JumpStart CuratedHub Launch #4748

malav-shastri · 2024-06-20T16:39:20Z

Issue #, if available:
CuratedHub launch
Description of changes:

Testing done:

Merge Checklist

Put an x in the boxes that apply. You can also fill these out after creating the PR. If you're unsure about any of them, don't hesitate to ask. We're here to help! This is simply a reminder of what we are going to look for before merging your pull request.

General

I have read the CONTRIBUTING doc
I certify that the changes I am introducing will be backward compatible, and I have discussed concerns about this, if any, with the Python SDK team
I used the commit message format described in CONTRIBUTING
I have passed the region in to all S3 and STS clients that I've initialized as part of this change.
I have updated any necessary documentation, including READMEs and API docs (if appropriate)

Tests

I have added tests that prove my fix is effective or that my feature works (if appropriate)
I have added unit and/or integration tests as appropriate to ensure backward compatibility of the changes
I have checked that my tests are not configured for a specific region or account (if appropriate)
I have used unique_name_from_base to create resource names in integ tests (if appropriate)

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.

* Implement CuratedHub Admin APIs * making some parameters optional in create_hub_content_reference as per the API design * add describe_hub and list_hubs APIs * implement delete_hub API * Implement list_hub_contents API * create CuratedHub class and supported utils * implement list_models and address comments * Add unit tests * add describe_model function * cache retrieval for describeHubContent changes * fix curated hub class unit tests * add utils needed for curatedHub * Cache retrieval * implement get_hub_model_reference() * cleanup HUB type datatype * cleanup constants * rename list_public_models to list_jumpstart_service_hub_models * implement describe_model_reference * Rename CuratedHub to Hub * address nit * address nits and fix failing tests --------- Co-authored-by: Malav Shastri <malavhs@amazon.com>

…umpStart public hub models (aws#1456) * Implement CuratedHub Admin APIs * making some parameters optional in create_hub_content_reference as per the API design * add describe_hub and list_hubs APIs * implement delete_hub API * Implement list_hub_contents API * create CuratedHub class and supported utils * implement list_models and address comments * Add unit tests * add describe_model function * cache retrieval for describeHubContent changes * fix curated hub class unit tests * add utils needed for curatedHub * Cache retrieval * implement get_hub_model_reference() * cleanup HUB type datatype * cleanup constants * rename list_public_models to list_jumpstart_service_hub_models * implement describe_model_reference * Rename CuratedHub to Hub * address nit * address nits and fix failing tests * implement list_jumpstart_service_hub_models function --------- Co-authored-by: Malav Shastri <malavhs@amazon.com>

* get_model_spec() changes to support hub_arn and hub_content_type * implement get_hub_model_reference() * support hub_arn and hub_content_type for specs retrieval * add support for hub_arn and hub_content_type for serializers, deserializers, estimators, models, predictors and various spec retrieval functionalities * address nits and test failures * remove hub_content_type support --------- Co-authored-by: Malav Shastri <malavhs@amazon.com>

* implement HubContentDocument parser * modify the parser to remove aliases for hubcontent documents * bug fix * update boto3 * Bug Fix in the parser * Improve Hub Class and related functionalities * Bug Fix and parser updates * add missing hub_arn support * Add model reference deployment support and other minor bug fixes * fix: retrieve correct image_uri (parser update) * fix: retrieve correct model URI and model data path from HubContentDocument (parser update) * Add model reference deployment support * Model accessor and cache retrival bug fixes * fix: curated hub model training workflow * fix: pass sagemaker sessions object to retrieve model specs from describe_hub_content call * fix: fix payload retrieval for curated hub models * modify constants, enums * fix: update parser * Address nits in the parser * Add unit tests for parser * implement pagination for list_models utility * feat: support wildcard chars for model versions * Address nits and comments * Add Hub Content Arn Tag to training and hosting * Add Hub Content Arn Tag to training and hosting * fix: HubContentDocument schema version * fix broken unit tests * fix prepare_container_def unit tests to include ModelReferenceArn * fix unit tests for test_session.py * revert boto version changes * Fix unit tests * support wildcard model versions for training workflow * Add test cases for get_model_versions * Add/fix unit tests --------- Co-authored-by: Malav Shastri <malavhs@amazon.com>

benieric · 2024-06-20T18:48:08Z

For this one, check the failing unit and codestyle-doc tests. The integ tests failing are unrelated to this change (fixing in our infra account)

…_models()

AWS-pratab

Lets address these nit comments in next/later release

AWS-pratab · 2024-06-21T02:13:58Z

src/sagemaker/jumpstart/accessors.py

+                )
+                model_specs.set_hub_content_type(HubContentType.MODEL)
+                return model_specs
+            except:  # noqa: E722


nit: We can catch the specific exception here.

nit: We can check model-ref first, assuming there are more chances Hub has more refs than models.

AWS-pratab · 2024-06-21T02:31:58Z

src/sagemaker/jumpstart/factory/model.py

@@ -489,6 +549,9 @@ def _add_tags_to_kwargs(kwargs: JumpStartModelDeployKwargs) -> Dict[str, Any]:
            kwargs.tags, kwargs.model_id, full_model_version, kwargs.model_type
        )

+    if kwargs.hub_arn:
+        kwargs.tags = add_hub_content_arn_tags(kwargs.tags, kwargs.hub_arn)


It seems it is passing the hub_arn instead of hub_content_arn to the method.

Can be fixed after initial release, since for revenue tracking we will be needing the key part only.

benieric · 2024-06-21T06:31:06Z

src/sagemaker/jumpstart/accessors.py

@@ -10,6 +10,7 @@
 # distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF
 # ANY KIND, either express or implied. See the License for the specific
 # language governing permissions and limitations under the License.
+# pylint: skip-file


Lets make sure we follow up on fixing these lints instead of doing pylint: skip-file

benieric · 2024-06-21T06:35:00Z

Failing integs not an issue, saw these failing earlier and need to run cleaning scripts and should work during release. Everything else looks good

* Implement CuratedHub APIs (aws#1449) * Implement CuratedHub Admin APIs * making some parameters optional in create_hub_content_reference as per the API design * add describe_hub and list_hubs APIs * implement delete_hub API * Implement list_hub_contents API * create CuratedHub class and supported utils * implement list_models and address comments * Add unit tests * add describe_model function * cache retrieval for describeHubContent changes * fix curated hub class unit tests * add utils needed for curatedHub * Cache retrieval * implement get_hub_model_reference() * cleanup HUB type datatype * cleanup constants * rename list_public_models to list_jumpstart_service_hub_models * implement describe_model_reference * Rename CuratedHub to Hub * address nit * address nits and fix failing tests --------- Co-authored-by: Malav Shastri <malavhs@amazon.com> * feat: implement list_jumpstart_service_hub_models function to fetch JumpStart public hub models (aws#1456) * Implement CuratedHub Admin APIs * making some parameters optional in create_hub_content_reference as per the API design * add describe_hub and list_hubs APIs * implement delete_hub API * Implement list_hub_contents API * create CuratedHub class and supported utils * implement list_models and address comments * Add unit tests * add describe_model function * cache retrieval for describeHubContent changes * fix curated hub class unit tests * add utils needed for curatedHub * Cache retrieval * implement get_hub_model_reference() * cleanup HUB type datatype * cleanup constants * rename list_public_models to list_jumpstart_service_hub_models * implement describe_model_reference * Rename CuratedHub to Hub * address nit * address nits and fix failing tests * implement list_jumpstart_service_hub_models function --------- Co-authored-by: Malav Shastri <malavhs@amazon.com> * Feat/Curated Hub hub_arn and hub_content_type support (aws#1453) * get_model_spec() changes to support hub_arn and hub_content_type * implement get_hub_model_reference() * support hub_arn and hub_content_type for specs retrieval * add support for hub_arn and hub_content_type for serializers, deserializers, estimators, models, predictors and various spec retrieval functionalities * address nits and test failures * remove hub_content_type support --------- Co-authored-by: Malav Shastri <malavhs@amazon.com> * feat: implement curated hub parser and bug bash fixes (aws#1457) * implement HubContentDocument parser * modify the parser to remove aliases for hubcontent documents * bug fix * update boto3 * Bug Fix in the parser * Improve Hub Class and related functionalities * Bug Fix and parser updates * add missing hub_arn support * Add model reference deployment support and other minor bug fixes * fix: retrieve correct image_uri (parser update) * fix: retrieve correct model URI and model data path from HubContentDocument (parser update) * Add model reference deployment support * Model accessor and cache retrival bug fixes * fix: curated hub model training workflow * fix: pass sagemaker sessions object to retrieve model specs from describe_hub_content call * fix: fix payload retrieval for curated hub models * modify constants, enums * fix: update parser * Address nits in the parser * Add unit tests for parser * implement pagination for list_models utility * feat: support wildcard chars for model versions * Address nits and comments * Add Hub Content Arn Tag to training and hosting * Add Hub Content Arn Tag to training and hosting * fix: HubContentDocument schema version * fix broken unit tests * fix prepare_container_def unit tests to include ModelReferenceArn * fix unit tests for test_session.py * revert boto version changes * Fix unit tests * support wildcard model versions for training workflow * Add test cases for get_model_versions * Add/fix unit tests --------- Co-authored-by: Malav Shastri <malavhs@amazon.com> * address unit tests failures in codebuild * change list_jumpstart_service_hub_models to list_sagemaker_public_hub_models() * fix: Changing list input output shapes * fix: gated model training bug * run black -l 100 * flake 8 * address formatting issues * black -l * DocStyle issues * address flake8, pylint * blake -l * pass model type down * disabling pylint for release * disable pylint --------- Co-authored-by: Malav Shastri <malavhs@amazon.com> Co-authored-by: chrstfu <chrstfu@amazon.com> Co-authored-by: Erick Benitez-Ramos <141277478+benieric@users.noreply.github.com>

malav-shastri and others added 4 commits June 20, 2024 12:30

malav-shastri requested a review from a team as a code owner June 20, 2024 16:39

malav-shastri requested a review from knikure June 20, 2024 16:39

malav-shastri temporarily deployed to auto-approve June 20, 2024 16:39 — with GitHub Actions Inactive

address unit tests failures in codebuild

8af8d92

malav-shastri temporarily deployed to auto-approve June 20, 2024 19:58 — with GitHub Actions Inactive

change list_jumpstart_service_hub_models to list_sagemaker_public_hub…

6e8550d

…_models()

malav-shastri temporarily deployed to auto-approve June 20, 2024 20:03 — with GitHub Actions Inactive

fix: Changing list input output shapes

156b724

chrstfu temporarily deployed to auto-approve June 20, 2024 23:02 — with GitHub Actions Inactive

fix: gated model training bug

89b70fe

malav-shastri temporarily deployed to auto-approve June 21, 2024 00:25 — with GitHub Actions Inactive

Malav Shastri added 2 commits June 20, 2024 20:44

run black -l 100

20419c1

flake 8

f94a15b

malav-shastri temporarily deployed to auto-approve June 21, 2024 02:04 — with GitHub Actions Inactive

malav-shastri requested a review from AWS-pratab June 21, 2024 02:05

AWS-pratab approved these changes Jun 21, 2024

View reviewed changes

Merge branch 'master' into curated_hub_ga

10c5d92

benieric temporarily deployed to auto-approve June 21, 2024 02:56 — with GitHub Actions Inactive

address formatting issues

552f1f4

malav-shastri temporarily deployed to auto-approve June 21, 2024 04:23 — with GitHub Actions Inactive

AWS-pratab approved these changes Jun 21, 2024

View reviewed changes

black -l

97633cb

malav-shastri temporarily deployed to auto-approve June 21, 2024 04:34 — with GitHub Actions Inactive

DocStyle issues

7ee4a1a

malav-shastri temporarily deployed to auto-approve June 21, 2024 04:36 — with GitHub Actions Inactive

address flake8, pylint

608c62c

malav-shastri temporarily deployed to auto-approve June 21, 2024 04:53 — with GitHub Actions Inactive

blake -l

f27e5ea

malav-shastri temporarily deployed to auto-approve June 21, 2024 04:59 — with GitHub Actions Inactive

pass model type down

2495538

malav-shastri temporarily deployed to auto-approve June 21, 2024 05:24 — with GitHub Actions Inactive

disabling pylint for release

12d7cf7

malav-shastri temporarily deployed to auto-approve June 21, 2024 05:33 — with GitHub Actions Inactive

disable pylint

1433430

malav-shastri temporarily deployed to auto-approve June 21, 2024 05:42 — with GitHub Actions Inactive

benieric reviewed Jun 21, 2024

View reviewed changes

benieric approved these changes Jun 21, 2024

View reviewed changes

benieric merged commit 3f9acac into aws:master Jun 21, 2024
10 of 11 checks passed

malav-shastri mentioned this pull request Jun 27, 2024

feat: Curated hub improvements #4760

Open

9 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

JumpStart CuratedHub Launch #4748

JumpStart CuratedHub Launch #4748

malav-shastri commented Jun 20, 2024 •

edited

Loading

benieric commented Jun 20, 2024

AWS-pratab left a comment

AWS-pratab Jun 21, 2024

AWS-pratab Jun 21, 2024

AWS-pratab Jun 21, 2024

benieric Jun 21, 2024

benieric commented Jun 21, 2024

JumpStart CuratedHub Launch #4748

JumpStart CuratedHub Launch #4748

Conversation

malav-shastri commented Jun 20, 2024 • edited Loading

Merge Checklist

General

Tests

benieric commented Jun 20, 2024

AWS-pratab left a comment

Choose a reason for hiding this comment

AWS-pratab Jun 21, 2024

Choose a reason for hiding this comment

AWS-pratab Jun 21, 2024

Choose a reason for hiding this comment

AWS-pratab Jun 21, 2024

Choose a reason for hiding this comment

benieric Jun 21, 2024

Choose a reason for hiding this comment

benieric commented Jun 21, 2024

malav-shastri commented Jun 20, 2024 •

edited

Loading