TensorFlow 简介

借助 TensorFlow,初学者和专家可以在桌面、移动、网络和云端环境下创建机器学习模型。请参阅以下几部分,了解如何开始使用。

TensorFlow

通过面向初学者和专家的教程学习 TensorFlow 的基础知识,运用这些知识创建下一个机器学习项目。

针对 JavaScript

通过 TensorFlow.js 使用 JavaScript 创建新的机器学习模型和部署现有模型。

针对移动设备和 IoT 设备

使用 TensorFlow Lite 在 Android、iOS、Edge TPU 和 Raspberry Pi 等移动设备和嵌入式设备上进行推断。

针对生产

使用 TensorFlow Extended (TFX) 部署可正式投入使用的机器学习流水线,以进行训练和推断。

Swift for TensorFlow

Integrate directly with Swift for TensorFlow, the next generation platform for deep learning and differentiable programming.

TensorFlow 生态系统

TensorFlow 提供了一系列工作流程,以供您使用 Python、JavaScript 或 Swift 开发和训练模型,并在云端、本地、浏览器中或设备上轻松地部署模型,无论您采用什么语言,都能提供支持。

Load & preprocess data
Build, train & reuse models
部署
TensorFlow
Build TensorFlow Input Pipelines
The tf.data API enables you to build complex input pipelines from simple, reusable pieces.
Explore
TensorFlow
Build and train models using Keras
tf.keras is a high-level API to build and train models. It supports TensorFlow-specific functionality, such as eager execution, tf.data pipelines, and estimators.
Explore
TensorFlow
Deploy using Python
Deploy on a mobile or edge device, in browser, or at scale using TensorFlow Serving.
TensorFlow.js
Import a Python model, or write one in JavaScript
Learn to convert pretrained models from Python to TensorFlow.js, as well as how to build and train models directly in JavaScript.
Explore
TensorFlow.js
Deploy in browser or Node.js
Learn how to deploy TensorFlow.js models in the browser, on node.js, or on the Google Cloud platform.
Explore
Swift for TensorFlow (in Beta)
Develop models natively in Swift (beta)
Using Swift differentiable programming allows for first-class support in a general-purpose programming language. Take derivatives of functions, and make custom data structures differentiable in an instant. Learn how Swift APIs give you transparent access to all low-level TensorFlow operators.
Explore
TensorFlow Lite
Deploy on mobile or embedded devices, like Android, iOS, and Raspberry Pi
Read the developer guide and pick a new model or retrain an existing one, convert it to a compressed file, load it on an edge device, and then optimize it.
Explore
TFX
Validate input data with TF Data Validation
See how to use TFX components to analyze and transform your data before you even train a model.
Explore
TFX
Feature engineering with TF Transform
Learn how to define a preprocessing function that transforms raw data into the data used to train a machine learning model, and see how the Apache Beam implementation is used to transform data by converting the preprocessing function into a Beam pipeline.
Explore
TFX
Modeling and training
Learn how to train your models in a TFX pipeline as a managed process.
Explore
TFX
Understanding model performance with TF model analysis
See how TensorFlow Model Analysis allows you to perform model evaluations in the TFX pipeline and visualize the results in a Jupyter notebook.
Explore
TFX
Serve models with a REST API with TF Serving
Learn how TensorFlow Serving allows you to deploy new algorithms and experiments while keeping the same server architecture and APIs.
Explore
TensorBoard
TensorBoard is a tool to visualize training and results
With TensorBoard you can track experiment metrics like loss and accuracy, visualize the model graph, project embeddings to a lower dimensional space, and more.
Explore
TensorFlow Hub
TensorFlow Hub is an extensive library of existing models
TensorFlow Hub is a library for the publication, discovery, and consumption of reusable parts of machine learning models called modules.
Explore

产品路线图和 RFC

探索即将发布的 TensorFlow 版本中的优先目标、侧重点和预期功能。查看即将发布的评论征集 (RFC) 通知,以进行技术方面的深入探讨并参与设计决策。其中许多方面都是由社区用例推动的,我们欢迎各方人员进一步做出贡献。

Looking to expand your ML knowledge?

TensorFlow is easier to use with a basic understanding of machine learning principles and core concepts. Learn and apply fundamental machine learning practices to develop your skills.

Learn ML

Begin with curated curriculums to improve your skills in foundational ML areas.