创建任何 Vertex AI Search 应用时,系统还会自动创建默认的投放配置。对于推荐应用,服务配置的 ID 与引擎 ID 相同。
对于搜索应用,默认投放配置的 ID 为 default_search。在进行 recommend 或 search API 调用时,系统会指定默认的投放配置 ID。
[[["易于理解","easyToUnderstand","thumb-up"],["解决了我的问题","solvedMyProblem","thumb-up"],["其他","otherUp","thumb-up"]],[["很难理解","hardToUnderstand","thumb-down"],["信息或示例代码不正确","incorrectInformationOrSampleCode","thumb-down"],["没有我需要的信息/示例","missingTheInformationSamplesINeed","thumb-down"],["翻译问题","translationIssue","thumb-down"],["其他","otherDown","thumb-down"]],["最后更新时间 (UTC):2025-08-19。"],[[["\u003cp\u003eServing configs determine how media recommendations are generated and are referenced during serving time.\u003c/p\u003e\n"],["\u003cp\u003eWhile a default serving config is automatically created for each Vertex AI Search app, only media recommendation apps allow for the creation and modification of multiple serving configs.\u003c/p\u003e\n"],["\u003cp\u003eMedia recommendation serving configs can be customized with settings for boosting/burying, demotion, and diversity, which affect the recommendations provided.\u003c/p\u003e\n"],["\u003cp\u003eYou can create, edit, and delete serving configs for media recommendation apps via the Google Cloud console, under the Agent Builder page and the app's configurations tab.\u003c/p\u003e\n"],["\u003cp\u003eTo obtain recommendations tailored to specific settings, you must specify a serving config's ID in the \u003ccode\u003erecommend\u003c/code\u003e request.\u003c/p\u003e\n"]]],[],null,["# Create and manage media serving configs\n\nThis page describes serving configs for media recommendation apps and how to\ncreate, edit, and delete them. A serving config contains configurations\nthat are used when generating your media recommendations results.\n\nAbout serving configs\n---------------------\n\nServing configs are referenced at serving time to determine what recommendations\nto generate.\n\nWhen any Vertex AI Search app is created, a default serving config is\nalso automatically created.\nFor recommendations apps, the serving config's ID is the same\nas the engine ID.\nFor search apps, the default serving config's ID is\n`default_search`.\nThe default serving config ID is specified when making\n`recommend` or `search` API calls.\n\nHowever, search and custom recommendations apps don't allow you to\nmodify serving configs or create additional serving configs. Only media\nrecommendation apps have modifiable serving configs and allow multiple serving\nconfigs.\n\nFor any media recommendations app, you can create multiple serving configs with\ndifferent settings for recommendations boost/bury, demotion, and diversity.\nTo get recommendations from a specific serving config with the boost/bury,\ndemotion, and diversity settings that you need, specify that serving config's ID\nin the `recommend` request.\n\nServing config settings\n-----------------------\n\nThe following settings are available for serving configs:\n\n- **Recommendation boost/bury** . You can boost or bury media recommendations based on\n a filter and boost value. For information about boosting and burying media\n recommendations,\n [Boost and bury media recommendations](/generative-ai-app-builder/docs/media-recommendations-boost-bury).\n\n | **Note:** Boost/bury controls can only be created through the API. They can't be created using Google Cloud console.\n- **Recommendation demotion** . You can demote media recommendations based on\n viewer data or content age. For information about demotion settings and how to\n set them, see [Demote media recommendations](/generative-ai-app-builder/docs/demote-recommendations).\n\n- **Recommendation diversity** . You can diversify recommendations to reduce the\n likelihood that similar content is shown in a recommendation panel. For\n information about diversification settings and how to set them, see\n [Diversify media recommendations](/generative-ai-app-builder/docs/diversify-recommendations).\n\nCreate a serving config\n-----------------------\n\nTo create a serving config, follow these steps:\n\n1. In the Google Cloud console, go to the **AI Applications** page.\n\n [AI Applications](https://console.cloud.google.com/gen-app-builder/engines)\n2. Click the name of the app for which you will create a serving config.\n\n3. Go to the **Configurations** \\\u003e **Serving** tab.\n\n4. Click **Create new serving config**.\n\n5. Enter a name for your serving config.\n\n6. Set configurations for recommendations:\n demotion and diversity.\n\n \u003cbr /\u003e\n\n For more information about these settings, see [Demote media\n recommendations](/generative-ai-app-builder/docs/demote-recommendations), and [Diversify media\n recommendations](/generative-ai-app-builder/docs/diversify-recommendations).\n7. Click **Save and publish**.\n\nEdit a serving config\n---------------------\n\nTo edit a serving config, follow these steps:\n\n1. In the Google Cloud console, go to the **AI Applications** page.\n\n [AI Applications](https://console.cloud.google.com/gen-app-builder/engines)\n2. Click the name of the app whose serving config you want to edit.\n\n3. Go to the **Configurations** \\\u003e **Serving** tab.\n\n4. If you have more than one serving config, click the name of the serving\n config that you want to edit.\n\n5. Update your configurations for recommendation demotion or\n diversity.\n\n For more information about these settings, see\n [Demote media recommendations](/generative-ai-app-builder/docs/demote-recommendations) and\n [Diversify media recommendations](/generative-ai-app-builder/docs/diversify-recommendations).\n\n \u003cbr /\u003e\n\n6. Click **Save and publish**.\n\nDelete a serving config\n-----------------------\n\nYou can delete a serving config only if you have multiple serving configs.\n\nTo delete a serving config, follow these steps:\n\n1. In the Google Cloud console, go to the **AI Applications** page.\n\n [AI Applications](https://console.cloud.google.com/gen-app-builder/engines)\n2. Click the name of the app whose serving config you will delete.\n\n3. Go to the **Configurations** \\\u003e **Serving** tab.\n\n4. For the serving config you that want to delete, click\n more_vert, and then click **Delete**.\n\n5. In the confirmation dialog, type the name of the serving config to delete,\n and then click **Confirm**."]]