Buckets:
| # Quickstart | |
| <img class="rounded-xl border border-gray-200 dark:border-gray-800" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/sagemaker/hub-deploy-aws-modal.png" | |
| alt="The Deploy on AWS modal on a Hugging Face model page" | |
| width="640"> | |
| The easiest way to deploy a model from the Hugging Face Hub to AWS starts on the model page itself. Every model page has a **Deploy** button that opens the **Deploy on AWS** modal, which offers two paths depending on the model: | |
| - **One-click deployment with JumpStart** for eligible models. | |
| - **Copy-paste-ready SageMaker Python SDK snippets** for every other model. | |
| ## Deploy JumpStart models with one click | |
| <img class="rounded-xl border border-gray-200 dark:border-gray-800" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/sagemaker/hub-one-click-aws-deploy.gif" | |
| alt="One-click deployment of a Hugging Face model to SageMaker AI Studio with the JumpStart integration" | |
| width="640"> | |
| Models available on [JumpStart](https://aws.amazon.com/sagemaker-ai/jumpstart/) show a one-click integration in the modal. Clicking **Deploy** takes you straight to SageMaker AI Studio with a preconfigured deployment: the model artifacts, the serving container, and a sensible default instance type are resolved for you. Before confirming, you can customize the deployment in Studio, for example the instance type, instance count, or endpoint name. | |
| The recording below shows the whole flow, from the model page on the Hub to a running endpoint in SageMaker AI Studio: | |
| ## Deploy any model with the SageMaker Python SDK | |
| <img class="rounded-xl border border-gray-200 dark:border-gray-800" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/sagemaker/hub-deploy-aws-sdk-tab.png" | |
| alt="The SageMaker SDK tab in the Deploy on AWS modal" | |
| width="640"> | |
| If you still prefer to deploy your models manually, the modal includes a SageMaker SDK tab with curated snippets for deployment and customization. The snippets are copy-paste ready: they use the [SageMaker Python SDK](https://github.com/aws/sagemaker-python-sdk) with the Hugging Face [Deep Learning Containers](./dlcs) and work as-is from a local environment or a SageMaker notebook. Adjust the model ID, instance type, or container configuration to fit your workload. | |
| ## Next steps | |
| This page covers the fastest paths from the Hub to a running endpoint. For more in-depth workflows, head to the dedicated sections: | |
| - [SageMaker SDK](../tutorials/sagemaker-sdk/sagemaker-sdk-quickstart): programmatic deployment and training with the Python SDK, from quickstart to production setups. | |
| - [JumpStart](../tutorials/jumpstart/jumpstart-quickstart): the JumpStart model catalog, deployment options, and customization in detail. | |
Xet Storage Details
- Size:
- 2.84 kB
- Xet hash:
- a2e0dba4a0b96abccd56d7c882668208262fc84ea3543637a5ca6164e8f631b9
·
Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.