Posted by the TensorFlow team
Lightweight
Enables inference of on-device machine learning
models with a small binary size and fast initialization/startup
Cross-platform
A runtime designed to run on many different
platforms, starting with Android and iOS
Fast
Optimized for mobile devices, including dramatically
improved model loading times, and supporting hardware acceleration
More and more mobile devices today incorporate purpose-built custom hardware to
process ML workloads more efficiently. TensorFlow Lite supports the Android
Neural Networks API to take advantage of these new accelerators as they come
available.
Architecture
The following diagram shows the architectural design of TensorFlow Lite:
TensorFlow Model : A trained TensorFlow model saved on disk.
TensorFlow Lite Converter : A program that converts the
model to the TensorFlow Lite file format.
TensorFlow Lite Model File : A model file format based on FlatBuffers , that has been
optimized for maximum speed and minimum size.
The TensorFlow Lite Model File is then deployed within a Mobile App, where:
Java API: A convenience wrapper around the C++ API on
Android
C++ API : Loads the TensorFlow Lite Model File and invokes
the Interpreter. The same library is available on both Android and iOS
Interpreter : Executes the model using a set of operators.
The interpreter supports selective operator loading; without operators it is
only 70KB, and 300KB with all the operators loaded. This is a significant
reduction from the 1.5M required by TensorFlow Mobile (with a normal set of
operators).
On select Android devices, the Interpreter will use the Android
Neural Networks API
Developers can also implement custom kernels using the C++ API, that can be used
by the Interpreter.
Models
TensorFlow Lite already has support for a number of models that have been
trained and optimized for mobile:
MobileNet :
A class of vision models able to identify across 1000 different object classes,
specifically designed for efficient execution on mobile and embedded devices
Inception v3 : An image
recognition model, similar in functionality to MobileNet, that offers higher
accuracy but also has a larger size
Smart Reply : An on-device conversational model that
provides one-touch replies to incoming conversational chat messages. First-party
and third-party messaging apps use this feature on Android
Wear .
Inception v3 and MobileNets have been trained on the ImageNet dataset. You can
easily retrain these on your own image datasets through transfer
learning .
What About TensorFlow Mobile?
As you may know, TensorFlow already supports mobile and embedded deployment of
models through the TensorFlow
Mobile API. Going forward, TensorFlow Lite should be seen as the evolution
of TensorFlow Mobile, and as it matures it will become the recommended solution
for deploying models on mobile and embedded devices. With this announcement,
TensorFlow Lite is made available as a developer preview, and TensorFlow Mobile
is still there to support production apps.
TensorFlow Lite documentation
pages .
Thank you! re: "solution for mobile and embedded devices", can we expect to see early support for rpi & piZero-w? The process for getting TF on a piZero right now is pretty painful. Many of us who are using TF on rpis for research in the self driving car space would *really* make good use of TF on a pi zero-w due to the cost difference and wifi support .
ReplyDeleteThank you! How about the inference speed compared to other mobile DL framework such as https://github.com/Tencent/ncnn ?
ReplyDeleteawesome idea thanks for this new soft.. https://mastaram.net what do you think
ReplyDeleteawesome !!!
ReplyDeleteGood news. Thank you.
ReplyDeleteI got two questions.
1. Is it possible to retrain the model using Tensorflow Lite direct on the Android device?
2. Are you planning to support Tensorflow Lite on Raspberry Pi 3?
1. It is not yet possible to do training with TensorFlow lite on-device. That is an obvious avenue of future work that we are excited about.
Delete2. It should be pretty easy to build TensorFlow Lite on Raspberry PI. We plan to make sure this path works seamlessly soon, but for the initial release we have tested tflite on Android and iOS.
This comment has been removed by the author.
Deleterfjhk,j
ReplyDeleteyuj;lo':/
ReplyDeletewww.sendcash9.com
ReplyDeletei͓̽ q͓̽u͓̽i͓̽t͓̽ w͓̽o͓̽r͓̽k͓̽i͓̽n͓̽g͓̽ a͓̽t͓̽ s͓̽h͓̽o͓̽p͓̽r͓̽i͓̽t͓̽e͓̽ a͓̽n͓̽d͓̽ n͓̽o͓̽w͓̽ i͓̽ m͓̽a͓̽k͓̽e͓̽ $35h͓̽ - $80h͓̽...h͓̽o͓̽w͓̽? i͓̽'m͓̽ w͓̽o͓̽r͓̽k͓̽i͓̽n͓̽g͓̽ o͓̽n͓̽l͓̽i͓̽n͓̽e͓̽! m͓̽y͓̽ w͓̽o͓̽r͓̽k͓̽ d͓̽i͓̽d͓̽n͓̽'t͓̽ e͓̽x͓̽a͓̽c͓̽t͓̽l͓̽y͓̽ ★★ www.sendcash9.com 。
ReplyDeleteWonderful
ReplyDeleteGood Job!
ReplyDeleteFor TF-integrated solutions that run on mobile and desktop platforms, would you recommend migrating to TensorFlow Lite across all platforms, or is there a benefit to using TensorFlow Lite on mobile and "regular" TensorFlow on desktop?
ReplyDeleteawesome. can't wait to start using the APIs
ReplyDeleteCan anyone please show an example of how to use this API on Android?
ReplyDeleteMaybe, take a simple tick-tack-toe game, to show how it learns what to do