No video :(

Deploy a model on AWS SageMaker with vetiver

Подписаться 15 тыс.

Просмотров 2,7 тыс.

50% 1

AWS SageMaker is a fully managed machine learning service that lots of organizations use for their ML tasks, but it hasn’t always been easy for R users to go about their work on this platform. In this screencast, learn how to train and deploy a model with R and vetiver on SageMaker infrastructure. Check out the code on my blog: juliasilge.com...

Опубликовано:

28 авг 2024

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист

Посмотреть позже

Комментарии : 12

@Andres-yf4wb Год назад

I love this channel, I hope it never dies!

@MartyAckerman310 Год назад

Remember the book "Crossing the Chasm"? Well in DS, the chasm is going from Jupyter notebooks to running in production. Thanks for another great vid on a super relevant topic. MLOps is often treated as someone else's problem.

@TURALOWEN Год назад

Julia, thank you so much for this particular video and all of the videos you put here. I have learned tons of tidymodels and vetiver deployments just watching these. Best, Tural.

@brittnyfreeman3650 Год назад

Do you need to pay for an Rstudio workbench license to use this?

@JuliaSilge Год назад

Yep, to use the RStudio IDE on SageMaker, you need a license. There's a pretty recent post on the Posit blog about using RStudio on SageMaker: posit.co/blog/getting-started-rstudio-sagemaker/

@sobohobo12123 Год назад

Is the plan to make similar functions for deployments on other cloud platforms (Azure, GCP, etc)?

@JuliaSilge Год назад

We've got GCP's Vertex AI on our "to do" list! If folks are interested in using this on Azure, we would love to hear more about their use cases.

@Levy957 Год назад

what If i want to use the endpoint from my local rstudio?

@JuliaSilge Год назад

As of today what you'll want to do is set up a SageMaker runtime (put your AWS credentials here): paws-r.github.io/docs/sagemakerruntime/ And then use `invoke_endpoint()`: paws-r.github.io/docs/sagemakerruntime/invoke_endpoint.html We could perhaps consider adjusting the vetiver `predict` method to work better locally, as well as on SageMaker. You can look here to see how it works now: github.com/rstudio/vetiver-r/blob/main/R/sagemaker.R#L407 If you decide to try any of this out, please let me know how it went for you.