Enterprise Computer Vision

Good afternoon,

Are there any resources about deploying our Computer Vision server to an azure VM? I have looked at https://activities.uipath.com/docs/deploying-a-local-machine-learning-model, but am wondering if there is anything more specific.

Thank you!

No, there’s nothing more specific because those steps should be enough. Are you looking for something in particular?

I guess I am looking to confirm the requirements for hardware other than video card. Microsoft docs have the min and recommended GPUs listed for their NC and ND series VMs. Will the minimum-spec versions of those instances work?

Additionally, it looks like we would want to have this instance in an Azure region that does not offer their VM with a P40 GPU. They do however offer VMs with M60 GPUs and plan to roll out V100 GPUs.

According to NVIDIA’s website both of these cards are “CUDA-enabled” and are listed as having a higher “Compute Capability” than the K80 cards. Is there anything else we would need to look for to confirm that these cards would be sufficient to run the computer vision machine learning model?

Hi @cypher1014

Welcome to the uipath community forum

cheers :smiley:
Happy learning :smiley:

M60 should be fine

Nvidia publishes a list of CUDA enabled GPUs…

I also wish UI Path’s document on imepleting computer vision had more detail on hardware requirements and troubleshooting installation.

There are no more details on hardware because those are the only officially supported GPUs at the moment.