高度なユースケース向けにカスタム設定を使用して Gemini をファインチューニングする
コレクションでコンテンツを整理
必要に応じて、コンテンツの保存と分類を行います。
高度なパラメータを使用して Vertex AI 教師ありファインチューニングで生成 AI モデルをチューニングする。
もっと見る
このコードサンプルを含む詳細なドキュメントについては、以下をご覧ください。
コードサンプル
特に記載のない限り、このページのコンテンツはクリエイティブ・コモンズの表示 4.0 ライセンスにより使用許諾されます。コードサンプルは Apache 2.0 ライセンスにより使用許諾されます。詳しくは、Google Developers サイトのポリシーをご覧ください。Java は Oracle および関連会社の登録商標です。
[[["わかりやすい","easyToUnderstand","thumb-up"],["問題の解決に役立った","solvedMyProblem","thumb-up"],["その他","otherUp","thumb-up"]],[["わかりにくい","hardToUnderstand","thumb-down"],["情報またはサンプルコードが不正確","incorrectInformationOrSampleCode","thumb-down"],["必要な情報 / サンプルがない","missingTheInformationSamplesINeed","thumb-down"],["翻訳に関する問題","translationIssue","thumb-down"],["その他","otherDown","thumb-down"]],[],[],[],null,["Tune a Generative AI model using Vertex AI Supervised Fine-tuning with advanced parameters.\n\nExplore further\n\n\nFor detailed documentation that includes this code sample, see the following:\n\n- [Tuning API](/vertex-ai/generative-ai/docs/model-reference/tuning)\n\nCode sample \n\nPython\n\n\nBefore trying this sample, follow the Python setup instructions in the\n[Vertex AI quickstart using\nclient libraries](/vertex-ai/docs/start/client-libraries).\n\n\nFor more information, see the\n[Vertex AI Python API\nreference documentation](/python/docs/reference/aiplatform/latest).\n\n\nTo authenticate to Vertex AI, set up Application Default Credentials.\nFor more information, see\n\n[Set up authentication for a local development environment](/docs/authentication/set-up-adc-local-dev-environment).\n\n\n import time\n\n import https://cloud.google.com/python/docs/reference/vertexai/latest/\n from vertexai.tuning import https://cloud.google.com/python/docs/reference/vertexai/latest/vertexai.preview.tuning.sft.html\n\n # TODO(developer): Update and un-comment below line\n # PROJECT_ID = \"your-projecthttps://cloud.google.com/python/docs/reference/vertexai/latest/;\n vertexai.init(project=PROJECT_ID, location=\"us-central1\")\n\n # Initialize Vertex AI with your service account for BYOSA (Bring Your Own Service Account).\n # Uncomment the following and replace \"your-service-account\"\n # vertexai.init(service_account=\"your-service-account\")\n\n # Initialize Vertex AI with your CMEK (Customer-Managed Encryption Key).\n # Un-comment the following line and replace \"your-kms-key\"\n # vertexai.init(encryptiohttps://cloud.google.com/python/docs/reference/vertexai/latest/vertexai.preview.tuning.sft.htmlphttps://cloud.google.com/python/docs/reference/vertexai/latest/vertexai.preview.tuning.sft.htmly_name=\"your-kms-key\")\n\n sft_tuning_job = sft.train(\n source_model=\"gemini-2.0-flash-001\",\n # 1.5 and 2.0 models use the same JSONL format\n train_dataset=\"gs://cloud-samples-data/ai-platform/generative_ai/gemini-1_5/text/sft_train_data.jsonl\",\n # The following parameters are optional\n validation_dataset=\"gs://cloud-samples-data/ai-platform/generative_ai/gemini-1_5/text/sft_validation_data.jsonl\",\n tuned_model_display_name=\"tuned_gemini_2_0_flash\",\n # Advanced use only below. It is recommended to use auto-selection and leave them unset\n # epochs=4,\n # adapter_size=4,\n # learning_rate_multiplier=1.0,\n )\n\n # Polling for job completion\n while not sft_tuning_job.has_ended:\n time.sleep(60)\n sft_tuning_job.refresh()\n\n print(sft_tuning_job.tuned_model_name)\n print(sft_tuning_job.tuned_model_endpoint_name)\n print(sft_tuning_job.experiment)\n # Example response:\n # projects/123456789012/locations/us-cen\u003ctral1/models/1234567890@1\n # projects/123456789012/locations/us-central1/endpoints/1234567\u003e89012345\n # google.cloud.aiplatform.metadata.experiment_resources.Experiment object at 0x7b5b4ae07af0\n\nWhat's next\n\n\nTo search and filter code samples for other Google Cloud products, see the\n[Google Cloud sample browser](/docs/samples?product=generativeaionvertexai)."]]