有没有办法在网络浏览器中获得 ML Kit 的功能? [英] Is there a way to get the functionality of ML Kit in the web browser?

查看:23
本文介绍了有没有办法在网络浏览器中获得 ML Kit 的功能?的处理方法,对大家解决问题具有一定的参考价值,需要的朋友们下面随着小编来一起学习吧!

问题描述

我正在开发跨平台应用(iOS/Android/网络),并且非常喜欢移动设备上机器学习套件的快速、廉价的设备端图像标记功能.有没有办法复制网络上的行为?ML Kit 模型是否可以与不同的 ML 库重用,以便重新调整用途?

I'm developing a cross-platform app (iOS/Android/web) and am loving the fast, cheap on-device image labeling feature of ML Kit on mobile. Is there a way to replicate the behavior on the web? Are the ML Kit models available for re-use with a different ML library so it can be repurposed?

推荐答案

不幸的是,ML Kit 似乎不允许您导出使用它创建的模型,只能导入模型.但是,tensorflow.js 允许您在网络上运行 TensorFlow 模型.如果您正在寻找一种创建模型的简单方法,有几个基于 Web 的程序可让您轻松创建 ML 模型并导出为 TensorFlow Lite(可以在 tensorflow.js 中运行,甚至可以托管在 Firebase 上).我听说过的几个是:lobe.ai 和 ml5.js.希望这会有所帮助.

Unfortunately, it does not seem like ML Kit allows you to export models created using it, only import models. However, tensorflow.js lets you run TensorFlow models on the web. If you are looking for an easy way to create models there are several web-based programs which allow you to easily create ML models and export as TensorFlow Lite (which can be run in tensorflow.js or even hosted on Firebase). A couple I have heard of are: lobe.ai and ml5.js. Hope this helps.

这篇关于有没有办法在网络浏览器中获得 ML Kit 的功能?的文章就介绍到这了,希望我们推荐的答案对大家有所帮助,也希望大家多多支持IT屋!

查看全文
登录 关闭
扫码关注1秒登录
发送“验证码”获取 | 15天全站免登陆