sagemaker

library3.21.0pypypi✓ verified 29d ago

Amazon SageMaker is an open-source library for training and deploying models on Amazon SageMaker. The current version is 3.7.1, and it is released regularly with new features and bug fixes. It requires Python version 3.9 or higher.

aws ai-ml devops

pip install sagemaker

github ↗package ↗homepage ↗

INSTALL

IMPORT

SIG · SAGEMAKER

sagemaker

awspythonv3.21.0

Install

—

Import

8376ms

Disk

426MB

Pass rate

5/ 10

Env Coverage5 / 10

glibc

3.9–3.13

musl

3.9–3.13

Install & Compatibility

Where this runs

tested against v3.0 · pip install

no network on importno background threads

Install × environment matrix

Each cell = how many times install + import succeeded across repeated harness runs. Partial = flaky.
glibc = Debian/Ubuntu slim · musl = Alpine Linux

musl

py 3.10–3.95 runs

installs and imports cleanly · install 0.0s · import 8.376s · 415.6MB

glibc

py 3.10–3.95 runs

timeout

426MB installed

● package 426MB

Code

Verified usage

Verified import paths — ran on the pinned version, not inferred.

sagemaker

✓ import sagemaker

The sagemaker library should be imported directly.

SageMaker

✓ from sagemaker import SageMaker

✗ from sagemaker.sagemaker import SageMaker

The SageMaker class should be imported from the top-level sagemaker module.

Quickstart example to get the execution role using Amazon SageMaker.

import sagemaker from sagemaker import get_execution_role role = get_execution_role() print(role)

Debug

Known issues

breakingIn version 3.7.1, the ModelBuilder's LoRA deployment path now respects the user-provided `accept_eula` value. If not set to `True`, it raises a `ValueError`.

fix

Explicitly set `accept_eula` to `True` in the ModelBuilder's LoRA deployment path.

affects: >=3.7.1

gotchaWhen using the `Evaluate` class, it's essential to set the IAM role correctly to avoid permission errors.

fix

Set the IAM role correctly using the `role` parameter in the `Evaluate` class.

affects: all

deprecatedThe `Estimator` class is deprecated in favor of the `ModelTrainer` class.

fix

Use the `ModelTrainer` class instead of the `Estimator` class.

affects: >=3.4.0

Errors

Common errors & fixes

ModuleNotFoundError: No module named 'sagemaker'

The Amazon SageMaker Python SDK is not installed in your current Python environment or the selected kernel.

fix

pip install sagemaker

botocore.exceptions.NoCredentialsError: Unable to locate credentials

The SageMaker SDK, which relies on `boto3`, cannot find AWS credentials configured in your environment to authenticate with AWS services.

fix

Configure AWS credentials using `aws configure` via the AWS CLI, set `AWS_ACCESS_KEY_ID` and `AWS_SECRET_ACCESS_KEY` environment variables, or ensure your execution environment (e.g., SageMaker notebook instance, EC2) has an attached IAM role with appropriate permissions.

ValueError: RoleArn must be an arn for an IAM role

The `role` parameter in SageMaker SDK functions (like `Estimator` or `Session`) expects a full IAM Role ARN (Amazon Resource Name), not just a role name.

fix

Use `sagemaker.get_execution_role()` to automatically fetch the ARN if running in a SageMaker environment, or provide the complete ARN string, e.g., 'arn:aws:iam::123456789012:role/SageMakerExecutionRole'.

ValueError: Framework version X.Y is not supported. Please use one of the following versions: [list of versions]

The specified framework version (e.g., for TensorFlow, PyTorch, MXNet estimators) is not supported by the SageMaker Python SDK or the underlying SageMaker service for the chosen image.

fix

Update the `framework_version` parameter in your estimator definition to one of the supported versions listed in the error message, for example: `sagemaker.tensorflow.TensorFlow(framework_version='2.12', ...)`.

Upgrade

Version history

3.21.0latest on PyPI · released Aug 25, 2026

Audit

Dependencies

boto3requiredAmazon SageMaker is built on top of AWS services and uses boto3 for API calls

Agent activity

20 hits · last 30 days

node

OpenAI (training)

Resources

githubhttps://github.com/aws/sagemaker-python-sdk ↗

packagehttps://pypi.org/project/sagemaker/ ↗

homepagehttps://aws.amazon.com/sagemaker/ ↗