All notable changes to this project will be documented in this file.
The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.
=======
- added
ray-on-eks
, andmanifests/ray-on-eks
manifests - added a
sagemaker-model-monitoring-module
module with an example of data quality, model quality, model bias, and model explainability monitoring of a SageMaker Endpoint - added an option to enable data capture in the
sagemaker-endpoint-module
- added a
personas
example module to deploy various roles required for an AI/ML project
- remove explicit module manifest account/region mappings from
fmops-qna-rag
- changed ECR encryption to KMS_MANAGED
- changed encryption for each bucket to KMS_MANAGED
- refactor
airflow-dags
module to use Pydantic - fix inputs for
bedrock-finetuning
module not working - add
retention-type
argument for the bucket in thebedrock-finetuning
module - fix broken dependencies for
examples/airflow-dags
- use
add_dependency
to avoid deprecation warnings from CDK.
- added multi-acc sagemaker-mlops manifest example
- fixed model deploy cross-account permissions
- added bucket and model package group names as stack outputs in the
sagemaker-templates
module - refactor inputs for the following modules to use Pydantic:
mlflow-fargate
mlflow-image
sagemaker-studio
sagemaker-endpoint
sagemaker-templates-service-catalog
sagemaker-custom-kernel
qna-rag
- add CDK nag to
qna-rag
module - rename seedfarmer project name to
aiops
- chore: adding some missing auto_delete attributes
- chore: Add
auto_delete
tomlflow-fargate
elb access logs bucket - updating
storage/ecr
module to latest pendingv1.8.0
of IDF - enabled ECR image scan on push
- added managed autoscaling config to
sagemaker-endpoint
module - added SSO support in
sagemaker-studio
module - added VPC/subnets/sg config for multi-account project template to
sagemaker-templates-service-catalog
module - added
sagemaker-custom-kernel
module - added batch inference project template to
sagemaker-templates-service-catalog
module - added EFS removal policy to
mlflow-fargate
module - added
mwaa
module with example dag which demonstrates the MLOps in Airflow - added
sagemaker-model-event-bus
module. - added
sagemaker-model-package-group
module. - added
sagemaker-model-package-promote-pipeline
module. - added
sagemaker-hugging-face-endpoint
module - added
hf_import_models
template to import hugging face models - added
qna-rag
module - added
bedrock-finetuning
module
- reogranized manifests by use-case
- add account/region props for project templates in
sagemaker-templates-service-catalog
module - fix
sagemaker-templates-service-catalog
model deploy role lookup issue & abalone_xgboost model registry permissions - update
sagemaker-custom-kernel
module IAM permissions - split
xgboost_abalone
andmodel_deploy
project templates insagemaker-templates-service-catalog
module - add support for other AWS partitions
- update MySQL instance to use T3 instance type
- upgrade
cdk_ecr_deployment
version to fix the deprecatedgo1.x
lambda runtime
- remove AmazonSageMakerFullAccess from
multi_account_basic
template in thesagemaker-templates-service-catalog
module - remove AmazonSageMakerFullAccess from
sagemaker-endpoint
module
- added
sagemaker-templates-service-catalog
module withmulti_account_basic
project template - bump cdk & ecr deployment version to fix deprecated custom resource runtimes issue in
mlflow-image
- added
sagemaker-jumpstart-fm-endpoint
module - added RDS persistence layer to MLFlow modules
- added
mlflow-image
andmlflow-fargate
modules - added
sagemaker-studio
module - added
sagemaker-endpoint
module - added
sagemaker-notebook
module
- refactor validation script to use
ruff
instead ofblack
andisort