Transformer.transform() raises pydantic ValidationError ("tags: extra_forbidden") after submitting the job when tags are set

[Alex-Mesmer]

PySDK Version

• PySDK V2 (2.x)
• PySDK V3 (3.x)

Describe the bug
Transformer.transform() raises a pydantic.ValidationError for TransformJob whenever the Transformer was created with non-empty tags. The error is raised after the transform job has already been submitted to SageMaker: transform() builds and submits the CreateTransformJob request successfully, then reconstructs a local sagemaker.core.resources.TransformJob from the request dict but that resource model declares no tags field and sets extra="forbid", so the round-trip fails when the request contains tags.

Another major issue is that there are no tags attached to the created transformation job.

Consequences: the transform job is actually created on SageMaker, but the SDK call raises and the caller never gets the job handle. The failure is independent of the tag value/shape, any non-empty tags triggers it, because the tags field itself is rejected on the resource model.

A related issue, the shape of the tags between ModelTrainer and Transformer are not the same.

To reproduce
This is exactly the construction transform() performs:

from sagemaker.core.resources import TransformJob
TransformJob(transform_job_name="x", tags=[{"key": "team", "value": "ml"}])

Real-world trigger:

from sagemaker.core.transformer import Transformer
t = Transformer(model_name="my-model", instance_count=1, instance_type="ml.m5.large",
                output_path="s3://bucket/out/", tags=[{"Key": "team", "Value": "ml"}])
t.transform(data="s3://bucket/in/", content_type="text/csv")   # job submits, then raises

Actual behavior:

pydantic_core._pydantic_core.ValidationError: 1 validation error for TransformJob
tags
  Extra inputs are not permitted [type=extra_forbidden, input_value=[{'key': 'team', 'value': 'ml'}], input_type=list]
    For further information visit https://errors.pydantic.dev/2.12/v/extra_forbidden

Expected behavior
Transformer.transform() succeeds with tags set and returns the job handle; tags are applied to the transform job.

Root Cause
In sagemaker-core/src/sagemaker/core/transformer.py, Transformer.transform():

• L717: transform_request["Tags"] = tags
• L405: create_transform_job(**request) - job is submitted
• L411–L412:

transformed = transform_util(serialized_request, "CreateTransformJobRequest")
self.latest_transform_job = TransformJob(**transformed)   # sagemaker.core.resources.TransformJob

transformed includes a tags key, but sagemaker.core.resources.TransformJob has no tags/Tags field and model_config["extra"] == "forbid" → extra_forbidden.

Suggested fix
Any of: (1) add a tags field to sagemaker.core.resources.TransformJob; (2) drop tags from transformed before TransformJob(**transformed) at L412; or (3) build the post-submit local object via TransformJob.get(...) instead of TransformJob(**transformed).

Screenshots or logs
If applicable, add screenshots or logs to help explain your problem.

System information
A description of your system. Please provide:

• SageMaker Python SDK version: sagemaker==3.15.1, sagemaker-core==2.16.0
• Framework name (eg. PyTorch) or algorithm (eg. KMeans):
• Framework version:
• Python version:
• CPU or GPU:
• Custom Docker image (Y/N):

Additional context
Add any other context about the problem here.