Add MaaP Finetuning Job implementation.

sdk/ml/azure-ai-ml/azure/ai/ml/_ml_client.py

@@ -709,6 +709,7 @@ def __init__(
             _service_client_kwargs=kwargs,
             requests_pipeline=self._requests_pipeline,
             service_client_01_2024_preview=self._service_client_01_2024_preview,
+            service_client_10_2024_preview=self._service_client_10_2024_preview,
             **ops_kwargs,
         )
         self._operation_container.add(AzureMLResourceType.JOB, self._jobs)
@@ -746,7 +747,8 @@ def __init__(
                 **ops_kwargs,  # type: ignore[arg-type]
             )
             self._operation_container.add(
-                AzureMLResourceType.VIRTUALCLUSTER, self._virtual_clusters  # type: ignore[arg-type]
+                AzureMLResourceType.VIRTUALCLUSTER,
+                self._virtual_clusters,  # type: ignore[arg-type]
             )
         except Exception as ex:  # pylint: disable=broad-except
             module_logger.debug("Virtual Cluster operations could not be initialized due to %s ", ex)

sdk/ml/azure-ai-ml/azure/ai/ml/_schema/_finetuning/finetuning_job.py

@@ -7,10 +7,17 @@
 from azure.ai.ml._schema.job import BaseJobSchema
 from azure.ai.ml._schema.job.input_output_fields_provider import OutputsField
 from azure.ai.ml._utils._experimental import experimental
+from azure.ai.ml._schema.core.fields import (
+    NestedField,
+)
+from ..queue_settings import QueueSettingsSchema
+from ..job_resources import JobResourcesSchema
 
 # This is meant to match the yaml definition NOT the models defined in _restclient
 
 
 @experimental
 class FineTuningJobSchema(BaseJobSchema):
     outputs = OutputsField()
+    queue_settings = NestedField(QueueSettingsSchema)
+    resources = NestedField(JobResourcesSchema)

sdk/ml/azure-ai-ml/azure/ai/ml/_schema/job_resources.py

@@ -0,0 +1,21 @@
+# ---------------------------------------------------------
+# Copyright (c) Microsoft Corporation. All rights reserved.
+# ---------------------------------------------------------
+
+# pylint: disable=unused-argument
+
+from marshmallow import fields, post_load
+
+from azure.ai.ml._schema.core.schema_meta import PatchedSchemaMeta
+
+
+class JobResourcesSchema(metaclass=PatchedSchemaMeta):
+    instance_types = fields.List(
+        fields.Str(), metadata={"description": "The instance type to make available to this job."}
+    )
+
+    @post_load
+    def make(self, data, **kwargs):
+        from azure.ai.ml.entities import JobResources
+
+        return JobResources(**data)

sdk/ml/azure-ai-ml/azure/ai/ml/constants/_finetuning.py

@@ -0,0 +1,17 @@
+# ---------------------------------------------------------
+# Copyright (c) Microsoft Corporation. All rights reserved.
+# ---------------------------------------------------------
+
+
+class FineTuningTaskType:
+    CHAT_COMPLETION = "ChatCompletion"
+    TEXT_COMPLETION = "TextCompletion"
+    TEXT_CLASSIFICATION = "TextClassification"
+    QUESTION_ANSWERING = "QuestionAnswering"
+    TEXT_SUMMARIZATION = "TextSummarization"
+    TOKEN_CLASSIFICATION = "TokenClassification"
+    TEXT_TRANSLATION = "TextTranslation"
+    IMAGE_CLASSIFICATION = "ImageClassification"
+    IMAGE_INSTANCE_SEGMENTATION = "ImageInstanceSegmentation"
+    IMAGE_OBJECT_DETECTION = "ImageObjectDetection"
+    VIDEO_MULTI_OBJECT_TRACKING = "VideoMultiObjectTracking"

sdk/ml/azure-ai-ml/azure/ai/ml/entities/__init__.py

@@ -52,15 +52,29 @@
 from ._component.pipeline_component import PipelineComponent
 from ._component.spark_component import SparkComponent
 from ._compute._aml_compute_node_info import AmlComputeNodeInfo
-from ._compute._custom_applications import CustomApplications, EndpointsSettings, ImageSettings, VolumeSettings
+from ._compute._custom_applications import (
+    CustomApplications,
+    EndpointsSettings,
+    ImageSettings,
+    VolumeSettings,
+)
 from ._compute._image_metadata import ImageMetadata
-from ._compute._schedule import ComputePowerAction, ComputeSchedules, ComputeStartStopSchedule, ScheduleState
+from ._compute._schedule import (
+    ComputePowerAction,
+    ComputeSchedules,
+    ComputeStartStopSchedule,
+    ScheduleState,
+)
 from ._compute._setup_scripts import ScriptReference, SetupScripts
 from ._compute._usage import Usage, UsageName
 from ._compute._vm_size import VmSize
 from ._compute.aml_compute import AmlCompute, AmlComputeSshSettings
 from ._compute.compute import Compute, NetworkSettings
-from ._compute.compute_instance import AssignedUserConfiguration, ComputeInstance, ComputeInstanceSshSettings
+from ._compute.compute_instance import (
+    AssignedUserConfiguration,
+    ComputeInstance,
+    ComputeInstanceSshSettings,
+)
 from ._compute.kubernetes_compute import KubernetesCompute
 from ._compute.synapsespark_compute import AutoPauseSettings, AutoScaleSettings, SynapseSparkCompute
 from ._compute.unsupported_compute import UnsupportedCompute
@@ -84,7 +98,11 @@
 from ._data_import.data_import import DataImport
 from ._data_import.schedule import ImportDataSchedule
 from ._datastore.adls_gen1 import AzureDataLakeGen1Datastore
-from ._datastore.azure_storage import AzureBlobDatastore, AzureDataLakeGen2Datastore, AzureFileDatastore
+from ._datastore.azure_storage import (
+    AzureBlobDatastore,
+    AzureDataLakeGen2Datastore,
+    AzureFileDatastore,
+)
 from ._datastore.datastore import Datastore
 from ._datastore.one_lake import OneLakeArtifact, OneLakeDatastore
 from ._deployment.batch_deployment import BatchDeployment
@@ -94,7 +112,11 @@
 from ._deployment.data_asset import DataAsset
 from ._deployment.data_collector import DataCollector
 from ._deployment.deployment_collection import DeploymentCollection
-from ._deployment.deployment_settings import BatchRetrySettings, OnlineRequestSettings, ProbeSettings
+from ._deployment.deployment_settings import (
+    BatchRetrySettings,
+    OnlineRequestSettings,
+    ProbeSettings,
+)
 from ._deployment.model_batch_deployment import ModelBatchDeployment
 from ._deployment.model_batch_deployment_settings import ModelBatchDeploymentSettings
 from ._deployment.online_deployment import (
@@ -106,7 +128,11 @@
 from ._deployment.pipeline_component_batch_deployment import PipelineComponentBatchDeployment
 from ._deployment.request_logging import RequestLogging
 from ._deployment.resource_requirements_settings import ResourceRequirementsSettings
-from ._deployment.scale_settings import DefaultScaleSettings, OnlineScaleSettings, TargetUtilizationScaleSettings
+from ._deployment.scale_settings import (
+    DefaultScaleSettings,
+    OnlineScaleSettings,
+    TargetUtilizationScaleSettings,
+)
 from ._endpoint.batch_endpoint import BatchEndpoint
 from ._endpoint.endpoint import Endpoint
 from ._endpoint.online_endpoint import (
@@ -136,11 +162,19 @@
 from ._indexes import ModelConfiguration as IndexModelConfiguration
 from ._job.command_job import CommandJob
 from ._job.compute_configuration import ComputeConfiguration
+from ._job.finetuning.custom_model_finetuning_job import CustomModelFineTuningJob
 from ._job.input_port import InputPort
 from ._job.job import Job
 from ._job.job_limits import CommandJobLimits
+from ._job.job_resources import JobResources
 from ._job.job_resource_configuration import JobResourceConfiguration
-from ._job.job_service import JobService, JupyterLabJobService, SshJobService, TensorBoardJobService, VsCodeJobService
+from ._job.job_service import (
+    JobService,
+    JupyterLabJobService,
+    SshJobService,
+    TensorBoardJobService,
+    VsCodeJobService,
+)
 from ._job.parallel.parallel_task import ParallelTask
 from ._job.parallel.retry_settings import RetrySettings
 from ._job.parameterized_command import ParameterizedCommand
@@ -156,7 +190,12 @@
 from ._monitoring.alert_notification import AlertNotification
 from ._monitoring.compute import ServerlessSparkCompute
 from ._monitoring.definition import MonitorDefinition
-from ._monitoring.input_data import FixedInputData, MonitorInputData, StaticInputData, TrailingInputData
+from ._monitoring.input_data import (
+    FixedInputData,
+    MonitorInputData,
+    StaticInputData,
+    TrailingInputData,
+)
 from ._monitoring.schedule import MonitorSchedule
 from ._monitoring.signals import (
     BaselineDataRange,
@@ -244,7 +283,11 @@
 from ._workspace.private_endpoint import EndpointConnection, PrivateEndpoint
 from ._workspace.serverless_compute import ServerlessComputeSettings
 from ._workspace.workspace import Workspace
-from ._workspace.workspace_keys import ContainerRegistryCredential, NotebookAccessKeys, WorkspaceKeys
+from ._workspace.workspace_keys import (
+    ContainerRegistryCredential,
+    NotebookAccessKeys,
+    WorkspaceKeys,
+)
 
 __all__ = [
     "Resource",
@@ -258,8 +301,10 @@
     "SparkJobEntryType",
     "CommandJobLimits",
     "ComputeConfiguration",
+    "CustomModelFineTuningJob",
     "CreatedByType",
     "ResourceConfiguration",
+    "JobResources",
     "JobResourceConfiguration",
     "QueueSettings",
     "JobService",

sdk/ml/azure-ai-ml/azure/ai/ml/entities/_job/_input_output_helpers.py

@@ -205,8 +205,7 @@ def validate_pipeline_input_key_characters(key: str) -> None:
     # so a valid pipeline key is: ^{single_key}([.]{single_key})*$
     if re.match(IOConstants.VALID_KEY_PATTERN, key) is None:
         msg = (
-            "Pipeline input key name {} must be composed letters, numbers, and underscores with optional "
-            "split by dots."
+            "Pipeline input key name {} must be composed letters, numbers, and underscores with optional split by dots."
         )
         raise ValidationException(
             message=msg.format(key),
@@ -262,7 +261,6 @@ def to_rest_dataset_literal_inputs(
                             uri=input_value.path,
                             mode=(INPUT_MOUNT_MAPPING_TO_REST[input_value.mode.lower()] if input_value.mode else None),
                         )
-
                     else:
                         msg = f"Job input type {input_value.type} is not supported as job input."
                         raise ValidationException(
@@ -415,7 +413,7 @@ def from_rest_data_outputs(outputs: Dict[str, RestJobOutput]) -> Dict[str, Outpu
                 path_on_compute=sourcePathOnCompute,
                 description=output_value.description,
                 name=output_value.asset_name,
-                version=output_value.asset_version,
+                version=(output_value.asset_version if hasattr(output_value, "asset_version") else None),
             )
         else:
             msg = "unsupported JobOutput type: {}".format(output_value.job_output_type)

sdk/ml/azure-ai-ml/azure/ai/ml/entities/_job/finetuning/custom_model_finetuning_job.py

@@ -6,7 +6,7 @@
 
 from typing import Any, Dict
 
-from azure.ai.ml._restclient.v2024_01_01_preview.models import (
+from azure.ai.ml._restclient.v2024_10_01_preview.models import (
     ModelProvider as RestModelProvider,
     CustomModelFineTuning as RestCustomModelFineTuningVertical,
     FineTuningJob as RestFineTuningJob,
@@ -16,6 +16,8 @@
     from_rest_data_outputs,
     to_rest_data_outputs,
 )
+from azure.ai.ml.entities._job.job_resources import JobResources
+from azure.ai.ml.entities._job.queue_settings import QueueSettings
 from azure.ai.ml.entities._inputs_outputs import Input
 from azure.ai.ml.constants._common import BASE_PATH_CONTEXT_KEY
 from azure.ai.ml.entities._job.finetuning.finetuning_vertical import FineTuningVertical
@@ -89,9 +91,14 @@ def _to_rest_object(self) -> "RestFineTuningJob":
             experiment_name=self.experiment_name,
             tags=self.tags,
             properties=self.properties,
+            compute_id=self.compute,
             fine_tuning_details=custom_finetuning_vertical,
             outputs=to_rest_data_outputs(self.outputs),
         )
+        if self.resources:
+            finetuning_job.resources = self.resources._to_rest_object()
+        if self.queue_settings:
+            finetuning_job.queue_settings = self.queue_settings._to_rest_object()
 
         result = RestJobBase(properties=finetuning_job)
         result.name = self.name
@@ -166,9 +173,15 @@ def _from_rest_object(cls, obj: RestJobBase) -> "CustomModelFineTuningJob":
             "status": properties.status,
             "creation_context": obj.system_data,
             "display_name": properties.display_name,
+            "compute": properties.compute_id,
             "outputs": from_rest_data_outputs(properties.outputs),
         }
 
+        if properties.resources:
+            job_args_dict["resources"] = JobResources._from_rest_object(properties.resources)
+        if properties.queue_settings:
+            job_args_dict["queue_settings"] = QueueSettings._from_rest_object(properties.queue_settings)
+
         custom_model_finetuning_job = cls(
             task=finetuning_details.task_type,
             model=finetuning_details.model,