... #define the program, cost, and create sgd optimizer
optimize_ops, params_grads = sgd_optimizer.minimize(avg_cost) #get optimize OPs and gradient parameters
t = fluid.DistributeTranspiler() # create the transpiler instance
# slice the program into 2 pieces with optimizer_ops and gradient parameters list, as well as pserver_endpoints, which is a comma separated list of [IP:PORT] and number of trainers
PaddlePaddle supports English documentation ``doc`` and Chinese documentation ``doc_cn``.
Both are compiled by `cmake`_ and `sphinx`_ , the compiled documentations will be stored under ``doc`` and ``doc_cn`` directories.
When using the PaddlePaddle.org to compile documentations, the compiled documentations will be stored under a consolidated directory: .ppo_workspace/content
PaddlePaddle's documentation includes both Chinese and English versions. The documentation is built using the ``cmake`` command to drive the ``sphinx`` compiler. The PaddlePaddle.org tool helps us to implement this compilation process and provides better preview results.
How to Build Documentations
============
How to build Documentation
===========================
We recommend using PaddlePaddle.org tool to build documentation
PaddlePaddle's documentation is built in two ways: using the PaddlePaddle.org tool and without using it. Both methods have their own advantages. The former facilitates previewing, while the latter facilitates debugging by the developer. We could choose to build the documentation with Docker or without it in each of the above ways.
We recommend using PaddlePaddle.org tool to build documentation.
Use PaddlePaddle.org tool
--------------
This is the recommended method to build documentation. It can compile documentation and preview the documentation in a web browser.
Using PaddlePaddle.org tool
-----------------------------
This is the recommended method to build documentation, because it can automatically compile the documentation and preview the documentation directly in a web page. Note that, although you can preview the documentation in other ways, its style may not be consistent with the official website. Compiling with the PaddlePaddle.org tool produces a preview that will be consistent with the official website documentation style.
The tool uses Docker, please install it on your system. Please check Docker official website on how to install Docker. You may use the following commands to activate the tool
The PaddlePaddle.org tool can be used with Docker and Docker needs to be installed first. Please refer to `Docker's official website <https://docs.docker.com/>`_ on how to install Docker. After installing Docker, you may use the following commands to activate the tool
..code-block:: bash
@ -32,8 +31,8 @@ The tool uses Docker, please install it on your system. Please check Docker offi
# Please specify the working directory through -v
docker run -it -p 8000:8000 -v `pwd`:/var/content paddlepaddle/paddlepaddle.org:latest
Note: PaddlePaddle.org will read the content repos specified in the -v (volume) flag of the docker run command
Use a web browser and navigate to http://localhost:8000, click the buttons to compile the documentation
Note: PaddlePaddle.org will read the content repos specified in the -v (volume) flag of the docker run commands
Use a web browser and navigate to http://localhost:8000. Click the buttons to compile the documentation.
The compiled documentations will be stored in <paddlepaddle working directory>/.ppo_workspace/content
@ -58,19 +57,62 @@ If you don't wish to use Docker, you can also activate the tool through Django.
pip install -r requirements.txt
python manage.py runserver
Use a web browser and navigate to http://localhost:8000, click the buttons to compile the documentation
Specify the PaddlePaddle working directory for the environment variable CONTENT_DIR so that the tool could find where the working directory is.
Use a web browser and navigate to http://localhost:8000. Click the buttons to compile the documentation
The compiled documentations will be stored in <paddlepaddle working directory>/.ppo_workspace/content
If you want to learn more on the PaddlePaddle.org, please `click here <https://github.com/PaddlePaddle/PaddlePaddle.org/blob/develop/README.md>`_ 。
Please `click here <https://github.com/PaddlePaddle/PaddlePaddle.org/blob/develop/README.md>`_ for more information about the PaddlePaddle.org tool.
Manually Building the Documentation
-------------------------------------
Build PaddlePaddle's documentation with Docker，you need to install Docker first. Please refer to `Docker's official website <https://docs.docker.com/>`_ on how to install Docker. After Docker is installed, you could use the scripts in the source directory to build the documentation.
[TBD]
If you do not wish to use Docker, you can also use the following commands to directly build the PaddlePaddle documentation.
# If you only need to build documents, use the following commands
make -j $processors gen_proto_py
make -j $processors paddle_docs paddle_docs_cn
# If you only need to build APIs, use the following commands
make -j $processors gen_proto_py framework_py_proto
make -j $processors copy_paddle_pybind
make -j $processors paddle_api_docs
$processors indicates that as many processes as the CPU cores are started to compile in parallel. It should be set according to the number of CPU cores of your machine.
After the compilation is complete, enter the ``doc/v2`` directory. If you chose to build documents, it will generate ``cn/html/`` and ``en/html`` subdirectories under this directory. If you chose to build APIs，it will generate``api/en/html`` subdirectory. Please enter these directories respectively and execute the following commands:
..code-block:: bash
python -m SimpleHTTPServer 8088
Use a web browser and navigate to http://localhost:8000, you could see the compiled Chinese/English documents page and the English APIs page. The following figure is an example of the built English documents home page. Note that due to the sphinx's original theme used in the example, the style of the page is not consistent with the official website, but this does not affect the developer's debugging.
How to write Documentations
============
..image:: src/doc_en.png
:align:center
:scale:60 %
PaddlePaddle uses `sphinx`_ to compile documentations，Please check sphinx official website for more detail.
How to write Documentation
===========================
PaddlePaddle uses `sphinx`_ to compile documentation，Please check sphinx official website for more detail.
How to update www.paddlepaddle.org
============================
===================================
Please create PRs and submit them to github, please check `Contribute Code <http://www.paddlepaddle.org/docs/develop/documentation/en/howto/dev/contribute_to_paddle_en.html>`_ 。
PaddlePaddle develop branch will update the documentation once the PR is merged. User may check latest `Chinese Docs <http://www.paddlepaddle.org/docs/develop/documentation/zh/getstarted/index_cn.html>`_ and
After we train a neural network, we use it to do inference. Inference is the process of preparing input data and propagating it through the model to produce the result.
Compared with model training, prediction has the following features:
#. Inference does not require backpropagation and parameter updates, as required during training.
#. Labels are not needed in prediction.
#. Most of the time, predictions need to be integrated with the user system.
Therefore, the model prediction SDK needs to be designed separately and has the following features:
#. The predictive SDK does not include backpropagation and parameter updates to reduce the size of the SDK.
#. The predictive SDK needs a simple user interface for ease of use.
#. Since the input data may have a variety of structures, the format of the input data is clearly and compactly packaged.
#. In order to be compatible with user's system, the SDK's interface must conform to the C-standard interface.
PaddlePaddle provides C-API to solve the above problem. Following are the guidelines to use the C-API:
PaddlePaddle provides the users the ability to flexibly set various command line parameters to control the model training and inference process. Please refer to the following instructions on using PaddlePaddle:
..toctree::
:maxdepth:1
cmd_parameter/index_cn.rst
PaddlePaddle supports distributed training tasks on fabric clusters, MPI clusters, and Kubernetes clusters. For detailed configuration and usage instructions, refer to:
..toctree::
:maxdepth:1
cluster/index_cn.rst
PaddlePaddle provides a C-API for inference. We provide the following guidelines for using the C-API:
..toctree::
:maxdepth:1
capi/index_cn.rst
PaddlePaddle supports a variety of flexible and efficient recurrent neural networks. For details, please refer to：
..toctree::
:maxdepth:1
rnn/index_cn.rst
How to use the built-in timing tool, nvprof, or nvvp to run performance analysis and tuning, please refer to：
guosheng
7 years ago