Merge branch 'main' of github.com:oracle/accelerated-data-science into feature/ms_backup_retention

Author: iamishaan

ads/aqua/config/config.py

@@ -29,19 +29,3 @@ def get_evaluation_service_config(
         .get(ContainerSpec.CONTAINER_SPEC, {})
         .get(container, {})
     )
-
-
-# TODO: move this to global config.json in object storage
-def get_finetuning_config_defaults():
-    """Generate and return the fine-tuning default configuration dictionary."""
-    return {
-        "shape": {
-            "VM.GPU.A10.1": {"batch_size": 1, "replica": "1-10"},
-            "VM.GPU.A10.2": {"batch_size": 1, "replica": "1-10"},
-            "BM.GPU.A10.4": {"batch_size": 1, "replica": 1},
-            "BM.GPU4.8": {"batch_size": 4, "replica": 1},
-            "BM.GPU.L40S-NC.4": {"batch_size": 4, "replica": 1},
-            "BM.GPU.A100-v2.8": {"batch_size": 6, "replica": 1},
-            "BM.GPU.H100.8": {"batch_size": 6, "replica": 1},
-        }
-    }

ads/aqua/config/deployment_config_defaults.json

@@ -27,6 +27,7 @@
 LIFECYCLE_DETAILS_MISSING_JOBRUN = "The associated JobRun resource has been deleted."
 READY_TO_DEPLOY_STATUS = "ACTIVE"
 READY_TO_FINE_TUNE_STATUS = "TRUE"
+PRIVATE_ENDPOINT_TYPE = "MODEL_DEPLOYMENT"
 AQUA_GA_LIST = ["id19sfcrra6z"]
 AQUA_MODEL_TYPE_SERVICE = "service"
 AQUA_MODEL_TYPE_CUSTOM = "custom"

ads/aqua/config/resource_limit_names.json

@@ -1,5 +1,4 @@
 #!/usr/bin/env python
-# -*- coding: utf-8 -*-
 # Copyright (c) 2024 Oracle and/or its affiliates.
 # Licensed under the Universal Permissive License v 1.0 as shown at https://oss.oracle.com/licenses/upl/
 
@@ -9,8 +8,8 @@
 from tornado.web import HTTPError
 
 from ads.aqua.common.decorator import handle_exceptions
-from ads.aqua.extension.errors import Errors
 from ads.aqua.extension.base_handler import AquaAPIhandler
+from ads.aqua.extension.errors import Errors
 from ads.aqua.extension.utils import validate_function_parameters
 from ads.aqua.finetuning import AquaFineTuningApp
 from ads.aqua.finetuning.entities import CreateFineTuningDetails

ads/aqua/constants.py

@@ -1,5 +1,4 @@
 #!/usr/bin/env python
-# -*- coding: utf-8 -*-
 # Copyright (c) 2024 Oracle and/or its affiliates.
 # Licensed under the Universal Permissive License v 1.0 as shown at https://oss.oracle.com/licenses/upl/
 
@@ -10,8 +9,9 @@
 
 from ads.aqua.common.decorator import handle_exceptions
 from ads.aqua.common.enums import Tags
-from ads.aqua.extension.errors import Errors
+from ads.aqua.constants import PRIVATE_ENDPOINT_TYPE
 from ads.aqua.extension.base_handler import AquaAPIhandler
+from ads.aqua.extension.errors import Errors
 from ads.aqua.extension.utils import validate_function_parameters
 from ads.aqua.model.entities import ImportModelDetails
 from ads.aqua.ui import AquaUIApp
@@ -72,6 +72,8 @@ def get(self, id=""):
             return self.list_vcn()
         elif paths.startswith("aqua/subnets"):
             return self.list_subnets()
+        elif paths.startswith("aqua/privateendpoints"):
+            return self.list_private_endpoints()
         elif paths.startswith("aqua/shapes/limit"):
             return self.get_shape_availability()
         elif paths.startswith("aqua/bucket/versioning"):
@@ -175,15 +177,31 @@ def list_subnets(self, **kwargs):
             )
         )
 
+    def list_private_endpoints(self, **kwargs):
+        """Lists the private endpoints in the specified compartment."""
+        compartment_id = self.get_argument("compartment_id", default=COMPARTMENT_OCID)
+        resource_type = self.get_argument(
+            "resource_type", default=PRIVATE_ENDPOINT_TYPE
+        )
+        return self.finish(
+            AquaUIApp().list_private_endpoints(
+                compartment_id=compartment_id, resource_type=resource_type, **kwargs
+            )
+        )
+
     def get_shape_availability(self, **kwargs):
         """For a given compartmentId, resource limit name, and scope, returns the number of available resources associated
         with the given limit."""
         compartment_id = self.get_argument("compartment_id", default=COMPARTMENT_OCID)
         instance_shape = self.get_argument("instance_shape")
+        limit_name = self.get_argument("limit_name")
 
         return self.finish(
             AquaUIApp().get_shape_availability(
-                compartment_id=compartment_id, instance_shape=instance_shape, **kwargs
+                compartment_id=compartment_id,
+                instance_shape=instance_shape,
+                limit_name=limit_name,
+                **kwargs,
             )
         )
 
@@ -244,6 +262,7 @@ def post(self, *args, **kwargs):
     ("job/shapes/?([^/]*)", AquaUIHandler),
     ("vcn/?([^/]*)", AquaUIHandler),
     ("subnets/?([^/]*)", AquaUIHandler),
+    ("privateendpoints/?([^/]*)", AquaUIHandler),
     ("shapes/limit/?([^/]*)", AquaUIHandler),
     ("bucket/versioning/?([^/]*)", AquaUIHandler),
     ("containers/?([^/]*)", AquaUIHandler),

ads/aqua/extension/finetune_handler.py

@@ -1,5 +1,4 @@
 #!/usr/bin/env python
-# -*- coding: utf-8 -*-
 # Copyright (c) 2024 Oracle and/or its affiliates.
 # Licensed under the Universal Permissive License v 1.0 as shown at https://oss.oracle.com/licenses/upl/
 from dataclasses import dataclass, field
@@ -14,16 +13,18 @@ class AquaFineTuningParams(DataClassSerializable):
     epochs: int
     learning_rate: Optional[float] = None
     sample_packing: Optional[bool] = "auto"
-    batch_size: Optional[
-        int
-    ] = None  # make it batch_size for user, but internally this is micro_batch_size
+    batch_size: Optional[int] = (
+        None  # make it batch_size for user, but internally this is micro_batch_size
+    )
     sequence_len: Optional[int] = None
     pad_to_sequence_len: Optional[bool] = None
     lora_r: Optional[int] = None
     lora_alpha: Optional[int] = None
     lora_dropout: Optional[float] = None
     lora_target_linear: Optional[bool] = None
     lora_target_modules: Optional[List] = None
+    early_stopping_patience: Optional[int] = None
+    early_stopping_threshold: Optional[float] = None
 
 
 @dataclass(repr=False)

ads/aqua/extension/ui_handler.py

@@ -1,11 +1,10 @@
 #!/usr/bin/env python
-# -*- coding: utf-8 -*-
 # Copyright (c) 2024 Oracle and/or its affiliates.
 # Licensed under the Universal Permissive License v 1.0 as shown at https://oss.oracle.com/licenses/upl/
 
 import json
 import os
-from dataclasses import asdict, fields, MISSING
+from dataclasses import MISSING, asdict, fields
 from typing import Dict
 
 from oci.data_science.models import (
@@ -14,7 +13,7 @@
     UpdateModelProvenanceDetails,
 )
 
-from ads.aqua import ODSC_MODEL_COMPARTMENT_OCID, logger
+from ads.aqua import logger
 from ads.aqua.app import AquaApp
 from ads.aqua.common.enums import Resource, Tags
 from ads.aqua.common.errors import AquaFileExistsError, AquaValueError
@@ -31,7 +30,6 @@
     UNKNOWN,
     UNKNOWN_DICT,
 )
-from ads.aqua.config.config import get_finetuning_config_defaults
 from ads.aqua.data import AquaResourceIdentifier
 from ads.aqua.finetuning.constants import *
 from ads.aqua.finetuning.entities import *
@@ -132,7 +130,7 @@ def create(
             or create_fine_tuning_details.validation_set_size >= 1
         ):
             raise AquaValueError(
-                f"Fine tuning validation set size should be a float number in between [0, 1)."
+                "Fine tuning validation set size should be a float number in between [0, 1)."
             )
 
         if create_fine_tuning_details.replica < DEFAULT_FT_REPLICA:
@@ -394,7 +392,7 @@ def create(
         )
         # track shapes that were used for fine-tune creation
         self.telemetry.record_event_async(
-            category=f"aqua/service/finetune/create/shape/",
+            category="aqua/service/finetune/create/shape/",
             action=f"{create_fine_tuning_details.shape_name}x{create_fine_tuning_details.replica}",
             **telemetry_kwargs,
         )
@@ -533,6 +531,12 @@ def _build_oci_launch_cmd(
                     oci_launch_cmd += f"--num_{key} {value} "
                 elif key == "lora_target_modules":
                     oci_launch_cmd += f"--{key} {','.join(str(k) for k in value)} "
+                elif key == "early_stopping_patience":
+                    if value != 0:
+                        oci_launch_cmd += f"--{key} {value} "
+                elif key == "early_stopping_threshold":
+                    if "early_stopping_patience" in oci_launch_cmd:
+                        oci_launch_cmd += f"--{key} {value} "
                 else:
                     oci_launch_cmd += f"--{key} {value} "
 
@@ -558,8 +562,9 @@ def get_finetuning_config(self, model_id: str) -> Dict:
 
         config = self.get_config(model_id, AQUA_MODEL_FINETUNING_CONFIG)
         if not config:
-            logger.info(f"Fetching default fine-tuning config for model: {model_id}")
-            config = get_finetuning_config_defaults()
+            logger.debug(
+                f"Fine-tuning config for custom model: {model_id} is not available."
+            )
         return config
 
     @telemetry(