diff --git a/.gitignore b/.gitignore index 1d74e21965c4f858f5f818a270e64e1bfad7d843..36f2b0bac954023a45832337ff1c2d09426e9600 100644 --- a/.gitignore +++ b/.gitignore @@ -1 +1,3 @@ .vscode/ +/doc/fluid/menu.zh.json +/doc/fluid/menu.en.json diff --git a/doc/fluid/advanced_usage/best_practice/cpu_train_best_practice.rst b/doc/fluid/advanced_usage/best_practice/cpu_train_best_practice.rst index 7461346a96de470b023295087efdf5fdbe23209a..1c54dba9f2ff034408ee54a074762b68875463aa 100644 --- a/doc/fluid/advanced_usage/best_practice/cpu_train_best_practice.rst +++ b/doc/fluid/advanced_usage/best_practice/cpu_train_best_practice.rst @@ -46,7 +46,7 @@ API详细使用方法参考 :ref:`cn_api_fluid_ParallelExecutor` ,简单实例 提高通信速度 ========== -要减少通信数据量,提高通信速度,主要是使用稀疏更新 ,目前支持 `稀疏更新 <../layers/sparse_update.html>`_ 的主要是 :ref:`cn_api_fluid_layers_embedding` 。 +要减少通信数据量,提高通信速度,主要是使用稀疏更新 ,目前支持 :ref:`api_guide_sparse_update` 的主要是 :ref:`cn_api_fluid_layers_embedding` 。 .. code-block:: python diff --git a/doc/fluid/advanced_usage/best_practice/index_cn.rst b/doc/fluid/advanced_usage/best_practice/index_cn.rst index f3555a3e83ea92a30a8c0917ef748d86b2871c0b..64cbb5a0ccd0c71157ad48b102dbece6223f3056 100644 --- a/doc/fluid/advanced_usage/best_practice/index_cn.rst +++ b/doc/fluid/advanced_usage/best_practice/index_cn.rst @@ -3,7 +3,7 @@ ######### .. toctree:: - :hidden: + :maxdepth: 1 cpu_train_best_practice.rst dist_training_gpu.rst diff --git a/doc/fluid/advanced_usage/deploy/inference/paddle_gpu_benchmark.md b/doc/fluid/advanced_usage/deploy/inference/paddle_gpu_benchmark.md index c5b746290f7d445b68cca393e3c2e0d2f31e8128..b7ac92c5ca7d2a07384d086c83f0c96b7e84ca8e 100644 --- a/doc/fluid/advanced_usage/deploy/inference/paddle_gpu_benchmark.md +++ b/doc/fluid/advanced_usage/deploy/inference/paddle_gpu_benchmark.md @@ -6,15 +6,15 @@ - 测试模型 ResNet50,MobileNet,ResNet101, Inception V3. ## 测试对象 -**PaddlePaddle, Pytorch, Tensorflow** +**PaddlePaddle, Pytorch, Tensorflow** -- 在测试中,PaddlePaddle使用子图优化的方式集成了TensorRT, 模型[地址](https://github.com/PaddlePaddle/models/tree/develop/fluid/PaddleCV/image_classification/models)。 +- 在测试中,PaddlePaddle使用子图优化的方式集成了TensorRT, 模型[地址](https://github.com/PaddlePaddle/models/tree/develop/PaddleCV/image_classification/models)。 - Pytorch使用了原生的实现, 模型[地址1](https://github.com/pytorch/vision/tree/master/torchvision/models)、[地址2](https://github.com/marvis/pytorch-mobilenet)。 - 对TensorFlow测试包括了对TF的原生的测试,和对TF—TRT的测试,**对TF—TRT的测试并没有达到预期的效果,后期会对其进行补充**, 模型[地址](https://github.com/tensorflow/models)。 -### ResNet50 - +### ResNet50 + |batch_size|PaddlePaddle(ms)|Pytorch(ms)|TensorFlow(ms)| |---|---|---|---| |1|4.64117 |16.3|10.878| diff --git a/doc/fluid/advanced_usage/design_idea/fluid_design_idea.md b/doc/fluid/advanced_usage/design_idea/fluid_design_idea.md index 189ea5c0fd0bd6cf2ad940aa96f7e411c78e358f..e1df7964c5b80e6422bea6e3fe889141c4b0da77 100644 --- a/doc/fluid/advanced_usage/design_idea/fluid_design_idea.md +++ b/doc/fluid/advanced_usage/design_idea/fluid_design_idea.md @@ -359,5 +359,5 @@ Fluid使用Executor.run来运行一段Program。 [6.099215 ]], dtype=float32), array([1.6935859], dtype=float32)] ``` -至此您已经了解了Fluid 内部的执行流程的核心概念,更多框架使用细节请参考[使用指南](../../user_guides/index.html)相关内容,[模型库](../../user_guides/models/index_cn.html +至此您已经了解了Fluid 内部的执行流程的核心概念,更多框架使用细节请参考[使用指南](../../user_guides/index_cn.html)相关内容,[模型库](../../user_guides/models/index_cn.html )中也为您提供了丰富的模型示例以供参考。 diff --git a/doc/fluid/advanced_usage/development/contribute_to_paddle/submit_pr_guide.md b/doc/fluid/advanced_usage/development/contribute_to_paddle/submit_pr_guide.md index ce6c6e408667dbe2def4d9c51ad1a8f9a4ff8bfe..a43d38ee5470494757f59b94e9248010f1c7c775 100644 --- a/doc/fluid/advanced_usage/development/contribute_to_paddle/submit_pr_guide.md +++ b/doc/fluid/advanced_usage/development/contribute_to_paddle/submit_pr_guide.md @@ -1,4 +1,4 @@ -# Github提交PR指南 +# 提交PR注意事项 ## 建立 Issue 并完成 Pull Request diff --git a/doc/fluid/advanced_usage/development/new_op/index_cn.rst b/doc/fluid/advanced_usage/development/new_op/index_cn.rst index fc56472ce9186b24698c836e18de63c3dc0693f3..73776132a21180c2f0e5a6d79a1a8ef322ce3e88 100644 --- a/doc/fluid/advanced_usage/development/new_op/index_cn.rst +++ b/doc/fluid/advanced_usage/development/new_op/index_cn.rst @@ -1,5 +1,5 @@ ############# -新增Operator +新增OP ############# 本部分将指导您如何新增Operator,也包括一些必要的注意事项 diff --git a/doc/fluid/advanced_usage/development/new_op/new_op.md b/doc/fluid/advanced_usage/development/new_op/new_op.md index b1d89954e7dfb71dc081eb778a158f267b0cbbc3..6fd22ab378f29afe9b2fc591b500cf4f801457f3 100644 --- a/doc/fluid/advanced_usage/development/new_op/new_op.md +++ b/doc/fluid/advanced_usage/development/new_op/new_op.md @@ -1,4 +1,4 @@ -# 如何写新的op +# 如何写新的OP ## 概念简介 diff --git a/doc/fluid/advanced_usage/development/new_op/op_notes.md b/doc/fluid/advanced_usage/development/new_op/op_notes.md index 846038075a1e933d90b7a0f9fa75b82749d50241..86244b9ed5235eae5273f61117a848bfbba6875c 100644 --- a/doc/fluid/advanced_usage/development/new_op/op_notes.md +++ b/doc/fluid/advanced_usage/development/new_op/op_notes.md @@ -1,4 +1,4 @@ -# op相关注意事项 +# OP相关注意事项 ## Fluid中Op的构建逻辑 ### 1.Fluid中Op的构建逻辑 diff --git a/doc/fluid/advanced_usage/development/new_op/op_notes_en.md b/doc/fluid/advanced_usage/development/new_op/op_notes_en.md index 79804846eba01c51f3ae1d0d888cc3829861ad9f..82b3ac93d608f1957b7f8f14791f046a2113c671 100644 --- a/doc/fluid/advanced_usage/development/new_op/op_notes_en.md +++ b/doc/fluid/advanced_usage/development/new_op/op_notes_en.md @@ -11,7 +11,7 @@ The Fluid framework is designed to run on a variety of devices and third-party l Operator inheritance diagram: ![op_inheritance_relation_diagram](../../pics/op_inheritance_relation_diagram.png) -For further information, please refer to: [multi_devices](https://github.com/PaddlePaddle/FluidDoc/tree/develop/doc/fluid/design/multi_devices) , [scope](https://github.com/PaddlePaddle/FluidDoc/Blob/develop/doc/fluid/design/concepts/scope.md) , [Developer's_Guide_to_Paddle_Fluid](https://github.com/PaddlePaddle/FluidDoc/blob/release/1.2/doc/fluid/getstarted/Developer's_Guide_to_Paddle_Fluid.md) +For further information, please refer to: [multi_devices](https://github.com/PaddlePaddle/FluidDoc/tree/develop/doc/fluid/design/multi_devices) , [scope](https://github.com/PaddlePaddle/FluidDoc/blob/develop/doc/fluid/design/concepts/scope.md) , [Developer's_Guide_to_Paddle_Fluid](https://github.com/PaddlePaddle/FluidDoc/blob/release/1.2/doc/fluid/getstarted/Developer's_Guide_to_Paddle_Fluid.md) ### 2.Op's registration logic The registration entries for each Operator include: diff --git a/doc/fluid/advanced_usage/index.rst b/doc/fluid/advanced_usage/index_cn.rst similarity index 66% rename from doc/fluid/advanced_usage/index.rst rename to doc/fluid/advanced_usage/index_cn.rst index 51ed37fdac48c3270fbb68d778f381ae3e4daffe..b45b0ea3d56c355f585a8e250ae65d87d1f4dd7c 100644 --- a/doc/fluid/advanced_usage/index.rst +++ b/doc/fluid/advanced_usage/index_cn.rst @@ -2,17 +2,19 @@ 进阶使用 ######## -.. todo:: - 如果您非常熟悉 Fluid,期望获得更高效的模型或者定义自己的Operator,请阅读: - `Fluid 设计思想 <../advanced_usage/design_idea/fluid_design_idea.html>`_:介绍 Fluid 底层的设计思想,帮助您更好的理解框架运作过程 - - `预测部署 <../advanced_usage/deploy/index_cn.html>`_ :介绍如何应用训练好的模型进行预测 + - `预测部署 <../advanced_usage/deploy/index_cn.html>`_ :介绍如何应用训练好的模型进行预测 + + - `新增OP <../advanced_usage/development/new_op/index_cn.html>`_ :介绍新增operator的方法及注意事项 - - `新增operator <../advanced_usage/development/new_op/index_cn.html>`_ :介绍新增operator的方法及注意事项 + - `性能调优 <../advanced_usage/development/profiling/index_cn.html>`_ :介绍 Fluid 使用过程中的调优方法 - - `性能调优 <../advanced_usage/development/profiling/index_cn.html>`_ :介绍 Fluid 使用过程中的调优方法 + - `最佳实践 <../advanced_usage/best_practice/index_cn.html>`_ + + - `模型压缩工具库 <../advanced_usage/paddle_slim/paddle_slim.html>`_ 非常欢迎您为我们的开源社区做出贡献,关于如何贡献您的代码或文档,请阅读: @@ -27,7 +29,7 @@ deploy/index_cn.rst development/new_op/index_cn.rst development/profiling/index_cn.rst - development/contribute_to_paddle/index_cn.rst - development/write_docs_cn.md best_practice/index_cn.rst paddle_slim/paddle_slim.md + development/contribute_to_paddle/index_cn.rst + development/write_docs_cn.md diff --git a/doc/fluid/api_cn/data/data_reader_cn.rst b/doc/fluid/api_cn/data/data_reader_cn.rst index b9ba3c60b8b91fa93392f53505b5721cc493ccd0..630f82e357ed967bbdc52bc116f17bd5923db9f2 100644 --- a/doc/fluid/api_cn/data/data_reader_cn.rst +++ b/doc/fluid/api_cn/data/data_reader_cn.rst @@ -26,7 +26,7 @@ DataFeeder将reader返回的数据转换为可以输入Executor和ParallelExecut result = feeder.feed([([0] * 784, [9]), ([1] * 784, [1])]) -如果您想在使用多个GPU训练模型时预先将数据单独输入GPU端,可以使用decorate_reader函数。 +如果您想在使用多个GPU训练模型时预先将数据单独输入GPU端,可以使用decorate_reader函数。 **代码示例** @@ -90,7 +90,7 @@ DataFeeder将reader返回的数据转换为可以输入Executor和ParallelExecut -.. note:: +.. note:: 设备数量和mini-batches数量必须一致。 @@ -121,7 +121,7 @@ Reader - reader是一个读取数据(从文件、网络、随机数生成器等)并生成数据项的函数。 - reader creator是返回reader函数的函数。 - reader decorator是一个函数,它接受一个或多个reader,并返回一个reader。 - - batch reader是一个函数,它读取数据(从reader、文件、网络、随机数生成器等)并生成一批数据项。 + - batch reader是一个函数,它读取数据(从reader、文件、网络、随机数生成器等)并生成一批数据项。 Data Reader Interface @@ -133,10 +133,10 @@ Data Reader Interface iterable = data_reader() -从iterable生成的元素应该是单个数据条目,而不是mini batch。数据输入可以是单个项目,也可以是项目的元组,但应为 `支持的类型 `_ (如, numpy 1d array of float32, int, list of int) +从iterable生成的元素应该是单个数据条目,而不是mini batch。数据输入可以是单个项目,也可以是项目的元组,但应为 `支持的类型 <../../user_guides/howto/prepare_data/feeding_data.html#fluid>`_ (如, numpy 1d array of float32, int, list of int) -单项目数据读取器创建者的示例实现: +单项目数据读取器创建者的示例实现: .. code-block:: python @@ -147,7 +147,7 @@ Data Reader Interface return reader -多项目数据读取器创建者的示例实现: +多项目数据读取器创建者的示例实现: .. code-block:: python @@ -194,11 +194,11 @@ Data Reader Interface 参数: - **readers** - 将被组合的多个读取器。 - - **check_alignment** (bool) - 如果为True,将检查输入reader是否正确对齐。如果为False,将不检查对齐,将丢弃跟踪输出。默认值True。 + - **check_alignment** (bool) - 如果为True,将检查输入reader是否正确对齐。如果为False,将不检查对齐,将丢弃跟踪输出。默认值True。 返回:新的数据读取器 -抛出异常: ``ComposeNotAligned`` – reader的输出不一致。 当check_alignment设置为False,不会升高。 +抛出异常: ``ComposeNotAligned`` – reader的输出不一致。 当check_alignment设置为False,不会升高。 @@ -220,7 +220,7 @@ Data Reader Interface 创建数据读取器,该reader的数据输出将被无序排列。 -由原始reader创建的迭代器的输出将被缓冲到shuffle缓冲区,然后进行打乱。打乱缓冲区的大小由参数buf_size决定。 +由原始reader创建的迭代器的输出将被缓冲到shuffle缓冲区,然后进行打乱。打乱缓冲区的大小由参数buf_size决定。 参数: - **reader** (callable) – 输出会被打乱的原始reader @@ -257,7 +257,7 @@ Data Reader Interface PipeReader通过流从一个命令中读取数据,将它的stdout放到管道缓冲区中,并将其重定向到解析器进行解析,然后根据需要的格式生成数据。 -您可以使用标准Linux命令或调用其他Program来读取数据,例如通过HDFS、CEPH、URL、AWS S3中读取: +您可以使用标准Linux命令或调用其他Program来读取数据,例如通过HDFS、CEPH、URL、AWS S3中读取: **代码示例** @@ -340,7 +340,7 @@ Creator包包含一些简单的reader creator,可以在用户Program中使用 .. py:function:: paddle.reader.creator.np_array(x) -如果是numpy向量,则创建一个生成x个元素的读取器。或者,如果它是一个numpy矩阵,创建一个生成x行元素的读取器。或由最高维度索引的任何子超平面。 +如果是numpy向量,则创建一个生成x个元素的读取器。或者,如果它是一个numpy矩阵,创建一个生成x行元素的读取器。或由最高维度索引的任何子超平面。 参数: - **x** – 用于创建reader的numpy数组。 @@ -359,7 +359,7 @@ Creator包包含一些简单的reader creator,可以在用户Program中使用 .. py:function:: paddle.reader.creator.recordio(paths, buf_size=100) -从给定的recordio文件路径创建数据reader,用“,”分隔“,支持全局模式。 +从给定的recordio文件路径创建数据reader,用“,”分隔“,支持全局模式。 路径:recordio文件的路径,可以是字符串或字符串列表。 diff --git a/doc/fluid/api_cn/index_cn.rst b/doc/fluid/api_cn/index_cn.rst index 0b0eded620268daafa90baf305cf9025fbc7e424..f5c38bc9d6b5e1b1dc269724e78ee66b767ea23b 100644 --- a/doc/fluid/api_cn/index_cn.rst +++ b/doc/fluid/api_cn/index_cn.rst @@ -5,12 +5,13 @@ API .. toctree:: :maxdepth: 1 - ../api_guides/index.rst + ../api_guides/index_cn.rst fluid_cn.rst average_cn.rst backward_cn.rst clip_cn.rst data_feeder_cn.rst + dataset_cn.rst executor_cn.rst initializer_cn.rst io_cn.rst @@ -21,6 +22,5 @@ API profiler_cn.rst regularizer_cn.rst transpiler_cn.rst - dataset_cn.rst data/dataset_cn.rst data/data_reader_cn.rst diff --git a/doc/fluid/api_cn/profiler_cn.rst b/doc/fluid/api_cn/profiler_cn.rst index 2647650b7b25fb1e5e05c10d3f8f462d637f2192..11aeb457ab38661577fffdbed4c9728d424a2f66 100644 --- a/doc/fluid/api_cn/profiler_cn.rst +++ b/doc/fluid/api_cn/profiler_cn.rst @@ -12,7 +12,7 @@ cuda_profiler .. py:function:: paddle.fluid.profiler.cuda_profiler(output_file, output_mode=None, config=None) -CUDA分析器。通过CUDA运行时应用程序编程接口对CUDA程序进行性能分析。分析结果将以键-值对格式或逗号分隔的格式写入output_file。用户可以通过output_mode参数设置输出模式,并通过配置参数设置计数器/选项。默认配置是[' gpustarttimestamp ', ' gpustarttimestamp ', ' gridsize3d ', ' threadblocksize ', ' streamid ', ' enableonstart 0 ', ' conckerneltrace ']。然后,用户可使用 `NVIDIA Visual Profiler `_ 工具来加载这个输出文件以可视化结果。 +CUDA分析器。通过CUDA运行时应用程序编程接口对CUDA程序进行性能分析。分析结果将以键-值对格式或逗号分隔的格式写入output_file。用户可以通过output_mode参数设置输出模式,并通过配置参数设置计数器/选项。默认配置是[' gpustarttimestamp ', ' gpustarttimestamp ', ' gridsize3d ', ' threadblocksize ', ' streamid ', ' enableonstart 0 ', ' conckerneltrace ']。然后,用户可使用 `NVIDIA Visual Profiler `_ 工具来加载这个输出文件以可视化结果。 参数: @@ -28,7 +28,7 @@ CUDA分析器。通过CUDA运行时应用程序编程接口对CUDA程序进行 .. code-block:: python - + import paddle.fluid as fluid import paddle.fluid.profiler as profiler @@ -46,7 +46,7 @@ CUDA分析器。通过CUDA运行时应用程序编程接口对CUDA程序进行 for i in range(epoc): input = np.random.random(dshape).astype('float32') exe.run(fluid.default_main_program(), feed={'data': input}) - + # 之后可以使用 NVIDIA Visual Profile 可视化结果 @@ -67,20 +67,20 @@ profiler profile interface 。与cuda_profiler不同,此profiler可用于分析CPU和GPU程序。默认情况下,它记录CPU和GPU kernel,如果想分析其他程序,可以参考教程来在c++代码中添加更多代码。 -如果 state== ' All ',在profile_path 中写入文件 profile proto 。该文件记录执行期间的时间顺序信息。然后用户可以看到这个文件的时间轴,请参考 `https://github.com/PaddlePaddle/Paddle/blob/develop/doc/fluid/howto/optimization/timeline.md `_ +如果 state== ' All ',在profile_path 中写入文件 profile proto 。该文件记录执行期间的时间顺序信息。然后用户可以看到这个文件的时间轴,请参考 `这里 <../advanced_usage/development/profiling/timeline_cn.html>`_ 参数: - **state** (string) – profiling state, 取值为 'CPU' 或 'GPU', profiler 使用 CPU timer 或GPU timer 进行 profiling. 虽然用户可能在开始时指定了执行位置(CPUPlace/CUDAPlace),但是为了灵活性,profiler不会使用这个位置。 - **sorted_key** (string) – 如果为None,prfile的结果将按照事件的第一次结束时间顺序打印。否则,结果将按标志排序。标志取值为"call"、"total"、"max"、"min" "ave"之一,根据调用着的数量进行排序。total表示按总执行时间排序,max 表示按最大执行时间排序。min 表示按最小执行时间排序。ave表示按平均执行时间排序。 - **profile_path** (string) – 如果 state == 'All', 结果将写入文件 profile proto. - + 抛出异常: - ``ValueError`` – 如果state 取值不在 ['CPU', 'GPU', 'All']中. 如果 sorted_key 取值不在 ['calls', 'total', 'max', 'min', 'ave'] - + **代码示例** .. code-block:: python - + import paddle.fluid.profiler as profiler with profiler.profiler('All', 'total', '/tmp/profile') as prof: @@ -110,7 +110,7 @@ reset_profiler **代码示例** .. code-block:: python - + import paddle.fluid.profiler as profiler with profiler.profiler(state, 'total', '/tmp/profile'): for iter in range(10): @@ -133,10 +133,10 @@ start_profiler .. py:function:: paddle.fluid.profiler.start_profiler(state) 激活使用 profiler, 用户可以使用 ``fluid.profiler.start_profiler`` 和 ``fluid.profiler.stop_profiler`` 插入代码 -不能使用 ``fluid.profiler.profiler`` +不能使用 ``fluid.profiler.profiler`` -如果 state== ' All ',在profile_path 中写入文件 profile proto 。该文件记录执行期间的时间顺序信息。然后用户可以看到这个文件的时间轴,请参考 `https://github.com/PaddlePaddle/Paddle/blob/develop/doc/fluid/howto/optimization/timeline.md `_ +如果 state== ' All ',在profile_path 中写入文件 profile proto 。该文件记录执行期间的时间顺序信息。然后用户可以看到这个文件的时间轴,请参考 `这里 <../advanced_usage/development/profiling/timeline_cn.html>`_ 参数: - **state** (string) – profiling state, 取值为 'CPU' 或 'GPU' 或 'All', 'CPU' 代表只分析 cpu. 'GPU' 代表只分析 GPU . 'All' 会产生 timeline. @@ -147,7 +147,7 @@ start_profiler **代码示例** .. code-block:: python - + import paddle.fluid.profiler as profiler profiler.start_profiler('GPU') @@ -174,12 +174,12 @@ stop_profiler .. py:function:: paddle.fluid.profiler.stop_profiler(sorted_key=None, profile_path='/tmp/profile') 停止 profiler, 用户可以使用 ``fluid.profiler.start_profiler`` 和 ``fluid.profiler.stop_profiler`` 插入代码 -不能使用 ``fluid.profiler.profiler`` +不能使用 ``fluid.profiler.profiler`` 参数: - **sorted_key** (string) – 如果为None,prfile的结果将按照事件的第一次结束时间顺序打印。否则,结果将按标志排序。标志取值为"call"、"total"、"max"、"min" "ave"之一,根据调用着的数量进行排序。total表示按总执行时间排序,max 表示按最大执行时间排序。min 表示按最小执行时间排序。ave表示按平均执行时间排序。 - **profile_path** (string) - 如果 state == 'All', 结果将写入文件 profile proto. - + 抛出异常: - ``ValueError`` – 如果state 取值不在 ['CPU', 'GPU', 'All']中 @@ -187,7 +187,7 @@ stop_profiler **代码示例** .. code-block:: python - + import paddle.fluid.profiler as profiler profiler.start_profiler('GPU') diff --git a/doc/fluid/api_guides/index.rst b/doc/fluid/api_guides/index_cn.rst similarity index 65% rename from doc/fluid/api_guides/index.rst rename to doc/fluid/api_guides/index_cn.rst index 2b10c21e3be3a5e9dc48378cc34965ce139dfcf8..01e9d93ba8cff959ec4eea50584ed50638c5fd20 100755 --- a/doc/fluid/api_guides/index.rst +++ b/doc/fluid/api_guides/index_cn.rst @@ -1,8 +1,8 @@ =========== -API快速检索 +API分类检索 =========== -API快速检索分功能向您介绍PaddlePaddle Fluid的API体系和用法,帮助您快速了解PaddlePaddle Fluid API的全貌,包括以下几个模块: +本模块分功能向您介绍PaddlePaddle Fluid的API体系和用法,提高您的查找效率,帮助您快速了解PaddlePaddle Fluid API的全貌,包括以下几个模块: .. toctree:: :maxdepth: 1 @@ -18,5 +18,6 @@ API快速检索分功能向您介绍PaddlePaddle Fluid的API体系和用法, low_level/memory_optimize.rst low_level/executor.rst low_level/parallel_executor.rst + low_level/compiled_program.rst low_level/parameter.rst low_level/distributed/index.rst diff --git a/doc/fluid/api_guides/index_en.rst b/doc/fluid/api_guides/index_en.rst index f358f887af2b8d57d4c0118f37f527d59a2d418a..012a3a401c79937169de355d1a15e27027321a5f 100755 --- a/doc/fluid/api_guides/index_en.rst +++ b/doc/fluid/api_guides/index_en.rst @@ -18,5 +18,6 @@ This section introduces the Fluid API structure and usage, to help you quickly g low_level/memory_optimize_en.rst low_level/executor_en.rst low_level/parallel_executor_en.rst + low_level/compiled_program_en.rst low_level/parameter_en.rst low_level/distributed/index_en.rst diff --git a/doc/fluid/api_guides/low_level/compiled_program_cn.rst b/doc/fluid/api_guides/low_level/compiled_program.rst similarity index 100% rename from doc/fluid/api_guides/low_level/compiled_program_cn.rst rename to doc/fluid/api_guides/low_level/compiled_program.rst diff --git a/doc/fluid/api_guides/low_level/distributed/async_training.rst b/doc/fluid/api_guides/low_level/distributed/async_training.rst index 0a360e40c02f105fa09484d65debbece7172f62c..61f2ee0f79b271dc4d339d1659adbd3264475d80 100644 --- a/doc/fluid/api_guides/low_level/distributed/async_training.rst +++ b/doc/fluid/api_guides/low_level/distributed/async_training.rst @@ -4,7 +4,7 @@ 分布式异步训练 ############ -Fluid支持数据并行的分布式异步训练,API使用 :code:`DistributedTranspiler` 将单机网络配置转换成可以多机执行的 +Fluid支持数据并行的分布式异步训练,API使用 :code:`DistributeTranspiler` 将单机网络配置转换成可以多机执行的 :code:`pserver` 端程序和 :code:`trainer` 端程序。用户在不同的节点执行相同的一段代码,根据环境变量或启动参数, 可以执行对应的 :code:`pserver` 或 :code:`trainer` 角色。Fluid异步训练只支持pserver模式,异步训练和 `同步训练 <../distributed/sync_training.html>`_ 的主要差异在于:异步训练每个trainer的梯度是单独更新到参数上的, 而同步训练是所有trainer的梯度合并之后统一更新到参数上,因此,同步训练和异步训练的超参数需要分别调节。 @@ -16,10 +16,10 @@ API详细使用方法参考 :ref:`cn_api_fluid_DistributeTranspiler` ,简单 .. code-block:: python - config = fluid.DistributedTranspilerConfig() + config = fluid.DistributeTranspilerConfig() # 配置策略config config.slice_var_up = False - t = fluid.DistributedTranspiler(config=config) + t = fluid.DistributeTranspiler(config=config) t.transpile(trainer_id, program=main_program, pservers="192.168.0.1:6174,192.168.0.2:6174", diff --git a/doc/fluid/api_guides/low_level/executor.rst b/doc/fluid/api_guides/low_level/executor.rst index a8a9d8c1ae36ad8c69412a90073ab0a0f49ba566..c4368d513182e4b4ea3b88448039e250b751ff3a 100644 --- a/doc/fluid/api_guides/low_level/executor.rst +++ b/doc/fluid/api_guides/low_level/executor.rst @@ -4,11 +4,11 @@ 执行引擎 ########## -:code:`Executor` 实现了一个简易的执行器,所有的操作在其中顺序执行。你可以在Python脚本中运行:code:`Executor`。PaddlePaddle Fluid中有两种执行器。一种是:code:`Executor` 默认的单线程执行器,另一种是并行计算执行器,在:ref:`api_guide_parallel_executor`中进行了解释。`Executor`和:ref:`api_guide_parallel_executor`的配置不同,这可能会给部分用户带来困惑。为使执行器更加灵活,我们引入了:ref:`api_guide_compiled_program`,:ref:`api_guide_compiled_program`用于把一个程序转换为不同的优化组合,可以通过:code:`Executor`运行。 +:code:`Executor` 实现了一个简易的执行器,所有的操作在其中顺序执行。你可以在Python脚本中运行 :code:`Executor` 。PaddlePaddle Fluid中有两种执行器。一种是 :code:`Executor` 默认的单线程执行器,另一种是并行计算执行器,在 :ref:`api_guide_parallel_executor` 中进行了解释。``Executor`` 和 :ref:`api_guide_parallel_executor` 的配置不同,这可能会给部分用户带来困惑。为使执行器更加灵活,我们引入了 :ref:`api_guide_compiled_program` , :ref:`api_guide_compiled_program` 用于把一个程序转换为不同的优化组合,可以通过 :code:`Executor` 运行。 -:code:`Executor`的逻辑非常简单。建议在调试阶段用:code:`Executor`在一台计算机上完整地运行模型,然后转向多设备或多台计算机计算。 + :code:`Executor` 的逻辑非常简单。建议在调试阶段用 :code:`Executor` 在一台计算机上完整地运行模型,然后转向多设备或多台计算机计算。 -:code:`Executor`在构造时接受一个:code:`Place`,它既可能是:ref:`api_fluid_CPUPlace`也可能是:ref:`api_fluid_CUDAPlace`。 + :code:`Executor` 在构造时接受一个 :code:`Place` ,它既可能是 :ref:`api_fluid_CPUPlace` 也可能是 :ref:`api_fluid_CUDAPlace` 。 .. code-block:: python # 首先创建Executor。 @@ -16,12 +16,12 @@ exe = fluid.Executor(place) # 运行启动程序仅一次。 exe.run(fluid.default_startup_program()) - + # 直接运行主程序。 loss, = exe.run(fluid.default_main_program(), feed=feed_dict, fetch_list=[loss.name]) -简单样例请参照 `quick_start_fit_a_line `_ +简单样例请参照 `basics_fit_a_line <../../beginners_guide/basics/fit_a_line/README.cn.html>`_ - 相关API : - - :ref:`cn_api_fluid_Executor` + - :ref:`cn_api_fluid_Executor` diff --git a/doc/fluid/api_guides/low_level/executor_en.rst b/doc/fluid/api_guides/low_level/executor_en.rst index 5e4fdf16a9e9816da29832bfbea948ec87befe6d..eb92d2dd66510e982b8e8ca3b827ae66a695f2ce 100755 --- a/doc/fluid/api_guides/low_level/executor_en.rst +++ b/doc/fluid/api_guides/low_level/executor_en.rst @@ -8,7 +8,7 @@ Executor The logic of :code:`Executor` is very simple. It is suggested to thoroughly run the model with :code:`Executor` in debugging phase on one computer and then switch to mode of multiple devices or multiple computers to compute. -:code:`Executor` receives a :code:`Place` at construction, which can either be :ref:`api_fluid_CPUPlace` or :ref:`api_fluid_CUDAPlace`. +:code:`Executor` receives a :code:`Place` at construction, which can either be :ref:`api_fluid_CPUPlace` or :ref:`api_fluid_CUDAPlace`. .. code-block:: python @@ -18,14 +18,14 @@ The logic of :code:`Executor` is very simple. It is suggested to thoroughly run # Run the startup program once and only once. exe.run(fluid.default_startup_program()) - + # Run the main program directly. loss, = exe.run(fluid.default_main_program(), feed=feed_dict, fetch_list=[loss.name]) -For simple example please refer to `quick_start_fit_a_line `_ +For simple example please refer to `basics_fit_a_line <../../beginners_guide/basics/fit_a_line/README.html>`_ - Related API : - :ref:`api_fluid_Executor` diff --git a/doc/fluid/api_guides/low_level/layers/conv.rst b/doc/fluid/api_guides/low_level/layers/conv.rst index 3bdc848988fd2506d8816203859a80c84fb5432d..bbe279da8a67a6cc97420648baad35d1e12c204b 100644 --- a/doc/fluid/api_guides/low_level/layers/conv.rst +++ b/doc/fluid/api_guides/low_level/layers/conv.rst @@ -14,30 +14,31 @@ --------------------- 卷积需要依据滑动步长(stride)、填充长度(padding)、卷积核窗口大小(filter size)、分组数(groups)、扩张系数(dilation rate)来决定如何计算。groups最早在 `AlexNet `_ 中引入, 可以理解为将原始的卷积分为独立若干组卷积计算。 - + **注意**: 同cuDNN的方式,Fluid目前只支持在特征图上下填充相同的长度,左右也是。 -- 输入输出Layout: +- 输入输出Layout: 2D卷积输入特征的Layout为[N, C, H, W]或[N, H, W, C], N即batch size,C是通道数,H、W是特征的高度和宽度,输出特征和输入特征的Layout一致。(相应的3D卷积输入特征的Layout为[N, C, D, H, W]或[N, D, H, W, C],但**注意**,Fluid的卷积当前只支持[N, C, H, W],[N, C, D, H, W]。) - -- 卷积核的Layout: - + +- 卷积核的Layout: + Fluid中2D卷积的卷积核(也称权重)的Layout为[C_o, C_in / groups, f_h, f_w],C_o、C_in表示输出、输入通道数,f_h、f_w表示卷积核窗口的高度和宽度,按行序存储。(相应的2D卷积的卷积核Layout为[C_o, C_in / groups, f_d, f_h, d_w],同样按行序存储。) - -- 深度可分离卷积(depthwise separable convolution): - + +- 深度可分离卷积(depthwise separable convolution): + 在深度可分离卷积中包括depthwise convolution和pointwise convolution两组,这两个卷积的接口和上述普通卷积接口相同。前者可以通过给普通卷积设置groups来做,后者通过设置卷积核filters的大小为1x1,深度可分离卷积减少参数的同时减少了计算量。 - + 对于depthwise convolution,可以设置groups等于输入通道数,此时,2D卷积的卷积核形状为[C_o, 1, f_h, f_w]。 对于pointwise convolution,卷积核的形状为[C_o, C_in, 1, 1]。 - - **注意**:Fluid针对depthwise convolution的GPU计算做了高度优化,您可以通过在 :code:`fluid.layers.conv2d`接口设置 :code:`use_cudnn=False`来使用Fluid自身优化的CUDA程序。 - + + **注意**:Fluid针对depthwise convolution的GPU计算做了高度优化,您可以通过在 + :code:`fluid.layers.conv2d` 接口设置 :code:`use_cudnn=False` 来使用Fluid自身优化的CUDA程序。 + - 空洞卷积(dilated convolution): - + 空洞卷积相比普通卷积而言,卷积核在特征图上取值时不在连续,而是间隔的,这个间隔数称作dilation,等于1时,即为普通卷积,空洞卷积相比普通卷积的感受野更大。 - + - API汇总: - :ref:`cn_api_fluid_layers_conv2d` - :ref:`cn_api_fluid_layers_conv3d` @@ -50,14 +51,14 @@ Fluid可以表示变长的序列结构,这里的变长是指不同样本的时间步(step)数不一样,通常是一个2D的Tensor和一个能够区分的样本长度的辅助结构来表示。假定,2D的Tensor的形状是shape,shape[0]是所有样本的总时间步数,shape[1]是序列特征的大小。 -基于此数据结构的卷积在Fluid里称作序列卷积,也表示一维卷积。同图像卷积,序列卷积的输入参数有卷积核大小、填充大小、滑动步长,但与2D卷积不同的是,这些参数个数都为1。**注意**,目前仅支持stride为1的情况,输出序列的时间步数和输入序列相同。 +基于此数据结构的卷积在Fluid里称作序列卷积,也表示一维卷积。同图像卷积,序列卷积的输入参数有卷积核大小、填充大小、滑动步长,但与2D卷积不同的是,这些参数个数都为1。**注意**,目前仅支持stride为1的情况,输出序列的时间步数和输入序列相同。 假如:输入序列形状为(T, N), T即该序列的时间步数,N是序列特征大小;卷积核的上下文步长为K,输出序列长度为M,则卷积核权重形状为(K * N, M),输出序列形状为(T, M)。 - + 另外,参考DeepSpeech,Fluid实现了行卷积row convolution, 或称 `look ahead convolution `_ , 该卷积相比上述普通序列卷积可以减少参数。 - + - API汇总: - :ref:`cn_api_fluid_layers_sequence_conv` diff --git a/doc/fluid/api_guides/low_level/layers/sparse_update.rst b/doc/fluid/api_guides/low_level/layers/sparse_update.rst index 5a7405e908d0e7d187c7e98c8e003c89e2a9d500..0f4113504b927e4a26da61fe0dee3f1605845368 100644 --- a/doc/fluid/api_guides/low_level/layers/sparse_update.rst +++ b/doc/fluid/api_guides/low_level/layers/sparse_update.rst @@ -37,9 +37,9 @@ API详细使用方法参考 :ref:`cn_api_fluid_layers_embedding` ,以下是一 以上参数中: -- :code:`is_sparse` : 反向计算的时候梯度是否为sparse tensor。如果不设置,梯度是一个 `LodTensor `_ 。默认为False。 +- :code:`is_sparse` : 反向计算的时候梯度是否为sparse tensor。如果不设置,梯度是一个 `LodTensor `_ 。默认为False。 -- :code:`is_distributed` : 标志是否是用在分布式的场景下。一般大规模稀疏更新(embedding的第0维维度很大,比如几百万以上)才需要设置。具体可以参考大规模稀疏的API guide :ref:`api_guide_async_training` 。默认为False。 +- :code:`is_distributed` : 标志是否是用在分布式的场景下。一般大规模稀疏更新(embedding的第0维维度很大,比如几百万以上)才需要设置。具体可以参考大规模稀疏的API guide :ref:`cn_api_guide_async_training` 。默认为False。 - API汇总: - :ref:`cn_api_fluid_layers_embedding` diff --git a/doc/fluid/api_guides/low_level/layers/sparse_update_en.rst b/doc/fluid/api_guides/low_level/layers/sparse_update_en.rst index 75e541103fbb628886686917c1a4f6e90fa4faad..f97ef279b8f8bc8dd5fb5beb24d13dc9815eb942 100755 --- a/doc/fluid/api_guides/low_level/layers/sparse_update_en.rst +++ b/doc/fluid/api_guides/low_level/layers/sparse_update_en.rst @@ -37,9 +37,9 @@ API reference :ref:`api_fluid_layers_embedding` . Here is a simple example: The parameters: -- :code:`is_sparse` : Whether the gradient is a sparse tensor in the backward calculation. If not set, the gradient is a `LodTensor `_ . The default is False. +- :code:`is_sparse` : Whether the gradient is a sparse tensor in the backward calculation. If not set, the gradient is a `LodTensor `_ . The default is False. - :code:`is_distributed` : Whether the current training is in a distributed scenario. Generally, this parameter can only be set in large-scale sparse updates (the 0th dimension of embedding is very large, such as several million or more). For details, please refer to the large-scale sparse API guide :ref:`api_guide_async_training`. The default is False. - API : - - :ref:`api_fluid_layers_embedding` \ No newline at end of file + - :ref:`api_fluid_layers_embedding` diff --git a/doc/fluid/api_guides/low_level/nets.rst b/doc/fluid/api_guides/low_level/nets.rst index 134d9359f829f473c23e17e5cac261574cb7d20c..5e7f1fef7239b65e87a1cf3a7714aad6167deb8a 100644 --- a/doc/fluid/api_guides/low_level/nets.rst +++ b/doc/fluid/api_guides/low_level/nets.rst @@ -33,8 +33,9 @@ API Reference 请参考 :ref:`cn_api_fluid_nets_img_conv_group` :code:`sequence_conv_pool` 是由 :ref:`cn_api_fluid_layers_sequence_conv` 与 :ref:`cn_api_fluid_layers_sequence_pool` 串联而成。 该模块在 `自然语言处理 `_ 以及 `语音识别 `_ 等领域均有广泛应用, -比如 `文本分类模型 `_ , -`TagSpace `_ 以及 `Multi-view Simnet `_ 等模型。 +比如 `文本分类模型 `_ , +`TagSpace `_ 以及 +`Multi-view Simnet `_ 等模型。 API Reference 请参考 :ref:`cn_api_fluid_nets_sequence_conv_pool` @@ -55,7 +56,7 @@ API Reference 请参考 :ref:`cn_api_fluid_nets_glu` .. math:: Attention(Q, K, V)= softmax(QK^\mathrm{T})V -该模块广泛使用在 `机器翻译 `_ 的模型中,比如 `Transformer `_ 。 +该模块广泛使用在 `机器翻译 `_ 的模型中,比如 `Transformer `_ 。 API Reference 请参考 :ref:`cn_api_fluid_nets_scaled_dot_product_attention` diff --git a/doc/fluid/api_guides/low_level/nets_en.rst b/doc/fluid/api_guides/low_level/nets_en.rst index aaf4f3ce9b88fa5ddd7777632519790f2aaba85b..b31d834917dad9bc1223f511d14673ceb171bcb3 100644 --- a/doc/fluid/api_guides/low_level/nets_en.rst +++ b/doc/fluid/api_guides/low_level/nets_en.rst @@ -32,8 +32,8 @@ For API Reference, please refer to :ref:`api_fluid_nets_img_conv_group` -------------------- :code:`sequence_conv_pool` is got by concatenating :ref:`api_fluid_layers_sequence_conv` with :ref:`api_fluid_layers_sequence_pool`. -The module is widely used in the field of `natural language processing `_ and `speech recognition `_ . Models such as the `text classification model `_ , -`TagSpace `_ and `Multi view Simnet `_. +The module is widely used in the field of `natural language processing `_ and `speech recognition `_ . Models such as the `text classification model `_ , +`TagSpace `_ and `Multi view Simnet `_. For API Reference, please refer to :ref:`api_fluid_nets_sequence_conv_pool` @@ -54,6 +54,6 @@ For the input data :code:`Queries` , :code:`Key` and :code:`Values`, calculate t .. math:: Attention(Q, K, V)= softmax(QK^\mathrm{T})V -This module is widely used in the model of `machine translation `_, such as `Transformer `_ . +This module is widely used in the model of `machine translation `_, such as `Transformer `_ . For API Reference, please refer to :ref:`api_fluid_nets_scaled_dot_product_attention` diff --git a/doc/fluid/api_guides/low_level/parallel_executor.rst b/doc/fluid/api_guides/low_level/parallel_executor.rst index 629c0d445b652ff805b72fed93b369d88abcaf9f..f2ae56c032480d474d1ec263451a5971831f1343 100644 --- a/doc/fluid/api_guides/low_level/parallel_executor.rst +++ b/doc/fluid/api_guides/low_level/parallel_executor.rst @@ -29,7 +29,7 @@ **注意** :如果在Reduce模式下使用 :code:`CPU` 多线程执行 :code:`Program` , :code:`Program` 的参数在多个线程间是共享的,在某些模型上,Reduce模式可以大幅节省内存。 -鉴于模型的执行速率和模型结构及执行器的执行策略有关,:code:`ParallelExecutor` 允许你修改执行器的相关参数,例如线程池的规模( :code:`num_threads` )、为清除临时变量:code:`num_iteration_per_drop_scope`需要进行的循环次数。更多信息请参照:ref:`cn_api_fluid_ExecutionStrategy`。 +鉴于模型的执行速率和模型结构及执行器的执行策略有关,:code:`ParallelExecutor` 允许你修改执行器的相关参数,例如线程池的规模( :code:`num_threads` )、为清除临时变量 :code:`num_iteration_per_drop_scope` 需要进行的循环次数。更多信息请参照 :ref:`cn_api_fluid_ExecutionStrategy` 。 .. code-block:: python @@ -49,8 +49,8 @@ exec_strategy.num_threads = dev_count * 4 # the size of thread pool. build_strategy = fluid.BuildStrategy() build_strategy.memory_optimize = True if memory_opt else False - train_exe = fluid.ParallelExecutor(use_cuda=use_cuda, - main_program=train_program, + train_exe = fluid.ParallelExecutor(use_cuda=use_cuda, + main_program=train_program, build_strategy=build_strategy, exec_strategy=exec_strategy, loss_name=loss.name) diff --git a/doc/fluid/beginners_guide/quick_start/fit_a_line/README.cn.md b/doc/fluid/beginners_guide/basics/fit_a_line/README.cn.md similarity index 100% rename from doc/fluid/beginners_guide/quick_start/fit_a_line/README.cn.md rename to doc/fluid/beginners_guide/basics/fit_a_line/README.cn.md diff --git a/doc/fluid/beginners_guide/quick_start/fit_a_line/README.md b/doc/fluid/beginners_guide/basics/fit_a_line/README.md similarity index 100% rename from doc/fluid/beginners_guide/quick_start/fit_a_line/README.md rename to doc/fluid/beginners_guide/basics/fit_a_line/README.md diff --git a/doc/fluid/beginners_guide/quick_start/fit_a_line/image b/doc/fluid/beginners_guide/basics/fit_a_line/image similarity index 100% rename from doc/fluid/beginners_guide/quick_start/fit_a_line/image rename to doc/fluid/beginners_guide/basics/fit_a_line/image diff --git a/doc/fluid/beginners_guide/basics/index.rst b/doc/fluid/beginners_guide/basics/index_cn.rst similarity index 95% rename from doc/fluid/beginners_guide/basics/index.rst rename to doc/fluid/beginners_guide/basics/index_cn.rst index d68f549266305ea5a9b61d0d6282575fd0a7e3ed..55f9bc0e7116e9677ae05080bbbd4fdefcdba13c 100644 --- a/doc/fluid/beginners_guide/basics/index.rst +++ b/doc/fluid/beginners_guide/basics/index_cn.rst @@ -1,9 +1,9 @@ ################ -深度学习基础 +深度学习基础教程 ################ -本章由7篇文档组成,它们按照简单到难的顺序排列,将指导您如何使用PaddlePaddle完成基础的深度学习任务 +本章由9篇文档组成,它们按照简单到难的顺序排列,将指导您如何使用PaddlePaddle完成基础的深度学习任务 本章文档涉及大量了深度学习基础知识,也介绍了如何使用PaddlePaddle实现这些内容,请参阅以下说明了解如何使用: @@ -15,6 +15,8 @@ .. toctree:: :titlesonly: + fit_a_line/README.cn.md + recognize_digits/README.cn.md image_classification/index.md word2vec/index.md recommender_system/index.md diff --git a/doc/fluid/beginners_guide/basics/index_en.rst b/doc/fluid/beginners_guide/basics/index_en.rst index 97efb54d132e5a7185fe39fd0b0c6702a1e5a44e..8c810cf41c2caf5dd3c4a3abc7e4d60aac5dbc7f 100644 --- a/doc/fluid/beginners_guide/basics/index_en.rst +++ b/doc/fluid/beginners_guide/basics/index_en.rst @@ -1,8 +1,8 @@ -########################## +############################ Basic Deep Learning Models -########################## +############################ -This section collects six documents arranging from the simplest to the most challenging, which will guide you through the basic deep learning tasks in PaddlePaddle. +This section collects 8 documents arranging from the simplest to the most challenging, which will guide you through the basic deep learning tasks in PaddlePaddle. The documentation in this chapter covers a lot of deep learning basics and how to implement them with PaddlePaddle. See the instructions below for how to use: @@ -15,6 +15,8 @@ The book you are reading is an "interactive" e-book - each chapter can be run in .. toctree:: :titlesonly: + fit_a_line/README.md + recognize_digits/README.md image_classification/index_en.md word2vec/index_en.md recommender_system/index_en.md @@ -45,7 +47,7 @@ Just run these in shell: docker run -d -p 8888:8888 paddlepaddle/book -It downloads the Docker image for running books from DockerHub.com. +It downloads the Docker image for running books from DockerHub.com. To read and edit this book on-line, please visit http://localhost:8888 in your browser. If the Internet connection to DockerHub.com is compromised, try our spare docker image named docker.paddlepaddlehub.com: diff --git a/doc/fluid/beginners_guide/quick_start/recognize_digits/README.cn.md b/doc/fluid/beginners_guide/basics/recognize_digits/README.cn.md similarity index 100% rename from doc/fluid/beginners_guide/quick_start/recognize_digits/README.cn.md rename to doc/fluid/beginners_guide/basics/recognize_digits/README.cn.md diff --git a/doc/fluid/beginners_guide/quick_start/recognize_digits/README.md b/doc/fluid/beginners_guide/basics/recognize_digits/README.md similarity index 100% rename from doc/fluid/beginners_guide/quick_start/recognize_digits/README.md rename to doc/fluid/beginners_guide/basics/recognize_digits/README.md diff --git a/doc/fluid/beginners_guide/quick_start/recognize_digits/image b/doc/fluid/beginners_guide/basics/recognize_digits/image similarity index 100% rename from doc/fluid/beginners_guide/quick_start/recognize_digits/image rename to doc/fluid/beginners_guide/basics/recognize_digits/image diff --git a/doc/fluid/beginners_guide/index.rst b/doc/fluid/beginners_guide/index_cn.rst similarity index 73% rename from doc/fluid/beginners_guide/index.rst rename to doc/fluid/beginners_guide/index_cn.rst index a01ae5c40e20e266ed49f81b333c5e2aee0e3ba7..5eb973831c29206daecd811540ad051964048330 100644 --- a/doc/fluid/beginners_guide/index.rst +++ b/doc/fluid/beginners_guide/index_cn.rst @@ -6,23 +6,24 @@ PaddlePaddle (PArallel Distributed Deep LEarning)是一个易用、高效、灵 您可参考PaddlePaddle的 `Github `_ 了解详情,也可阅读 `版本说明 <../release_note.html>`_ 了解新版本的特性 -当您第一次来到PaddlePaddle,请您首先阅读以下文档,了解安装方法: +让我们从这里开始: - - `安装说明 <../beginners_guide/install/index_cn.html>`_:我们支持在Ubuntu/CentOS/Windows/MacOS环境上的安装 + - `快速开始 <../beginners_guide/quick_start.html>`_ -如果您已经具备一定的深度学习基础,第一次使用PaddlePaddle时,可以跟随下列简单的模型案例供您快速上手: +当您第一次来到PaddlePaddle,请您首先阅读以下文档,了解安装方法: - - `Fluid编程指南 <../beginners_guide/programming_guide/programming_guide.html>`_:介绍 Fluid 的基本概念和使用方法 + - `安装说明 <../beginners_guide/install/index_cn.html>`_:我们支持在Ubuntu/CentOS/Windows/MacOS环境上的安装 - - `快速入门 <../beginners_guide/quick_start/index.html>`_:提供线性回归和识别数字两个入门级模型,帮助您快速上手训练网络 +这里为您提供了更多学习资料: - `深度学习基础 <../beginners_guide/basics/index.html>`_:覆盖图像分类、个性化推荐、机器翻译等多个深度领域的基础知识,提供 Fluid 实现案例 + - `Fluid编程指南 <../beginners_guide/programming_guide/programming_guide.html>`_:介绍 Fluid 的基本概念和使用方法 .. toctree:: :hidden: + quick_start_cn.rst install/index_cn.rst - quick_start/index.rst - basics/index.rst + basics/index_cn.rst programming_guide/programming_guide.md diff --git a/doc/fluid/beginners_guide/index_en.rst b/doc/fluid/beginners_guide/index_en.rst index eaa0d1a0fa1bf4a63d30dbf5477a858b6aceab76..c76a0e2a5b41d37d5ef474b6fa935f2828c9aab8 100644 --- a/doc/fluid/beginners_guide/index_en.rst +++ b/doc/fluid/beginners_guide/index_en.rst @@ -15,8 +15,6 @@ If you have been armed with certain level of deep learning knowledge, and it hap - `Programming with Fluid <../beginners_guide/programming_guide/programming_guide_en.html>`_ : Core concepts and basic usage of Fluid - - `Quick Start <../beginners_guide/quick_start/index_en.html>`_: Two easy-to-go models, linear regression model and digit recognition model, are in place to speed up your study of training neural networks - - `Deep Learning Basics <../beginners_guide/basics/index_en.html>`_: This section encompasses various fields of fundamental deep learning knowledge, such as image classification, customized recommendation, machine translation, and examples implemented by Fluid are provided. @@ -24,6 +22,5 @@ If you have been armed with certain level of deep learning knowledge, and it hap :hidden: install/index_en.rst - quick_start/index_en.rst basics/index_en.rst programming_guide/programming_guide_en.md diff --git a/doc/fluid/beginners_guide/install/compile/compile_MacOS_en.md b/doc/fluid/beginners_guide/install/compile/compile_MacOS_en.md index ffef18edd835058285656daf70aa3879994acf9f..bb48e5f363628fe854f55afced5f437e6bd98a53 100644 --- a/doc/fluid/beginners_guide/install/compile/compile_MacOS_en.md +++ b/doc/fluid/beginners_guide/install/compile/compile_MacOS_en.md @@ -68,7 +68,7 @@ Once you have **properly installed Docker**, you can start **compiling PaddlePad 9. Execute cmake: - > For details on the compilation options, see the [compilation options table](../Tables.html/#Compile). + > For details on the compilation options, see the [compilation options table](../Tables_en.html/#Compile). * For users who need to compile the **CPU version PaddlePaddle**: diff --git a/doc/fluid/beginners_guide/install/install_Ubuntu_en.md b/doc/fluid/beginners_guide/install/install_Ubuntu_en.md index bff13faba002b02a79f62f3a277126aada390563..aeb9d8c5995c248c30e1ebfb191fde64f5a2b255 100644 --- a/doc/fluid/beginners_guide/install/install_Ubuntu_en.md +++ b/doc/fluid/beginners_guide/install/install_Ubuntu_en.md @@ -4,7 +4,7 @@ This instruction describes how to install PaddlePaddle on a *64-bit desktop or laptop* and Ubuntu system. The Ubuntu systems we support must meet the following requirements: -Please note: Attempts on other systems may cause the installation to fail. Please ensure that your environment meets the conditions. The installation we provide by default requires your computer processor to support the AVX instruction set. Otherwise, please select the version of `no_avx` in the [latest Release installation package list](./Tables.html/#ciwhls-release). +Please note: Attempts on other systems may cause the installation to fail. Please ensure that your environment meets the conditions. The installation we provide by default requires your computer processor to support the AVX instruction set. Otherwise, please select the version of `no_avx` in the [latest Release installation package list](./Tables_en.html/#ciwhls-release). Under Ubuntu, you can use `cat /proc/cpuinfo | grep avx` to check if your processor supports the AVX instruction set. @@ -80,9 +80,9 @@ Now let's install PaddlePaddle: * For users who need **the GPU version PaddlePaddle**: `pip install paddlepaddle-gpu` or `pip3 install paddlepaddle-gpu` > 1.In order to prevent problem "nccl.h cannot be found", please first install nccl2 according to the following command (here is ubuntu 16.04, CUDA9, ncDNN v7 nccl2 installation instructions), for more information about the installation information, please refer to [the NVIDIA official website](https://developer.nvidia.com/nccl/nccl-download): - + i. `Wget http://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1604/x86_64/nvidia-machine-learning-repo-ubuntu1604_1.0.0-1_amd64.deb` - ii. `dpkg -i nvidia-machine- Learning-repo-ubuntu1604_1.0.0-1_amd64.deb` + ii. `dpkg -i nvidia-machine- Learning-repo-ubuntu1604_1.0.0-1_amd64.deb` iii. `sudo apt-get install -y libnccl2=2.2.13-1+cuda9.0 libnccl-dev=2.2.13-1+cuda9.0` > 2.If you do not specify the pypi package version number, we will by default provide you with a version of PaddlePaddle that supports Cuda 9/cuDNN v7. diff --git a/doc/fluid/beginners_guide/programming_guide/programming_guide.md b/doc/fluid/beginners_guide/programming_guide/programming_guide.md index 5baa35490aa75c957313d6394fe5b567d5ccea55..d11d815578c899383be139c5fe63d91b4fd182dc 100644 --- a/doc/fluid/beginners_guide/programming_guide/programming_guide.md +++ b/doc/fluid/beginners_guide/programming_guide/programming_guide.md @@ -407,17 +407,17 @@ outs = exe.run( ``` 可以看到100次迭代后,预测值已经非常接近真实值了,损失值也从初始值9.05下降到了0.01。 - 恭喜您!已经成功完成了第一个简单网络的搭建,想尝试线性回归的进阶版——房价预测模型,请阅读:[线性回归](../../beginners_guide/quick_start/fit_a_line/README.cn.html)。更多丰富的模型实例可以在[模型库](../../user_guides/models/index_cn.html)中找到。 + 恭喜您!已经成功完成了第一个简单网络的搭建,想尝试线性回归的进阶版——房价预测模型,请阅读:[线性回归](../../beginners_guide/basics/fit_a_line/README.cn.html)。更多丰富的模型实例可以在[模型库](../../user_guides/models/index_cn.html)中找到。 ## What's next 如果您已经掌握了基本操作,可以进行下一阶段的学习了: -跟随这一教程将学习到如何对实际问题建模并使用fluid构建模型:[配置简单的网络](../../user_guides/howto/configure_simple_model/index.html)。 +跟随这一教程将学习到如何对实际问题建模并使用fluid构建模型:[配置简单的网络](../../user_guides/howto/configure_simple_model/index_cn.html)。 -完成网络搭建后,可以开始在单机或多机上训练您的网络了,详细步骤请参考[训练神经网络](../../user_guides/howto/training/index.html)。 +完成网络搭建后,可以开始在单机或多机上训练您的网络了,详细步骤请参考[训练神经网络](../../user_guides/howto/training/index_cn.html)。 -除此之外,使用文档模块根据开发者的不同背景划分了三个学习阶段:[新手入门](../../beginners_guide/index.html)、[使用指南](../../user_guides/index.html)和[进阶使用](../../advanced_usage/index.html)。 +除此之外,使用文档模块根据开发者的不同背景划分了三个学习阶段:[新手入门](../../beginners_guide/index_cn.html)、[使用指南](../../user_guides/index_cn.html)和[进阶使用](../../advanced_usage/index_cn.html)。 -如果您希望阅读更多场景下的应用案例,可以跟随导航栏进入[快速入门](../../beginners_guide/quick_start/index.html)和[深度学习基础知识](../../beginners_guide/basics/index.html)。已经具备深度学习基础知识的用户,可以从[使用指南](../../user_guides/index.html)开始阅读。 +如果您希望阅读更多场景下的应用案例,可以参考[深度学习基础教程](../../beginners_guide/basics/index_cn.html)。已经具备深度学习基础知识的用户,可以从[使用指南](../../user_guides/index_cn.html)开始阅读。 diff --git a/doc/fluid/beginners_guide/programming_guide/programming_guide_en.md b/doc/fluid/beginners_guide/programming_guide/programming_guide_en.md index 411e89f109c4c5bcbb5f0257a3bb0184608399ab..c7ee85d4f2c82c12f40cd79f4ed937dddfc03c1a 100644 --- a/doc/fluid/beginners_guide/programming_guide/programming_guide_en.md +++ b/doc/fluid/beginners_guide/programming_guide/programming_guide_en.md @@ -414,7 +414,7 @@ Firstly, define input data format, model structure,loss function and optimized a ``` Now we discover that predicted value is nearly close to real value and the loss value descends from original value 9.05 to 0.01 after iteration for 100 times. - Congratulations! You have succeed to create a simple network. If you want to try advanced linear regression —— predict model of housing price, please read [linear regression](../../beginners_guide/quick_start/fit_a_line/README.en.html). More examples of model can be found in [models](../../user_guides/models/index_en.html). + Congratulations! You have succeed to create a simple network. If you want to try advanced linear regression —— predict model of housing price, please read [linear regression](../../beginners_guide/basics/fit_a_line/README.en.html). More examples of model can be found in [models](../../user_guides/models/index_en.html). ## What's next @@ -427,4 +427,4 @@ After the construction of network, you can start training your network in single In addition, there are three learning levels in documentation according to developer's background and experience: [Beginner's Guide](../../beginners_guide/index_en.html) , [User Guides](../../user_guides/index_en.html) and [Advanced User Guides](../../advanced_usage/index_en.html). -If you want to read examples in more application scenarios, you can go to [quick start](../../beginners_guide/quick_start/index_en.html) and [basic knowledge of deep learning](../../beginners_guide/basics/index_en.html) .If you have learned basic knowledge of deep learning, you can read from [user guide](../../user_guides/index_en.html). +If you want to read examples in more application scenarios, you can go to [basic knowledge of deep learning](../../beginners_guide/basics/index_en.html) .If you have learned basic knowledge of deep learning, you can read from [user guide](../../user_guides/index_en.html). diff --git a/doc/fluid/beginners_guide/quick_start/index.rst b/doc/fluid/beginners_guide/quick_start/index.rst deleted file mode 100644 index 69e45881ab2efe93dab1a0b5dceca63dae9ef8f0..0000000000000000000000000000000000000000 --- a/doc/fluid/beginners_guide/quick_start/index.rst +++ /dev/null @@ -1,11 +0,0 @@ -######## -快速入门 -######## - -欢迎来到快速入门部分,在这里,我们将向您介绍如何通过PaddlePaddle Fluid实现经典的线性回归和手写识别的模型,以下两篇文档将指导您使用真实数据集搭建起模型、进行训练和预测: - -.. toctree:: - :titlesonly: - - fit_a_line/README.cn.md - recognize_digits/README.cn.md diff --git a/doc/fluid/beginners_guide/quick_start/index_en.rst b/doc/fluid/beginners_guide/quick_start/index_en.rst deleted file mode 100644 index 6964c43def5240c8ba02aae7b4278f0747fa2817..0000000000000000000000000000000000000000 --- a/doc/fluid/beginners_guide/quick_start/index_en.rst +++ /dev/null @@ -1,13 +0,0 @@ -############## -Quick Start -############## - -Welcome to Quick Start! - -This section will tutor you to invent your won models of classical *linear Regression* and *Handwritten Digits Recognition* tasks in PaddlePaddle Fluid. The following tutorials provide details on model definition, training, and inference in a friendly manner based on real-life datasets: - -.. toctree:: - :titlesonly: - - fit_a_line/README.md - recognize_digits/README.md diff --git a/doc/fluid/beginners_guide/quick_start.rst b/doc/fluid/beginners_guide/quick_start_cn.rst similarity index 100% rename from doc/fluid/beginners_guide/quick_start.rst rename to doc/fluid/beginners_guide/quick_start_cn.rst diff --git a/doc/fluid/index_cn.rst b/doc/fluid/index_cn.rst index d524b85496c0a999efeffeb1c271950ee974ceb3..9549801e133de616f031bcb6c1a61887f86215ff 100644 --- a/doc/fluid/index_cn.rst +++ b/doc/fluid/index_cn.rst @@ -11,8 +11,8 @@ :maxdepth: 1 - beginners_guide/index.rst - user_guides/index.rst - advanced_usage/index.rst + beginners_guide/index_cn.rst + user_guides/index_cn.rst + advanced_usage/index_cn.rst api_cn/index_cn.rst - release_note.rst + release_note_cn.rst diff --git a/doc/fluid/release_note.rst b/doc/fluid/release_note_cn.rst similarity index 100% rename from doc/fluid/release_note.rst rename to doc/fluid/release_note_cn.rst diff --git a/doc/fluid/user_guides/howto/configure_simple_model/index.rst b/doc/fluid/user_guides/howto/configure_simple_model/index_cn.rst similarity index 100% rename from doc/fluid/user_guides/howto/configure_simple_model/index.rst rename to doc/fluid/user_guides/howto/configure_simple_model/index_cn.rst diff --git a/doc/fluid/user_guides/howto/evaluation_and_debugging/index.rst b/doc/fluid/user_guides/howto/evaluation_and_debugging/index_cn.rst similarity index 100% rename from doc/fluid/user_guides/howto/evaluation_and_debugging/index.rst rename to doc/fluid/user_guides/howto/evaluation_and_debugging/index_cn.rst diff --git a/doc/fluid/user_guides/howto/prepare_data/index.rst b/doc/fluid/user_guides/howto/prepare_data/index_cn.rst similarity index 100% rename from doc/fluid/user_guides/howto/prepare_data/index.rst rename to doc/fluid/user_guides/howto/prepare_data/index_cn.rst diff --git a/doc/fluid/user_guides/howto/training/deploy_ctr_on_baidu_cloud_cn.rst b/doc/fluid/user_guides/howto/training/deploy_ctr_on_baidu_cloud_cn.rst index 23c11da195d990e9b23252fb30968b17982f9c5d..bc2d731acd866fb8f56175c72d5ac9574d5098d7 100644 --- a/doc/fluid/user_guides/howto/training/deploy_ctr_on_baidu_cloud_cn.rst +++ b/doc/fluid/user_guides/howto/training/deploy_ctr_on_baidu_cloud_cn.rst @@ -1,6 +1,6 @@ .. _deploy_ctr_on_baidu_cloud_cn: -百度云分布式训练CTR +在百度云分布式训练CTR ========================= Fluid支持数据并行的分布式训练,也支持基于Kubernetes的分布式部署。本文以百度云为例,说明如何通过在云服务器上分布式训练Click-Through-Rate(以下简称ctr)任务。 diff --git a/doc/fluid/user_guides/howto/training/index.rst b/doc/fluid/user_guides/howto/training/index_cn.rst similarity index 77% rename from doc/fluid/user_guides/howto/training/index.rst rename to doc/fluid/user_guides/howto/training/index_cn.rst index e517bf72e7a56889d3e1f9d2b812ea6debf5dc9e..d4ff8c5cc65d23d3578a84aafe45e4db2752516d 100644 --- a/doc/fluid/user_guides/howto/training/index.rst +++ b/doc/fluid/user_guides/howto/training/index_cn.rst @@ -7,6 +7,6 @@ PaddlePaddle Fluid支持单机训练和多节点训练。每种训练模式下 .. toctree:: :maxdepth: 1 - single_node - multi_node - save_load_variables + single_node.rst + multi_node.rst + save_load_variables.rst diff --git a/doc/fluid/user_guides/howto/training/multi_node.rst b/doc/fluid/user_guides/howto/training/multi_node.rst index 9008f47c1e88c9682ea715921ec3cd4ffee9e8bc..5835c14db00ef8dc3a0c6640b16ad80b54be387f 100644 --- a/doc/fluid/user_guides/howto/training/multi_node.rst +++ b/doc/fluid/user_guides/howto/training/multi_node.rst @@ -2,9 +2,10 @@ 多机训练 ######## -.. toctree:: - :maxdepth: 1 +.. toctree:: + :maxdepth: 1 - cluster_quick_start.rst - cluster_howto.rst - train_on_baidu_cloud_cn.rst + cluster_quick_start.rst + cluster_howto.rst + train_on_baidu_cloud_cn.rst + deploy_ctr_on_baidu_cloud_cn.rst diff --git a/doc/fluid/user_guides/howto/training/save_load_variables.rst b/doc/fluid/user_guides/howto/training/save_load_variables.rst index 09b2a17ea2e4137d7135f398c347d23830c2bcdc..cdb585cc26994b38bf413d13c3efa5b0717cb1cf 100644 --- a/doc/fluid/user_guides/howto/training/save_load_variables.rst +++ b/doc/fluid/user_guides/howto/training/save_load_variables.rst @@ -1,8 +1,8 @@ .. _user_guide_save_load_vars: -################## -模型/变量的保存、载入与增量训练 -################## +############################# +模型/变量的保存/载入与增量训练 +############################# 模型变量分类 ############ @@ -69,7 +69,7 @@ 载入模型用于对新样本的预测 ========================== -对于通过 :code:`fluid.io.save_params` 保存的模型,可以使用 :code:`fluid.io.load_params` +对于通过 :code:`fluid.io.save_params` 保存的模型,可以使用 :code:`fluid.io.load_params` 来进行载入。 例如: @@ -149,7 +149,7 @@ fluid.io.load_persistables(exe, path, startup_prog) main_prog = fluid.default_main_program() exe.run(main_prog) - + 上面的例子中,通过调用 :code:`fluid.io.load_persistables` 函数,PaddlePaddle Fluid会从默认 :code:`fluid.Program` 也就是 :code:`prog` 的所有模型变量中找出长期变量,从指定的 :code:`path` 目录中将它们一一加载, 然后再继续进行训练。 diff --git a/doc/fluid/user_guides/howto/training/single_node.rst b/doc/fluid/user_guides/howto/training/single_node.rst index 2c8e21029311cc7053c3240d07e58e7ac5e94463..298094f7fc3f99ceff32503732df4ff632da927f 100644 --- a/doc/fluid/user_guides/howto/training/single_node.rst +++ b/doc/fluid/user_guides/howto/training/single_node.rst @@ -77,25 +77,25 @@ 多卡训练 ####################### -在多卡训练中,你可以使用:code:`fluid.compiler.CompiledProgram`来编译:code:`fluid.Program`,然后调用:code:`with_data_parallel`。例如: +在多卡训练中,你可以使用 :code:`fluid.compiler.CompiledProgram` 来编译 :code:`fluid.Program` ,然后调用 :code:`with_data_parallel` 。例如: .. code-block:: python - + exe = fluid.Executor(...) - + compiled_prog = fluid.compiler.CompiledProgram( fluid.default_main_program()).with_data_parallel( loss_name=loss.name) - - result = exe.run(program=compiled_prog, - fetch_list=[loss.name], - feed={"image": ..., "label": ...}) + + result = exe.run(program=compiled_prog, + fetch_list=[loss.name], + feed={"image": ..., "label": ...}) 注释: -1. :ref:`cn_api_fluid_CompiledProgram`的构造函数需要经过:code:`fluid.Program`设置后运行,这在运行时内无法被修改。 -2. 如果:code:`exe`是用CUDAPlace来初始化的,模型会在GPU中运行。在显卡训练模式中,所有的显卡都将被占用。用户可以配置 `CUDA_VISIBLE_DEVICES `_ 以更改被占用的显卡。 -3. 如果:code:`exe`是用CPUPlace来初始化的,模型会在CPU中运行。在这种情况下,多线程用于运行模型,同时线程的数目和逻辑核的数目相等。用户可以配置`CPU_NUM`以更改使用中的线程数目。 +1. :ref:`cn_api_fluid_CompiledProgram` 的构造函数需要经过 :code:`fluid.Program` 设置后运行,这在运行时内无法被修改。 +2. 如果 :code:`exe` 是用CUDAPlace来初始化的,模型会在GPU中运行。在显卡训练模式中,所有的显卡都将被占用。用户可以配置 `CUDA_VISIBLE_DEVICES `_ 以更改被占用的显卡。 +3. 如果 :code:`exe` 是用CPUPlace来初始化的,模型会在CPU中运行。在这种情况下,多线程用于运行模型,同时线程的数目和逻辑核的数目相等。用户可以配置 ``CPU_NUM`` 以更改使用中的线程数目。 进阶使用 ############### diff --git a/doc/fluid/user_guides/howto/training/train_on_baidu_cloud_cn.rst b/doc/fluid/user_guides/howto/training/train_on_baidu_cloud_cn.rst index 71d610ad5342a049e1f67fb6650c5081bd128563..04767cec8dc4d90d0009b70b757d405292045769 100644 --- a/doc/fluid/user_guides/howto/training/train_on_baidu_cloud_cn.rst +++ b/doc/fluid/user_guides/howto/training/train_on_baidu_cloud_cn.rst @@ -1,6 +1,6 @@ .. _train_on_baidu_cloud_cn: -在百度云启动Fluid分布式训练 +在百度云启动分布式训练 ========================= PaddlePaddle Fluid分布式训练,可以不依赖集群系统(比如MPI,Kubernetes)启动分布式训练。 diff --git a/doc/fluid/user_guides/index.rst b/doc/fluid/user_guides/index.rst deleted file mode 100644 index d65f0c34565d5ab4a544a44bacb13f107b4cddcd..0000000000000000000000000000000000000000 --- a/doc/fluid/user_guides/index.rst +++ /dev/null @@ -1,31 +0,0 @@ -######## -使用指南 -######## - -.. todo:: - -如果您已经掌握了新手入门阶段的内容,期望可以针对实际问题建模、搭建自己网络,本模块提供了一些 Fluid 的使用细节供您参考: - - - `基本概念 <../user_guides/howto/basic_concept/index_cn.html>`_ :介绍了Fluid的基本使用概念 - - - `准备数据 <../user_guides/howto/prepare_data/index.html>`_ :介绍使用 Fluid 训练网络时,数据的支持类型及传输方法 - - - `配置简单的网络 <../user_guides/howto/configure_simple_model/index.html>`_: 介绍如何针对问题建模,并利用 Fluid 中相关算子搭建网络 - - - `训练神经网络 <../user_guides/howto/training/index.html>`_:介绍如何使用 Fluid 进行单机训练、多机训练、以及保存和载入模型变量 - - - `模型评估与调试 <../user_guides/howto/evaluation_and_debugging/index.html>`_:介绍在 Fluid 下进行模型评估和调试的方法,包括: - -基于 Fluid 复现的多领域经典模型: - - - `Fluid 模型库 <../user_guides/models/index_cn.html>`_ - -.. toctree:: - :hidden: - - howto/basic_concept/index_cn.rst - howto/prepare_data/index - howto/configure_simple_model/index - howto/training/index - howto/evaluation_and_debugging/index - models/index_cn.rst diff --git a/doc/fluid/user_guides/index_cn.rst b/doc/fluid/user_guides/index_cn.rst new file mode 100644 index 0000000000000000000000000000000000000000..ed44aea6d0cb387863b37509fe470c3e4f106fbe --- /dev/null +++ b/doc/fluid/user_guides/index_cn.rst @@ -0,0 +1,31 @@ +######## +使用指南 +######## + +.. todo:: + +如果您已经掌握了新手入门阶段的内容,期望可以针对实际问题建模、搭建自己网络,本模块提供了一些 Fluid 的使用细节供您参考: + + - `基本概念 <../user_guides/howto/basic_concept/index_cn.html>`_ :介绍了Fluid的基本使用概念 + + - `准备数据 <../user_guides/howto/prepare_data/index_cn.html>`_ :介绍使用 Fluid 训练网络时,数据的支持类型及传输方法 + + - `配置简单的网络 <../user_guides/howto/configure_simple_model/index_cn.html>`_: 介绍如何针对问题建模,并利用 Fluid 中相关算子搭建网络 + + - `训练神经网络 <../user_guides/howto/training/index_cn.html>`_:介绍如何使用 Fluid 进行单机训练、多机训练、以及保存和载入模型变量 + + - `模型评估与调试 <../user_guides/howto/evaluation_and_debugging/index_cn.html>`_:介绍在 Fluid 下进行模型评估和调试的方法,包括: + +基于 Fluid 复现的多领域经典模型: + + - `Fluid 模型库 <../user_guides/models/index_cn.html>`_ + +.. toctree:: + :hidden: + + howto/basic_concept/index_cn.rst + howto/prepare_data/index_cn.rst + howto/configure_simple_model/index_cn.rst + howto/training/index_cn.rst + howto/evaluation_and_debugging/index_cn.rst + models/index_cn.rst diff --git a/doc/fluid/user_guides/models/index_en.rst b/doc/fluid/user_guides/models/index_en.rst index a372fd6f95146e4901fd8204734ec595e70a9f5f..ce59eec780860b7bec9077e401c0d6e1f441c6ad 100644 --- a/doc/fluid/user_guides/models/index_en.rst +++ b/doc/fluid/user_guides/models/index_en.rst @@ -87,7 +87,7 @@ Automatic Speech Recognition (ASR) is a technique for transcribing vocabulary co Different from the end-to-end direct prediction for word distribution of the deep learning model `DeepSpeech `__ , this example is closer to the traditional language recognition process. With phoneme as the modeling unit, it focuses on the training of acoustic models in speech recognition, use `kaldi `__ for feature extraction and label alignment of audio data, and integrate kaldi's decoder to complete decoding. -- `DeepASR `__ +- `DeepASR `__ Machine Translation ---------------------