Tutorial: How to use raw tensorFlow.js pre-trained models in browser

Another way to use pre-made Web ML models Is simply just the raw trained model itself, with no JavaScript class wrapped around it that typically hides the inner workings, to make it easier to use for beginners. In this lesson you will learn to write the code to send data into the raw model, and then extract data from it yourself, using Tensors that you just learnt about in the prior lesson.

Individuals or researchers may share their raw models like this, which can be very useful for finding and using examples that may still be in development, or using models that solve less common problems, where an example may exist,
but does not have full documentation. Learn how to use raw pre-trained models from layers models to graph based models from any site including TensorFlow Hub - Google's model garden.

Connect with Jason Mayes to ask questions:

Use #WebML to share your learnings and creations from this course to meet your peers on social media!

[piotrgalas]

This is the best tutorial I have ever seen. You did amazing job. I thank you

[imtanuki]

It would be helpful if we had the url's for project source code and tf models, rather than having to manually copy links from the video. Like, seriously...

[dvjkizy]

Awesome & nice live code session..

[ateshirsizi]

where do you copy the link from storage googleapis?

[maumenvi]

The link to get model give me 'acess denied' how i can get this model?

[faraazkhan]

I am running my reactjs app on localhost and passing model.json file path to load model

I got error console .
Load local model not allowed

[melaniearcher]

Okay, perhaps I'm dim, but I can't find how you've run script.js on Glitch. Is it really sufficient to preview index.html to start up the loadModel() function?

[jaynouoliverbeck]

how can we use it in angular applications on GCP