Category:
VCM provides GPU-enabled environments for student projects and course work, packaged as JupyterLab notebooks with CUDA-enabled GPUs. These environments are typically used for machine learning applications using Pytorch or Tensorflow for computation. If you are an instructor and need GPUs for your course, please follow the instructions on requesting a custom container, including an explanation of how the GPU will be used and which packages you need.
When not in use for courses with priority access, the GPU cluster is also available to students via the "GPUscavenger" containers running JupyterLab. GPU scanvenger containers may have their sessions terminated without warning if capacity is needed for courses.
In addition, there is an undergrad-only GPU cluster running JupyterLab containers aimed at student projects. This cluster is access via the "UndergradGPU" containers.
Not all machine learning applications need a GPU, particularly if the ML model or training dataset is relatively small. The "MLwinter" containers were used for the Machine Leanring Winter school, and are available for those who want access to JupyterLab, Pytorch, and some common ML libraries.