RSVP for your your local TensorFlow Everywhere event today!

在行動裝置和 IoT 裝置上部署機器學習模型

TensorFlow Lite 是一種開放原始碼深度學習架構,可在裝置端執行推論。

參閱指南

說明 TensorFlow Lite 概念與元件的指南。

參閱範例

探索 TensorFlow Lite Android 版和 iOS 版應用程式。

查看教學課程

Learn how to use TensorFlow Lite for common use cases.

運作方式

選擇模型

選擇新模型或重新訓練現有模型。

轉換

使用 TensorFlow Lite Converter,將 TensorFlow 模型轉換成壓縮處理的一般緩衝區。

部署

將壓縮過的 .tflite 檔案載入到行動裝置或嵌入式裝置中。

最佳化

將 32 位元的浮點數轉換成更有效率的 8 位元整數,或改為在 GPU 上執行,以便進行量化。

常見問題的解決方案

探索對一般行動裝置與邊緣使用案例有助益的最佳化模型。

圖片分類

辨識數百個物件,包含人物、活動、動物、植物和地點。

物件偵測

使用定界框來偵測多個物件。沒錯,這也能偵測貓和狗。

問題回答

使用先進的自然語言模型,透過 BERT 根據特定文字段落的內容回答問題。

最新消息與公告

請前往我們的網誌查看其他最新消息,並訂閱我們每月發行的 TensorFlow 電子報,讓最新的公告直接寄到你的收件匣。

February 10, 2020  
Accelerated inference on Arm microcontrollers with TensorFlow Lite for Microcontrollers and CMSIS-NN

Arm’s engineers have developed optimized versions of the TensorFlow Lite kernels that use CMSIS-NN to deliver blazing fast performance on Arm Cortex-M cores.

December 18, 2020  
How to generate super resolution images using TensorFlow Lite on Android

The task of recovering a high resolution (HR) image from its low resolution counterpart is commonly referred to as Single Image Super Resolution (SISR). In this tutorial, we use a pre-trained ESRGAN model from TensorFlow Hub and generate super resolution images using...

December 2, 2020  
Build sound classification models for mobile apps with Teachable Machine and TFLite

We are excited to announce that Teachable Machine now allows you to train your own sound classification model and export it in the TensorFlow Lite (TFLite) format. Then you can integrate the TFLite model to your mobile applications or your IoT devices. This is an easy...

November 25, 2020  
Training and deploying ML models on edge devices (TF Fall 2020 Updates)

Learn how to train and deploy an ML model on an Android app in just a few lines of code with TensorFlow Lite Model Maker and Android Studio. From here you can then explore how to use various tools from Google to turn a prototype into a production app. Presented by...

Continue