https://github.com/radi-cho/tfjs-firebase

Train a model with data from Firestore, save it to Cloud Storage and make predictions in Cloud Functions - entirely using NodeJS
https://github.com/radi-cho/tfjs-firebase

bag-of-words cloud-firestore cloud-functions firebase google-cloud tensorflow tfjs

Last synced: 29 days ago
JSON representation

Train a model with data from Firestore, save it to Cloud Storage and make predictions in Cloud Functions - entirely using NodeJS

• Host: GitHub
• URL: https://github.com/radi-cho/tfjs-firebase
• Owner: radi-cho
• License: apache-2.0
• Created: 2018-08-25T08:24:00.000Z (over 6 years ago)
• Default Branch: master
• Last Pushed: 2023-03-01T07:13:17.000Z (almost 2 years ago)
• Last Synced: 2024-05-01T17:52:42.060Z (8 months ago)
• Topics: bag-of-words, cloud-firestore, cloud-functions, firebase, google-cloud, tensorflow, tfjs
• Language: JavaScript
• Homepage:
• Size: 541 KB
• Stars: 46
• Watchers: 7
• Forks: 13
• Open Issues: 33
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Bag of words text classification

This workshop covers:

- Training a model with Tensorflow.js inside Cloud Functions using data from Firestore (JavaScript only).

- Retraining a model; Saving and loading a model.

- Making predictions with a cloud function.

The developed model can predict if a comment is positive or negative. We train it with data from the Firestore database.

The code can be easily modified to predict star reviews, app statistics, etc.

# Demo

## Training

![Training the model, gif](https://i.imgur.com/p6x8Hea.gif)

## Predictions

![Doing predictions with the model, gif](https://i.imgur.com/CW94keL.gif)