#Tensorflow.js Universal Sentence Encoder

1 messages · Page 1 of 1 (latest)

shut pivot
#

Hey guys,
I am trying to train an intent classifier. As input I want to use the embeddings given by the Universal Sentence Encoder. I can't seem to use the output of the embed function although it is as Tensor2D.

I am using:

  • @tensorflow-models/universal-sentence-encoder": "^1.3.3"
  • @tensorflow/tfjs": "^4.2.0"

When I installed the universal-sentence-encoder I had a version conflict for tfjs-converter... Might this be a possible source of the problem? And how would I go about fixing it?

Thanks :)

Screenshot_2023-01-08_203806.png
crystal grail
#

Have you followed this tutorial by Yannick by chance? https://youtu.be/D7ZL45xS39I?t=520

YouTube
TensorFlow
Machine Learning Magic for Your JavaScript Application (Google I/O'19)

TensorFlow.js is a library for training and deploying machine learning models in the browser and in Node.js and offers unique opportunities for JavaScript developers. In this talk, you will learn about the TensorFlow.js ecosystem: how to bring an existing machine learning model into your JS app, re-train the model using your data and go beyond t...

▶ Play video