I am working on a flutter project that requires some machine learning. I have my model already trained and ready to be used.
How to integrate this Tensorflow lite model with flutter?
I have tried couple of packages that exists but unfortunately, I did not worked! Most of these packages are no longer maintained.
Related
I am working on an application where main screen has a live camera opened automatically.
Here I need faces to be detected.
Am getting confused about how to integrate ML kit.
Do I have to signup with firebase????
I tried working with google ml kit but got too many errors.
There are two separate ML libraries now - Firebase ML and ML Kit. Firebase ML runs on the cloud, and you need to use Firebase for it. You can find the documentation here.
ML Kit, however, runs entirely on the user's device, and it does not require you to use Firebase, unless you want to use custom models.
The easiest way to integrate it into your flutter app would be to use the google_ml_kit library, which provides modules for different purposes. For face recognition, you can use google_mlkit_face_detection.
I'm building a mobile app which basically should create a facemesh from a live camera video and apply some filters/makeup to it.
I wanted to use Flutter as the mobile framework in combination with Tensorflow lite.
I'm looking at Tensorflows Facemesh model but I can't find any good integration in Tensorflow lite.
Is there either a way to use Tensorflow lite with the facemesh model or use tensorflow.js (includes the facemesh feature) with Flutter?
Thanks in advance!
Pascal
I have developed a model in python and exported it to TFLite format but I tried to integrate it into my flutter application all in vain. The plugins available at pub.dev that deal with tflite models are all deprecated and therefore cant run with flutter v2. How can I overgo this? Thanks.
This worked for me!
Follow This great tutorial. It explains step by step how to use your tflite models in your flutter apps
Im going to start bluetooth project in flutter, can anyone suggest which package is best for production.
My case is to connect with smartwatches,weightscale,bp devices to get the medical data.
Flutter blue:https://pub.dev/packages/flutter_blue
Flutter reactive ble:https://pub.dev/packages/flutter_reactive_ble
Which ble package that is best is a personal opinion. But I have spent the last four years on making apps that use ble to talk to iOT devices. I have use three different packages:
I used flutter_blue in production about 4 years ago. I was not happy with it(I don't remember exactly why). They made breaking changes that included a switch to RxDart so there where no reason to stay with flutter _blue.
When I started to develop a new app I selected flutter_ble_lib. The reason for that was that the main developers come from the consultant company Polidea. But Polidea was bought by an other company and they stopped developing packages. So I had to change ble package to be able to keep my app upgraded.
I asked google if they had any plans on making a ble package, they said no. But they recommended flutter_reactive_ble. So I have been using flutter_reactive_ble since then. It is Philips Hue that develop it for their smart lamps. The reasons why I prefer flutter_reactive_ble:
I think its the more stable choice because of its developed by Philips hue. It is as good as it get when it comes to stability.
I use the library to tell iOT devises how to connect to internet. It's the probably what this package is made for.
Flutter reactive ble is better since It has lesser reported issues.
please read this disclaimer by flutter_blue team
This library is actively developed alongside production apps, and the
API will evolve as we continue our way to version 1.0.
Please be fully prepared to deal with breaking changes. This package
must be tested on a real device.
Flutter reactive ble is a lot more stable and it's not a newly started project. It's been tested and fixed over a long course of time.
Other good option is flutter_blue_plus. It's a fork of flutter_blue and it's actively maintained.
I started with flutter_blue but last commit was two years ago. Then I decided to switch to flutter_blue_plus
My team are developing an app that will involve some on device ML model that are in onnx format. Currently we considering Flutter & React Native. I prefer Flutter but couldn't find any plugin that support running on device onnx model. in RN we could use onnx.js from Microsoft.
I think there is no Flutter Plugin to do this but you only need to create one in your project (link) and take implement Android version and iOS version. The plugin will make the wrapper and you can use it in your Flutter code. Depending of the phone OS, flutter will execute Android or iOS code.