Placeholder Image

字幕列表 影片播放

  • did you know the tensorflow dot Js isn't just for JavaScript on the back end with node, but can also be used in the browser.

  • Yes, you can train models in the browser with JavaScript, And not only that, you can run inference to.

  • And if that wasn't enough, you could have models that you've trained elsewhere on run inference on those in the browser using JavaScript.

  • So today's tip is, how do I get my browser ready for tensorflow?

  • And this one's almost cheating.

  • It's a little too easy.

  • So let's start with the simplest Web page imaginable will save and run it, and we'll see that we just get an empty page.

  • So let's have a script tag to the head, and we'll add the tents of Florida J a script to that.

  • We can save and run it again.

  • And now look at the page.

  • There's not a lot to see right now, so let's add a div to the page, and we'll output to that later.

  • So now let's add a simple script for training a model tow.

  • Unfair.

  • A linear relationship will create a model on just at a single layer to it will specify the lost function, and the optimizer after that will specify a couple of chancers for the values.

  • One tensor has the X values, and the other has the Y.

  • There's a relationship between them, which is why equals two X minus one.

  • But the model is going to infer that we train the model with Modeled Outfit on.

  • Then we'll output the prediction for why, where X equals 10 now, of course, it should be about 19.

  • So let's take a look and see on.

  • We actually get pretty close.

  • So there's your start in using tensorflow JavaScript.

  • I do hope this was a useful primer for you.

  • For more great videos like this one, hit that subscribe button and we'll see you there.

did you know the tensorflow dot Js isn't just for JavaScript on the back end with node, but can also be used in the browser.

字幕與單字

單字即點即查 點擊單字可以查詢單字解釋

B1 中級

如何讓你的瀏覽器為TensorFlow做好準備(TensorFlow一週小技巧)。 (How to get your browser ready for TensorFlow (TensorFlow Tip of the Week))

  • 2 0
    林宜悉 發佈於 2021 年 01 月 14 日
影片單字