learning for research in Australia. MLeRP provides leading-edge infrastructure and enhances provision of services for Nectar, which supports the ARDC’s strategic Thematic Research Data Commons (Thematic RDCs) for health and medical research, earth and environmental sciences, and HASS and Indigenous studies. We’ll ensure the new national machine learning service can be utilised in our continued efforts to enable and accelerate Australian research.”
You can now use the open Beta version of MLeRP with your institutional account through AAF authentication. Sign up and learn more about the platform:
Introducing MLeRP: A Webinar
The Monash University team gave a quick introduction to the platform, a summary of its offerings, and a demonstration at a recent webinar. Find out what makes MLeRP different from other HPC platforms and how best to take advantage of its powerful NVIDIA A100 hardware by watching the recording below.
The webinar was hosted by the Machine Learning Community of Practice for Australia (ML4AU) and Monash University.
More from the National Machine Learning Service Project
As another part of the National Machine Learning Service project, The University of Queensland and QCIF are developing training materials for researchers to help them adopt deep learning. Stay tuned by subscribing to the ARDC Connect newsletter.
The ARDC is funded through the National Collaborative Research Infrastructure Strategy (NCRIS) to support national digital research infrastructure for Australian researchers.