|
Amazon SageMaker Canvas presents enterprise analysts with a visible interface to fix small business complications using device learning (ML) without composing a solitary line of code. Given that we launched SageMaker Canvas in 2021, several end users have requested us for an improved, seamless collaboration encounter that permits knowledge researchers to share properly trained products with their business enterprise analysts with a couple of uncomplicated clicks.
Nowadays, I’m excited to announce that you can now convey ML models created wherever into SageMaker Canvas and produce predictions.
New – Bring Your Own Design into SageMaker Canvas
As a knowledge scientist or ML practitioner, you can now seamlessly share types constructed any place, inside or outside Amazon SageMaker, with your small business teams. This gets rid of the heavy lifting for your engineering teams to develop a independent device or user interface to share ML designs and collaborate amongst the distinctive areas of your firm. As a business analyst, you can now leverage ML products shared by your details researchers inside minutes to make predictions.
Enable me display you how this works in exercise!
In this illustration, I share an ML model that has been qualified to establish buyers that are perhaps at hazard of churning with my advertising and marketing analyst. 1st, I register the model in the SageMaker model registry. SageMaker design registry allows you catalog styles and control model versions. I make a design team called 2022-shopper-churn-model-team and then select Create design model to register my model.
To register your product, present the spot of the inference impression in Amazon ECR, as well as the locale of your design.tar.gz file in Amazon S3. You can also increase design endpoint suggestions and additional model information and facts. At the time you’ve registered your model, find the model model and select Share.
You can now select the SageMaker Canvas consumer profile(s) in the very same SageMaker area you want to share your design with. Then, provide more design aspects, these types of as data about instruction and validation datasets, the ML difficulty form, and model output details. You can also insert a note for the SageMaker Canvas buyers you share the product with.
Likewise, you can now also share styles educated in SageMaker Autopilot and models accessible in SageMaker JumpStart with SageMaker Canvas buyers.
The enterprise analysts will receive an in-application notification in SageMaker Canvas that a model has been shared with them, together with any notes you additional.
My promoting analyst can now open, review, and start off using the product to deliver ML predictions in SageMaker Canvas.
Decide on Batch prediction to generate ML predictions for an complete dataset or One prediction to create predictions for a solitary input. You can download the final results in a .csv file.
New – Enhanced Model Sharing and Collaboration from SageMaker Canvas with SageMaker Studio People
We also improved the sharing and collaboration capabilities from SageMaker Canvas with information science and ML groups. As a enterprise analyst, you can now pick out which SageMaker Studio person profile(s) you want to share your normal-build models with.
Your details experts or ML practitioners will obtain a equivalent in-app notification in SageMaker Studio after a design has been shared with them, alongside with any notes from you. In addition to just examining the design, SageMaker Studio users can now also, if wanted, update the details transformations in SageMaker Knowledge Wrangler, retrain the product in SageMaker Autopilot, and share back the up-to-date product. SageMaker Studio users can also suggest an alternate model from the list of models in SageMaker Autopilot.
After SageMaker Studio end users share again a model, you receive a further notification in SageMaker Canvas that an current design has been shared back with you. This collaboration in between small business analysts and details experts will assist democratize ML throughout corporations by bringing transparency to automated selections, setting up have confidence in, and accelerating ML deployments.
Now Out there
The increased, seamless collaboration abilities for Amazon SageMaker Canvas, which include the potential to provide your ML products built anywhere, are out there nowadays in all AWS Regions in which SageMaker Canvas is out there with no changes to the existing SageMaker Canvas pricing.
Commence collaborating and carry your ML design to Amazon SageMaker Canvas currently!
— Antje
