How to convert your ML model to TensorFlow Lite (TensorFlow Tip of the Week)

preview_player
Показать описание
Welcome back to another episode of TensorFlow Tip of the Week! In this short episode, we’re going to create a simple machine learned model using Keras and convert it to TensorFlow Lite for use on mobile, or IoT devices. TensorFlow Lite is TensorFlow’s lightweight solution for mobile and embedded devices. Let us know your feedback in the comments below and subscribe to the channel for more TensorFlow tutorials!

Рекомендации по теме
Комментарии
Автор

Can you please post a video on Quantization + TFLite model creation from Tensorflow rather than from Keras?

TheAnilmaddala
Автор

Here is the code for saving the model as a tflite

converter =
tflite_model = converter.convert()

with tf.io.gfile.GFile('model.tflite', 'wb') as f:
f.write(tflite_model)

shuhulmujoo
Автор

Lawrence, I'm a big fan of your series, but I think you made a mistake on the 3rd x and y values. (2 * 1) - 1 = 1 but you put 0 as your 3rd y value.

JoeWong
Автор

Great videos, Lawrence. I would like to see more content about how this conversion changes the model graph, weights, model accuracy etc. If you train a model with GPUs and big data, but then convert (i.e. compress) the model so that it can run in lightweight mobile environments, then I assume there is also some drawbacks?

kai-waehner
Автор

Tensorflow.Contrib is no more supported in TF 2. Any Solution for that?

anagrarian
Автор

how can i use a simple trained svm model into android studio please make a video on this topic

ImPushpendraPal
Автор

It only works with Linux - NOT windows

INVESTDESK
Автор

Sir, which version of tensorFlow you are using?

SooryaGangaraj
Автор

How to use tensorflow lite with flutter??

sovanmondal
Автор

Hello Lawrence. When I run the code in the walk-through lesson, I get "AttributeError: module 'tensorflow.contrib.lite' has no attribute 'TocoConverter'" I have tried searching for the error cause but unfortunately haven't found a solution. I use tf version 1.10, please any recommendations or solutions will be profoundly welcomed.

Gospel_With_AI
Автор

I did transfer learning on the MobileNetV2 model using Keras and I am trying to implement the resulted trained model in the TFLite Object Detection demo. How do I convert correctly the model in tflite format from h5 format to be compatible with the demo?

grosugeorge
Автор

i got some error ImportError: cannot import name 'lite'

m.revaldianggara
Автор

How to convert any normal machine learning model like decision trees, and linear regressions to tflite format other than ANN? Is it possible?

pavithrasenthilkumar
Автор

can I do the same in jupyter notebook?

usamabaig
Автор

how to convert sklearn random forest classifier trained model to .tflite?

noumanrafiq
Автор

awesome tutorial..you are amazing...!! thanks

mobassir
Автор

please tell me how i can predict from this model after converting into .tflite, in flutter

nextgodlevel
Автор

How much time will it take to convert the model into TFlite

sriraj
Автор

While executing the program im getting this error: toco_from_protos' is not recognized as an internal or external command. Help me to solve this

naveenkumara
Автор

Not metadata on this model (android studio) How can fix it?

cerpenai