如要在「選取模型」選項中搭配 MediaPipe 使用自訂模型,請選擇「Choose a model file」(選擇模型檔案),並從檔案儲存空間中選取模型,如螢幕截圖所示。
您選擇的模型必須符合所用 MediaPipe Tasks API 的模型輸入和輸出要求,並包含相容的模型中繼資料。如要建立自己的模型並與 MediaPipe Tasks API 搭配使用,最快的方法就是使用 MediaPipe Model Maker 工具,以自己的資料修改相容的解決方案模型。詳情請參閱 MediaPipe 模型製作工具。
自訂輸入資料
在 MediaPipe Studio 的每個解決方案頁面中,您都可以使用自己的資料。如果是文字工作,您可以在提供的欄位中輸入文字。Vision 工作可讓您使用網路攝影機做為輸入,也可以使用下拉式選單中的「Choose an image file」選項來上傳圖片,如下所示:
[[["容易理解","easyToUnderstand","thumb-up"],["確實解決了我的問題","solvedMyProblem","thumb-up"],["其他","otherUp","thumb-up"]],[["缺少我需要的資訊","missingTheInformationINeed","thumb-down"],["過於複雜/步驟過多","tooComplicatedTooManySteps","thumb-down"],["過時","outOfDate","thumb-down"],["翻譯問題","translationIssue","thumb-down"],["示例/程式碼問題","samplesCodeIssue","thumb-down"],["其他","otherDown","thumb-down"]],["上次更新時間:2025-07-24 (世界標準時間)。"],[],[],null,["# MediaPipe Studio is a web-based application for evaluating and customizing\non-device ML models and pipelines for your applications. The app lets you\nquickly test MediaPipe solutions in your browser with your own data, and your\nown customized ML models. Each solution demo also lets you experiment with model\nsettings for the total number of results, minimum confidence threshold for\nreporting results, and more.\n\n[Try it!arrow_forward](https://goo.gle/mediapipe-studio)\n\nThis page provides some quick instructions on working with solutions in\nMediaPipe Studio.\n\n\n**Figure 1. MediaPipe Studio application home page.**\n\nCustom models\n-------------\n\n\nYou can use customized models with MediaPipe in the **Model selection** option,\nby choosing **Choose a model file** and selecting a model from your file\nstorage, as shown in the screenshot.\n\nThe model you choose must conform to the model input and output requirements of\nthe MediaPipe Tasks API you are using, and include compatible model metadata.\nThe quickest way to create your own model for use with a MediaPipe Tasks API is\nto use the MediaPipe Model Maker tool to modify a compatible solution model with\nyour own data. For more information, see\n[MediaPipe Model Maker](https://developers.google.com/mediapipe/solutions/model_maker).\n\nCustom input data\n-----------------\n\nYou can use your own data on each solution page in MediaPipe Studio. For\ntext-based tasks, you can enter text in the field provided. Vision tasks allow\nyou to use a web camera as input, and you can also upload images using the\n**Choose an image file** option in the drop-down menu, as shown below:\n\nGet started\n-----------\n\nYou can start using MediaPipe Studio by running one of the solution demos, such\nas Image Classification, and then use the related\n[developer guide](https://developers.google.com/mediapipe/solutions/vision/image_classifier)\nto build this functionality into your own application."]]