Open sunildarji811 opened 4 years ago
@sunildarji811 - the only way to do this right now is to use Azure Functions with Kubernetes, For example AKS supports GPU's
Any update?
Did the Azure function container support GPU instead of CPU such as the use case mentioned in the following article: https://towardsdatascience.com/searching-the-clouds-for-serverless-gpu-597b34c59d55
I copied the relevant description here for ease of reference:- A simple app that provides a user interface to Neural Style Transfer. Unfamiliar with this technique? In short, Neural Style Transfer (NST) is a Deep Learning technique that can be used to stylize images. The algorithm takes a content image and a style image as input, and uses these to generate a new image that shows the content of the content image, and the style of the style image.
Please advise. Thank you.
Any update here @anirudhgarg?
How do I setup a ML inference pipeline with Azure functions that can connect to a GPU and handle auto-scaling?
Thank you,
Tagging @pragnagopa , for further inputs
This repo may be useful: https://github.com/puthurr/python-azure-function-gpu
@BoPengGit hey. Could you please tell whether you were able to achieve your goal with what @ziyuang shared?
Also since 2021 have you found any better solution than to mess with Kuber?
Can I use GPU powered app with Azure functions?
Like I want to host external .exe with some GPU requirement on Azure function but it fails with runtime error.
Any solution or workaround?