feat(genai): Update tuning samples to include automatic evaluations (#13550)

* feat: Update create tuning job to include tuning with automatic evaluation

* Update genai/tuning/tuning_job_create.py

Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>

* Update genai/tuning/tuning_with_checkpoints_create.py

Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>

* Update genai/tuning/tuning_with_checkpoints_create.py

Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>

* Update genai/tuning/tuning_job_create.py

Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>

* Update tuning_job_create.py

* Update tuning_with_checkpoints_create.py

* Update tuning_job_create.py

* Update tuning_with_checkpoints_create.py

* Update requirements.txt

Update GenAI SDK version

* Update test_tuning_examples.py

* Update test_tuning_examples.py

* Update test_tuning_examples.py

* Update test_tuning_examples.py

* Update tuning_with_checkpoints_create.py

* Update tuning_job_create.py

* Update test_tuning_examples.py

* Update test_tuning_examples.py

* Update test_tuning_examples.py

---------

Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>

Author: lakeyk

genai/tuning/requirements.txt

@@ -1 +1 @@
-google-genai==1.27.0
+google-genai==1.30.0

genai/tuning/test_tuning_examples.py

@@ -12,9 +12,13 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+from datetime import datetime as dt
+
 from unittest.mock import call, MagicMock, patch
 
+from google.cloud import storage
 from google.genai import types
+import pytest
 
 import tuning_job_create
 import tuning_job_get
@@ -27,8 +31,24 @@
 import tuning_with_checkpoints_textgen_with_txt
 
 
+GCS_OUTPUT_BUCKET = "python-docs-samples-tests"
+
+
+@pytest.fixture(scope="session")
+def output_gcs_uri() -> str:
+    prefix = f"text_output/{dt.now()}"
+
+    yield f"gs://{GCS_OUTPUT_BUCKET}/{prefix}"
+
+    storage_client = storage.Client()
+    bucket = storage_client.get_bucket(GCS_OUTPUT_BUCKET)
+    blobs = bucket.list_blobs(prefix=prefix)
+    for blob in blobs:
+        blob.delete()
+
+
 @patch("google.genai.Client")
-def test_tuning_job_create(mock_genai_client: MagicMock) -> None:
+def test_tuning_job_create(mock_genai_client: MagicMock, output_gcs_uri: str) -> None:
     # Mock the API response
     mock_tuning_job = types.TuningJob(
         name="test-tuning-job",
@@ -40,9 +60,9 @@ def test_tuning_job_create(mock_genai_client: MagicMock) -> None:
     )
     mock_genai_client.return_value.tunings.tune.return_value = mock_tuning_job
 
-    response = tuning_job_create.create_tuning_job()
+    response = tuning_job_create.create_tuning_job(output_gcs_uri=output_gcs_uri)
 
-    mock_genai_client.assert_called_once_with(http_options=types.HttpOptions(api_version="v1"))
+    mock_genai_client.assert_called_once_with(http_options=types.HttpOptions(api_version="v1beta1"))
     mock_genai_client.return_value.tunings.tune.assert_called_once()
     assert response == "test-tuning-job"
 
@@ -121,7 +141,7 @@ def test_tuning_textgen_with_txt(mock_genai_client: MagicMock) -> None:
 
 
 @patch("google.genai.Client")
-def test_tuning_job_create_with_checkpoints(mock_genai_client: MagicMock) -> None:
+def test_tuning_job_create_with_checkpoints(mock_genai_client: MagicMock, output_gcs_uri: str) -> None:
     # Mock the API response
     mock_tuning_job = types.TuningJob(
         name="test-tuning-job",
@@ -137,9 +157,9 @@ def test_tuning_job_create_with_checkpoints(mock_genai_client: MagicMock) -> Non
     )
     mock_genai_client.return_value.tunings.tune.return_value = mock_tuning_job
 
-    response = tuning_with_checkpoints_create.create_with_checkpoints()
+    response = tuning_with_checkpoints_create.create_with_checkpoints(output_gcs_uri=output_gcs_uri)
 
-    mock_genai_client.assert_called_once_with(http_options=types.HttpOptions(api_version="v1"))
+    mock_genai_client.assert_called_once_with(http_options=types.HttpOptions(api_version="v1beta1"))
     mock_genai_client.return_value.tunings.tune.assert_called_once()
     assert response == "test-tuning-job"
 

genai/tuning/tuning_job_create.py

@@ -13,24 +13,46 @@
 # limitations under the License.
 
 
-def create_tuning_job() -> str:
+def create_tuning_job(output_gcs_uri: str) -> str:
     # [START googlegenaisdk_tuning_job_create]
     import time
 
     from google import genai
-    from google.genai.types import HttpOptions, CreateTuningJobConfig, TuningDataset
+    from google.genai.types import HttpOptions, CreateTuningJobConfig, TuningDataset, EvaluationConfig, OutputConfig, GcsDestination, Metric
 
-    client = genai.Client(http_options=HttpOptions(api_version="v1"))
+    # TODO(developer): Update and un-comment below line
+    # output_gcs_uri = "gs://your-bucket/your-prefix"
+
+    client = genai.Client(http_options=HttpOptions(api_version="v1beta1"))
 
     training_dataset = TuningDataset(
         gcs_uri="gs://cloud-samples-data/ai-platform/generative_ai/gemini/text/sft_train_data.jsonl",
     )
+    validation_dataset = TuningDataset(
+        gcs_uri="gs://cloud-samples-data/ai-platform/generative_ai/gemini/text/sft_validation_data.jsonl",
+    )
+
+    evaluation_config = EvaluationConfig(
+        metrics=[
+            Metric(
+                name="FLUENCY",
+                prompt_template="""Evaluate this {response}"""
+            )
+        ],
+        output_config=OutputConfig(
+            gcs_destination=GcsDestination(
+                output_uri_prefix=output_gcs_uri,
+            )
+        ),
+    )
 
     tuning_job = client.tunings.tune(
         base_model="gemini-2.5-flash",
         training_dataset=training_dataset,
         config=CreateTuningJobConfig(
             tuned_model_display_name="Example tuning job",
+            validation_dataset=validation_dataset,
+            evaluation_config=evaluation_config,
         ),
     )
 
@@ -64,4 +86,4 @@ def create_tuning_job() -> str:
 
 
 if __name__ == "__main__":
-    create_tuning_job()
+    create_tuning_job(output_gcs_uri="gs://your-bucket/your-prefix")

genai/tuning/tuning_with_checkpoints_create.py

@@ -13,18 +13,38 @@
 # limitations under the License.
 
 
-def create_with_checkpoints() -> str:
+def create_with_checkpoints(output_gcs_uri: str) -> str:
     # [START googlegenaisdk_tuning_with_checkpoints_create]
     import time
 
     from google import genai
-    from google.genai.types import HttpOptions, CreateTuningJobConfig, TuningDataset
+    from google.genai.types import HttpOptions, CreateTuningJobConfig, TuningDataset, EvaluationConfig, OutputConfig, GcsDestination, Metric
 
-    client = genai.Client(http_options=HttpOptions(api_version="v1"))
+    # TODO(developer): Update and un-comment below line
+    # output_gcs_uri = "gs://your-bucket/your-prefix"
+
+    client = genai.Client(http_options=HttpOptions(api_version="v1beta1"))
 
     training_dataset = TuningDataset(
         gcs_uri="gs://cloud-samples-data/ai-platform/generative_ai/gemini/text/sft_train_data.jsonl",
     )
+    validation_dataset = TuningDataset(
+        gcs_uri="gs://cloud-samples-data/ai-platform/generative_ai/gemini/text/sft_validation_data.jsonl",
+    )
+
+    evaluation_config = EvaluationConfig(
+        metrics=[
+            Metric(
+                name="FLUENCY",
+                prompt_template="""Evaluate this {response}"""
+            )
+        ],
+        output_config=OutputConfig(
+            gcs_destination=GcsDestination(
+                output_uri_prefix=output_gcs_uri,
+            )
+        ),
+    )
 
     tuning_job = client.tunings.tune(
         base_model="gemini-2.5-flash",
@@ -33,6 +53,8 @@ def create_with_checkpoints() -> str:
             tuned_model_display_name="Example tuning job",
             # Set to True to disable tuning intermediate checkpoints. Default is False.
             export_last_checkpoint_only=False,
+            validation_dataset=validation_dataset,
+            evaluation_config=evaluation_config,
         ),
     )
 
@@ -66,4 +88,4 @@ def create_with_checkpoints() -> str:
 
 
 if __name__ == "__main__":
-    create_with_checkpoints()
+    create_with_checkpoints(output_gcs_uri="gs://your-bucket/your-prefix")