Skip to content

A tutorial on how to train a TensorFlow Lite model and make it compatible with ML Kit

Notifications You must be signed in to change notification settings

flutter-ml/mlkit-custom-model

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

6 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Custom ML model for ML Kit

By default, ML Kit’s APIs make use of Google trained machine learning models. Both the Image Labeling and the Object Detection & Tracking API offer support for custom image classification models.

In this tutorial is shown how to create a TensorFlow Lite model and make it compatible with ML Kit.

NOTE: Before jumping into coding, make sure you read and understand the ML Kit's compatibility requirements for TensorFlow Lite models here.

You can run this tutorial in Google Colab.

Run in Google Colab

Or you can clone this repo and run this in your local terminal:

python3 ml_kit_custom_model.py

That will generate these files:

  • model.tflite
  • model_with_metadata.tflite
  • labels.txt

You will need model_with_metadata.tflite to test in your mobile app using ML Kit.

You can use and tweak our demo app to test your tflite model using google_mlkit_image_labeling and google_mlkit_object_detection in Flutter.

About

A tutorial on how to train a TensorFlow Lite model and make it compatible with ML Kit

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published