README.md 4.1 KB
Newer Older
Z
zhangjinchao01 已提交
1 2
# PaddlePaddle

G
gangliao 已提交
3 4 5
|  **`Linux`**   |
|----------------|
|[![Build Status](https://travis-ci.org/baidu/Paddle.svg?branch=master)](https://travis-ci.org/baidu/Paddle)|
E
emailweixu 已提交
6

G
gangliao 已提交
7
Welcome to the PaddlePaddle GitHub.
E
emailweixu 已提交
8

Z
zhangjinchao01 已提交
9 10 11 12 13
PaddlePaddle (PArallel Distributed Deep LEarning) is an easy-to-use,
efficient, flexible and scalable deep learning platform, which is originally
developed by Baidu scientists and engineers for the purpose of applying deep
learning to many products at Baidu.

G
gangliao 已提交
14 15 16
Our vision is to enable deep learning for everyone via PaddlePaddle.
Please refer to our [release log](https://github.com/baidu/Paddle/releases) to track the latest feature of PaddlePaddle. 

Z
zhangjinchao01 已提交
17 18 19 20
## Features

- **Flexibility**

G
gangliao 已提交
21 22 23 24
    PaddlePaddle supports a wide range of neural network architectures and
    optimization algorithms. It is easy to configure complex models such as
    neural machine translation model with attention mechanism or complex memory
    connection.
Z
zhangjinchao01 已提交
25 26 27

-  **Efficiency**
  
G
gangliao 已提交
28 29 30 31 32 33 34 35 36 37 38
    In order to unleash the power of heterogeneous computing resource,
    optimization occurs at different levels of PaddlePaddle, including
    computing, memory, architecture and communication. The following are some
    examples:

      - Optimized math operations through SSE/AVX intrinsics, BLAS libraries
      (e.g. MKL, ATLAS, cuBLAS) or customized CPU/GPU kernels. 
      - Highly optimized recurrent networks which can handle **variable-length** 
      sequence without padding.
      - Optimized local and distributed training for models with high dimensional
      sparse data.
Z
zhangjinchao01 已提交
39 40 41

- **Scalability**

G
gangliao 已提交
42 43 44
    With PaddlePaddle, it is easy to use many CPUs/GPUs and machines to speed
    up your training. PaddlePaddle can achieve high throughput and performance
    via optimized communication.
Z
zhangjinchao01 已提交
45 46 47

- **Connected to Products**

G
gangliao 已提交
48 49 50 51 52 53 54
    In addition, PaddlePaddle is also designed to be easily deployable. At Baidu,
    PaddlePaddle has been deployed into products or service with a vast number
    of users, including ad click-through rate (CTR) prediction, large-scale image
    classification, optical character recognition(OCR), search ranking, computer
    virus detection, recommendation, etc. It is widely utilized in products at
    Baidu and it has achieved a significant impact. We hope you can also exploit
    the capability of PaddlePaddle to make a huge impact for your product.
Z
zhangjinchao01 已提交
55 56

## Installation
G
gangliao 已提交
57 58 59 60
Check out the [Install Guide](http://paddlepaddle.org/doc/build/) to install from
pre-built packages (**docker image**, **deb package**) or 
directly build on **Linux** and **Mac OS X** from the source code.
 
Z
zhangjinchao01 已提交
61
## Documentation
G
gangliao 已提交
62
Both [English Docs](http://paddlepaddle.org/doc/) and [Chinese Docs](http://paddlepaddle.org/doc_cn/) are provided for our users and developers.
63

L
liaogang 已提交
64
- [Quick Start](http://paddlepaddle.org/doc/demo/quick_start/index_en) <br>
Z
zhangjinchao01 已提交
65 66 67
   You can follow the quick start tutorial to learn how use PaddlePaddle
   step-by-step.
    
68
- [Example and Demo](http://paddlepaddle.org/doc/demo/) <br>
Z
zhangjinchao01 已提交
69
   We provide five demos, including: image classification, sentiment analysis,
D
dangqingqing 已提交
70
   sequence to sequence model, recommendation, semantic role labeling. 
Z
zhangjinchao01 已提交
71
   
72
- [Distributed Training](http://paddlepaddle.org/doc/cluster) <br>
Z
zhangjinchao01 已提交
73 74 75
  This system supports training deep learning models on multiple machines
  with data parallelism.
   
76
- [Python API](http://paddlepaddle.org/doc/ui/) <br>
Z
zhangjinchao01 已提交
77 78 79 80 81
   PaddlePaddle supports using either Python interface or C++ to build your
   system. We also use SWIG to wrap C++ source code to create a user friendly
   interface for Python. You can also use SWIG to create interface for your
   favorite programming language.
 
82 83
- [How to Contribute](http://paddlepaddle.org/doc/build/contribute_to_paddle.html) <br>
   We sincerely appreciate your interest and contributions. If you would like to
Z
zhangjinchao01 已提交
84 85
   contribute, please read the contribution guide.   

86 87
- [Source Code Documents](http://paddlepaddle.org/doc/source/) <br>

Z
zhangjinchao01 已提交
88 89 90 91
## Ask Questions

If you want to ask questions and discuss about methods and models, welcome
to send email to paddle-dev@baidu.com. Framework development discussions and
Y
yuyang18 已提交
92
bug reports are collected on [Issues](https://github.com/baidu/paddle/issues).
Z
zhangjinchao01 已提交
93 94 95

## Copyright and License
PaddlePaddle is provided under the [Apache-2.0 license](LICENSE).