Microsoft: Using GPU-powered Virtual Machines in the Cloud for your Machine Learning and Deep Learning workloads
Until a few years ago running Machine Learning (ML) and Deep Neural Networks required owning really powerful computers with lots of CPUs or even clusters of machines. With the adoption of GPUs for non-graphics related workloads new opportunities became available. You can now run many of your workloads in a much shorter period of time making use of modern GPUs you already have instead of using CPUs. The good thing is, you do not even have to own your own servers equipped with pricy graphics hardware any more. As we have become used to for CPU-powered machines we can now also get GPU powered machines on-demand in the cloud.
Microsoft Azure offers two categories of GPU powered machines with our N-Series. NC VMs feature NVIDIA Tesla K80 GPUs for compute and machine learning workloads. NV VMs are tailored towards 3d workloads and rendering and feature NVIDIA Tesla M60 GPUs. Learn more in these VM sizes in this announcement.
So how do you get started for Machine Learning and Deep Learning workloads?
Read the entire article here, Using GPU-powered Virtual Machines in the Cloud for your Machine Learning and Deep Learning workloads – Malte Lantin
via the fine folks at Microsoft