Which Sagemaker DL container images are fully compliant with FedRamp?

Hittinh 100% may challenging, btw, may consider below:

1. AWS Deep Learning Containers (DLCs) – AWS maintains a set of prebuilt SageMaker images, including PyTorch-based ones, that undergo regular security scans for vulnerabilities. However, even supported images may contain some vulnerabilities, and AWS recommends customers build and host patched versions in their own Amazon Elastic Container Registry (ECR) for high-security environments.
2. FedRAMP Compliance Requirements – FedRAMP requires hardened images that meet security benchmarks defined by the National Institute of Standards and Technology (NIST). AWS DLCs are updated to address critical vulnerabilities, but organizations may need to apply additional security measures to meet FedRAMP standards.
3. Custom Hardened Containers – If the latest Triton and DJL DL containers contain vulnerabilities, you might consider building a custom container using a minimal base image and applying strict security policies. AWS provides guidance on container vulnerability scanning and best practices for compliance.
4. Checking Supported Images – You can review the AWS Deep Learning Containers GitHub repository to check available images and their security updates. AWS also provides a support policy for prebuilt SageMaker images.

Hey,

Hope you're keeping well.

AWS doesn’t publish any SageMaker deep learning containers that guarantee “zero vulnerabilities” or full FedRAMP compliance out of the box. The AWS Deep Learning Containers used in SageMaker are security scanned and updated regularly, but FedRAMP workloads often require you to take the base image and apply additional hardening and patching. For high-security environments, the common pattern is to start with the AWS DLC that matches your framework (e.g., PyTorch, Triton), rebuild it with only required packages, apply OS and library updates, and host it in your own Amazon ECR in the FedRAMP region.

Thanks and regards,
Taz