diff --git a/README.md b/README.md index bbb2d498589092de78b21a662f03171a0721f840..ceeb6d9e5193763293d3fce76e464340fbce533f 100644 --- a/README.md +++ b/README.md @@ -2,8 +2,8 @@ [![Build Status](https://travis-ci.org/PaddlePaddle/Paddle.svg?branch=develop)](https://travis-ci.org/PaddlePaddle/Paddle) -[![Documentation Status](https://img.shields.io/badge/docs-latest-brightgreen.svg?style=flat)](http://doc.paddlepaddle.org/develop/doc/) -[![Documentation Status](https://img.shields.io/badge/中文文档-最新-brightgreen.svg)](http://doc.paddlepaddle.org/develop/doc_cn/) +[![Documentation Status](https://img.shields.io/badge/docs-latest-brightgreen.svg?style=flat)](http://www.paddlepaddle.org/docs/develop/documentation/en/getstarted/index_en.html) +[![Documentation Status](https://img.shields.io/badge/中文文档-最新-brightgreen.svg)](http://www.paddlepaddle.org/docs/develop/documentation/zh/getstarted/index_cn.html) [![Coverage Status](https://coveralls.io/repos/github/PaddlePaddle/Paddle/badge.svg?branch=develop)](https://coveralls.io/github/PaddlePaddle/Paddle?branch=develop) [![Release](https://img.shields.io/github/release/PaddlePaddle/Paddle.svg)](https://github.com/PaddlePaddle/Paddle/releases) [![License](https://img.shields.io/badge/license-Apache%202-blue.svg)](LICENSE) diff --git a/benchmark/paddle/image/run_mkldnn_infer.sh b/benchmark/paddle/image/run_mkl_infer.sh similarity index 100% rename from benchmark/paddle/image/run_mkldnn_infer.sh rename to benchmark/paddle/image/run_mkl_infer.sh diff --git a/benchmark/paddle/image/run_mkldnn_train.sh b/benchmark/paddle/image/run_mkl_train.sh similarity index 85% rename from benchmark/paddle/image/run_mkldnn_train.sh rename to benchmark/paddle/image/run_mkl_train.sh index 320206239ae960bd088b05d3b10934a98da741b1..5335af5ac1b9a4a48ec107b8b6386b50ead8284c 100755 --- a/benchmark/paddle/image/run_mkldnn_train.sh +++ b/benchmark/paddle/image/run_mkl_train.sh @@ -28,6 +28,10 @@ function train() { --test_period=100 \ --config_args=$args \ 2>&1 | tee ${log} + + avg_time=`tail ${log} -n 1 | awk -F ' ' '{print $8}' | sed 's/avg=//'` + fps=`awk 'BEGIN{printf "%.2f",('$bs' / '$avg_time' * 1000)}'` + echo "FPS: $fps images/sec" 2>&1 | tee -a ${log} } if [ ! -f "train.list" ]; then diff --git a/benchmark/paddle/image/run_openblas_infer.sh b/benchmark/paddle/image/run_openblas_infer.sh new file mode 100755 index 0000000000000000000000000000000000000000..c1001d3a7c95a293d0b2b5b78fb7415e167b3e9f --- /dev/null +++ b/benchmark/paddle/image/run_openblas_infer.sh @@ -0,0 +1,62 @@ +set -e + +function clock_to_seconds() { + hours=`echo $1 | awk -F ':' '{print $1}'` + mins=`echo $1 | awk -F ':' '{print $2}'` + secs=`echo $1 | awk -F ':' '{print $3}'` + echo `awk 'BEGIN{printf "%.2f",('$secs' + '$mins' * 60 + '$hours' * 3600)}'` +} + +function infer() { + unset OMP_NUM_THREADS MKL_NUM_THREADS OMP_DYNAMIC KMP_AFFINITY + topology=$1 + layer_num=$2 + bs=$3 + thread=`nproc` + if [ $thread -gt $bs ]; then + thread=$bs + fi + log="logs/infer-${topology}-${layer_num}-${thread}openblas-${bs}.log" + + models_in="models/${topology}-${layer_num}/pass-00000/" + if [ ! -d $models_in ]; then + echo "./run_mkl_infer.sh to save the model first" + exit 0 + fi + log_period=$((256 / bs)) + paddle train --job=test \ + --config="${topology}.py" \ + --use_gpu=False \ + --trainer_count=$thread \ + --log_period=$log_period \ + --config_args="batch_size=${bs},layer_num=${layer_num},is_infer=True" \ + --init_model_path=$models_in \ + 2>&1 | tee ${log} + + # calculate the last 5 logs period time of 1280 samples, + # the time before are burning time. + start=`tail ${log} -n 7 | head -n 1 | awk -F ' ' '{print $2}' | xargs` + end=`tail ${log} -n 2 | head -n 1 | awk -F ' ' '{print $2}' | xargs` + start_sec=`clock_to_seconds $start` + end_sec=`clock_to_seconds $end` + fps=`awk 'BEGIN{printf "%.2f",(1280 / ('$end_sec' - '$start_sec'))}'` + echo "Last 1280 samples start: ${start}(${start_sec} sec), end: ${end}(${end_sec} sec;" >> ${log} + echo "FPS: $fps images/sec" 2>&1 | tee -a ${log} +} + +if [ ! -f "train.list" ]; then + echo " " > train.list +fi +if [ ! -f "test.list" ]; then + echo " " > test.list +fi +if [ ! -d "logs" ]; then + mkdir logs +fi + +# inference benchmark +for batchsize in 1 2 4 8 16; do + infer googlenet v1 $batchsize + infer resnet 50 $batchsize + infer vgg 19 $batchsize +done diff --git a/benchmark/paddle/image/run_openblas_train.sh b/benchmark/paddle/image/run_openblas_train.sh new file mode 100755 index 0000000000000000000000000000000000000000..b9494ce119523953a3360b2b67e2cb6f3e0f1643 --- /dev/null +++ b/benchmark/paddle/image/run_openblas_train.sh @@ -0,0 +1,39 @@ +set -e + +function train() { + unset OMP_NUM_THREADS MKL_NUM_THREADS OMP_DYNAMIC KMP_AFFINITY + topology=$1 + layer_num=$2 + bs=$3 + thread=`nproc` + # each trainer_count use only 1 core to avoid conflict + log="logs/train-${topology}-${layer_num}-${thread}openblas-${bs}.log" + args="batch_size=${bs},layer_num=${layer_num}" + config="${topology}.py" + paddle train --job=time \ + --config=$config \ + --use_gpu=False \ + --trainer_count=$thread \ + --log_period=10 \ + --test_period=100 \ + --config_args=$args \ + 2>&1 | tee ${log} + + avg_time=`tail ${log} -n 1 | awk -F ' ' '{print $8}' | sed 's/avg=//'` + fps=`awk 'BEGIN{printf "%.2f",('$bs' / '$avg_time' * 1000)}'` + echo "FPS: $fps images/sec" 2>&1 | tee -a ${log} +} + +if [ ! -f "train.list" ]; then + echo " " > train.list +fi +if [ ! -d "logs" ]; then + mkdir logs +fi + +# training benchmark +for batchsize in 64 128 256; do + train vgg 19 $batchsize + train resnet 50 $batchsize + train googlenet v1 $batchsize +done diff --git a/doc/api/index_cn.rst b/doc/api/index_cn.rst index 9be0b370ee5e301aee4a6e31b1cfa905754968e8..84f9097a6cdc2da269bd6a0685796e14e26da37e 100644 --- a/doc/api/index_cn.rst +++ b/doc/api/index_cn.rst @@ -7,3 +7,4 @@ API 模型配置 数据访问 训练与应用 + v2/fluid.rst diff --git a/doc/api/v2/fluid/layers.rst b/doc/api/v2/fluid/layers.rst index 89e5fec13bf9062dc7a7187b1334c8f5486a980b..9f3669e11583a4ed6467f1a1bb509481fdf0b9d1 100644 --- a/doc/api/v2/fluid/layers.rst +++ b/doc/api/v2/fluid/layers.rst @@ -300,3 +300,7 @@ conv2d_transpose .. autofunction:: paddle.v2.fluid.layers.conv2d_transpose :noindex: +sequence_expand +--------- +.. autofunction:: paddle.v2.fluid.layers.sequence_expand + :noindex: diff --git a/doc/design/executor.md b/doc/design/executor.md index b5fb6c5c3c1da3c112ce63878322083dd5c42b70..2d4b371cc56db82ce5747da6db07f05aa7f7e6c1 100644 --- a/doc/design/executor.md +++ b/doc/design/executor.md @@ -1,23 +1,29 @@ # Executor Design Doc ## Motivation +In [fluid](https://github.com/PaddlePaddle/Paddle/blob/develop/doc/design/fluid.md), we encourage the user to use deep learning programming paradigms to describe the training process. When the user-written Python program is executed, it will first create a protobuf message +[`ProgramDesc`](https://github.com/PaddlePaddle/Paddle/blob/a91efdde6910ce92a78e3aa7157412c4c88d9ee8/paddle/framework/framework.proto#L145) that describes the process and is conceptually like an [abstract syntax tree](https://en.wikipedia.org/wiki/Abstract_syntax_tree). -We use executor to do the runtime evaluation of a `ProgramDesc`. +The executor runs the `ProgramDesc` like an interpreter. `ProgramDesc` contains the intrinsics (operators in this case) and variables which will be used, executor explicitly executes the stored precompiled code. ## Overview -An executor takes a `ProgramDesc`, a `block_id` and a `Scope`. The `ProgramDesc` is a list of blocks and each block contains the protobuf definition of all the parameters and operators. The `block_id` specifies the entrance block. And the `Scope` is the container of all the variable instance, which is persistent throughout different runs. +An executor takes a `ProgramDesc`, a `block_id` and a `Scope`. The `ProgramDesc` is a list of blocks and each block contains the protobuf definition of all the parameters and operators in the block. The `block_id` specifies the entrance block. And the `Scope` is the container of all the variable instances, which is persistent throughout different runs. -### What does executor do? +## Executor -It evaluates all the operators in the `block_id`th block of a `ProgramDesc`. +The `Executor` explicitly executes all the intrinsics (operators here) in the `block_id`th block of a `ProgramDesc`. Essentially, it instantiates Variables and Operators, then runs all the operators in sequence one-by-one. +It is very similar to how a push stack frame works when entering a block, following which it cleans up all the temporary variables when a mini-batch is finished. It does not however, have the stack frame pop process. -### What does executor NOT do? +### The interface +```c++ + Executor(places); +``` +A executor does not own any computing resources, a user can only construct an executor using the specified places. -It does not do runtime optimization, meaning intelligently parse the dependency of each op a choose which one to be run and in which order they should be run. +### Running an Executor -It does not do graph partitioning, meaning dividing the `ProgramDesc` into several small pieces and executing them on different devices. - -## Implementation - -`Executor` evaluates a `ProgramDesc`. Essentially, it instantiates Variables and Operators, then run all the operators in sequence. [[code]](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/framework/executor.cc) +``` + void Run(ProgramDesc, Scope, block_id, create_local_scope); +``` +An `Executor` only provides a unified way to execute `ProgramDesc`. `ProgramDesc` is the target that will be executed, the `Scope` specifies the variable container, the `block_id` indicates the entrance block and `create_local_scope` is a boolean that states whether it will destroy the temporary variables after the execution is finished. diff --git a/doc/design/fluid-compiler.graffle b/doc/design/fluid-compiler.graffle new file mode 100644 index 0000000000000000000000000000000000000000..c933df2cb855462c52b2d25f7f9a99b95652961d Binary files /dev/null and b/doc/design/fluid-compiler.graffle differ diff --git a/doc/design/fluid-compiler.png b/doc/design/fluid-compiler.png new file mode 100644 index 0000000000000000000000000000000000000000..1b0ffed2039c91a3a00bbb719da08c91c3acf7bb Binary files /dev/null and b/doc/design/fluid-compiler.png differ diff --git a/doc/design/fluid.md b/doc/design/fluid.md new file mode 100644 index 0000000000000000000000000000000000000000..585dc8ef39c0cfb30f470d79f7b27a59ceb5e940 --- /dev/null +++ b/doc/design/fluid.md @@ -0,0 +1,122 @@ +# Design Doc: PaddlePaddle Fluid + +## Why Fluid + +When Baidu developed PaddlePaddle in 2013, the only well-known open source deep learning system at the time was Caffe. However, when PaddlePaddle was open-sourced in 2016, many other choices were available. There was a challenge -- what is the need for open sourcing yet another deep learning framework? + +Fluid is the answer. Fluid is similar to PyTorch and TensorFlow Eager Execution, which describes the "process" of training or inference using the concept of a model. In fact in PyTorch, TensorFlow Eager Execution and Fluid, there is no concept of a model at all. The details are covered in the sections below. Fluid is currently more extreme in the above mentioned idea than PyTorch and Eager Execution, and we are trying to push Fluid towards the directions of a compiler and a new programming language for deep learning. + +## The Evolution of Deep Learning Systems + +Deep learning infrastructure is one of the fastest evolving technologies. Within four years, there have already been three generations of technologies invented. + +| Existed since | model as sequence of layers | model as graph of operators | No model | +|--|--|--|--| +| 2013 | Caffe, Theano, Torch, PaddlePaddle | | | +| 2015 | | TensorFlow, MxNet, Caffe2, ONNX, n-graph | | +| 2016 | | | PyTorch, TensorFlow Eager Execution, PaddlePaddle Fluid | + +From the above table, we see that the deep learning technology is evolving towards getting rid of the concept of a model. To understand the reasons behind this direction, a comparison of the *programming paradigms* or the ways to program deep learning applications using these systems, would be helpful. The following section goes over these. + +## Deep Learning Programming Paradigms + +With the systems listed as the first or second generation, e.g., Caffe or TensorFlow, an AI application training program looks like the following: + +```python +x = layer.data("image") +l = layer.data("label") +f = layer.fc(x, W) +s = layer.softmax(f) +c = layer.mse(l, s) + +for i in xrange(1000): # train for 1000 iterations + m = read_minibatch() + forward({input=x, data=m}, minimize=c) + backward(...) + +print W # print the trained model parameters. +``` + +The above program includes two parts: + +1. The first part describes the model, and +2. The second part describes the training process (or inference process) for the model. + +This paradigm has a well-known problem that limits the productivity of programmers. If the programmer made a mistake in configuring the model, the error messages wouldn't show up until the second part is executed and `forward` and `backward` propagations are performed. This makes it difficult for the programmer to debug and locate a mistake that is located blocks away from the actual error prompt. + +This problem of being hard to debug and re-iterate fast on a program is the primary reason that programmers, in general, prefer PyTorch over the older systems. Using PyTorch, we would write the above program as following: + +```python +W = tensor(...) + +for i in xrange(1000): # train for 1000 iterations + m = read_minibatch() + x = m["image"] + l = m["label"] + f = layer.fc(x, W) + s = layer.softmax(f) + c = layer.mse(l, s) + backward() + +print W # print the trained model parameters. +``` + +We can see that the main difference is the moving the model configuration part (the first step) into the training loop. This change would allow the mistakes in model configuration to be reported where they actually appear in the programming block. This change also represents the model better, or its forward pass, by keeping the configuration process in the training loop. + +## Describe Arbitrary Models for the Future + +Describing the process instead of the model also brings Fluid, the flexibility to define different non-standard models that haven't been invented yet. + +As we write out the program for the process, we can write an RNN as a loop, instead of an RNN as a layer or as an operator. A PyTorch example would look like the following: + +```python +for i in xrange(1000): + m = read_minibatch() + x = m["sentence"] + for t in xrange x.len(): + h[t] = the_step(x[t]) +``` + +With Fluid, the training loop and the RNN in the above program are not really Python loops, but just a "loop structure" provided by Fluid and implemented in C++ as the following: + +```python +train_loop = layers.While(cond) +with train_loop.block(): + m = read_minibatch() + x = m["sentence"] + rnn = layers.While(...) + with rnn.block(): + h[t] = the_step(input[t]) +``` + +An actual Fluid example is described [here](https://github.com/PaddlePaddle/Paddle/blob/a91efdde6910ce92a78e3aa7157412c4c88d9ee8/python/paddle/v2/fluid/tests/test_while_op.py#L36-L44). + +From the example, the Fluid programs look very similar to their PyTorch equivalent programs, except that Fluid's loop structure, wrapped with Python's `with` statement, could run much faster than just a Python loop. + +We have more examples of the [`if-then-else`](https://github.com/PaddlePaddle/Paddle/blob/develop/doc/design/if_else_op.md) structure of Fluid. + +## Turing Completeness + +In computability theory, a system of data-manipulation rules, such as a programming language, is said to be Turing complete if it can be used to simulate any Turing machine. For a programming language, if it provides if-then-else and loop, it is Turing complete. From the above examples, Fluid seems to be Turing complete; however, it is noteworthy to notice that there is a slight difference between the `if-then-else` of Fluid and that of a programming language. The difference being that the former runs both of its branches and splits the input mini-batch into two -- one for the True condition and another for the False condition. This hasn't been researched in depth if this is equivalent to the `if-then-else` in programming languages that makes them Turing-complete. Based on a conversation with [Yuang Yu](https://research.google.com/pubs/104812.html), it seems to be the case but this needs to be looked into in-depth. + +## The Execution of a Fluid Program + +There are two ways to execute a Fluid program. When a program is executed, it creates a protobuf message [`ProgramDesc`](https://github.com/PaddlePaddle/Paddle/blob/a91efdde6910ce92a78e3aa7157412c4c88d9ee8/paddle/framework/framework.proto#L145) that describes the process and is conceptually like an [abstract syntax tree](https://en.wikipedia.org/wiki/Abstract_syntax_tree). + +There is a C++ class [`Executor`](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/framework/executor.h), which runs a `ProgramDesc`, similar to how an interpreter runs a Python program. + +Fluid is moving towards the direction of a compiler, which is explain in more detail later in this article. + +## Backward Compatibility of Fluid + +Given all the advantages from the removal of the concept of a *model*, hardware manufacturers might still prefer the existence of the concept of a model, so it would be easier for them to support multiple frameworks all at once and could run a trained model during inference. For example, Nervana, a startup company acquired by Intel, has been working on an XPU that reads the models in the format known as [n-graph](https://github.com/NervanaSystems/ngraph). Similarly, [Movidius](https://www.movidius.com/) is producing a mobile deep learning chip that reads and runs graphs of operators. The well-known [ONNX](https://github.com/onnx/onnx) is also a file format of graphs of operators. + +For Fluid, we can write a converter that extracts the parts in the `ProgramDesc` protobuf message, converts them into a graph of operators, and exports the graph into the ONNX or n-graph format. + +## Towards a Deep Learning Language and the Compiler + +We can change the `if-then-else` and loop structure a little bit in the above Fluid example programs, to make it into a new programming language, different than Python. + +Even if we do not invent a new language, as long as we get the `ProgramDesc` message filled in, we can write a transpiler, which translates each invocation to an operator, into a C++ call to a kernel function of that operator. For example, a transpiler that weaves the CUDA kernels outputs an NVIDIA-friendly C++ program, which can be built using `nvcc`. Another transpiler could generate MKL-friendly code that should be built using `icc` from Intel. More interestingly, we can translate a Fluid program into its distributed version of two `ProgramDesc` messages, one for running on the trainer process, and the other one for the parameter server. For more details of the last example, the [concurrent programming design](concurrent_programming.md) document would be a good pointer. The following figure explains the proposed two-stage process: + +![](fluid-compiler.png) diff --git a/doc/design/images/multigpu_allreduce.graffle b/doc/design/images/multigpu_allreduce.graffle new file mode 100644 index 0000000000000000000000000000000000000000..cb5bc420ceafe8ba4c87694d44ee4e5e4ad06779 Binary files /dev/null and b/doc/design/images/multigpu_allreduce.graffle differ diff --git a/doc/design/images/multigpu_allreduce.png b/doc/design/images/multigpu_allreduce.png new file mode 100644 index 0000000000000000000000000000000000000000..87a1b3e8f6dd4a713ec9df9f0037d1da04e9178a Binary files /dev/null and b/doc/design/images/multigpu_allreduce.png differ diff --git a/doc/design/images/multigpu_before_convert.graffle b/doc/design/images/multigpu_before_convert.graffle new file mode 100644 index 0000000000000000000000000000000000000000..6c35ab1b21fb76ceae82d3693ed0d085b5bc0855 Binary files /dev/null and b/doc/design/images/multigpu_before_convert.graffle differ diff --git a/doc/design/images/multigpu_before_convert.png b/doc/design/images/multigpu_before_convert.png new file mode 100644 index 0000000000000000000000000000000000000000..9c8f7711165d80a2fa3911280fdee91855a401b1 Binary files /dev/null and b/doc/design/images/multigpu_before_convert.png differ diff --git a/doc/design/mkldnn/image/engine.png b/doc/design/mkl/image/engine.png similarity index 100% rename from doc/design/mkldnn/image/engine.png rename to doc/design/mkl/image/engine.png diff --git a/doc/design/mkldnn/image/gradients.png b/doc/design/mkl/image/gradients.png similarity index 100% rename from doc/design/mkldnn/image/gradients.png rename to doc/design/mkl/image/gradients.png diff --git a/doc/design/mkldnn/image/layers.png b/doc/design/mkl/image/layers.png similarity index 100% rename from doc/design/mkldnn/image/layers.png rename to doc/design/mkl/image/layers.png diff --git a/doc/design/mkldnn/image/matrix.png b/doc/design/mkl/image/matrix.png similarity index 100% rename from doc/design/mkldnn/image/matrix.png rename to doc/design/mkl/image/matrix.png diff --git a/doc/design/mkldnn/image/overview.png b/doc/design/mkl/image/overview.png similarity index 100% rename from doc/design/mkldnn/image/overview.png rename to doc/design/mkl/image/overview.png diff --git a/doc/design/mkl/mkl_packed.md b/doc/design/mkl/mkl_packed.md new file mode 100644 index 0000000000000000000000000000000000000000..0123315ad4368e68b377f66119949bfd6c1c7860 --- /dev/null +++ b/doc/design/mkl/mkl_packed.md @@ -0,0 +1,108 @@ +# Intel® MKL Packed on PaddlePaddle: Design Doc + + +## Contents + +- [Overview](#overview) +- [Key Points](#key-points) + - [Background](#background) + - [Solution](#solution) +- [Actions](#actions) + - [CMake](#cmake) + - [Layers](#layers) + - [Unit Tests](#unit-tests) + - [Python API](#python-api) + - [Benchmarking](#benchmarking) + + +## Overview +我们计划将 Intel® MKL 中引入的 GEMM Packed APIs\[[1](#references)\] 集成到 PaddlePaddle 中，充分发挥英特尔平台的优势，有效提升PaddlePaddle在英特尔架构上的性能。 +现阶段的优化主要针对 Recurrent Neural Network（以下简称RNN）相关层（包括`RecurrentLayer`, `GatedRecurrentLayer`和`LstmLayer`）， 以及 PaddlePaddle V1 API。 + +## Key Points + +### Background +目前PaddlePaddle采用了 Intel® MKL库的[cblas_?gemm](https://software.intel.com/en-us/mkl-developer-reference-c-cblas-gemm)函数，这个函数本身会在计算前将原数据转换为更适合英特尔平台的内部格式。 + +1. 转换耗时 \ +这一数据格式的转换操作（Packing），在问题本身的计算量比较小的时候，显得相对来说较为耗时。例如在DeepSpeech2 \[[2](#references)\] 的Vanilla RNN部分中，矩阵大小是`batch_size * 2048`。 +2. 转换冗余 \ +由于在现有的某些情况下（例如RNN），多次调用 cblas_?gemm 会使用相同的原数据，因此，每次调用时对原数据的重复Packing便成为了冗余。 + +为了最大程度减少多次调用 cblas_?gemm 在Packing上的耗时，Intel® MKL 引入了以下四个API: + * [cblas_?gemm_alloc](https://software.intel.com/en-us/mkl-developer-reference-c-cblas-gemm-alloc) + * [cblas_?gemm_pack](https://software.intel.com/en-us/mkl-developer-reference-c-cblas-gemm-pack) + * [cblas_?gemm_compute](https://software.intel.com/en-us/mkl-developer-reference-c-cblas-gemm-compute) + * [cblas_?gemm_free](https://software.intel.com/en-us/mkl-developer-reference-c-cblas-gemm-free) + +通过使用这些API，我们可以先完成对原数据的Packing操作，再把已转换为Packed格式的数据传递给那些复用同一数据的gemm_compute函数，从而避免了Packing冗余。 + +### Solution +在RNN的情况下，同一次前向、后向（forward/backward）过程中所有时间步（time step）共享同一个权重（weight）。当只做推断（inference）时，各次前向之间也都使用了相同的权重，没有必要在每次前向中每个时间步的计算时对权重进行重复的Packing操作。 + +我们通过使用新引入的GEMM Packed APIs，在层初始化的时候，先完成对权重的Packing操作，然后在前向，后向时复用已经转换过的权重，并在每次权重更新后，对新的权重进行转换用于下次迭代。 + +* 优化前，对于序列长度（sequence length）为`T`的网络模型（model）, `N`次迭代执行的转换次数为： + - `inference`： `N * T` + - `training`： `2 * N * T` +* 优化后，对于同样设置的网络模型，其转换次数减少至： + - `inference`： `1` + - `training`： `2 * N` + +## Actions + +添加的相关文件和目录结构如下： + +```txt +PaddlePaddle/Paddle +├── ... +└── paddle/ + ├── ... + └── gserver/ + ├── ... + ├── layers/ + │ ├── ... + │ ├── MKLPackedRecurrentLayer.* + | ├── MKLPackedGatedRecurrentLayer.* + | ├── MKLPackedLstmLayer.* + | └── MKLPackedGemm.h + └── tests/ + ├── ... + └── test_MKLPacked.cpp +``` + +### CMake +在对应的`CMakeLists.txt`中根据`WITH_MKL`是否打开，来决定是否开启MKL Packed相关功能。 + +### Layers +所有的`MKLPacked*Layer`都继承于PaddlePaddle的基类`Layer`, 并添加头文件 `MKLPackedGemm.h`，该文件对相关GEMM Packed APIs做了封装。 + +### Unit Tests +我们会添加`test_MKLPacked.cpp`用于MKL Packed优化后layer的测试。 +对于每一个新加的RNN layer，我们会对比如下2个方面： +1. 对比优化后layer自身，sequence mode（`rnn_use_batch=false`）与batch mode(`rnn_use_batch=true`)的结果。 +2. 对比优化后layer与相对应的PaddlePaddle原有layer, 在batch mode下的结果。 + +### Python API +计划在`paddle/utils.Flags`中添加`use_mkl_packed`的flag，用于选择是否使用相关功能，并且当编译时`WITH_MKL=ON`的情况下，默认设置为`true`。 + +同时，在`python/paddle/trainer/config_parser.py`中对应的layer处，添加`use_mkl_packed`这个选择，方便用户在Python端选择是否启用这个功能。 + +具体实现方式比如： + +```python +use_mkl_packed = bool(int(g_command_config_args.get("use_mkl_packed", 0))) +if use_mkl_packed: + self.layer_type = mkl_packed_* +``` + +所有相关的`layer_type`会以*mkl_packed_*开头，这些会在`MKLPacked*Layer`注册layer的时候保证，以示区分。 + + +### Benchmarking +会添加相应的脚本用于测试和对比在使用MKL Packed recurrent layers 前后的网络性能。 + +## References +1. [Introducing the new Packed APIs for GEMM](https://software.intel.com/en-us/articles/introducing-the-new-packed-apis-for-gemm) +2. [DeepSpeech2 on PaddlePaddle](https://github.com/PaddlePaddle/DeepSpeech#deepspeech2-on-paddlepaddle) + diff --git a/doc/design/mkldnn/README.MD b/doc/design/mkl/mkldnn.md similarity index 99% rename from doc/design/mkldnn/README.MD rename to doc/design/mkl/mkldnn.md index 61d453de243c25defc56161641bc4a888a88a3b7..e2fe1e6b26ffa73fda81863abfadf697c0acbfcf 100644 --- a/doc/design/mkldnn/README.MD +++ b/doc/design/mkl/mkldnn.md @@ -208,4 +208,3 @@ if use_mkldnn 但是在PaddlePaddle中，无论是重构前的layer还是重构后的op，都不会想要知道next layer/op的信息。 4. MKL-DNN的高性能格式与PaddlePaddle原有的`NCHW`不同(PaddlePaddle中的cuDNN部分使用的也是`NCHW`，所以不存在这个问题)。 所以需要引入一个转换方法，并且只需要在必要的时候转换这种格式，才能更好的发挥MKL-DNN的性能。 - diff --git a/doc/design/paddle_nccl.md b/doc/design/paddle_nccl.md new file mode 100644 index 0000000000000000000000000000000000000000..c7dac70998a6cfec3a6d2fc72b698ff9722e6805 --- /dev/null +++ b/doc/design/paddle_nccl.md @@ -0,0 +1,65 @@ +# Design Doc: NCCL support in Paddle Fluid + +## Abstract + +This Design Doc refers to the NCCL feature in paddle. We propose an approach to support NCCL library both on a single machine and multiple machines. We wrapper the NCCL primitives `Broadcast`, `Allreduce`, `Reduce` as operators to utilize Multi-GPU powers in one script. + + +## Motivation + +[NCCL](https://developer.nvidia.com/nccl) is a NVIDIA library support Multi-GPU communicating and optimized for NVIDIA GPUs, it provides routines such as all-gather, all-reduce, broadcast, reduce, reduce-scatter, that can achieve high bandwidth over PCIe and NVLink high-speed interconnect. With NCCL library, we can easily accelerate the training in parallel. + +- Pros +1. easily plug-in with [NCCL2](https://developer.nvidia.com/nccl) library. +1. high performance in NVIDIA GPUs. +1. MPI like primitives, which have low learning cost for users. + +- Cons +1. Only design for NVIDIA GPUs, not a general multi-device solution. +1. Although NCCL1 is opensourced under BSD license, but NCCL2 is not opensourced anymore. + +At the beginning of training, the framework needs to distribute the same parameters to every GPU, and merge the gradients at any time user interests. + +As a result, during training, we need the operations of peer to peer copy between different GPUs, aggregating gradients/parameters from GPUs, and broadcasting parameters to GPUs. Every GPU only need to run the operator with correct place information. + +Besides, it needs interfaces to synchronize model update with each different GPU Cards. + +## Implementation + +As mentioned above, we wrap the NCCL routines as several kinds of operators. Need to note that NCCL need to create Communicator between gpu at the beginning, so there is a NCCLInit operator created. + +### Transpiler + +To be compatible with [parameter server design doc](https://github.com/PaddlePaddle/Paddle/blob/develop/doc/design/ops/dist_train.md), the transpiler compiles the user defined operation graph into sub-graphs to be executed on different devices. + +1. The user-defined model will be a single device program + +2. Broadcast/Reduce operators between GPUs will be inserted into the program, even for the multi-node, may insert the `Send`, `Recv` operator. + + *Broadcast, AllReduce in a single machine. And Broadcast, AllReduce, [Send, Recv](https://github.com/PaddlePaddle/Paddle/blob/develop/doc/design/ops/dist_train.md#graph-converter) in multiple machines* + + + +After compiling, the graph as shows + + + +Operators are added to the sub-graphs. Every GPU assigned a role of `rank0`, `rank1` etc. + +- **Broadcast**. Broadcast operator distribute initialized parameter to all the GPUs from the GPU who owns it. e.g. from`rank0` GPU. +- **AllReduce**. AllReduce operator synchronizes parameters/gradients between GPUs. AllReduce implemented in the Ring-Based communicating method, avoid of the bottle neck in a single GPU. + +Need to notice that AllReduce operator force GPUs synchronized at that point. The whole training process in asynchronous or synchronous mode depends on the AllReduce point in the graph. + +As it shown in the picture, when each GPU compute the gradient of `W`, followed with a `AllReduce` operator, accumulate the `dW` to full batch of data, then run the optimize process individually and apply the gradient to its `W`. + +- **AllReduce** + Need to note that our AllReduce operator is a ring-base AllReduce implementation. If we use the NCCL2 AllReduce primitive, every GPU optimized full batch of data, wasted (n-1) GPU compute resources. In addition, NCCL2 built-in AllReduce will only utilize the communicating resource during synchronization, then update the gradient will be a subsequent phase. In fact, we can amortize the update gradient time cost into the communicating phase. The process is +1. Every parameter has its root card. That card will responsible for aggregating the gradients from GPUs. +2. The whole model's parameter will be hashed to different root card, ensure the load balance between GPUs. +3. Logically neighberhood card will start send parameter to the next one. After one round, the parameter main card will aggregate the full gradients. +4. Then the root card will optimize the parameter. +5. This parameter card will send its optimized result to its neighberhood, then the neighberhood will send parameter to its next one. +6. Finish the sychronization round. + +The total time cost will be 2 * (n-1) * per-parameter-send-time, we reach the goal of amortize the upgrade time into communicating phase. diff --git a/doc/design/support_new_device.md b/doc/design/support_new_device.md index 92443e43927ddb184ed51199a3a8c548ae607b3f..fd23dc211a35fdc9d87bc9233fcf4e90254da748 100644 --- a/doc/design/support_new_device.md +++ b/doc/design/support_new_device.md @@ -1,33 +1,33 @@ -# Design Doc: Support new Device/Library +# Design Doc: Supporting new Device/Library ## Background -Deep learning has a high demand for computing resources. New high-performance device and computing library are coming constantly. The deep learning framework has to integrate these high-performance device and computing library flexibly. +Deep learning has a high demand for computing resources. New high-performance devices and computing libraries are appearing very frequently. Deep learning frameworks have to integrate these high-performance devices and computing libraries flexibly and efficiently. -On the one hand, hardware and computing library are not usually one-to-one coresponding relations. For example, in Intel CPU, there are Eigen and MKL computing library. And in Nvidia GPU, there are Eigen and cuDNN computing library. We have to implement specific kernels for an operator for each computing library. +On one hand, hardware and computing libraries usually do not have a one-to-one correspondence. For example,Intel CPUs support Eigen and MKL computing libraries while Nvidia GPUs support Eigen and cuDNN computing libraries. We have to implement operator specific kernels for each computing library. -On the other hand, users usually do not want to care about the low-level hardware and computing library when writing a neural network configuration. In Fluid, `Layer` is exposed in `Python`, and `Operator` is exposed in `C++`. Both `Layer` and `Operator` are independent on hardwares. +On the other hand, users usually do not want to care about the low-level hardware and computing libraries when writing a neural network configuration. In Fluid, `Layer` is exposed in `Python`, and `Operator` is exposed in `C++`. Both `Layer` and `Operator` are hardware independent. So, how to support a new Device/Library in Fluid becomes a challenge. ## Basic: Integrate A New Device/Library -For a general overview of fluid, please refer to [overview doc](https://github.com/PaddlePaddle/Paddle/blob/develop/doc/howto/read_source.md). +For a general overview of fluid, please refer to the [overview doc](https://github.com/PaddlePaddle/Paddle/blob/develop/doc/howto/read_source.md). -There are mainly there parts we have to consider in integrating a new device/library: +There are mainly three parts that we have to consider while integrating a new device/library: - Place and DeviceContext: indicates the device id and manages hardware resources - Memory and Tensor: malloc/free data on certain device -- Math Functor and OpKernel: implement computing unit on certain device/library +- Math Functor and OpKernel: implement computing unit on certain devices/libraries ### Place and DeviceContext #### Place -Fluid use class [Place](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/platform/place.h#L55) to represent specific device and computing library. There are inheritance relationships between different kinds of `Place`. +Fluid uses class [Place](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/platform/place.h#L55) to represent different devices and computing libraries. There are inheritance relationships between different kinds of `Place`. ``` | CPUPlace --> MKLDNNPlace @@ -43,7 +43,7 @@ typedef boost::variant Place; #### DeviceContext -Fluid use class [DeviceContext](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/platform/device_context.h#L30) to manage the resources in certain hardware, such as CUDA stream in `CDUADeviceContext`. There are also inheritance relationships between different kinds of `DeviceContext`. +Fluid uses class [DeviceContext](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/platform/device_context.h#L30) to manage the resources in different hardwares, such as CUDA stream in `CDUADeviceContext`. There are also inheritance relationships between different kinds of `DeviceContext`. ``` @@ -52,7 +52,7 @@ DeviceContext ----> CUDADeviceContext --> CUDNNDeviceContext \-> FPGADeviceContext ``` -A example of Nvidia GPU is as follows: +An example of Nvidia GPU is as follows: - DeviceContext @@ -93,7 +93,7 @@ class CUDNNDeviceContext : public CUDADeviceContext { #### memory module -Fluid provide following [memory interfaces](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/memory/memory.h#L36): +Fluid provides the following [memory interfaces](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/memory/memory.h#L36): ``` template @@ -106,12 +106,12 @@ template size_t Used(Place place); ``` -To implementing these interfaces, we have to implement MemoryAllocator for specific Device +To implementing these interfaces, we have to implement MemoryAllocator for different Devices #### Tensor -[Tensor](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/framework/tensor.h#L36) holds data with some shape in certain Place. +[Tensor](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/framework/tensor.h#L36) holds data with some shape in a specific Place. ```cpp class Tensor { @@ -168,7 +168,7 @@ t.mutable_data(place); ### Math Functor and OpKernel -Fluid implements computing unit based on different DeviceContext. Some computing unit is shared between operators. These common part will be put in operators/math directory as basic Functors. +Fluid implements computing units based on different DeviceContexts. Some computing units are shared between operators. This common part will be put in operators/math directory as basic Functors. Let's take [MaxOutFunctor](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/operators/math/maxouting.h#L27) as an example: @@ -183,7 +183,7 @@ class MaxOutFunctor { }; ``` -CPU implement in .cc file +CPU implemention is in .cc file ``` template @@ -197,7 +197,7 @@ class MaxOutFunctor { }; ``` -CUDA implement in .cu file +CUDA implemention is in .cu file ``` template @@ -212,11 +212,11 @@ class MaxOutFunctor { ``` -We get computing handle from concrete DeviceContext, and make compution on tensors. +We get computing handle from a concrete DeviceContext, and make compution on tensors. -The implement of `OpKernel` is similar to math functors, the extra thing we need to do is registering the OpKernel to global map. +The implemention of `OpKernel` is similar to math functors, the extra thing we need to do is to register the OpKernel in a global map. -Fluid provides different register interface in op_registry.h +Fluid provides different register interfaces in op_registry.h Let's take [Crop](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/operators/crop_op.cc#L134) operator as an example: @@ -240,7 +240,7 @@ REGISTER_OP_CUDA_KERNEL( ## Advanced topics: How to switch between different Device/Library -Generally, we will impelement OpKernel for all Device/Library of an Operator. We can easily train a Convolutional Neural Network in GPU. However, some OpKernel is not sutibale in a specific Device. For example, crf operator can be only run at CPU, whereas most other operators can be run at GPU. To achieve high performance in such circumstance, we have to switch between different Device/Library. +Generally, we will impelement OpKernel for all Device/Library of an Operator. We can easily train a Convolutional Neural Network in GPU. However, some OpKernel is not sutibale on a specific Device. For example, crf operator can only run on CPU, whereas most other operators can run at GPU. To achieve high performance in such circumstance, we have to switch between different Device/Library. We will discuss how to implement an efficient OpKernel switch policy. diff --git a/doc/faq/build_and_install/index_cn.rst b/doc/faq/build_and_install/index_cn.rst index f1677e216f31d79b53ac29a0afbf6fbb886a0dcd..a2bdeead7841393fdfe90c78e5b91d9e61678a24 100644 --- a/doc/faq/build_and_install/index_cn.rst +++ b/doc/faq/build_and_install/index_cn.rst @@ -14,7 +14,7 @@ $ export CUDA_SO="$(\ls usr/lib64/libcuda* | xargs -I{} echo '-v {}:{}') $(\ls /usr/lib64/libnvidia* | xargs -I{} echo '-v {}:{}')" $ export DEVICES=$(\ls /dev/nvidia* | xargs -I{} echo '--device {}:{}') - $ docker run ${CUDA_SO} ${DEVICES} -it paddledev/paddlepaddle:latest-gpu + $ docker run ${CUDA_SO} ${DEVICES} -it paddlepaddle/paddle:latest-gpu 更多关于Docker的安装与使用, 请参考 `PaddlePaddle Docker 文档 `_ 。 diff --git a/doc/getstarted/build_and_install/docker_install_cn.rst b/doc/getstarted/build_and_install/docker_install_cn.rst index f78b1fb0e11aa028a4b7abb5270740b97f8039e9..1eb06e4182d40c3be20d71e37b34009905eaf9d6 100644 --- a/doc/getstarted/build_and_install/docker_install_cn.rst +++ b/doc/getstarted/build_and_install/docker_install_cn.rst @@ -114,7 +114,7 @@ PaddlePaddle Book是为用户和开发者制作的一个交互式的Jupyter Note .. code-block:: bash - nvidia-docker run -it -v $PWD:/work paddledev/paddle:latest-gpu /bin/bash + nvidia-docker run -it -v $PWD:/work paddlepaddle/paddle:latest-gpu /bin/bash **注: 如果没有安装nvidia-docker，可以尝试以下的方法，将CUDA库和Linux设备挂载到Docker容器内：** @@ -122,7 +122,7 @@ PaddlePaddle Book是为用户和开发者制作的一个交互式的Jupyter Note export CUDA_SO="$(\ls /usr/lib64/libcuda* | xargs -I{} echo '-v {}:{}') $(\ls /usr/lib64/libnvidia* | xargs -I{} echo '-v {}:{}')" export DEVICES=$(\ls /dev/nvidia* | xargs -I{} echo '--device {}:{}') - docker run ${CUDA_SO} ${DEVICES} -it paddledev/paddle:latest-gpu + docker run ${CUDA_SO} ${DEVICES} -it paddlepaddle/paddle:latest-gpu **关于AVX：** diff --git a/doc/getstarted/build_and_install/docker_install_en.rst b/doc/getstarted/build_and_install/docker_install_en.rst index d7acc7aeb744b19d83acb520d07c8551168dd096..5a46c598f2248c7912169a9e77b16851230c1d2e 100644 --- a/doc/getstarted/build_and_install/docker_install_en.rst +++ b/doc/getstarted/build_and_install/docker_install_en.rst @@ -122,7 +122,7 @@ GPU driver installed before move on. .. code-block:: bash - nvidia-docker run -it -v $PWD:/work paddledev/paddle:latest-gpu /bin/bash + nvidia-docker run -it -v $PWD:/work paddlepaddle/paddle:latest-gpu /bin/bash **NOTE: If you don't have nvidia-docker installed, try the following method to mount CUDA libs and devices into the container.** @@ -130,7 +130,7 @@ GPU driver installed before move on. export CUDA_SO="$(\ls /usr/lib64/libcuda* | xargs -I{} echo '-v {}:{}') $(\ls /usr/lib64/libnvidia* | xargs -I{} echo '-v {}:{}')" export DEVICES=$(\ls /dev/nvidia* | xargs -I{} echo '--device {}:{}') - docker run ${CUDA_SO} ${DEVICES} -it paddledev/paddle:latest-gpu + docker run ${CUDA_SO} ${DEVICES} -it paddlepaddle/paddle:latest-gpu **About AVX:** diff --git a/doc/getstarted/concepts/src/infer.py b/doc/getstarted/concepts/src/infer.py new file mode 100644 index 0000000000000000000000000000000000000000..4cc58dfee0bd6dade0340b4fd0ee1adb49ffebf6 --- /dev/null +++ b/doc/getstarted/concepts/src/infer.py @@ -0,0 +1,18 @@ +import paddle.v2 as paddle +import numpy as np + +paddle.init(use_gpu=False) +x = paddle.layer.data(name='x', type=paddle.data_type.dense_vector(2)) +y_predict = paddle.layer.fc(input=x, size=1, act=paddle.activation.Linear()) + +# loading the model which generated by training +with open('params_pass_90.tar', 'r') as f: + parameters = paddle.parameters.Parameters.from_tar(f) + +# Input multiple sets of data，Output the infer result in a array. +i = [[[1, 2]], [[3, 4]], [[5, 6]]] +print paddle.infer(output_layer=y_predict, parameters=parameters, input=i) +# Will print: +# [[ -3.24491572] +# [ -6.94668722] +# [-10.64845848]] diff --git a/doc/getstarted/concepts/src/train.py b/doc/getstarted/concepts/src/train.py index 8aceb23406a476f08639cc6223cdf730b728a705..4bccbfca3c70c12aec564e2cae3b8ca174b68777 100644 --- a/doc/getstarted/concepts/src/train.py +++ b/doc/getstarted/concepts/src/train.py @@ -26,6 +26,11 @@ def event_handler(event): if event.batch_id % 1 == 0: print "Pass %d, Batch %d, Cost %f" % (event.pass_id, event.batch_id, event.cost) + # product model every 10 pass + if isinstance(event, paddle.event.EndPass): + if event.pass_id % 10 == 0: + with open('params_pass_%d.tar' % event.pass_id, 'w') as f: + trainer.save_parameter_to_tar(f) # define training dataset reader diff --git a/doc/getstarted/concepts/use_concepts_cn.rst b/doc/getstarted/concepts/use_concepts_cn.rst index c243083794bb3c4659242de99b3b2715af9d7c24..e695ff283e2e806377a51c559b37e8068360a4ff 100644 --- a/doc/getstarted/concepts/use_concepts_cn.rst +++ b/doc/getstarted/concepts/use_concepts_cn.rst @@ -147,4 +147,9 @@ PaddlePaddle支持不同类型的输入数据，主要包括四种类型，和 .. literalinclude:: src/train.py :linenos: +使用以上训练好的模型进行预测，取其中一个模型params_pass_90.tar，输入需要预测的向量组，然后打印输出： + +.. literalinclude:: src/infer.py + :linenos: + 有关线性回归的实际应用，可以参考PaddlePaddle book的 `第一章节 `_。 diff --git a/doc/howto/index_cn.rst b/doc/howto/index_cn.rst index 991b9e2596a3b499846b963152c838d66260265d..ccd909770253bb85dbc8a5a2560594076c2f68b0 100644 --- a/doc/howto/index_cn.rst +++ b/doc/howto/index_cn.rst @@ -9,9 +9,6 @@ usage/cmd_parameter/index_cn.rst usage/cluster/cluster_train_cn.md - usage/k8s/k8s_basis_cn.md - usage/k8s/k8s_cn.md - usage/k8s/k8s_distributed_cn.md 开发标准 -------- diff --git a/doc/howto/index_en.rst b/doc/howto/index_en.rst index 61bf25ccd12eeedffc747fdd4ce84fa4adde07ee..6d1bf7dfc003da6de31410ee0a7959233adfaf76 100644 --- a/doc/howto/index_en.rst +++ b/doc/howto/index_en.rst @@ -9,8 +9,6 @@ Usage usage/cmd_parameter/index_en.rst usage/cluster/cluster_train_en.md - usage/k8s/k8s_en.md - usage/k8s/k8s_aws_en.md Development ------------ diff --git a/doc/howto/read_source.md b/doc/howto/read_source.md index 383acb0c8251043c3c6bbf309d2e07bf0074cd4f..e4211abb3be9cace80bc14dbe3db3e0a31221dd0 100644 --- a/doc/howto/read_source.md +++ b/doc/howto/read_source.md @@ -6,10 +6,10 @@ Core: https://github.com/PaddlePaddle/Paddle/tree/develop/paddle/framework Operator: https://github.com/PaddlePaddle/Paddle/tree/develop/paddle/operators -Optimizer: https://github.com/PaddlePaddle/Paddle/tree/develop/paddle/optimizer - Memory: https://github.com/PaddlePaddle/Paddle/tree/develop/paddle/memory +Platform: https://github.com/PaddlePaddle/Paddle/tree/develop/paddle/platform + # Compile Time The following **defines** the NN. The definition goes into this [protocol buffer](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/framework/framework.proto). diff --git a/doc/howto/usage/cluster/cluster_train_cn.md b/doc/howto/usage/cluster/cluster_train_cn.md index 2e98b3de3fe2284375f87e883ff4bac19255dbeb..c9f90538a669d4705d18c3cd9b6dbf4a535c35b8 100644 --- a/doc/howto/usage/cluster/cluster_train_cn.md +++ b/doc/howto/usage/cluster/cluster_train_cn.md @@ -1,25 +1,8 @@ # PaddlePaddle分布式训练 -* [概述](#概述) -* [环境准备](#环境准备) -* [启动参数说明](#启动参数说明) - * [启动参数服务器](#启动参数服务器) - * [启动计算节点](#启动计算节点) - * [准备数据集](#准备数据集) - * [准备训练程序](#准备训练程序) -* [使用分布式计算平台或工具](#使用分布式计算平台或工具) - * [使用Fabric启动集群作业](#使用fabric启动集群作业) - * [准备一个Linux集群](#准备一个linux集群) - * [启动集群作业](#启动集群作业) - * [终止集群作业](#终止集群作业) - * [检查集群训练结果](#检查集群训练结果) - * [检查模型输出](#检查模型输出) - * [在OpenMPI集群中提交训练作业](#在openmpi集群中提交训练作业) - * [准备OpenMPI集群](#准备OpenMPI集群) - * [启动集群作业](#启动集群作业-1) - * [在Kubernetes集群中提交训练作业](#在kubernetes集群中提交训练作业) ## 概述 + 本文将介绍如何使用PaddlePaddle在不同的集群框架下完成分布式训练。分布式训练架构如下图所示： @@ -32,10 +15,11 @@ 在使用同步SGD训练神经网络时，PaddlePaddle使用同步屏障（barrier），使梯度的提交和参数的更新按照顺序方式执行。在异步SGD中，则并不会等待所有trainer提交梯度才更新参数，这样极大地提高了计算的并行性：参数服务器之间不相互依赖，并行地接收梯度和更新参数，参数服务器也不会等待计算节点全部都提交梯度之后才开始下一步，计算节点之间也不会相互依赖，并行地执行模型的训练。可以看出，虽然异步SGD方式会提高参数更新并行度, 但是并不能保证参数同步更新，在任意时间某一台参数服务器上保存的参数可能比另一台要更新，与同步SGD相比，梯度会有噪声。 + ## 环境准备 1. 准备您的计算集群。计算集群通常由一组（几台到几千台规模）的Linux服务器组成。服务器之间可以通过局域网（LAN）联通，每台服务器具有集群中唯一的IP地址（或者可被DNS解析的主机名）。集群中的每台计算机通常被成为一个“节点”。 -1. 我们需要在集群的所有节点上安装 PaddlePaddle。 如果要启用GPU，还需要在节点上安装对应的GPU驱动以及CUDA。PaddlePaddle的安装可以参考[build_and_install](https://github.com/PaddlePaddle/Paddle/tree/develop/doc/getstarted/build_and_install)的多种安装方式。我们推荐使用[Docker](https://github.com/PaddlePaddle/Paddle/blob/develop/doc/getstarted/build_and_install/docker_install_cn.rst)安装方式来快速安装PaddlePaddle。 +1. 我们需要在集群的所有节点上安装 PaddlePaddle。 如果要启用GPU，还需要在节点上安装对应的GPU驱动以及CUDA。PaddlePaddle的安装可以参考[build_and_install](http://www.paddlepaddle.org/docs/develop/documentation/zh/getstarted/build_and_install/index_cn.html)的多种安装方式。我们推荐使用[Docker](http://www.paddlepaddle.org/docs/develop/documentation/zh/getstarted/build_and_install/docker_install_cn.html)安装方式来快速安装PaddlePaddle。 安装完成之后，执行下面的命令可以查看已经安装的版本（docker安装方式可以进入docker容器执行：`docker run -it paddlepaddle/paddle:[tag] /bin/bash`）： ```bash @@ -63,12 +47,12 @@ $ paddle pserver --port=7164 --ports_num=1 --ports_num_for_sparse=1 --num_gradie $ stdbuf -oL /usr/bin/nohup paddle pserver --port=7164 --ports_num=1 --ports_num_for_sparse=1 --num_gradient_servers=1 &> pserver.log ``` -| 参数 | 是否必选 | 默认值 | 说明 | -| ------------- | ------------- | ------------- | ------------- | -| port | 必选 | 7164 | pserver监听的起始端口，根据ports_num决定
总端口个数，从起始端口监听多个端口用于通信 | -| ports_num | 必选 | 1 | 监听的端口个数 | -| ports_num_for_sparse | 必选 | 1 | 用于稀疏类型参数通信的端口个数 | -| num_gradient_servers | 必选 | 1 | 当前训练任务pserver总数 | +参数说明 + +- port：**必选，默认7164**，pserver监听的起始端口，根据ports_num决定总端口个数，从起始端口监听多个端口用于通信 +- ports_num：**必选，默认1**，监听的端口个数 +- ports_num_for_sparse：**必选，默认1**，用于稀疏类型参数通信的端口个数 +- num_gradient_servers：**必选，默认1**，当前训练任务pserver总数 ### 启动计算节点 执行以下命令启动使用python编写的trainer程序（文件名为任意文件名，如train.py） @@ -105,16 +89,16 @@ paddle.init( pservers="127.0.0.1") ``` -| 参数 | 是否必选 | 默认 | 说明 | -| ------------- | ------------- | ------------- | ------------- | -| use_gpu | 可选 | False | 是否启用GPU训练 | -| trainer_count | 必选 | 1 | 当前训练任务trainer总个数 | -| port | 必选 | 7164 | 连接到pserver的端口 | -| ports_num | 必选 | 1 | 连接到pserver的端口个数 | -| ports_num_for_sparse | 必选 | 1 | 和pserver之间用于稀疏类型参数通信的端口个数 | -| num_gradient_servers | 必选 | 1 | 当前训练任务pserver总数 | -| trainer_id | 必选 | 0 | 每个trainer的唯一ID，从0开始的整数 | -| pservers | 必选 | 127.0.0.1 | 当前训练任务启动的pserver的IP列表，多个IP使用“,”隔开 | +参数说明 + +- use_gpu： **可选，默认False**，是否启用GPU训练 +- trainer_count：**必选，默认1**，当前训练任务trainer总个数 +- port：**必选，默认7164**，连接到pserver的端口 +- ports_num：**必选，默认1**，连接到pserver的端口个数 +- ports_num_for_sparse：**必选，默认1**，和pserver之间用于稀疏类型参数通信的端口个数 +- num_gradient_servers：**必选，默认1**，当前训练任务pserver总数 +- trainer_id：**必选，默认0**，每个trainer的唯一ID，从0开始的整数 +- pservers：**必选，默认127.0.0.1**，当前训练任务启动的pserver的IP列表，多个IP使用“,”隔开 ### 准备数据集 @@ -171,7 +155,7 @@ test.txt-00002 - `my_lib.py`：会被`train.py`调用的一些用户定义的库函数，比如PIL库等。 - `word_dict.pickle`：在`train.py`中会使用到的字典数据文件。 -- `train.py`：训练程序，代码参考[api_train_v2_cluster.py](https://github.com/PaddlePaddle/Paddle/tree/develop/doc/howto/usage/cluster/src/word2vec/prepare.py)。***注意：*** 对于本样例代码，在使用不同的分布式计算平台时，您可能需要修改`train.py`开头的部分（如下），以便获得训练数据的位置和获取环境变量配置： +- `train.py`：训练程序，代码参考[api_train_v2_cluster.py](https://github.com/PaddlePaddle/Paddle/tree/develop/doc/howto/usage/cluster/src/word2vec/api_train_v2_cluster.py)。***注意：*** 对于本样例代码，在使用不同的分布式计算平台时，您可能需要修改`train.py`开头的部分（如下），以便获得训练数据的位置和获取环境变量配置： ```python cluster_train_file = "./train_data_dir/train/train.txt" @@ -195,91 +179,10 @@ PaddlePaddle可以使用多种分布式计算平台构建分布式计算任务 在使用分布式计算平台进行训练时，任务被调度在集群中时，分布式计算平台通常会通过API或者环境变量提供任务运行需要的参数，比如节点的ID、IP和任务节点个数等。 -### 使用Fabric启动集群作业 - -#### 准备一个Linux集群 -可以在`paddle/scripts/cluster_train_v2/fabric/docker_cluster`目录下，执行`kubectl -f ssh_servers.yaml`启动一个测试集群，并使用`kubectl get po -o wide`获得这些节点的IP地址。 - -#### 启动集群作业 - -`paddle.py` 提供了自动化脚本来启动不同节点中的所有 PaddlePaddle 集群进程。默认情况下，所有命令行选项可以设置为 `paddle.py` 命令选项并且 `paddle.py` 将透明、自动地将这些选项应用到 PaddlePaddle 底层进程。 - -`paddle.py` 为方便作业启动提供了两个独特的命令选项。 - -- `job_dispatch_package` 设为本地 `workspace` 目录，它将被分发到 `conf.py` 中设置的所有节点。它有助于帮助频繁修改和访问工作区文件的用户减少负担，否则频繁的多节点工作空间部署可能会很麻烦。 -- `job_workspace` 设为已部署的工作空间目录，`paddle.py` 将跳过分发阶段直接启动所有节点的集群作业。它可以帮助减少分发延迟。 - -`cluster_train/run.sh` 提供了命令样例来运行 `doc/howto/usage/cluster/src/word2vec` 集群任务，只需用您定义的目录修改 `job_dispatch_package` 和 `job_workspace`，然后： -``` -sh run.sh -``` - -集群作业将会在几秒后启动。 - -#### 终止集群作业 -`paddle.py`能获取`Ctrl + C` SIGINT 信号来自动终止它启动的所有进程。只需中断 `paddle.py` 任务来终止集群作业。如果程序崩溃你也可以手动终止。 - -#### 检查集群训练结果 -详细信息请检查 $workspace/log 里的日志，每一个节点都有相同的日志结构。 - -`paddle_trainer.INFO` -提供几乎所有训练的内部输出日志，与本地训练相同。这里检验运行时间模型的收敛。 - -`paddle_pserver2.INFO` -提供 pserver 运行日志，有助于诊断分布式错误。 - -`server.log` -提供 parameter server 进程的 stderr 和 stdout。训练失败时可以检查错误日志。 - -`train.log` -提供训练过程的 stderr 和 stdout。训练失败时可以检查错误日志。 - -#### 检查模型输出 -运行完成后，模型文件将被写入节点 0 的 `output` 目录中。 -工作空间中的 `nodefile` 表示当前集群作业的节点 ID。 - -### 在OpenMPI集群中提交训练作业 - -#### 准备OpenMPI集群 - -执行下面的命令以启动3个节点的OpenMPI集群和一个"head"节点： - -```bash -paddle/scripts/cluster_train_v2/openmpi/docker_cluster -kubectl create -f head.yaml -kubectl create -f mpi-nodes.yaml -``` - -然后可以从head节点ssh无密码登录到OpenMPI的每个节点上。 - -#### 启动集群作业 - -您可以按照下面的步骤在OpenMPI集群中提交paddle训练任务： - -```bash -# 获得head和node节点的IP地址 -kubectl get po -o wide -# 将node节点的IP地址保存到machines文件中 -kubectl get po -o wide | grep nodes | awk '{print $6}' > machines -# 拷贝必要的文件到head节点 -scp -i ssh/id_rsa.mpi.pub machines prepare.py train.py start_mpi_train.sh tutorial@[headIP]:~ -# ssh 登录到head节点 -ssh -i ssh/id_rsa.mpi.pub tutorial@[headIP] -# --------------- 以下操作均在head节点中执行 --------------- -# 准备训练数据 -python prepare.py -# 拷贝训练程序和字典文件到每台MPI节点 -cat machines | xargs -i scp word_dict.pickle train.py start_mpi_train.sh machines {}:/home/tutorial -# 创建日志目录 -mpirun -hostfile machines -n 3 mkdir /home/tutorial/logs -# 拷贝训练数据到各自的节点 -scp train.txt-00000 test.txt-00000 [node1IP]:/home/tutorial -scp train.txt-00001 test.txt-00001 [node2IP]:/home/tutorial -scp train.txt-00002 test.txt-00002 [node3IP]:/home/tutorial -# 启动训练任务 -mpirun -hostfile machines -n 3 /home/tutorial/start_mpi_train.sh -``` - -### 在Kubernetes集群中提交训练作业 +## 在不同集群中运行 -此部分的使用方法可以参考[here](../k8s/k8s_distributed_cn.md)。 + - [fabric](fabric_cn.md) + - [openmpi](openmpi_cn.md) + - [kubernetes](k8s_cn.md) + - [kubernetes distributed](k8s_distributed_cn.md) + - [kubernetes on AWS](k8s_aws_cn.md) diff --git a/doc/howto/usage/cluster/cluster_train_en.md b/doc/howto/usage/cluster/cluster_train_en.md index baa97c0c02ae490fff8587071bd2d4adfb5325e3..f9819470c0c622b4bc0ea064303d742385603230 100644 --- a/doc/howto/usage/cluster/cluster_train_en.md +++ b/doc/howto/usage/cluster/cluster_train_en.md @@ -1,24 +1,5 @@ # PaddlePaddle Distributed Training -* [Introduction](#introduction) -* [Preparations](#preparations) -* [Command-line arguments](#command-line-arguments) - * [Starting parameter server](#starting-parameter-server) - * [Starting trainer](#starting-trainer) - * [Prepare Training Dataset](#prepare-training-dataset) - * [Prepare Training program](#prepare-training-program) -* [Use cluster platforms or cluster management tools](#use-cluster-platforms-or-cluster-management-tools) - * [Cluster Training Using Fabric](#cluster-training-using-fabric) - * [Prepare a Linux cluster](#prepare-a-linux-cluster) - * [Launching Cluster Job](#launching-cluster-job) - * [Kill Cluster Job](#kill-cluster-job) - * [Check Cluster Training Result](#check-cluster-training-result) - * [Check Model Output](#check-model-output) - * [Cluster Training Using OpenMPI](#cluster-training-using-openmpi) - * [Prepare an OpenMPI cluster](#prepare-an-openmpi-cluster) - * [Launching Cluster Job](#launching-cluster-job-1) - * [Cluster Training Using Kubernetes](#cluster-training-using-kubernetes) - ## Introduction In this article, we'll explain how to run distributed training jobs with PaddlePaddle on different types of clusters. The diagram below shows the main architecture of a distributed trainning job: @@ -35,7 +16,7 @@ When training with synchronize SGD, PaddlePaddle uses an internal "synchronize b ## Preparations 1. Prepare your computer cluster. It's normally a bunch of Linux servers connected by LAN. Each server will be assigned a unique IP address. The computers in the cluster can be called "nodes". -2. Install PaddlePaddle on every node. If you are going to take advantage of GPU cards, you'll also need to install proper driver and CUDA libraries. To install PaddlePaddle please read [this build and install](https://github.com/PaddlePaddle/Paddle/tree/develop/doc/getstarted/build_and_install) document. We strongly recommend using [Docker installation](https://github.com/PaddlePaddle/Paddle/blob/develop/doc/getstarted/build_and_install/docker_install_en.rst). +2. Install PaddlePaddle on every node. If you are going to take advantage of GPU cards, you'll also need to install proper driver and CUDA libraries. To install PaddlePaddle please read [this build and install](http://www.paddlepaddle.org/docs/develop/documentation/en/getstarted/build_and_install/index_en.html) document. We strongly recommend using [Docker installation](http://www.paddlepaddle.org/docs/develop/documentation/en/getstarted/build_and_install/docker_install_en.html). After installation, you can check the version by typing the below command (run a docker container if using docker: `docker run -it paddlepaddle/paddle:[tag] /bin/bash`): @@ -67,12 +48,12 @@ If you wish to run parameter servers in background, and save a log file, you can $ stdbuf -oL /usr/bin/nohup paddle pserver --port=7164 --ports_num=1 --ports_num_for_sparse=1 --num_gradient_servers=1 &> pserver.log ``` -| param | required | default | description | -| ------------- | ------------- | ------------- | ------------- | -| port | required | 7164 | port which parameter server will listen on. If ports_num greater than 1, parameter server will listen on multiple ports for more network throughput | -| ports_num | required | 1 | total number of ports will listen on | -| ports_num_for_sparse | required | 1 | number of ports which serves sparse parameter update | -| num_gradient_servers | required | 1 | total number of gradient servers | +Parameter Description + +- port: **required, default 7164**, port which parameter server will listen on. If ports_num greater than 1, parameter server will listen on multiple ports for more network throughput. +- ports_num: **required, default 1**, total number of ports will listen on. +- ports_num_for_sparse: **required, default 1**, number of ports which serves sparse parameter update. +- num_gradient_servers: **required, default 1**, total number of gradient servers. ### Starting trainer Type the command below to start the trainer(name the file whatever you want, like "train.py") @@ -111,16 +92,16 @@ paddle.init( pservers="127.0.0.1") ``` -| param | required | default | description | -| ------------- | ------------- | ------------- | ------------- | -| use_gpu | optional | False | set to "True" to enable GPU training | -| trainer_count | required | 1 | total count of trainers in the training job | -| port | required | 7164 | port to connect to parameter server | -| ports_num | required | 1 | number of ports for communication | -| ports_num_for_sparse | required | 1 | number of ports for sparse type caculation | -| num_gradient_servers | required | 1 | total number of gradient server | -| trainer_id | required | 0 | ID for every trainer, start from 0 | -| pservers | required | 127.0.0.1 | list of IPs of parameter servers, separated by "," | +Parameter Description + +- use_gpu: **optional, default False**, set to "True" to enable GPU training. +- trainer_count: **required, default 1**, total count of trainers in the training job. +- port: **required, default 7164**, port to connect to parameter server. +- ports_num: **required, default 1**, number of ports for communication. +- ports_num_for_sparse: **required, default 1**, number of ports for sparse type caculation. +- num_gradient_servers: **required, default 1**, total number of gradient server. +- trainer_id: **required, default 0**, ID for every trainer, start from 0. +- pservers: **required, default 127.0.0.1**, list of IPs of parameter servers, separated by ",". ### Prepare Training Dataset @@ -178,7 +159,7 @@ Your workspace may looks like: - `my_lib.py`: user defined libraries, like PIL libs. This is optional. - `word_dict.pickle`: dict file for training word embeding. -- `train.py`: training program. Sample code: [api_train_v2_cluster.py](https://github.com/PaddlePaddle/Paddle/tree/develop/doc/howto/usage/cluster/src/word2vec/prepare.py). ***NOTE:*** You may need to modify the head part of `train.py` when using different cluster platform to retrive configuration environment variables: +- `train.py`: training program. Sample code: [api_train_v2_cluster.py](https://github.com/PaddlePaddle/Paddle/tree/develop/doc/howto/usage/cluster/src/word2vec/api_train_v2_cluster.py). ***NOTE:*** You may need to modify the head part of `train.py` when using different cluster platform to retrive configuration environment variables: ```python cluster_train_file = "./train_data_dir/train/train.txt" @@ -202,92 +183,10 @@ We'll introduce cluster job management on these platforms. The examples can be f These cluster platforms provide API or environment variables for training processes, when the job is dispatched to different nodes. Like node ID, IP or total number of nodes etc. -### Cluster Training Using Fabric - -#### Prepare a Linux cluster - -Run `kubectl -f ssh_servers.yaml` under the directory: `paddle/scripts/cluster_train_v2/fabric/docker_cluster` will launch a demo cluster. Run `kubectl get po -o wide` to get IP addresses of these nodes. - -#### Launching Cluster Job -`paddle.py` provides automatical scripts to start all PaddlePaddle cluster processes in different nodes. By default, all command line options can be set as `paddle.py` command options and `paddle.py` will transparently and automatically set these options to PaddlePaddle lower level processes. - -`paddle.py`provides two distinguished command option for easy job launching. - -- `job_dispatch_package` set it with local `workspace` directory, it will be dispatched to all nodes which is set in `conf.py`. It could be helpful for frequently manipulating workspace files. otherwise, frequent multi-nodes workspace deployment is very annoying. -- `job_workspace` set it with already deployed workspace directory, `paddle.py` will skip dispatch stage to directly launch cluster job with all nodes. It could help to reduce heavy -dispatch latency. - -`cluster_train/run.sh` provides command line sample to run `demo/recommendation` cluster job, just modify `job_dispatch_package` and `job_workspace` with your defined directory, then: -``` -sh run.sh -``` - -The cluster Job will start in several seconds. - -#### Kill Cluster Job -`paddle.py` can capture `Ctrl + C` SIGINT signal to automatically kill all processes launched by it. So just stop `paddle.py` to kill cluster job. You should manually kill the job if the program crashed. - -#### Check Cluster Training Result -Check log in $workspace/log for details, each node owns same log structure. - -`paddle_trainer.INFO` -It provides almost all internal output log for training, same as local training. Check runtime model convergence here. - -`paddle_pserver2.INFO` -It provides parameter server running log, which could help to diagnose distributed error. - -`server.log` -It provides stderr and stdout of parameter server process. Check error log if training crashes. - -`train.log` -It provides stderr and stdout of trainer process. Check error log if training crashes. - -#### Check Model Output -After one pass finished, model files will be written in `output` directory in node 0. -`nodefile` in workspace indicates the node id of current cluster job. - -### Cluster Training Using OpenMPI - -#### Prepare an OpenMPI cluster - -Run the following command to start a 3-node MPI cluster and one "head" node. - -```bash -cd paddle/scripts/cluster_train_v2/openmpi/docker_cluster -kubectl create -f head.yaml -kubectl create -f mpi-nodes.yaml -``` - -Then you can log in to every OpenMPI node using ssh without input any passwords. - -#### Launching Cluster Job - -Follow the steps to launch a PaddlePaddle training job in OpenMPI cluster:\ - -```bash -# find out node IP addresses -kubectl get po -o wide -# generate a "machines" file containing node IP addresses -kubectl get po -o wide | grep nodes | awk '{print $6}' > machines -# copy necessary files onto "head" node -scp -i ssh/id_rsa.mpi.pub machines prepare.py train.py start_mpi_train.sh tutorial@[headIP]:~ -# login to head node using ssh -ssh -i ssh/id_rsa.mpi.pub tutorial@[headIP] -# --------------- in head node --------------- -# prepare training data -python prepare.py -# copy training data and dict file to MPI nodes -cat machines | xargs -i scp word_dict.pickle train.py start_mpi_train.sh machines {}:/home/tutorial -# creat a directory for storing log files -mpirun -hostfile machines -n 3 mkdir /home/tutorial/logs -# copy training data to every node -scp train.txt-00000 test.txt-00000 [node1IP]:/home/tutorial -scp train.txt-00001 test.txt-00001 [node2IP]:/home/tutorial -scp train.txt-00002 test.txt-00002 [node3IP]:/home/tutorial -# start the job -mpirun -hostfile machines -n 3 /home/tutorial/start_mpi_train.sh -``` - -### Cluster Training Using Kubernetes +## Use different clusters -The details can be found [here](../k8s/k8s_cn.md) + - [fabric](fabric_en.md) + - [openmpi](openmpi_en.md) + - [kubernetes](k8s_en.md) + - kubernetes distributed + - [kubernetes on AWS](k8s_aws_en.md) diff --git a/doc/howto/usage/cluster/fabric_cn.md b/doc/howto/usage/cluster/fabric_cn.md new file mode 100644 index 0000000000000000000000000000000000000000..0385e401b399a51fad112e604dc56cb2f84c0a4b --- /dev/null +++ b/doc/howto/usage/cluster/fabric_cn.md @@ -0,0 +1,42 @@ +# 使用fabric启动集群训练 + +## 准备一个Linux集群 +可以在`paddle/scripts/cluster_train_v2/fabric/docker_cluster`目录下，执行`kubectl -f ssh_servers.yaml`启动一个测试集群，并使用`kubectl get po -o wide`获得这些节点的IP地址。 + +## 启动集群作业 + +`paddle.py` 提供了自动化脚本来启动不同节点中的所有 PaddlePaddle 集群进程。默认情况下，所有命令行选项可以设置为 `paddle.py` 命令选项并且 `paddle.py` 将透明、自动地将这些选项应用到 PaddlePaddle 底层进程。 + +`paddle.py` 为方便作业启动提供了两个独特的命令选项。 + +- `job_dispatch_package` 设为本地 `workspace` 目录，它将被分发到 `conf.py` 中设置的所有节点。它有助于帮助频繁修改和访问工作区文件的用户减少负担，否则频繁的多节点工作空间部署可能会很麻烦。 +- `job_workspace` 设为已部署的工作空间目录，`paddle.py` 将跳过分发阶段直接启动所有节点的集群作业。它可以帮助减少分发延迟。 + +`cluster_train/run.sh` 提供了命令样例来运行 `doc/howto/usage/cluster/src/word2vec` 集群任务，只需用您定义的目录修改 `job_dispatch_package` 和 `job_workspace`，然后： +``` +sh run.sh +``` + +集群作业将会在几秒后启动。 + +## 终止集群作业 +`paddle.py`能获取`Ctrl + C` SIGINT 信号来自动终止它启动的所有进程。只需中断 `paddle.py` 任务来终止集群作业。如果程序崩溃你也可以手动终止。 + +## 检查集群训练结果 +详细信息请检查 $workspace/log 里的日志，每一个节点都有相同的日志结构。 + +`paddle_trainer.INFO` +提供几乎所有训练的内部输出日志，与本地训练相同。这里检验运行时间模型的收敛。 + +`paddle_pserver2.INFO` +提供 pserver 运行日志，有助于诊断分布式错误。 + +`server.log` +提供 parameter server 进程的 stderr 和 stdout。训练失败时可以检查错误日志。 + +`train.log` +提供训练过程的 stderr 和 stdout。训练失败时可以检查错误日志。 + +## 检查模型输出 +运行完成后，模型文件将被写入节点 0 的 `output` 目录中。 +工作空间中的 `nodefile` 表示当前集群作业的节点 ID。 diff --git a/doc/howto/usage/cluster/fabric_en.md b/doc/howto/usage/cluster/fabric_en.md new file mode 100644 index 0000000000000000000000000000000000000000..bf270d89ab8514801ca4629cf412f73257429df9 --- /dev/null +++ b/doc/howto/usage/cluster/fabric_en.md @@ -0,0 +1,43 @@ +# Cluster Training Using Fabric + +## Prepare a Linux cluster + +Run `kubectl -f ssh_servers.yaml` under the directory: `paddle/scripts/cluster_train_v2/fabric/docker_cluster` will launch a demo cluster. Run `kubectl get po -o wide` to get IP addresses of these nodes. + +## Launching Cluster Job +`paddle.py` provides automatical scripts to start all PaddlePaddle cluster processes in different nodes. By default, all command line options can be set as `paddle.py` command options and `paddle.py` will transparently and automatically set these options to PaddlePaddle lower level processes. + +`paddle.py`provides two distinguished command option for easy job launching. + +- `job_dispatch_package` set it with local `workspace` directory, it will be dispatched to all nodes which is set in `conf.py`. It could be helpful for frequently manipulating workspace files. otherwise, frequent multi-nodes workspace deployment is very annoying. +- `job_workspace` set it with already deployed workspace directory, `paddle.py` will skip dispatch stage to directly launch cluster job with all nodes. It could help to reduce heavy +dispatch latency. + +`cluster_train/run.sh` provides command line sample to run `demo/recommendation` cluster job, just modify `job_dispatch_package` and `job_workspace` with your defined directory, then: +``` +sh run.sh +``` + +The cluster Job will start in several seconds. + +## Kill Cluster Job +`paddle.py` can capture `Ctrl + C` SIGINT signal to automatically kill all processes launched by it. So just stop `paddle.py` to kill cluster job. You should manually kill the job if the program crashed. + +## Check Cluster Training Result +Check log in $workspace/log for details, each node owns same log structure. + +`paddle_trainer.INFO` +It provides almost all internal output log for training, same as local training. Check runtime model convergence here. + +`paddle_pserver2.INFO` +It provides parameter server running log, which could help to diagnose distributed error. + +`server.log` +It provides stderr and stdout of parameter server process. Check error log if training crashes. + +`train.log` +It provides stderr and stdout of trainer process. Check error log if training crashes. + +## Check Model Output +After one pass finished, model files will be written in `output` directory in node 0. +`nodefile` in workspace indicates the node id of current cluster job. diff --git a/doc/howto/usage/cluster/k8s_aws_cn.md b/doc/howto/usage/cluster/k8s_aws_cn.md new file mode 120000 index 0000000000000000000000000000000000000000..c44cd9a731bed7067cdf19aa2f714abdce6c736a --- /dev/null +++ b/doc/howto/usage/cluster/k8s_aws_cn.md @@ -0,0 +1 @@ +k8s_aws_en.md \ No newline at end of file diff --git a/doc/howto/usage/k8s/k8s_aws_en.md b/doc/howto/usage/cluster/k8s_aws_en.md similarity index 100% rename from doc/howto/usage/k8s/k8s_aws_en.md rename to doc/howto/usage/cluster/k8s_aws_en.md diff --git a/doc/howto/usage/k8s/k8s_cn.md b/doc/howto/usage/cluster/k8s_cn.md similarity index 100% rename from doc/howto/usage/k8s/k8s_cn.md rename to doc/howto/usage/cluster/k8s_cn.md diff --git a/doc/howto/usage/k8s/k8s_distributed_cn.md b/doc/howto/usage/cluster/k8s_distributed_cn.md similarity index 91% rename from doc/howto/usage/k8s/k8s_distributed_cn.md rename to doc/howto/usage/cluster/k8s_distributed_cn.md index a9bebf09558b06993119803458977abedbbfbdd0..0fc9e37a990104e942636fc807f67a99f0df9da8 100644 --- a/doc/howto/usage/k8s/k8s_distributed_cn.md +++ b/doc/howto/usage/cluster/k8s_distributed_cn.md @@ -1,6 +1,6 @@ # Kubernetes分布式训练 -前一篇文章介绍了如何在Kubernetes集群上启动一个单机PaddlePaddle训练作业 (Job)。在这篇文章里，我们介绍如何在Kubernetes集群上进行分布式PaddlePaddle训练作业。关于PaddlePaddle的分布式训练，文章 [Cluster Training](https://github.com/baidu/Paddle/blob/develop/doc/cluster/opensource/cluster_train.md)介绍了一种通过SSH远程分发任务，进行分布式训练的方法，与此不同的是，本文将介绍在Kubernetes容器管理平台上快速构建PaddlePaddle容器集群，进行分布式训练的方案。 +前一篇文章介绍了如何在Kubernetes集群上启动一个单机PaddlePaddle训练作业 (Job)。在这篇文章里，我们介绍如何在Kubernetes集群上进行分布式PaddlePaddle训练作业。关于PaddlePaddle的分布式训练，文章 [Cluster Training](http://www.paddlepaddle.org/docs/develop/documentation/zh/howto/usage/cluster/cluster_train_cn.html)介绍了一种通过SSH远程分发任务，进行分布式训练的方法，与此不同的是，本文将介绍在Kubernetes容器管理平台上快速构建PaddlePaddle容器集群，进行分布式训练的方案。 有关Kubernetes相关概念以及如何搭建和配置Kubernetes集群，可以参考[k8s_basis](./k8s_basis_cn.md)。 @@ -28,7 +28,7 @@ PaddlePaddle镜像需要提供`paddle pserver`与`paddle train`进程的运行 - 拷贝训练文件到容器内 - 生成`paddle pserver`与`paddle train`进程的启动参数，并且启动训练 -因为官方镜像 `paddledev/paddle:cpu-latest` 内已经包含PaddlePaddle的执行程序但是还没上述功能，所以我们可以在这个基础上，添加启动脚本，制作新镜像来完成以上的工作。参考镜像的[*Dockerfile*](https://github.com/PaddlePaddle/Paddle/blob/develop/doc/howto/usage/cluster/k8s/src/k8s_train/Dockerfile)。 +因为官方镜像 `paddledev/paddle:cpu-latest` 内已经包含PaddlePaddle的执行程序但是还没上述功能，所以我们可以在这个基础上，添加启动脚本，制作新镜像来完成以上的工作。参考镜像的[*Dockerfile*](https://github.com/PaddlePaddle/Paddle/blob/develop/doc/howto/usage/cluster/src/k8s_train/Dockerfile)。 ```bash $ cd doc/howto/usage/k8s/src/k8s_train @@ -149,20 +149,19 @@ spec: 文件中，`metadata`下的`name`表示这个job的名字。`parallelism，completions`字段表示这个job会同时开启3个PaddlePaddle节点，成功训练且退出的pod数目为3时，这个job才算成功结束。然后申明一个存储卷`jobpath`，代表宿主机目录`/home/work/mfs`，在对容器的描述`containers`字段中，将此目录挂载为容器的`/home/jobpath`目录，这样容器的`/home/jobpath`目录就成为了共享存储，放在这个目录里的文件其实是保存到了MFS上。 -`env`字段表示容器的环境变量，我们将`paddle`运行的一些参数通过这种方式传递到容器内。 +`env`字段表示容器的环境变量，我们将`paddle`运行的一些参数通过这种方式传递到容器内： -环境变量 | 说明 ---- | --- -JOB_PATH | 共享存储挂在的路径 -JOB_NAME | Job的名字 -TRAIN_CONFIG_DIR | 本次训练文件所在目录，与JOB_PATH,JOB_NAME组合可以找到本次训练需要的文件路径 -CONF_PADDLE_NIC | `paddle pserver`进程需要的`--nics`参数，即网卡名 -CONF_PADDLE_PORT | `paddle paserver`的`--port`参数 -CONF_PADDLE_PORTS_NUM | 稠密更新的端口数量，即`--ports_num`参数 -CONF_PADDLE_PORTS_NUM_SPARSE | 稀疏更新的端口数量，即`--ports_num_for_sparse`参数 -CONF_PADDLE_GRADIENT_NUM | 训练节点数量，即`--num_gradient_servers参数` -这些参数的具体描述，读者可以查看[这里](http://www.paddlepaddle.org/doc/ui/cmd_argument/detail_introduction.html#parameter-server-and-distributed-communication)。 +- JOB_PATH：共享存储挂在的路径 +- JOB_NAME：Job的名字 +- TRAIN_CONFIG_DIR：本次训练文件所在目录，与JOB_PATH,JOB_NAME组合可以找到本次训练需要的文件路径 +- CONF_PADDLE_NIC：`paddle pserver`进程需要的`--nics`参数，即网卡名 +- CONF_PADDLE_PORT：`paddle paserver`的`--port`参数 +- CONF_PADDLE_PORTS_NUM：稠密更新的端口数量，即`--ports_num`参数 +- CONF_PADDLE_PORTS_NUM_SPARSE：稀疏更新的端口数量，即`--ports_num_for_sparse`参数 +- CONF_PADDLE_GRADIENT_NUM：训练节点数量，即`--num_gradient_servers参数` + +这些参数的具体描述，读者可以查看[这里](http://www.paddlepaddle.org/docs/develop/documentation/zh/howto/usage/cmd_parameter/detail_introduction_cn.html)。 编写完YAML文件后，可以使用Kubernetes的命令行工具创建job。 diff --git a/doc/howto/usage/k8s/k8s_en.md b/doc/howto/usage/cluster/k8s_en.md similarity index 100% rename from doc/howto/usage/k8s/k8s_en.md rename to doc/howto/usage/cluster/k8s_en.md diff --git a/doc/howto/usage/cluster/openmpi_cn.md b/doc/howto/usage/cluster/openmpi_cn.md new file mode 100644 index 0000000000000000000000000000000000000000..831cafdc03c6a908f31769d0467de022df42dab5 --- /dev/null +++ b/doc/howto/usage/cluster/openmpi_cn.md @@ -0,0 +1,41 @@ +# 在OpenMPI集群中提交训练作业 + +## 准备OpenMPI集群 + +执行下面的命令以启动3个节点的OpenMPI集群和一个"head"节点： + +```bash +paddle/scripts/cluster_train_v2/openmpi/docker_cluster +kubectl create -f head.yaml +kubectl create -f mpi-nodes.yaml +``` + +然后可以从head节点ssh无密码登录到OpenMPI的每个节点上。 + +## 启动集群作业 + +您可以按照下面的步骤在OpenMPI集群中提交paddle训练任务： + +```bash +# 获得head和node节点的IP地址 +kubectl get po -o wide +# 将node节点的IP地址保存到machines文件中 +kubectl get po -o wide | grep nodes | awk '{print $6}' > machines +# 拷贝必要的文件到head节点 +scp -i ssh/id_rsa.mpi.pub machines prepare.py train.py start_mpi_train.sh tutorial@[headIP]:~ +# ssh 登录到head节点 +ssh -i ssh/id_rsa.mpi.pub tutorial@[headIP] +# --------------- 以下操作均在head节点中执行 --------------- +# 准备训练数据 +python prepare.py +# 拷贝训练程序和字典文件到每台MPI节点 +cat machines | xargs -i scp word_dict.pickle train.py start_mpi_train.sh machines {}:/home/tutorial +# 创建日志目录 +mpirun -hostfile machines -n 3 mkdir /home/tutorial/logs +# 拷贝训练数据到各自的节点 +scp train.txt-00000 test.txt-00000 [node1IP]:/home/tutorial +scp train.txt-00001 test.txt-00001 [node2IP]:/home/tutorial +scp train.txt-00002 test.txt-00002 [node3IP]:/home/tutorial +# 启动训练任务 +mpirun -hostfile machines -n 3 /home/tutorial/start_mpi_train.sh +``` diff --git a/doc/howto/usage/cluster/openmpi_en.md b/doc/howto/usage/cluster/openmpi_en.md new file mode 100644 index 0000000000000000000000000000000000000000..09af46e25ebe1f843dc7c7be0997dc706413b65c --- /dev/null +++ b/doc/howto/usage/cluster/openmpi_en.md @@ -0,0 +1,41 @@ +# Cluster Training Using OpenMPI + +## Prepare an OpenMPI cluster + +Run the following command to start a 3-node MPI cluster and one "head" node. + +```bash +cd paddle/scripts/cluster_train_v2/openmpi/docker_cluster +kubectl create -f head.yaml +kubectl create -f mpi-nodes.yaml +``` + +Then you can log in to every OpenMPI node using ssh without input any passwords. + +## Launching Cluster Job + +Follow the steps to launch a PaddlePaddle training job in OpenMPI cluster:\ + +```bash +# find out node IP addresses +kubectl get po -o wide +# generate a "machines" file containing node IP addresses +kubectl get po -o wide | grep nodes | awk '{print $6}' > machines +# copy necessary files onto "head" node +scp -i ssh/id_rsa.mpi.pub machines prepare.py train.py start_mpi_train.sh tutorial@[headIP]:~ +# login to head node using ssh +ssh -i ssh/id_rsa.mpi.pub tutorial@[headIP] +# --------------- in head node --------------- +# prepare training data +python prepare.py +# copy training data and dict file to MPI nodes +cat machines | xargs -i scp word_dict.pickle train.py start_mpi_train.sh machines {}:/home/tutorial +# creat a directory for storing log files +mpirun -hostfile machines -n 3 mkdir /home/tutorial/logs +# copy training data to every node +scp train.txt-00000 test.txt-00000 [node1IP]:/home/tutorial +scp train.txt-00001 test.txt-00001 [node2IP]:/home/tutorial +scp train.txt-00002 test.txt-00002 [node3IP]:/home/tutorial +# start the job +mpirun -hostfile machines -n 3 /home/tutorial/start_mpi_train.sh +``` diff --git a/doc/howto/usage/k8s/src/Dockerfile b/doc/howto/usage/cluster/src/Dockerfile similarity index 100% rename from doc/howto/usage/k8s/src/Dockerfile rename to doc/howto/usage/cluster/src/Dockerfile diff --git a/doc/howto/usage/k8s/src/add_security_group.png b/doc/howto/usage/cluster/src/add_security_group.png similarity index 100% rename from doc/howto/usage/k8s/src/add_security_group.png rename to doc/howto/usage/cluster/src/add_security_group.png diff --git a/doc/howto/usage/k8s/src/create_efs.png b/doc/howto/usage/cluster/src/create_efs.png similarity index 100% rename from doc/howto/usage/k8s/src/create_efs.png rename to doc/howto/usage/cluster/src/create_efs.png diff --git a/doc/howto/usage/k8s/src/efs_mount.png b/doc/howto/usage/cluster/src/efs_mount.png similarity index 100% rename from doc/howto/usage/k8s/src/efs_mount.png rename to doc/howto/usage/cluster/src/efs_mount.png diff --git a/doc/howto/usage/cluster/src/k8s-paddle-arch.png b/doc/howto/usage/cluster/src/k8s-paddle-arch.png new file mode 100644 index 0000000000000000000000000000000000000000..b3800c4fe81302d35e49f7dbacb9221c4dfa5cde Binary files /dev/null and b/doc/howto/usage/cluster/src/k8s-paddle-arch.png differ diff --git a/doc/howto/usage/k8s/src/k8s_data/Dockerfile b/doc/howto/usage/cluster/src/k8s_data/Dockerfile similarity index 100% rename from doc/howto/usage/k8s/src/k8s_data/Dockerfile rename to doc/howto/usage/cluster/src/k8s_data/Dockerfile diff --git a/doc/howto/usage/k8s/src/k8s_data/README.md b/doc/howto/usage/cluster/src/k8s_data/README.md similarity index 100% rename from doc/howto/usage/k8s/src/k8s_data/README.md rename to doc/howto/usage/cluster/src/k8s_data/README.md diff --git a/doc/howto/usage/k8s/src/k8s_data/get_data.sh b/doc/howto/usage/cluster/src/k8s_data/get_data.sh similarity index 100% rename from doc/howto/usage/k8s/src/k8s_data/get_data.sh rename to doc/howto/usage/cluster/src/k8s_data/get_data.sh diff --git a/doc/howto/usage/k8s/src/k8s_train/Dockerfile b/doc/howto/usage/cluster/src/k8s_train/Dockerfile similarity index 100% rename from doc/howto/usage/k8s/src/k8s_train/Dockerfile rename to doc/howto/usage/cluster/src/k8s_train/Dockerfile diff --git a/doc/howto/usage/k8s/src/k8s_train/README.md b/doc/howto/usage/cluster/src/k8s_train/README.md similarity index 100% rename from doc/howto/usage/k8s/src/k8s_train/README.md rename to doc/howto/usage/cluster/src/k8s_train/README.md diff --git a/doc/howto/usage/k8s/src/k8s_train/start.sh b/doc/howto/usage/cluster/src/k8s_train/start.sh similarity index 100% rename from doc/howto/usage/k8s/src/k8s_train/start.sh rename to doc/howto/usage/cluster/src/k8s_train/start.sh diff --git a/doc/howto/usage/k8s/src/k8s_train/start_paddle.py b/doc/howto/usage/cluster/src/k8s_train/start_paddle.py similarity index 100% rename from doc/howto/usage/k8s/src/k8s_train/start_paddle.py rename to doc/howto/usage/cluster/src/k8s_train/start_paddle.py diff --git a/doc/howto/usage/k8s/src/managed_policy.png b/doc/howto/usage/cluster/src/managed_policy.png similarity index 100% rename from doc/howto/usage/k8s/src/managed_policy.png rename to doc/howto/usage/cluster/src/managed_policy.png diff --git a/doc/howto/usage/k8s/src/pserver_and_trainer.png b/doc/howto/usage/cluster/src/pserver_and_trainer.png similarity index 100% rename from doc/howto/usage/k8s/src/pserver_and_trainer.png rename to doc/howto/usage/cluster/src/pserver_and_trainer.png diff --git a/doc/howto/usage/k8s/src/route53_create_recordset.png b/doc/howto/usage/cluster/src/route53_create_recordset.png similarity index 100% rename from doc/howto/usage/k8s/src/route53_create_recordset.png rename to doc/howto/usage/cluster/src/route53_create_recordset.png diff --git a/doc/howto/usage/k8s/src/route53_create_zone.png b/doc/howto/usage/cluster/src/route53_create_zone.png similarity index 100% rename from doc/howto/usage/k8s/src/route53_create_zone.png rename to doc/howto/usage/cluster/src/route53_create_zone.png diff --git a/doc/howto/usage/k8s/src/worker_security_group.png b/doc/howto/usage/cluster/src/worker_security_group.png similarity index 100% rename from doc/howto/usage/k8s/src/worker_security_group.png rename to doc/howto/usage/cluster/src/worker_security_group.png diff --git a/doc/howto/usage/k8s/k8s_basis_cn.md b/doc/howto/usage/k8s/k8s_basis_cn.md deleted file mode 100644 index 4c3dc81ed38f239c1f4a83d22b49cf57b5d16a8b..0000000000000000000000000000000000000000 --- a/doc/howto/usage/k8s/k8s_basis_cn.md +++ /dev/null @@ -1,75 +0,0 @@ -# Kubernetes 简介 - -[*Kubernetes*](http://kubernetes.io/)是Google开源的容器集群管理系统，其提供应用部署、维护、扩展机制等功能，利用Kubernetes能方便地管理跨机器运行容器化的应用。Kubernetes可以在物理机或虚拟机上运行，且支持部署到[AWS](http://kubernetes.io/docs/getting-started-guides/aws)，[Azure](http://kubernetes.io/docs/getting-started-guides/azure/)，[GCE](http://kubernetes.io/docs/getting-started-guides/gce)等多种公有云环境。介绍分布式训练之前，需要对[Kubernetes](http://kubernetes.io/)有一个基本的认识，下面先简要介绍一下本文用到的几个Kubernetes概念。 - -- [*Node*](http://kubernetes.io/docs/admin/node/) 表示一个Kubernetes集群中的一个工作节点，这个节点可以是物理机或者虚拟机，Kubernetes集群就是由node节点与master节点组成的。 - -- [*Pod*](http://kubernetes.io/docs/user-guide/pods/) 是一组(一个或多个)容器，pod是Kubernetes的最小调度单元，一个pod中的所有容器会被调度到同一个node上。Pod中的容器共享NET，PID，IPC，UTS等Linux namespace。由于容器之间共享NET namespace，所以它们使用同一个IP地址，可以通过*localhost*互相通信。不同pod之间可以通过IP地址访问。 - -- [*Job*](http://kubernetes.io/docs/user-guide/jobs/) 描述Kubernetes上运行的作业，一次作业称为一个job，通常每个job包括一个或者多个pods，job启动后会创建这些pod并开始执行一个程序，等待这个程序执行成功并返回0则成功退出，如果执行失败，也可以配置不同的重试机制。 - -- [*Volume*](http://kubernetes.io/docs/user-guide/volumes/) 存储卷，是pod内的容器都可以访问的共享目录，也是容器与node之间共享文件的方式，因为容器内的文件都是暂时存在的，当容器因为各种原因被销毁时，其内部的文件也会随之消失。通过volume，就可以将这些文件持久化存储。Kubernetes支持多种volume，例如hostPath(宿主机目录)，gcePersistentDisk，awsElasticBlockStore等。 - -- [*Namespaces*](https://kubernetes.io/docs/user-guide/namespaces/) 命名空间，在kubernetes中创建的所有资源对象(例如上文的pod，job)等都属于一个命名空间，在同一个命名空间中，资源对象的名字是唯一的，不同空间的资源名可以重复，命名空间主要为了对象进行逻辑上的分组便于管理。本文只使用了默认命名空间。 - -- [*PersistentVolume*](https://kubernetes.io/docs/user-guide/persistent-volumes/): 和[*PersistentVolumeClaim*](https://kubernetes.io/docs/user-guide/persistent-volumes/#persistentvolumeclaims)结合，将外部的存储服务在Kubernetes中描述成为统一的资源形式，便于存储资源管理和Pod引用。 - -## 部署Kubernetes集群 - -Kubernetes提供了多种集群部署的方案，本文档内不重复介绍。这里给出集中常见的部署方法： - -- [*minikube*](https://kubernetes.io/docs/getting-started-guides/minikube/): 快速在本地启动一个单机的kubernetes服务器，便于本地验证和测试。 -- [*kubeadm*](http://kubernetes.io/docs/getting-started-guides/kubeadm/): 在不同操作系统，不同主机(Bare-Metal, AWS, GCE)条件下，快速部署集群。 -- [*AWS EC2*](https://kubernetes.io/docs/getting-started-guides/aws/): 在aws上快速部署集群。 -- [*Bare-Metal*](https://kubernetes.io/docs/getting-started-guides/centos/centos_manual_config/): 在物理机上手动部署。 - -可以参考[这个表格](https://kubernetes.io/docs/getting-started-guides/#table-of-solutions)选择适合您的场景的合适方案。 - -## 选择存储方案 - -容器不会保留在运行时生成的数据，job或者应用程序在容器中运行时生成的数据会在容器销毁时消失。为了完成分布式机器学习训练任务，需要有一个外部的存储服务来保存训练所需数据和训练输出。 -常见的可选存储服务包括： - -- [*NFS*](https://github.com/kubernetes/kubernetes/tree/master/examples/volumes/nfs): 可以将磁盘上某个目录共享给网络中其他机器访问。部署和配置比较简单，可以用于小量数据的验证。不提供分布式存储，高可用，冗余等功能。NFS的部署方法可以参考[这里](http://www.tecmint.com/how-to-setup-nfs-server-in-linux/)。 -- [*GlusterFS*](http://gluster.readthedocs.io/en/latest/Quick-Start-Guide/Quickstart/): 网络分布式文件系统，可以在Kubernetes中按照[这个](https://github.com/kubernetes/kubernetes/tree/master/examples/volumes/glusterfs)例子使用。 -- [*Ceph*](http://docs.ceph.com/docs/master/): 分布式文件系统，支持rbd，POSIX API接口(ceph fs)和对象存储API，参考[这里](https://kubernetes.io/docs/user-guide/volumes/#rbd)。 -- [*MooseFS*](https://moosefs.com/documentation.html): 一个分布式的存储系统。需要先挂载到服务器Node上再通过kubernetes hostPath Volume挂载到容器中。 - -## 配置kubectl - -### 安装kubectl -``` -# OS X -curl -LO https://storage.googleapis.com/kubernetes-release/release/$(curl -s https://storage.googleapis.com/kubernetes-release/release/stable.txt)/bin/darwin/amd64/kubectl - -# Linux -curl -LO https://storage.googleapis.com/kubernetes-release/release/$(curl -s https://storage.googleapis.com/kubernetes-release/release/stable.txt)/bin/linux/amd64/kubectl - -# Windows -curl -LO https://storage.googleapis.com/kubernetes-release/release/$(curl -s https://storage.googleapis.com/kubernetes-release/release/stable.txt)/bin/windows/amd64/kubectl.exe -``` - -### 配置kubectl访问你的kubernetes集群 - -编辑`~/.kube/config`这个配置文件，修改`Master-IP`的地址。如果使用SSL认证，则需要配置`certificate-authority`和`users`中的用户证书。如果是使用非SSL方式访问（比如通过8080端口），也可以去掉这些证书的配置。 -``` -apiVersion: v1 -clusters: -- cluster: - certificate-authority: /path/to/ca.crt - server: https://[Master-IP]:443 - name: minikube -contexts: -- context: - cluster: minikube - user: minikube - name: minikube -current-context: minikube -kind: Config -preferences: {} -users: -- name: minikube - user: - client-certificate: /path/to/apiserver.crt - client-key: /Users/wuyi/.minikube/apiserver.key -``` diff --git a/doc/howto/usage/k8s/src/k8s-paddle-arch.png b/doc/howto/usage/k8s/src/k8s-paddle-arch.png deleted file mode 100644 index 2183a232ad402b76f82a67234a5c93e13ce97ac3..0000000000000000000000000000000000000000 Binary files a/doc/howto/usage/k8s/src/k8s-paddle-arch.png and /dev/null differ diff --git a/doc/mobile/cross_compiling_for_ios_cn.md b/doc/mobile/cross_compiling_for_ios_cn.md index 9da48e7f2119ce901fbb3abab73400df27be16d2..d5196d9a4c93c7692d2a624ec7d0650e32806338 100644 --- a/doc/mobile/cross_compiling_for_ios_cn.md +++ b/doc/mobile/cross_compiling_for_ios_cn.md @@ -18,11 +18,11 @@ PaddlePaddle为交叉编译提供了工具链配置文档[cmake/cross_compiling/ - `CMAKE_SYSTEM_NAME`，CMake编译的目标平台，必须设置为`iOS`。在设置`CMAKE_SYSTEM_NAME=iOS`后，PaddlePaddle的CMake系统会自动编译所有的第三方依赖库，并且强制设置一些PaddlePaddle参数的值（`WITH_C_API=ON`、`WITH_GPU=OFF`、`WITH_AVX=OFF`、`WITH_PYTHON=OFF`、`WITH_RDMA=OFF`）。 - `WITH_C_API`，是否编译C-API预测库，必须设置为ON。在iOS平台上只支持使用C-API来预测。 -- `WITH_SWIG_PY`，必须设置为ON。在iOS平台上不支持通过swig调用来训练或者预测。 +- `WITH_SWIG_PY`，必须设置为`OFF`。在iOS平台上不支持通过swig调用来训练或者预测。 iOS平台可选配置参数： -- `IOS_PLATFORM`，可设置为`OS/SIMULATOR`，默认值为`OS`。 +- `IOS_PLATFORM`，可设置为`OS`（默认值）或`SIMULATOR`。 - `OS`，构建目标为`arm`架构的iPhone或者iPad等物理设备。 - `SIMULATOR`，构建目标为`x86`架构的模拟器平台。 - `IOS_ARCH`，目标架构。针对不同的`IOS_PLATFORM`，可设置的目标架构如下表所示，默认编译所有架构： diff --git a/doc/mobile/cross_compiling_for_ios_en.md b/doc/mobile/cross_compiling_for_ios_en.md new file mode 100644 index 0000000000000000000000000000000000000000..aa390cd61f3fbd75e5a3b342f3559e76da35a918 --- /dev/null +++ b/doc/mobile/cross_compiling_for_ios_en.md @@ -0,0 +1,120 @@ +# PaddlePaddle Compiling Guide for iOS + +This tutorial will walk you through cross compiling the PaddlePaddle library for iOS from the source in MacOS. + +## Preparation + +Apple provides Xcode for cross-compiling and IDE for iOS development. Download from App store or [here](https://developer.apple.com/cn/xcode/). To verify your installation, run command as follows + +```bash +$ xcodebuild -version +Xcode 9.0 +Build version 9A235 +``` + +## Cross-compiling configurations + +PaddlePaddle provides cross-compiling toolchain configuration documentation [cmake/cross_compiling/ios.cmake](https://github.com/PaddlePaddle/Paddle/blob/develop/cmake/cross_compiling/ios.cmake), which has some default settings for frequently used compilers. + +There are some mandatory environment variables need to be set before cross compiling PaddlePaddle for iOS: + +- `CMAKE_SYSTEM_NAME`, CMake compiling target platform name, has to be `iOS`. PaddlePaddle CMake will compile all the third party dependencies and enforce some parameters (`WITH_C_API=ON`, `WITH_GPU=OFF`, `WITH_AVX=OFF`, `WITH_PYTHON=OFF`,`WITH_RDMA=OFF`) when this variable is set with value `iOS`. + +- `WITH_C_API`, Whether to compile inference C-API library, has to be `ON`, since C-API is the only supported interface for inferencing in iOS. +- `WITH_SWIG_PY`, has to be `OFF`. It's not supported to inference or train via swig in iOS. + +Optional environment variables for iOS are: + +- `IOS_PLATFORM`, either `OS` (default) or `SIMULATOR`. + - `OS`, build targets ARM-based physical devices like iPhone or iPad. + - `SIMULATOR`, build targets x86 architecture simulators. +- `IOS_ARCH`, target architecture. By default, all architecture types will be compiled. If you need to specify the architecture to compile for, please find valid values for different `IOS_PLATFORM` settings from the table below: + + + + + + + + + + + + + + + + + + + + + + +
IOS_PLATFORMIOS_ARCH
OSarmv7, armv7s, arm64
SIMULATORi386, x86_64
+ +- `IOS_DEPLOYMENT_TARGET`, minimum iOS version to deployment, `7.0` by default. +- `IOS_ENABLE_BITCODE`, whether to enable [Bitcode](https://developer.apple.com/library/content/documentation/IDEs/Conceptual/AppDistributionGuide/AppThinning/AppThinning.html#//apple_ref/doc/uid/TP40012582-CH35-SW3), values can be `ON/OFF`, `ON` by default. +- `IOS_USE_VECLIB_FOR_BLAS`, whether to use [vecLib](https://developer.apple.com/documentation/accelerate/veclib) framework for BLAS computing. values can be `ON/OFF`, `OFF` by default. +- `IOS_DEVELOPMENT_ROOT`, the path to `Developer` directory, can be explicitly set with your `/path/to/platform/Developer`. If left blank, PaddlePaddle will automatically pick the Xcode corresponding `platform`'s `Developer` directory based on your `IOS_PLATFORM` value. +- `IOS_SDK_ROOT`, the path to `SDK` root, can be explicitly set with your `/path/to/platform/Developer/SDKs/SDK`. if left black, PaddlePaddle will pick the latest SDK in the directory of `IOS_DEVELOPMENT_ROOT`. + +other settings： + +- `USE_EIGEN_FOR_BLAS`, whether to use Eigen for matrix computing. effective when `IOS_USE_VECLIB_FOR_BLAS=OFF`. Values can be `ON/OFF`, `OFF` by default. +- `HOST_C/CXX_COMPILER`, host C/C++ compiler. Uses value from environment variable `CC/CXX` by default or `cc/c++` if `CC/CXX` doesn't exist. + +some typical cmake configurations: + +```bash +cmake -DCMAKE_SYSTEM_NAME=iOS \ + -DIOS_PLATFORM=OS \ + -DIOS_ARCH="armv7;arm64" \ + -DIOS_ENABLE_BITCODE=ON \ + -DIOS_USE_VECLIB_FOR_BLAS=ON \ + -DCMAKE_INSTALL_PREFIX=your/path/to/install \ + -DWITH_C_API=ON \ + -DWITH_TESTING=OFF \ + -DWITH_SWIG_PY=OFF \ + .. +``` + +```bash +cmake -DCMAKE_SYSTEM_NAME=iOS \ + -DIOS_PLATFORM=SIMULATOR \ + -DIOS_ARCH="x86_64" \ + -DIOS_USE_VECLIB_FOR_BLAS=ON \ + -DCMAKE_INSTALL_PREFIX=your/path/to/install \ + -DWITH_C_API=ON \ + -DWITH_TESTING=OFF \ + -DWITH_SWIG_PY=OFF \ + .. +``` + +You can set other compiling parameters for your own need. I.E. if you are trying to minimize the library size, set `CMAKE_BUILD_TYPE` with `MinSizeRel`; or if the performance is your concern, set `CMAKE_BUILD_TYPE` with `Release`. You can even manipulate the PaddlePaddle compiling procedure by manually set `CMAKE_C/CXX_FLAGS` values. + +**TIPS for a better performance**: + +- set `CMAKE_BUILD_TYPE` with `Release` +- set `IOS_USE_VECLIB_FOR_BLAS` with `ON` + +## Compile and install + +After CMake, run following commands, PaddlePaddle will download the compile 3rd party dependencies, compile and install PaddlePaddle inference library. + +``` +$ make +$ make install +``` + +Please Note: if you compiled PaddlePaddle in the source directory for other platforms, do remove `third_party` and `build` directory within the source with `rm -rf` to ensure that all the 3rd party libraries dependencies and PaddlePaddle is newly compiled with current CMake configuration. + +`your/path/to/install` directory will have following directories after `compile` and `install`: + +- `include`, contains all the C-API header files. +- `lib`, contains PaddlePaddle C-API static library. +- `third_party` contains all the 3rd party libraries. + +Please note: if PaddlePaddle library need to support both physical devices and simulators, you will need to compile correspondingly, then merge fat library with `lipo`. + +Now you will have PaddlePaddle library compiled and installed, the fat library can be used in deep learning related iOS APPs. Please refer to C-API documentation for usage guides. diff --git a/doc/mobile/index_en.rst b/doc/mobile/index_en.rst index 3c08d736717cfe8d5fdf449dc58015086befbe60..ef421dacad458828cadf8cf505375d6c4bfd9dde 100644 --- a/doc/mobile/index_en.rst +++ b/doc/mobile/index_en.rst @@ -5,4 +5,5 @@ MOBILE :maxdepth: 1 cross_compiling_for_android_en.md + cross_compiling_for_ios_en.md cross_compiling_for_raspberry_en.md diff --git a/paddle/capi/error.cpp b/paddle/capi/error.cpp index 169b65f92104336d9ec12e2a5a6778db25080270..96ce31b45fc3f83237146443cbe4289af7bfa239 100644 --- a/paddle/capi/error.cpp +++ b/paddle/capi/error.cpp @@ -14,7 +14,7 @@ limitations under the License. */ #include "error.h" -const char* paddle_error_string(paddle_error err) { +extern "C" const char* paddle_error_string(paddle_error err) { switch (err) { case kPD_NULLPTR: return "nullptr error"; diff --git a/paddle/capi/error.h b/paddle/capi/error.h index 9d9d0ed63a5276c6b9a8747e1ee1fce6872bdc9e..2da9e0a3ef604fbcd53bac271c72ef33b3105152 100644 --- a/paddle/capi/error.h +++ b/paddle/capi/error.h @@ -29,9 +29,17 @@ typedef enum { kPD_UNDEFINED_ERROR = -1, } paddle_error; +#ifdef __cplusplus +extern "C" { +#endif + /** * Error string for Paddle API. */ PD_API const char* paddle_error_string(paddle_error err); +#ifdef __cplusplus +} +#endif + #endif diff --git a/paddle/framework/CMakeLists.txt b/paddle/framework/CMakeLists.txt index 4b0eff3adb6fff0c9599b8613c5f19daea840674..206e298eb27a2daaec5c674d45cfe4b81a6b522d 100644 --- a/paddle/framework/CMakeLists.txt +++ b/paddle/framework/CMakeLists.txt @@ -58,3 +58,6 @@ cc_test(var_type_inference_test SRCS var_type_inference_test.cc DEPS op_registry proto_desc) cc_library(selected_rows SRCS selected_rows.cc DEPS tensor) cc_test(selected_rows_test SRCS selected_rows_test.cc DEPS selected_rows) + +cc_library(init SRCS init.cc DEPS gflags executor place stringpiece) +cc_test(init_test SRCS init_test.cc DEPS init) diff --git a/paddle/framework/backward.cc b/paddle/framework/backward.cc index a17036c6527da3a4a32f021a57542b6b6d68a395..faf6e60cbd1bcda9864c12696b336998ea7606b7 100644 --- a/paddle/framework/backward.cc +++ b/paddle/framework/backward.cc @@ -430,14 +430,14 @@ std::vector> MakeBlockBackward( std::vector> op_grads; if ((*it)->Type() == "recurrent" || (*it)->Type() == "while") { - int step_block_idx = (*it)->GetBlockAttr("step_block"); + int step_block_idx = (*it)->GetBlockAttr("sub_block"); BlockDescBind* backward_block = CreateStepBlock( program_desc, no_grad_vars, grad_to_var, step_block_idx); op_grads = MakeOpGrad(*it, no_grad_vars, grad_to_var, {backward_block}); } else if ((*it)->Type() == "conditional_block") { BlockDescBind* backward_block = CreateStepBlock(program_desc, no_grad_vars, grad_to_var, - (*it)->GetBlockAttr("block")); + (*it)->GetBlockAttr("sub_block")); op_grads = MakeOpGrad(*it, no_grad_vars, grad_to_var, {backward_block}); } else { op_grads = MakeOpGrad(*it, no_grad_vars, grad_to_var); diff --git a/paddle/framework/ddim_test.cc b/paddle/framework/ddim_test.cc index 756232b1b56a49d2c91cc2cac950ca508c54fb3f..bd5ea09d7da700479aa387283d7bde77c64c1293 100644 --- a/paddle/framework/ddim_test.cc +++ b/paddle/framework/ddim_test.cc @@ -1,3 +1,16 @@ +/* Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserve. + + Licensed under the Apache License, Version 2.0 (the "License"); + you may not use this file except in compliance with the License. + You may obtain a copy of the License at + + http://www.apache.org/licenses/LICENSE-2.0 + + Unless required by applicable law or agreed to in writing, software + distributed under the License is distributed on an "AS IS" BASIS, + WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + See the License for the specific language governing permissions and + limitations under the License. */ #include #include diff --git a/paddle/framework/executor.cc b/paddle/framework/executor.cc index 83aa927c293676c3800ed945c175e4f3dc5629d6..a8b8a6f8e82525bd9a1f709516483de6f44142dc 100644 --- a/paddle/framework/executor.cc +++ b/paddle/framework/executor.cc @@ -33,32 +33,12 @@ namespace framework { const std::string kFeedOpType = "feed"; const std::string kFetchOpType = "fetch"; -Executor::Executor(const std::vector& places) : own_(true) { - PADDLE_ENFORCE_GT(places.size(), 0); - device_contexts_.resize(places.size()); - for (size_t i = 0; i < places.size(); i++) { - if (platform::is_cpu_place(places[i])) { - device_contexts_[i] = new platform::CPUDeviceContext( - boost::get(places[i])); - } else if (platform::is_gpu_place(places[i])) { -#ifdef PADDLE_WITH_CUDA - device_contexts_[i] = new platform::CUDADeviceContext( - boost::get(places[i])); -#else - PADDLE_THROW( - "'GPUPlace' is not supported, Please re-compile with WITH_GPU " - "option"); -#endif - } - } -} +DeviceContextPool* DeviceContextPool::pool = nullptr; -Executor::~Executor() { - if (own_) { - for (auto& device_context : device_contexts_) { - delete device_context; - } - } +Executor::Executor(const std::vector& places) { + DeviceContextPool& pool = DeviceContextPool::Get(); + auto borrowed_contexts = pool.Borrow(places); + device_contexts_.swap(borrowed_contexts); } static void CreateTensor(Variable* var, VarDesc::VarType var_type) { @@ -132,8 +112,5 @@ void Executor::Run(const ProgramDescBind& pdesc, Scope* scope, int block_id, } } -Executor::Executor(const platform::DeviceContext& device) - : device_contexts_({&device}), own_(false) {} - } // namespace framework } // namespace paddle diff --git a/paddle/framework/executor.h b/paddle/framework/executor.h index b745f4f6474ef688774f4c833a3958942e9aa8cb..073e04729b1166f1cabd16709d161fda0d580f1c 100644 --- a/paddle/framework/executor.h +++ b/paddle/framework/executor.h @@ -14,19 +14,98 @@ limitations under the License. */ #pragma once +#include +#include + #include "paddle/framework/op_info.h" #include "paddle/framework/program_desc.h" #include "paddle/framework/scope.h" #include "paddle/framework/tensor.h" +#include "paddle/platform/device_context.h" namespace paddle { namespace framework { +class DeviceContextPool { + public: + static DeviceContextPool& Get() { + PADDLE_ENFORCE_NOT_NULL(pool, "Need to Create DeviceContextPool first!"); + return *pool; + } + + static DeviceContextPool& Create(const std::vector& places) { + if (pool == nullptr) { + pool = new DeviceContextPool(places); + } + return *pool; + } + + std::vector Borrow( + const std::vector& places) { + PADDLE_ENFORCE_GT(places.size(), 0); + PADDLE_ENFORCE_LE(places.size(), device_contexts_.size()); + std::vector borrowed_contexts; + for (auto& place : places) { + auto range = device_contexts_.equal_range(place); + if (range.first == range.second) { + PADDLE_THROW( + "'Place' is not supported, Please re-compile with WITH_GPU " + "option"); + } + // TODO(dzhwinter) : assign the first found device. Will enhanced later. + // device load balancer maybe useful here. + borrowed_contexts.emplace_back(range.first->second); + } + return borrowed_contexts; + } + + explicit DeviceContextPool(const std::vector& places) { + PADDLE_ENFORCE_GT(places.size(), 0); + for (size_t i = 0; i < places.size(); i++) { + if (platform::is_cpu_place(places[i])) { + device_contexts_.emplace( + places[i], new platform::CPUDeviceContext( + boost::get(places[i]))); + } else if (platform::is_gpu_place(places[i])) { +#ifdef PADDLE_WITH_CUDA + device_contexts_.emplace( + places[i], new platform::CUDADeviceContext( + boost::get(places[i]))); +#else + PADDLE_THROW( + "'GPUPlace' is not supported, Please re-compile with WITH_GPU " + "option"); +#endif + } + } + } + + ~DeviceContextPool() {} + + private: + static DeviceContextPool* pool; + struct Hash { + std::hash hash_; + size_t operator()(const platform::Place& place) const { + return hash_(place.which()); + } + }; + std::unordered_multimap + device_contexts_; + DISABLE_COPY_AND_ASSIGN(DeviceContextPool); +}; + class Executor { public: + // TODO(dzhwinter) : Do not rely on this function, it will be removed + explicit Executor(const platform::DeviceContext& device) + : Executor(std::vector({device.GetPlace()})) {} + + explicit Executor(const platform::Place& place) + : Executor(std::vector({place})) {} + explicit Executor(const std::vector& places); - explicit Executor(const platform::DeviceContext& devices); - ~Executor(); /* @Brief * Runtime evaluation of the given ProgramDesc under certain Scope @@ -39,7 +118,6 @@ class Executor { private: std::vector device_contexts_; - bool own_; }; } // namespace framework diff --git a/paddle/framework/init.cc b/paddle/framework/init.cc new file mode 100644 index 0000000000000000000000000000000000000000..1c4476f4b30aebf094eb27b45fb435c24a9061c1 --- /dev/null +++ b/paddle/framework/init.cc @@ -0,0 +1,80 @@ +/* Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserve. + + Licensed under the Apache License, Version 2.0 (the "License"); + you may not use this file except in compliance with the License. + You may obtain a copy of the License at + + http://www.apache.org/licenses/LICENSE-2.0 + + Unless required by applicable law or agreed to in writing, software + distributed under the License is distributed on an "AS IS" BASIS, + WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + See the License for the specific language governing permissions and + limitations under the License. */ +#include +#include + +#include "paddle/framework/executor.h" +#include "paddle/framework/init.h" +#include "paddle/platform/place.h" +#include "paddle/string/piece.h" + +namespace paddle { +namespace framework { + +std::once_flag gflags_init_flag; + +// TODO(qijun) move init gflags to init.cc +void InitGflags(std::vector &argv) { + std::call_once(gflags_init_flag, [&]() { + int argc = argv.size(); + char **arr = new char *[argv.size()]; + std::string line; + for (size_t i = 0; i < argv.size(); i++) { + arr[i] = &argv[i][0]; + line += argv[i]; + line += ' '; + } + google::ParseCommandLineFlags(&argc, &arr, true); + VLOG(1) << "Init commandline: " << line; + }); +} + +bool InitDevices(const std::vector &devices) { + // device format + // CPU + // GPU:1 + // TODO(dzhwinter) : add device format annotation for users. + std::vector places; + for (auto &device : devices) { + auto p = string::Piece(device); + if (string::Find(p, ':', 0) == string::Piece::npos) { + places.emplace_back(platform::CPUPlace()); + } else if (string::HasPrefix(p, "GPU")) { +#ifdef PADDLE_WITH_CUDA + auto pos = string::RFind(p, ':', string::Piece::npos); + auto number = device.substr(pos + 1); + places.emplace_back(platform::GPUPlace(std::stoi(number))); +#else + LOG(WARNING) + << "'GPU' is not supported, Please re-compile with WITH_GPU option"; +#endif + } else { + return false; + } + } + + if (std::find_if(places.begin(), places.end(), + [&](const platform::Place &place) { + return platform::is_cpu_place(place); + }) == places.end()) { + places.emplace_back(platform::CPUPlace()); + LOG(WARNING) << "Not specified any device, use CPU by Default."; + } + DeviceContextPool::Create(places); + return true; + return true; +} + +} // namespace framework +} // namespace paddle diff --git a/paddle/framework/init.h b/paddle/framework/init.h new file mode 100644 index 0000000000000000000000000000000000000000..1715cd81e6647158e269e39d4d91fbe065cd0008 --- /dev/null +++ b/paddle/framework/init.h @@ -0,0 +1,28 @@ +/* Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserve. + + Licensed under the Apache License, Version 2.0 (the "License"); + you may not use this file except in compliance with the License. + You may obtain a copy of the License at + + http://www.apache.org/licenses/LICENSE-2.0 + + Unless required by applicable law or agreed to in writing, software + distributed under the License is distributed on an "AS IS" BASIS, + WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + See the License for the specific language governing permissions and + limitations under the License. */ +#pragma once +#include + +#include "gflags/gflags.h" +#include "glog/logging.h" + +namespace paddle { +namespace framework { + +void InitGflags(std::vector &argv); + +bool InitDevices(const std::vector &devices); + +} // namespace framework +} // namespace paddle diff --git a/paddle/framework/init_test.cc b/paddle/framework/init_test.cc new file mode 100644 index 0000000000000000000000000000000000000000..f65e881a761e0a546d595eced26dd5b12475a763 --- /dev/null +++ b/paddle/framework/init_test.cc @@ -0,0 +1,27 @@ +/* Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserve. + + Licensed under the Apache License, Version 2.0 (the "License"); + you may not use this file except in compliance with the License. + You may obtain a copy of the License at + + http://www.apache.org/licenses/LICENSE-2.0 + + Unless required by applicable law or agreed to in writing, software + distributed under the License is distributed on an "AS IS" BASIS, + WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + See the License for the specific language governing permissions and + limitations under the License. */ +#include "gtest/gtest.h" + +#include "paddle/framework/init.h" + +TEST(Init, InitDevices) { + using paddle::framework::InitDevices; + std::vector ds1 = {"CPU"}; + ASSERT_EQ(InitDevices(ds1), true); + +#ifdef PADDLE_WITH_CUDA + std::vector ds2 = {"CPU", "GPU:0", "GPU:1"}; + ASSERT_EQ(InitDevices(ds2), true); +#endif +} diff --git a/paddle/function/GemmConvOp.cpp b/paddle/function/GemmConvOp.cpp index 8d34eee886a6202691e5dec2ab62e7c5b0ac7fb1..de7b70e271b38ebe3a4c38704d0cced47d010788 100644 --- a/paddle/function/GemmConvOp.cpp +++ b/paddle/function/GemmConvOp.cpp @@ -126,6 +126,11 @@ public: inputData += inputChannels * inputHeight * inputWidth; outputData += outputChannels * outputHeight * outputWidth; } +#ifdef PADDLE_MOBILE_INFERENCE + if (Device == DEVICE_TYPE_CPU) { + memory_.reset(); + } +#endif } }; diff --git a/paddle/gserver/layers/ROIPoolLayer.cpp b/paddle/gserver/layers/ROIPoolLayer.cpp index 2c8256b91c97b513ce7237b8174c522430094926..7d7c30b4d89e2dd137e7fc7de3159c07bbab9fb4 100644 --- a/paddle/gserver/layers/ROIPoolLayer.cpp +++ b/paddle/gserver/layers/ROIPoolLayer.cpp @@ -84,12 +84,15 @@ void ROIPoolLayer::forward(PassType passType) { size_t poolChannelOffset = pooledHeight_ * pooledWidth_; real* outputData = outputValue->getData(); - Matrix::resizeOrCreate(maxIdxs_, - numROIs, - channels_ * pooledHeight_ * pooledWidth_, - false, - false); - real* argmaxData = maxIdxs_->getData(); + real* argmaxData = nullptr; + if (passType != PASS_TEST) { + Matrix::resizeOrCreate(maxIdxs_, + numROIs, + channels_ * pooledHeight_ * pooledWidth_, + false, + false); + argmaxData = maxIdxs_->getData(); + } for (size_t n = 0; n < numROIs; ++n) { // the first five elememts of each RoI should be: @@ -128,14 +131,18 @@ void ROIPoolLayer::forward(PassType passType) { bool isEmpty = (hend <= hstart) || (wend <= wstart); size_t poolIndex = ph * pooledWidth_ + pw; outputData[poolIndex] = isEmpty ? 0 : -FLT_MAX; - argmaxData[poolIndex] = -1; + if (argmaxData) { + argmaxData[poolIndex] = -1; + } for (size_t h = hstart; h < hend; ++h) { for (size_t w = wstart; w < wend; ++w) { size_t index = h * width_ + w; if (batchData[index] > outputData[poolIndex]) { outputData[poolIndex] = batchData[index]; - argmaxData[poolIndex] = index; + if (argmaxData) { + argmaxData[poolIndex] = index; + } } } } @@ -143,7 +150,9 @@ void ROIPoolLayer::forward(PassType passType) { } batchData += channelOffset; outputData += poolChannelOffset; - argmaxData += poolChannelOffset; + if (argmaxData) { + argmaxData += poolChannelOffset; + } } bottomROIs += roiOffset; } diff --git a/paddle/gserver/layers/SequenceToBatch.cpp b/paddle/gserver/layers/SequenceToBatch.cpp index 5fa7b6f4881b9582b540a5b1bfe849220cc2a4ea..6b769378d24838364701d0f128a7308c6195cc41 100644 --- a/paddle/gserver/layers/SequenceToBatch.cpp +++ b/paddle/gserver/layers/SequenceToBatch.cpp @@ -171,12 +171,31 @@ void SequenceToBatch::sequence2BatchCopy(Matrix &batch, hl_sequence2batch_copy( batchData, seqData, idxData, seqWidth, batchCount, seq2batch); } else { - for (int i = 0; i < batchCount; ++i) { - if (seq2batch) { + if (seq2batch) { +#ifdef PADDLE_USE_MKLML + const int blockMemSize = 8 * 1024; + const int blockSize = blockMemSize / sizeof(real); +#pragma omp parallel for collapse(2) + for (int i = 0; i < batchCount; ++i) { + for (int j = 0; j < seqWidth; j += blockSize) { + memcpy(batch.rowBuf(i) + j, + sequence.rowBuf(idxData[i]) + j, + (j + blockSize > seqWidth) ? (seqWidth - j) * sizeof(real) + : blockMemSize); + } + } +#else + for (int i = 0; i < batchCount; ++i) { memcpy(batch.rowBuf(i), sequence.rowBuf(idxData[i]), seqWidth * sizeof(real)); - } else { + } +#endif + } else { +#ifdef PADDLE_USE_MKLML +#pragma omp parallel for +#endif + for (int i = 0; i < batchCount; ++i) { memcpy(sequence.rowBuf(idxData[i]), batch.rowBuf(i), seqWidth * sizeof(real)); diff --git a/paddle/math/float16.h b/paddle/math/float16.h index 76ad3a01239e409caeefc36a3d562ed5e388dc92..efebbce50405018c6b7ce2049f8d55c33680469f 100644 --- a/paddle/math/float16.h +++ b/paddle/math/float16.h @@ -79,7 +79,7 @@ public: #ifdef PADDLE_CUDA_FP16 HOSTDEVICE inline explicit float16(const half& h) { #if CUDA_VERSION >= 9000 - x = reinterpret_cast<__half_raw*>(&h)->x; + x = reinterpret_cast<__half_raw*>(const_cast(&h))->x; #else x = h.x; #endif // CUDA_VERSION >= 9000 @@ -145,7 +145,7 @@ public: #ifdef PADDLE_CUDA_FP16 HOSTDEVICE inline float16& operator=(const half& rhs) { #if CUDA_VERSION >= 9000 - x = reinterpret_cast<__half_raw*>(&rhs)->x; + x = reinterpret_cast<__half_raw*>(const_cast(&rhs))->x; #else x = rhs.x; #endif diff --git a/paddle/memory/detail/system_allocator.cc b/paddle/memory/detail/system_allocator.cc index 6a815a1b57db1d833781ca224f34e4559af9b9a5..509250debc2b2fd2e87078ab5f233ae2db6fd898 100644 --- a/paddle/memory/detail/system_allocator.cc +++ b/paddle/memory/detail/system_allocator.cc @@ -19,6 +19,7 @@ limitations under the License. */ #include // for malloc and free #include // for mlock and munlock +#include // for std::max #include "gflags/gflags.h" @@ -28,7 +29,7 @@ limitations under the License. */ // of memory available to the system for paging. So, by default, we // should set false to use_pinned_memory. DEFINE_bool(use_pinned_memory, true, "If set, allocate cpu pinned memory."); - +DECLARE_double(fraction_of_gpu_memory_to_use); namespace paddle { namespace memory { namespace detail { @@ -77,45 +78,20 @@ void* GPUAllocator::Alloc(size_t& index, size_t size) { // CUDA documentation doesn't explain if cudaMalloc returns nullptr // if size is 0. We just make sure it does. if (size <= 0) return nullptr; - - size_t available = 0; - size_t capacity = 0; - paddle::platform::GpuMemoryUsage(available, capacity); - - // Reserve memory for page tables, etc. - size_t reserving = 0.05 * capacity + paddle::platform::GpuMinChunkSize(); - size_t usable = available > reserving ? available - reserving : 0; - - // If remaining size no less than expected size, using general - // cudaMalloc to allocate GPU memory. - void* p = 0; - if (size <= usable) { - cudaError_t result = cudaMalloc(&p, size); - if (result == cudaSuccess) { - index = 0; - gpu_alloc_size_ += size; - return p; - } - } - - // If remaining size less than expected size or cudaMalloc failed, - // cudaMallocHost will be considered as a fallback allocator. - // - // NOTE: here, we use GpuMaxAllocSize() as the maximum memory size - // of host fallback allocation. Allocates too much would reduce - // the amount of memory available to the underlying system for paging. - usable = paddle::platform::GpuMaxAllocSize() - fallback_alloc_size_; - - if (size > usable) return nullptr; - - cudaError_t result = cudaMallocHost(&p, size); + void* p; + cudaError_t result = cudaMalloc(&p, size); if (result == cudaSuccess) { - index = 1; - fallback_alloc_size_ += size; + index = 0; + gpu_alloc_size_ += size; return p; + } else { + LOG(WARNING) + << "Cannot malloc " << size / 1024.0 / 1024.0 + << " MB GPU memory. Please shrink FLAGS_fraction_of_gpu_memory_to_use " + "environment variable to a lower value. Current value is " + << FLAGS_fraction_of_gpu_memory_to_use; + return nullptr; } - - return nullptr; } void GPUAllocator::Free(void* p, size_t size, size_t index) { diff --git a/paddle/operators/chunk_eval_op.cc b/paddle/operators/chunk_eval_op.cc index 94127ab33e51d5529b63b5e3696032ef8adcf03e..894f355deb9d764ef72d452f362e6b42f8831667 100644 --- a/paddle/operators/chunk_eval_op.cc +++ b/paddle/operators/chunk_eval_op.cc @@ -32,6 +32,13 @@ class ChunkEvalOp : public framework::OperatorWithKernel { "Output(Recall) of ChunkEvalOp should not be null."); PADDLE_ENFORCE(ctx->HasOutput("F1-Score"), "Output(F1-Score) of ChunkEvalOp should not be null."); + PADDLE_ENFORCE(ctx->HasOutput("NumInferChunks"), + "Output(NumInferChunks) of ChunkEvalOp should not be null."); + PADDLE_ENFORCE(ctx->HasOutput("NumLabelChunks"), + "Output(NumLabelChunks) of ChunkEvalOp should not be null."); + PADDLE_ENFORCE( + ctx->HasOutput("NumCorrectChunks"), + "Output(NumCorrectChunks) of ChunkEvalOp should not be null."); auto inference_dim = ctx->GetInputDim("Inference"); auto label_dim = ctx->GetInputDim("Label"); @@ -42,6 +49,9 @@ class ChunkEvalOp : public framework::OperatorWithKernel { ctx->SetOutputDim("Precision", {1}); ctx->SetOutputDim("Recall", {1}); ctx->SetOutputDim("F1-Score", {1}); + ctx->SetOutputDim("NumInferChunks", {1}); + ctx->SetOutputDim("NumLabelChunks", {1}); + ctx->SetOutputDim("NumCorrectChunks", {1}); } protected: @@ -70,6 +80,16 @@ class ChunkEvalOpMaker : public framework::OpProtoAndCheckerMaker { "sensitivity) of chunks on the given mini-batch."); AddOutput("F1-Score", "(float). The evaluated F1-Score on the given mini-batch."); + AddOutput("NumInferChunks", + "(int64_t). The number of chunks in Inference on the given " + "mini-batch."); + AddOutput( + "NumLabelChunks", + "(int64_t). The number of chunks in Label on the given mini-batch."); + AddOutput( + "NumCorrectChunks", + "(int64_t). The number of chunks both in Inference and Label on the " + "given mini-batch."); AddAttr("num_chunk_types", "(int). The number of chunk type. See below for details."); AddAttr( diff --git a/paddle/operators/chunk_eval_op.h b/paddle/operators/chunk_eval_op.h index 9cd758a8253914515437b480e17a94d5d6b21fd2..74ab435c860b22b2ee3f485743540976a7a31b96 100644 --- a/paddle/operators/chunk_eval_op.h +++ b/paddle/operators/chunk_eval_op.h @@ -111,9 +111,7 @@ class ChunkEvalKernel : public framework::OpKernel { std::vector label_segments; std::vector output_segments; std::set excluded_chunk_types; - int64_t num_output_segments = 0; - int64_t num_label_segments = 0; - int64_t num_correct = 0; + if (context.Attr("chunk_scheme") == "IOB") { num_tag_types = 2; tag_begin = 0; @@ -151,12 +149,24 @@ class ChunkEvalKernel : public framework::OpKernel { auto* precision = context.Output("Precision"); auto* recall = context.Output("Recall"); auto* f1 = context.Output("F1-Score"); + auto* num_infer_chunks = context.Output("NumInferChunks"); + auto* num_label_chunks = context.Output("NumLabelChunks"); + auto* num_correct_chunks = context.Output("NumCorrectChunks"); const int64_t* inference_data = inference->data(); const int64_t* label_data = label->data(); T* precision_data = precision->mutable_data(context.GetPlace()); T* racall_data = recall->mutable_data(context.GetPlace()); T* f1_data = f1->mutable_data(context.GetPlace()); + int64_t* num_infer_chunks_data = + num_infer_chunks->mutable_data(context.GetPlace()); + int64_t* num_label_chunks_data = + num_label_chunks->mutable_data(context.GetPlace()); + int64_t* num_correct_chunks_data = + num_correct_chunks->mutable_data(context.GetPlace()); + *num_infer_chunks_data = 0; + *num_label_chunks_data = 0; + *num_correct_chunks_data = 0; auto lod = label->lod(); PADDLE_ENFORCE_EQ(lod.size(), 1UL, "Only support one level sequence now."); @@ -166,17 +176,23 @@ class ChunkEvalKernel : public framework::OpKernel { for (int i = 0; i < num_sequences; ++i) { int seq_length = lod[0][i + 1] - lod[0][i]; EvalOneSeq(inference_data + lod[0][i], label_data + lod[0][i], seq_length, - output_segments, label_segments, num_output_segments, - num_label_segments, num_correct, num_chunk_types, - num_tag_types, other_chunk_type, tag_begin, tag_inside, - tag_end, tag_single, excluded_chunk_types); + output_segments, label_segments, *num_infer_chunks_data, + *num_label_chunks_data, *num_correct_chunks_data, + num_chunk_types, num_tag_types, other_chunk_type, tag_begin, + tag_inside, tag_end, tag_single, excluded_chunk_types); } - *precision_data = !num_output_segments ? 0 : static_cast(num_correct) / - num_output_segments; - *racall_data = !num_label_segments ? 0 : static_cast(num_correct) / - num_label_segments; - *f1_data = !num_correct ? 0 : 2 * (*precision_data) * (*racall_data) / - ((*precision_data) + (*racall_data)); + *precision_data = !(*num_infer_chunks_data) + ? 0 + : static_cast(*num_correct_chunks_data) / + (*num_infer_chunks_data); + *racall_data = !(*num_label_chunks_data) + ? 0 + : static_cast(*num_correct_chunks_data) / + (*num_label_chunks_data); + *f1_data = !(*num_correct_chunks_data) + ? 0 + : 2 * (*precision_data) * (*racall_data) / + ((*precision_data) + (*racall_data)); } void EvalOneSeq(const int64_t* output, const int64_t* label, int length, diff --git a/paddle/operators/conditional_block_op.cc b/paddle/operators/conditional_block_op.cc index 03c58a7eab8b2071a3a0b75ac0c665e32ef39876..6f2ef9174e84a0c0ae096956c04039435e6583c6 100644 --- a/paddle/operators/conditional_block_op.cc +++ b/paddle/operators/conditional_block_op.cc @@ -65,7 +65,7 @@ class ConditionalBlockOp : public ConditionalOp { scopes->front() = &scope.NewScope(); auto &cur_scope = *scopes->front(); - auto *block = Attr("block"); + auto *block = Attr("sub_block"); framework::Executor exec(dev_ctx); exec.Run(*block->Program(), &cur_scope, block->ID(), false); } @@ -88,7 +88,7 @@ class ConditionalBlockOpProtoMaker : public framework::OpProtoAndCheckerMaker { "unify the conditional block, rnn and while op, the type of " "scope is std::vector"); AddAttr( - "block", "The step block of conditional block operator"); + "sub_block", "The step block of conditional block operator"); AddComment(R"DOC(Conditional block operator Run the sub-block if X is not empty. Params is the other inputs and Out is the @@ -117,7 +117,7 @@ class ConditionalBlockGradOp : public ConditionalOp { auto &scopes = scope_var->Get>(); framework::Scope &cur_scope = *scopes[0]; - auto *block = Attr("block"); + auto *block = Attr("sub_block"); framework::Executor exec(dev_ctx); exec.Run(*block->Program(), &cur_scope, block->ID(), false); @@ -181,7 +181,7 @@ class ConditionalBlockGradMaker : public framework::SingleGradOpDescMaker { grad_op->SetInput("Scope", Output("Scope")); grad_op->SetOutput(framework::GradVarName("X"), InputGrad("X")); grad_op->SetOutput(framework::GradVarName("Params"), InputGrad("Params")); - grad_op->SetBlockAttr("block", *this->grad_block_[0]); + grad_op->SetBlockAttr("sub_block", *this->grad_block_[0]); return std::unique_ptr(grad_op); } }; diff --git a/paddle/operators/conv_op.h b/paddle/operators/conv_op.h index 749258183ba058cf0ed8d91c4406813694314b85..d2de4e80f751d4938ac9cad60871b470fccf225c 100644 --- a/paddle/operators/conv_op.h +++ b/paddle/operators/conv_op.h @@ -261,8 +261,12 @@ class GemmConvGradKernel : public framework::OpKernel { if (input_grad) { input_grad->mutable_data(context.GetPlace()); - set_zero(dev_ctx, input_grad, static_cast(0)); + // if is_expand is false, the operation of set_zero is unnecessary, + // because math::matmul will reset input_grad. + if (is_expand) { + set_zero(dev_ctx, input_grad, static_cast(0)); + } math::Col2VolFunctor col2vol; math::Col2ImFunctor col2im; diff --git a/paddle/operators/conv_transpose_op.h b/paddle/operators/conv_transpose_op.h index 80600b53614994ba0c740aed0d75c9944333fecc..1171b0435fd2b1abe541043e8283a8fc09dc13c7 100644 --- a/paddle/operators/conv_transpose_op.h +++ b/paddle/operators/conv_transpose_op.h @@ -225,7 +225,6 @@ class GemmConvTransposeGradKernel : public framework::OpKernel { if (input_grad) { input_grad->mutable_data(context.GetPlace()); - set_zero(dev_ctx, input_grad, static_cast(0)); } if (filter_grad) { // filter size (m, c, k_h, k_w) filter_grad->mutable_data(context.GetPlace()); diff --git a/paddle/operators/crop_op.cc b/paddle/operators/crop_op.cc index 7c2a0ac7a705e5aac3d181545f8dfc8881e811f2..5c973fbb3cf9513d82a5b87719cb947466082424 100644 --- a/paddle/operators/crop_op.cc +++ b/paddle/operators/crop_op.cc @@ -88,7 +88,8 @@ There are two ways to set shape: The input should be a k-D tensor(k > 0 and k < 7). As an example: -Given: +Case 1: +Given X = [[0, 1, 2, 0, 0] [0, 3, 4, 0, 0] @@ -107,6 +108,27 @@ we get: Out = [[1, 2], [3, 4]]. + +Case 2: +Given + + X = [[0, 1, 2, 5, 0] + [0, 3, 4, 6, 0] + [0, 0, 0, 0, 0]], + +and + + offsets = [0, 1], + +and + + Y = [[0, 0, 0] + [0, 0, 0]], + +we get: + + Out = [[1, 2, 5], + [3, 4, 6]]. )DOC"); } }; diff --git a/paddle/operators/math/math_function.cc b/paddle/operators/math/math_function.cc index 2b35e4532a9c9f72f473020d472244234af24248..a05810d7781f5286e70b53005ef0b193c945c54c 100644 --- a/paddle/operators/math/math_function.cc +++ b/paddle/operators/math/math_function.cc @@ -277,6 +277,14 @@ void set_constant_with_place( TensorSetConstantCPU(tensor, value)); } +template <> +void set_constant_with_place( + const platform::DeviceContext& context, framework::Tensor* tensor, + float value) { + framework::VisitDataType(framework::ToDataType(tensor->type()), + TensorSetConstantCPU(tensor, value)); +} + struct TensorSetConstantWithPlace : public boost::static_visitor { TensorSetConstantWithPlace(const platform::DeviceContext& context, framework::Tensor* tensor, float value) diff --git a/paddle/operators/math/math_function.cu b/paddle/operators/math/math_function.cu index 1b560a7e2d29c1b63a25d4ec9bbd82d5960a279d..7852bb53a9035f71f52a51529c8e3cea22b0d4aa 100644 --- a/paddle/operators/math/math_function.cu +++ b/paddle/operators/math/math_function.cu @@ -273,6 +273,13 @@ void set_constant_with_place( TensorSetConstantGPU(context, tensor, value)); } +template <> +void set_constant_with_place( + const platform::DeviceContext& context, framework::Tensor* tensor, + float value) { + set_constant_with_place(context, tensor, value); +} + template struct RowwiseAdd; template struct RowwiseAdd; template struct ColwiseSum; diff --git a/paddle/operators/recurrent_op.cc b/paddle/operators/recurrent_op.cc index 29f91636438449f90ea3ffee8adc21595aabe202..232d926f7b975c3b8ebecad983d0f1cc54b9486f 100644 --- a/paddle/operators/recurrent_op.cc +++ b/paddle/operators/recurrent_op.cc @@ -25,7 +25,7 @@ constexpr char kOutputs[] = "outputs"; constexpr char kStepScopes[] = "step_scopes"; constexpr char kExStates[] = "ex_states"; constexpr char kStates[] = "states"; -constexpr char kStepBlock[] = "step_block"; +constexpr char kStepBlock[] = "sub_block"; constexpr char kReverse[] = "reverse"; constexpr char kIsTrain[] = "is_train"; #define GRAD_SUFFIX "@GRAD" diff --git a/paddle/operators/reduce_op.cc b/paddle/operators/reduce_op.cc index b754637bf29225615f129d7423d60518e053ca18..fedc2a5c37ff84ffdf8ebd2f19296db92e256e5b 100644 --- a/paddle/operators/reduce_op.cc +++ b/paddle/operators/reduce_op.cc @@ -37,18 +37,23 @@ class ReduceOp : public framework::OperatorWithKernel { PADDLE_ENFORCE_LT( dim, x_rank, "The dim should be in the range [-rank(input), rank(input))."); - bool keep_dim = ctx->Attrs().Get("keep_dim"); - auto dims_vector = vectorize(x_dims); - if (keep_dim || x_rank == 1) { - dims_vector[dim] = 1; + bool reduce_all = ctx->Attrs().Get("reduce_all"); + if (reduce_all) { + ctx->SetOutputDim("Out", {1}); } else { - dims_vector.erase(dims_vector.begin() + dim); - } - auto out_dims = framework::make_ddim(dims_vector); - ctx->SetOutputDim("Out", out_dims); - if (dim != 0) { - // Only pass LoD when not reducing on the first dim. - ctx->ShareLoD("X", /*->*/ "Out"); + bool keep_dim = ctx->Attrs().Get("keep_dim"); + auto dims_vector = vectorize(x_dims); + if (keep_dim || x_rank == 1) { + dims_vector[dim] = 1; + } else { + dims_vector.erase(dims_vector.begin() + dim); + } + auto out_dims = framework::make_ddim(dims_vector); + ctx->SetOutputDim("Out", out_dims); + if (dim != 0) { + // Only pass LoD when not reducing on the first dim. + ctx->ShareLoD("X", /*->*/ "Out"); + } } } }; @@ -95,11 +100,16 @@ class ReduceOpMaker : public framework::OpProtoAndCheckerMaker { "(bool, default false) " "If true, retain the reduced dimension with length 1.") .SetDefault(false); + AddAttr("reduce_all", + "(bool, default false) " + "If true, output a scalar reduced along all dimensions.") + .SetDefault(false); comment_ = R"DOC( {ReduceOp} Operator. This operator computes the {reduce} of input tensor along the given dimension. The result tensor has 1 fewer dimension than the input unless keep_dim is true. +If reduce_all is true, just reduce along all dimensions and output a scalar. )DOC"; AddComment(comment_); diff --git a/paddle/operators/reduce_op.h b/paddle/operators/reduce_op.h index 47ce910f2821467c701a7f5e22a8dbe5c8c95c92..7bd99cb1e6d532963ef648202f460f363baad9b5 100644 --- a/paddle/operators/reduce_op.h +++ b/paddle/operators/reduce_op.h @@ -26,10 +26,12 @@ using DDim = framework::DDim; template using EigenTensor = framework::EigenTensor; - template using EigenScalar = framework::EigenScalar; +template +using EigenVector = framework::EigenVector; struct SumFunctor { template @@ -95,26 +97,41 @@ template class ReduceKernel : public framework::OpKernel { public: void Compute(const framework::ExecutionContext& context) const override { - int rank = context.Input("X")->dims().size(); - switch (rank) { - case 1: - ReduceCompute<1>(context); - break; - case 2: - ReduceCompute<2>(context); - break; - case 3: - ReduceCompute<3>(context); - break; - case 4: - ReduceCompute<4>(context); - break; - case 5: - ReduceCompute<5>(context); - break; - case 6: - ReduceCompute<6>(context); - break; + bool reduce_all = context.Attr("reduce_all"); + if (reduce_all) { + // Flatten and reduce 1-D tensor + auto* input = context.Input("X"); + auto* output = context.Output("Out"); + output->mutable_data(context.GetPlace()); + auto x = EigenVector::Flatten(*input); + auto out = EigenScalar::From(*output); + auto& place = + *context.template device_context().eigen_device(); + auto reduce_dim = Eigen::array({{0}}); + Functor functor; + functor(place, x, out, reduce_dim); + } else { + int rank = context.Input("X")->dims().size(); + switch (rank) { + case 1: + ReduceCompute<1>(context); + break; + case 2: + ReduceCompute<2>(context); + break; + case 3: + ReduceCompute<3>(context); + break; + case 4: + ReduceCompute<4>(context); + break; + case 5: + ReduceCompute<5>(context); + break; + case 6: + ReduceCompute<6>(context); + break; + } } } @@ -157,26 +174,46 @@ template class ReduceGradKernel : public framework::OpKernel { public: void Compute(const framework::ExecutionContext& context) const override { - int rank = context.Input("X")->dims().size(); - switch (rank) { - case 1: - ReduceGradCompute<1>(context); - break; - case 2: - ReduceGradCompute<2>(context); - break; - case 3: - ReduceGradCompute<3>(context); - break; - case 4: - ReduceGradCompute<4>(context); - break; - case 5: - ReduceGradCompute<5>(context); - break; - case 6: - ReduceGradCompute<6>(context); - break; + bool reduce_all = context.Attr("reduce_all"); + if (reduce_all) { + auto* input0 = context.Input("X"); + auto* input1 = context.Input("Out"); + auto* input2 = context.Input(framework::GradVarName("Out")); + auto* output = context.Output(framework::GradVarName("X")); + output->mutable_data(context.GetPlace()); + auto x = EigenVector::Flatten(*input0); + auto x_reduce = EigenVector::From(*input1); + auto x_reduce_grad = EigenVector::From(*input2); + auto x_grad = EigenVector::Flatten(*output); + auto& place = + *context.template device_context().eigen_device(); + auto broadcast_dim = + Eigen::array({{static_cast(input0->numel())}}); + Functor functor; + functor(place, x, x_reduce, x_grad, x_reduce_grad, broadcast_dim, + broadcast_dim[0]); + } else { + int rank = context.Input("X")->dims().size(); + switch (rank) { + case 1: + ReduceGradCompute<1>(context); + break; + case 2: + ReduceGradCompute<2>(context); + break; + case 3: + ReduceGradCompute<3>(context); + break; + case 4: + ReduceGradCompute<4>(context); + break; + case 5: + ReduceGradCompute<5>(context); + break; + case 6: + ReduceGradCompute<6>(context); + break; + } } } diff --git a/paddle/operators/reshape_op.cc b/paddle/operators/reshape_op.cc index 7fd33bf662a1d0b7b6fa4e772bdadbf34b2f4fdd..d82d828747c0c822195b699359b8e62d1cf7e92d 100644 --- a/paddle/operators/reshape_op.cc +++ b/paddle/operators/reshape_op.cc @@ -34,21 +34,33 @@ class ReshapeOp : public framework::OperatorWithKernel { auto shape = ctx->Attrs().Get>("shape"); PADDLE_ENFORCE(shape.size() > 0, "Attr(shape) shouldn't be empty."); auto x_dims = ctx->GetInputDim("X"); - // TODO(qiao) change batch_size - for (size_t i = 1; i < shape.size(); ++i) { - PADDLE_ENFORCE(shape[i] > 0, - "Each dimension of Attr(shape) " - "must be positive except the first one."); - } - if (shape[0] < 0) { - shape[0] = x_dims[0]; + + std::vector neg_dims_idx; + // set some dimension to -1 if it is unknown + const int unknown_size = -1; + for (size_t i = 0; i < shape.size(); ++i) { + PADDLE_ENFORCE(shape[i] > 0 || shape[i] == unknown_size, + "Each dimension of Attr(shape) must be positive or %d.", + unknown_size); + if (shape[i] == unknown_size) { + neg_dims_idx.push_back(i); + PADDLE_ENFORCE(neg_dims_idx.size() <= 1, + "Only one dimension of Attr(shape) can be unknown."); + } } - // capacity check + int64_t capacity = std::accumulate(shape.begin(), shape.end(), 1, std::multiplies()); int64_t in_size = framework::product(x_dims); - PADDLE_ENFORCE_EQ(capacity, in_size, - "The size of Input(X) mismatches with Attr(shape)."); + if (neg_dims_idx.size() == 1) { + // dim infer + shape[neg_dims_idx[0]] = in_size / (-capacity); + // recalculate capacity + capacity = shape[neg_dims_idx[0]] * (-capacity); + } + // capacity check + PADDLE_ENFORCE(capacity == in_size, + "The size of Input(X) mismatches with Attr(shape)."); // resize output std::vector shape_int64(shape.size(), 0); std::transform(shape.begin(), shape.end(), shape_int64.begin(), @@ -88,6 +100,9 @@ the tensor X into a 2-D tensor: [[1, 2, 3, 4]] +One dimension in the target shape can be set -1, representing that its +size is unknown. In this case, the real dimension will be infered from +the original shape of Input(X) and other dimensions in the target shape. )DOC"); } }; diff --git a/paddle/operators/seq_expand_op.cc b/paddle/operators/sequence_expand_op.cc similarity index 82% rename from paddle/operators/seq_expand_op.cc rename to paddle/operators/sequence_expand_op.cc index ede9754697429a4d24c51cf494b0ea8f4e408b44..770161b593e232f2f2cf4a2ccb952391557b9a3d 100644 --- a/paddle/operators/seq_expand_op.cc +++ b/paddle/operators/sequence_expand_op.cc @@ -12,14 +12,14 @@ See the License for the specific language governing permissions and limitations under the License. */ -#include "paddle/operators/seq_expand_op.h" +#include "paddle/operators/sequence_expand_op.h" namespace paddle { namespace operators { using framework::Tensor; -class SeqExpandOp : public framework::OperatorWithKernel { +class SequenceExpandOp : public framework::OperatorWithKernel { public: using framework::OperatorWithKernel::OperatorWithKernel; @@ -35,25 +35,25 @@ class SeqExpandOp : public framework::OperatorWithKernel { } }; -class SeqExpandOpMaker : public framework::OpProtoAndCheckerMaker { +class SequenceExpandOpMaker : public framework::OpProtoAndCheckerMaker { public: - SeqExpandOpMaker(framework::OpProto* proto, - framework::OpAttrChecker* op_checker) + SequenceExpandOpMaker(framework::OpProto* proto, + framework::OpAttrChecker* op_checker) : OpProtoAndCheckerMaker(proto, op_checker) { AddInput("X", "(Tensor or LoDTensor) The input(X) of this operator can be a " "LoDTensor or a base Tensor."); AddInput("Y", - "(LoDTensor)The reference input(Y) of seq_expand op." + "(LoDTensor)The reference input(Y) of sequence_expand op." "It must be a LoDTensor with k-level(k>0)." "The input(X) will be expanded according to LOD of input(Y)." "The element numbers of last level in input(Y) " "must be equal to dims[0] of input(X)."); AddOutput("Out", - "(LodTensor)The output of seq_expand op." + "(LodTensor)The output of sequence_expand op." "The lod of output will be as same as input(Y)'s lod."); AddComment(R"DOC( -Seq Expand Operator. +Sequence Expand Operator. This operator expands input(X) according to LOD of input(Y). Following are cases to better explain how this works: @@ -124,7 +124,7 @@ then we get 2-level LoDTensor } }; -class SeqExpandOpGrad : public framework::OperatorWithKernel { +class SequenceExpandOpGrad : public framework::OperatorWithKernel { public: using framework::OperatorWithKernel::OperatorWithKernel; @@ -146,11 +146,11 @@ class SeqExpandOpGrad : public framework::OperatorWithKernel { } // namespace paddle namespace ops = paddle::operators; -REGISTER_OP(seq_expand, ops::SeqExpandOp, ops::SeqExpandOpMaker, - seq_expand_grad, ops::SeqExpandOpGrad); +REGISTER_OP(sequence_expand, ops::SequenceExpandOp, ops::SequenceExpandOpMaker, + sequence_expand_grad, ops::SequenceExpandOpGrad); REGISTER_OP_CPU_KERNEL( - seq_expand, - ops::SeqExpandKernel); + sequence_expand, + ops::SequenceExpandKernel); REGISTER_OP_CPU_KERNEL( - seq_expand_grad, - ops::SeqExpandGradKernel); + sequence_expand_grad, + ops::SequenceExpandGradKernel); diff --git a/paddle/operators/seq_expand_op.cu b/paddle/operators/sequence_expand_op.cu similarity index 74% rename from paddle/operators/seq_expand_op.cu rename to paddle/operators/sequence_expand_op.cu index 8e67ce9ccb29497a957508a9ecdc6b810a7de543..f79c84dff8bf4f0e97f89d5c8bb23655abd73d46 100644 --- a/paddle/operators/seq_expand_op.cu +++ b/paddle/operators/sequence_expand_op.cu @@ -13,12 +13,12 @@ limitations under the License. */ #define EIGEN_USE_GPU -#include "paddle/operators/seq_expand_op.h" +#include "paddle/operators/sequence_expand_op.h" namespace ops = paddle::operators; REGISTER_OP_CUDA_KERNEL( - seq_expand, - ops::SeqExpandKernel); + sequence_expand, + ops::SequenceExpandKernel); REGISTER_OP_CUDA_KERNEL( - seq_expand_grad, - ops::SeqExpandGradKernel); + sequence_expand_grad, + ops::SequenceExpandGradKernel); diff --git a/paddle/operators/seq_expand_op.h b/paddle/operators/sequence_expand_op.h similarity index 96% rename from paddle/operators/seq_expand_op.h rename to paddle/operators/sequence_expand_op.h index fbee0db454f9701e3f58a41008efd24e728d0600..411b819c6563ec95b87881082caef5f5eb471d3b 100644 --- a/paddle/operators/seq_expand_op.h +++ b/paddle/operators/sequence_expand_op.h @@ -24,7 +24,7 @@ namespace operators { using LoDTensor = framework::LoDTensor; template -class SeqExpandKernel : public framework::OpKernel { +class SequenceExpandKernel : public framework::OpKernel { public: void Compute(const framework::ExecutionContext& context) const override { auto* x = context.Input("X"); @@ -71,7 +71,7 @@ class SeqExpandKernel : public framework::OpKernel { * * */ template -class SeqExpandGradKernel : public framework::OpKernel { +class SequenceExpandGradKernel : public framework::OpKernel { public: void Compute(const framework::ExecutionContext& context) const override { auto* d_out = context.Input(framework::GradVarName("Out")); diff --git a/paddle/operators/spp_op.cc b/paddle/operators/spp_op.cc new file mode 100644 index 0000000000000000000000000000000000000000..b1807b62616b80ea8a9e48409e0760c1c7b36a38 --- /dev/null +++ b/paddle/operators/spp_op.cc @@ -0,0 +1,99 @@ +/* Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserve. + +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +Indicesou may obtain a copy of the License at + + http://www.apache.org/licenses/LICENSE-2.0 + +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. */ + +#include "paddle/operators/spp_op.h" +namespace paddle { +namespace operators { + +class SppOpMaker : public framework::OpProtoAndCheckerMaker { + public: + SppOpMaker(framework::OpProto* proto, framework::OpAttrChecker* op_checker) + : OpProtoAndCheckerMaker(proto, op_checker) { + AddInput( + "X", + "(Tensor) The input tensor of spp operator. " + "The format of input tensor is NCHW. Where N is batch size, C is the " + "number of channels, H and W is the height and width of feature."); + AddOutput("Out", + "(Tensor) The output tensor of spp operator." + "N * M." + "M = C * H * W"); + AddAttr("pyramid_height", "(int), multi level pooling"); + AddAttr( + "pooling_type", + "(string), pooling type, can be \"max\" for max-pooling " + "and \"avg\" for average-pooling.") + .InEnum({"max", "avg"}); + AddComment(R"DOC( + "With spatial pyramid pooling, the input image can + be of any sizes. This not only allows arbitrary aspect + ratios, but also allows arbitrary scales. We can resize + the input image to any scale (e.g., min(w, h)=180, 224, + ...) and apply the same deep network. When the + input image is at different scales, the network (with + the same filter sizes) will extract features at different + scales. The scales play important roles in traditional + methods. + Input shape: $(N, C_{in}, H_{in}, W_{in})$ + Output shape: $(H_{out}, W_{out})$ + Where + $$ + H_{out} = N \\ + W_{out} = (((4^pyramid_height) - 1) / (4 - 1))$ * C_{in} + $$ + paper https://arxiv.org/pdf/1406.4729v4.pdf + )DOC"); + } +}; + +class SppOp : public framework::OperatorWithKernel { + public: + using framework::OperatorWithKernel::OperatorWithKernel; + void InferShape(framework::InferShapeContext* ctx) const override { + PADDLE_ENFORCE(ctx->HasInput("X"), + "Input(X) of SppOp" + "should not be null."); + PADDLE_ENFORCE(ctx->HasOutput("Out"), + "Output(Out) of SppOp should not be null."); + auto in_x_dims = ctx->GetInputDim("X"); + int pyramid_height = ctx->Attrs().Get("pyramid_height"); + PADDLE_ENFORCE(in_x_dims.size() == 4, + "Spping intput must be of 4-dimensional."); + int outlen = ((std::pow(4, pyramid_height) - 1) / (4 - 1)) * in_x_dims[1]; + std::vector output_shape({in_x_dims[0], outlen}); + ctx->SetOutputDim("Out", framework::make_ddim(output_shape)); + } +}; + +class SppOpGrad : public framework::OperatorWithKernel { + public: + using framework::OperatorWithKernel::OperatorWithKernel; + void InferShape(framework::InferShapeContext* ctx) const override { + PADDLE_ENFORCE(ctx->HasInput("X"), "Input(X) must not be null."); + PADDLE_ENFORCE(ctx->HasOutput(framework::GradVarName("X")), + "Input(X@GRAD) should not be null."); + ctx->SetOutputDim(framework::GradVarName("X"), ctx->GetInputDim("X")); + } +}; +} // namespace operators +} // namespace paddle + +namespace ops = paddle::operators; +REGISTER_OP(spp, ops::SppOp, ops::SppOpMaker, spp_grad, ops::SppOpGrad); +REGISTER_OP_CPU_KERNEL( + spp, ops::SppKernel, + ops::SppKernel); +REGISTER_OP_CPU_KERNEL( + spp_grad, ops::SppGradKernel, + ops::SppGradKernel); diff --git a/paddle/operators/spp_op.cu.cc b/paddle/operators/spp_op.cu.cc new file mode 100644 index 0000000000000000000000000000000000000000..761e4d6c4a9639898ba548d56bed3c8817720c1b --- /dev/null +++ b/paddle/operators/spp_op.cu.cc @@ -0,0 +1,23 @@ +/* Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserve. + +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +Indicesou may obtain a copy of the License at + + http://www.apache.org/licenses/LICENSE-2.0 + +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. */ + +#include "paddle/operators/spp_op.h" + +namespace ops = paddle::operators; +REGISTER_OP_CUDA_KERNEL( + spp, ops::SppKernel, + ops::SppKernel); +REGISTER_OP_CUDA_KERNEL( + spp_grad, ops::SppGradKernel, + ops::SppGradKernel); diff --git a/paddle/operators/spp_op.h b/paddle/operators/spp_op.h new file mode 100644 index 0000000000000000000000000000000000000000..f35b305d02c73bcae6e72b8afa5ce55148ea98b8 --- /dev/null +++ b/paddle/operators/spp_op.h @@ -0,0 +1,161 @@ +/* Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserve. + +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +Indicesou may obtain a copy of the License at + + http://www.apache.org/licenses/LICENSE-2.0 + +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. */ + +#pragma once +#include "paddle/framework/op_registry.h" +#include "paddle/operators/math/math_function.h" +#include "paddle/operators/math/pooling.h" +#include "paddle/operators/strided_memcpy.h" + +namespace paddle { +namespace operators { +template +class SppKernel : public framework::OpKernel { + public: + void Compute(const framework::ExecutionContext& context) const override { + const framework::Tensor* in_x = context.Input("X"); + auto* out = context.Output("Out"); + int pyramid_height = context.template Attr("pyramid_height"); + std::string pooling_type = + context.template Attr("pooling_type"); + out->mutable_data(context.GetPlace()); + auto out_stride = framework::stride(out->dims()); + int input_h = in_x->dims()[2]; + int input_w = in_x->dims()[3]; + size_t output_offset = 0; + for (int p = 0; p < pyramid_height; ++p) { + int bins = std::pow(2, p); + int kernel_size_h = std::ceil(input_h / static_cast(bins)); + int kernel_size_w = std::ceil(input_w / static_cast(bins)); + int padding_h = (kernel_size_h * bins - input_h + 1) / 2; + int padding_w = (kernel_size_w * bins - input_w + 1) / 2; + std::vector kernel_size({kernel_size_h, kernel_size_w}); + std::vector strides({kernel_size_h, kernel_size_w}); + std::vector paddings({padding_h, padding_w}); + // pooling output shape + framework::Tensor out_level; + std::vector output_shape_vec( + {in_x->dims()[0], in_x->dims()[1], bins, bins}); + framework::DDim output_shape(framework::make_ddim(output_shape_vec)); + out_level.mutable_data(output_shape, context.GetPlace()); + // pooling + if (pooling_type == "max") { + math::Pool2dFunctor, T> pool_forward; + math::MaxPool max_process; + pool_forward(context.template device_context(), *in_x, + kernel_size, strides, paddings, max_process, &out_level); + } else if (pooling_type == "avg") { + math::Pool2dFunctor, T> pool_forward; + math::AvgPool avg_process; + pool_forward(context.template device_context(), *in_x, + kernel_size, strides, paddings, avg_process, &out_level); + } + // flatten pooling output shape + int output_flatten_w = in_x->dims()[1] * bins * bins; + std::vector output_flatten_shape_vec( + {in_x->dims()[0], output_flatten_w}); + framework::DDim output_flatten_shape( + framework::make_ddim(output_flatten_shape_vec)); + out_level.Resize(output_flatten_shape); + // concat + auto out_level_stride = framework::stride(out_level.dims()); + StridedMemcpy(context.template device_context(), + out_level.data(), out_level_stride, out_level.dims(), + out_stride, out->data() + output_offset); + output_offset += out_level.dims()[1] * out_level_stride[1]; + } + } +}; +template +class SppGradKernel : public framework::OpKernel { + public: + void Compute(const framework::ExecutionContext& context) const override { + const framework::Tensor* in_x = context.Input("X"); + const framework::Tensor* out = context.Input("Out"); + const framework::Tensor* out_grad = + context.Input(framework::GradVarName("Out")); + framework::Tensor* in_x_grad = + context.Output(framework::GradVarName("X")); + int pyramid_height = context.template Attr("pyramid_height"); + std::string pooling_type = + context.template Attr("pooling_type"); + auto& device_ctx = context.template device_context(); + math::SetConstant zero; + in_x_grad->mutable_data(context.GetPlace()); + zero(device_ctx, in_x_grad, static_cast(0)); + auto out_stride = framework::stride(out->dims()); + int input_h = in_x->dims()[2]; + int input_w = in_x->dims()[3]; + size_t out_offset = 0; + for (int p = 0; p < pyramid_height; ++p) { + int bins = std::pow(2, p); + int kernel_size_h = std::ceil(input_h / static_cast(bins)); + int kernel_size_w = std::ceil(input_w / static_cast(bins)); + int padding_h = (kernel_size_h * bins - input_h + 1) / 2; + int padding_w = (kernel_size_w * bins - input_w + 1) / 2; + std::vector kernel_size({kernel_size_h, kernel_size_w}); + std::vector strides({kernel_size_h, kernel_size_w}); + std::vector paddings({padding_h, padding_w}); + // split out and outgrad ... to flatten + framework::Tensor out_level; + framework::Tensor outgrad_level; + int out_flatten_w = in_x->dims()[1] * bins * bins; + std::vector out_flatten_shape_vec( + {in_x->dims()[0], out_flatten_w}); + framework::DDim out_flatten_shape( + framework::make_ddim(out_flatten_shape_vec)); + out_level.mutable_data(out_flatten_shape, context.GetPlace()); + outgrad_level.mutable_data(out_flatten_shape, context.GetPlace()); + auto flatten_stride = framework::stride(out_level.dims()); + // memcpy + StridedMemcpy(context.template device_context(), + out->data() + out_offset, out_stride, + out_level.dims(), flatten_stride, out_level.data()); + + StridedMemcpy(context.template device_context(), + out_grad->data() + out_offset, out_stride, + outgrad_level.dims(), flatten_stride, + outgrad_level.data()); + out_offset += out_level.dims()[1] * out_stride[1]; + // flatten backward to nchw + + std::vector out_shape_vec({in_x->dims()[0], in_x->dims()[1]}); + out_shape_vec.push_back( + (input_h - kernel_size_h + 2 * padding_h) / kernel_size_h + 1); + out_shape_vec.push_back( + (input_w - kernel_size_w + 2 * padding_w) / kernel_size_w + 1); + framework::DDim out_shape(framework::make_ddim(out_shape_vec)); + out_level.ShareDataWith(out_level); + out_level.Resize(out_shape); + outgrad_level.ShareDataWith(outgrad_level); + outgrad_level.Resize(out_shape); + // pooling backward + if (pooling_type == "max") { + math::MaxPool2dGradFunctor pool2d_backward; + pool2d_backward(context.template device_context(), *in_x, + *&out_level, *&outgrad_level, kernel_size, strides, + paddings, in_x_grad); + } else if (pooling_type == "avg") { + math::Pool2dGradFunctor, T> + pool_backward; + math::AvgPoolGrad avg_process; + pool_backward(context.template device_context(), *in_x, + *&out_level, *&outgrad_level, kernel_size, strides, + paddings, avg_process, in_x_grad); + } + } + } +}; +} // namespace operators +} // namespace paddle diff --git a/paddle/operators/while_op.cc b/paddle/operators/while_op.cc index b8e44bcc5a99380fdf08cc2819b20045695eaf87..9a092a570ff1f3f529413cd44dff55147adbaadc 100644 --- a/paddle/operators/while_op.cc +++ b/paddle/operators/while_op.cc @@ -25,7 +25,7 @@ namespace operators { using StepScopeVar = std::vector; using LoDTensor = framework::LoDTensor; -constexpr char kStepBlock[] = "step_block"; +constexpr char kStepBlock[] = "sub_block"; constexpr char kCondition[] = "Condition"; constexpr char kStepScopes[] = "StepScopes"; constexpr char kParameters[] = "X"; diff --git a/paddle/platform/device_context.cc b/paddle/platform/device_context.cc index 2c7f96421621b9a34d1ec96c13d9c354a0d4012c..8cdc5f43403b0c54d3f1f01a3e97405fd5b2f434 100644 --- a/paddle/platform/device_context.cc +++ b/paddle/platform/device_context.cc @@ -125,6 +125,22 @@ cudnnHandle_t CUDADeviceContext::cudnn_handle() const { return cudnn_handle_; } cudaStream_t CUDADeviceContext::stream() const { return stream_; } +CUDNNDeviceContext::CUDNNDeviceContext(CUDNNPlace place) + : CUDADeviceContext(place), place_(place) { + PADDLE_ENFORCE(dynload::cudnnCreate(&cudnn_handle_)); + PADDLE_ENFORCE(dynload::cudnnSetStream(cudnn_handle_, stream())); +} + +CUDNNDeviceContext::~CUDNNDeviceContext() { + SetDeviceId(place_.device); + Wait(); + PADDLE_ENFORCE(dynload::cudnnDestroy(cudnn_handle_)); +} + +Place CUDNNDeviceContext::GetPlace() const { return CUDNNPlace(); } + +cudnnHandle_t CUDNNDeviceContext::cudnn_handle() const { return cudnn_handle_; } + #endif } // namespace platform diff --git a/paddle/platform/device_context.h b/paddle/platform/device_context.h index 596d9d0bba420a47fc10cc9dd96a755daa35dbac..56813a1d5b3c2a7f4ff7b4eddc6fa47ed861700c 100644 --- a/paddle/platform/device_context.h +++ b/paddle/platform/device_context.h @@ -86,6 +86,22 @@ class CUDADeviceContext : public DeviceContext { cublasHandle_t cublas_handle_; }; +class CUDNNDeviceContext : public CUDADeviceContext { + public: + explicit CUDNNDeviceContext(CUDNNPlace place); + virtual ~CUDNNDeviceContext(); + + /*! \brief Return place in the device context. */ + Place GetPlace() const final; + + /*! \brief Return cudnn handle in the device context. */ + cudnnHandle_t cudnn_handle() const; + + private: + cudnnHandle_t cudnn_handle_; + CUDNNPlace place_; +}; + #endif } // namespace platform diff --git a/paddle/platform/device_context_test.cc b/paddle/platform/device_context_test.cc index 4893cd92f6a74f7992c279ebd51232049f29e853..109c13a8812dffac10d202cbc9d85c4e601bf197 100644 --- a/paddle/platform/device_context_test.cc +++ b/paddle/platform/device_context_test.cc @@ -46,3 +46,19 @@ TEST(Device, CUDADeviceContext) { delete device_context; } } + +TEST(Device, CUDNNDeviceContext) { + using paddle::platform::CUDNNDeviceContext; + using paddle::platform::CUDNNPlace; + if (paddle::platform::dynload::HasCUDNN()) { + int count = paddle::platform::GetCUDADeviceCount(); + for (int i = 0; i < count; ++i) { + CUDNNDeviceContext* device_context = + new CUDNNDeviceContext(CUDNNPlace(i)); + cudnnHandle_t cudnn_handle = device_context->cudnn_handle(); + ASSERT_NE(nullptr, cudnn_handle); + ASSERT_NE(nullptr, device_context->stream()); + delete device_context; + } + } +} diff --git a/paddle/platform/dynload/nccl.cc b/paddle/platform/dynload/nccl.cc index 8f92b8d94d56047b7d3fb43b15e3c06575c8d57b..91168f37effff3f8b864b6bb2ede070cb0a976fa 100644 --- a/paddle/platform/dynload/nccl.cc +++ b/paddle/platform/dynload/nccl.cc @@ -25,6 +25,11 @@ void *nccl_dso_handle; NCCL_RAND_ROUTINE_EACH(DEFINE_WRAP); +void LoadNCCLDSO() { + platform::call_once(nccl_dso_flag, + [] { GetNCCLDsoHandle(&nccl_dso_handle); }); +} + } // namespace dynload } // namespace platform } // namespace paddle diff --git a/paddle/platform/dynload/nccl.h b/paddle/platform/dynload/nccl.h index 981b2ab258a34ce92f02ee12b5957f88ba61d1c0..11007c1031c6d224c475e9ef4f11e7797decd78e 100644 --- a/paddle/platform/dynload/nccl.h +++ b/paddle/platform/dynload/nccl.h @@ -28,18 +28,18 @@ extern std::once_flag nccl_dso_flag; extern void* nccl_dso_handle; #ifdef PADDLE_USE_DSO -#define DECLARE_DYNAMIC_LOAD_NCCL_WRAP(__name) \ - struct DynLoad__##__name { \ - template \ - auto operator()(Args... args) -> decltype(__name(args...)) { \ - using nccl_func = decltype(__name(args...)) (*)(Args...); \ - platform::call_once(nccl_dso_flag, \ - paddle::platform::dynload::GetNCCLDsoHandle, \ - &nccl_dso_handle); \ - void* p_##__name = dlsym(nccl_dso_handle, #__name); \ - return reinterpret_cast(p_##__name)(args...); \ - } \ - }; \ +extern void LoadNCCLDSO(); + +#define DECLARE_DYNAMIC_LOAD_NCCL_WRAP(__name) \ + struct DynLoad__##__name { \ + template \ + auto operator()(Args... args) -> decltype(__name(args...)) { \ + using nccl_func = decltype(__name(args...)) (*)(Args...); \ + paddle::platform::dynload::LoadNCCLDSO(); \ + void* p_##__name = dlsym(nccl_dso_handle, #__name); \ + return reinterpret_cast(p_##__name)(args...); \ + } \ + }; \ extern DynLoad__##__name __name #else #define DECLARE_DYNAMIC_LOAD_NCCL_WRAP(__name) \ diff --git a/paddle/platform/gpu_info.cc b/paddle/platform/gpu_info.cc index 4fa2eaed31c6e9368459c2da6f8b0667b453d58c..541eca5f39c2e6a4b464aec79fd8a920ab4c7732 100644 --- a/paddle/platform/gpu_info.cc +++ b/paddle/platform/gpu_info.cc @@ -73,19 +73,20 @@ size_t GpuMaxChunkSize() { size_t available = 0; GpuMemoryUsage(available, total); - - // Reserving the rest memory for page tables, etc. - size_t reserving = 0.05 * total; - + VLOG(10) << "GPU Usage " << available / 1024 / 1024 << "M/" + << total / 1024 / 1024 << "M"; + size_t reserving = static_cast(0.05 * total); // If available less than minimum chunk size, no usable memory exists. available = - std::max(std::max(available, GpuMinChunkSize()) - GpuMinChunkSize(), - reserving) - - reserving; + std::min(std::max(available, GpuMinChunkSize()) - GpuMinChunkSize(), + total - reserving); + + // Reserving the rest memory for page tables, etc. - size_t allocating = FLAGS_fraction_of_gpu_memory_to_use * total; + size_t allocating = static_cast(FLAGS_fraction_of_gpu_memory_to_use * + (total - reserving)); - PADDLE_ENFORCE_LT(allocating, available); + PADDLE_ENFORCE_LE(allocating, available); return allocating; } diff --git a/paddle/platform/nccl_test.cu b/paddle/platform/nccl_test.cu index c99dae68bef67c58d3efea42fef45e84bb3d9255..94ab360a1967d22e73bc6aefc0301487537c97f7 100644 --- a/paddle/platform/nccl_test.cu +++ b/paddle/platform/nccl_test.cu @@ -31,7 +31,7 @@ namespace platform { TEST(NCCL, init) { std::vector comms; comms.resize(dev_count); - PADDLE_ENFORCE(dynload::ncclCommInitAll(comms.data(), dev_count, nullptr)); + dynload::ncclCommInitAll(comms.data(), dev_count, nullptr); for (int i = 0; i < dev_count; ++i) { dynload::ncclCommDestroy(comms[i]); } @@ -62,7 +62,7 @@ TEST(NCCL, all_reduce) { std::vector comms; comms.resize(dev_count); VLOG(1) << "Initializing ncclComm"; - PADDLE_ENFORCE(dynload::ncclCommInitAll(comms.data(), dev_count, nullptr)); + dynload::ncclCommInitAll(comms.data(), dev_count, nullptr); VLOG(1) << "ncclComm initialized"; VLOG(1) << "Creating thread data"; std::vector>> data; diff --git a/paddle/platform/place.cc b/paddle/platform/place.cc index 856e54df89c1c18ade040957188a2fbda0901473..25fe8d21b49b07a6afe2938245906dc1bdd90398 100644 --- a/paddle/platform/place.cc +++ b/paddle/platform/place.cc @@ -23,6 +23,7 @@ class PlacePrinter : public boost::static_visitor<> { public: explicit PlacePrinter(std::ostream &os) : os_(os) {} void operator()(const CPUPlace &) { os_ << "CPUPlace"; } + void operator()(const MKLDNNPlace &) { os_ << "MKLDNNPlace"; } void operator()(const GPUPlace &p) { os_ << "GPUPlace(" << p.device << ")"; } private: @@ -38,12 +39,17 @@ const Place &get_place() { return the_default_place; } const GPUPlace default_gpu() { return GPUPlace(0); } const CPUPlace default_cpu() { return CPUPlace(); } +const MKLDNNPlace default_mkldnn() { return MKLDNNPlace(); } bool is_gpu_place(const Place &p) { return boost::apply_visitor(IsGPUPlace(), p); } bool is_cpu_place(const Place &p) { - return !boost::apply_visitor(IsGPUPlace(), p); + return !is_gpu_place(p) && !is_mkldnn_place(p); +} + +bool is_mkldnn_place(const Place &p) { + return boost::apply_visitor(IsMKLDNNPlace(), p); } bool places_are_same_class(const Place &p1, const Place &p2) { diff --git a/paddle/platform/place.h b/paddle/platform/place.h index 5370360a7de26e409a1545182a12d3df1f37658b..ca98920d414bc87ce243995a42e5672d0e61e108 100644 --- a/paddle/platform/place.h +++ b/paddle/platform/place.h @@ -31,6 +31,14 @@ struct CPUPlace { inline bool operator!=(const CPUPlace &) const { return false; } }; +struct MKLDNNPlace { + MKLDNNPlace() {} + + // needed for variant equality comparison + inline bool operator==(const MKLDNNPlace &) const { return true; } + inline bool operator!=(const MKLDNNPlace &) const { return false; } +}; + struct GPUPlace { GPUPlace() : GPUPlace(0) {} explicit GPUPlace(int d) : device(d) {} @@ -43,16 +51,28 @@ struct GPUPlace { int device; }; +struct CUDNNPlace : public GPUPlace { + CUDNNPlace() : GPUPlace() {} + explicit CUDNNPlace(int d) : GPUPlace(d) {} +}; + struct IsGPUPlace : public boost::static_visitor { bool operator()(const CPUPlace &) const { return false; } + bool operator()(const MKLDNNPlace &) const { return false; } bool operator()(const GPUPlace &gpu) const { return true; } }; +struct IsMKLDNNPlace : public boost::static_visitor { + bool operator()(const MKLDNNPlace &) const { return true; } + bool operator()(const CPUPlace &) const { return false; } + bool operator()(const GPUPlace &) const { return false; } +}; + // Define the max number of Place in bit length. i.e., the max number of places // should be less equal than 2^(NUM_PLACE_TYPE_LIMIT_IN_BIT) #define NUM_PLACE_TYPE_LIMIT_IN_BIT 4 -typedef boost::variant Place; +typedef boost::variant Place; // static check number of place types is less equal than // 2^(NUM_PLACE_TYPE_LIMIT_IN_BIT) @@ -65,9 +85,11 @@ const Place &get_place(); const GPUPlace default_gpu(); const CPUPlace default_cpu(); +const MKLDNNPlace default_mkldnn(); bool is_gpu_place(const Place &); bool is_cpu_place(const Place &); +bool is_mkldnn_place(const Place &); bool places_are_same_class(const Place &, const Place &); std::ostream &operator<<(std::ostream &, const Place &); diff --git a/paddle/platform/place_test.cc b/paddle/platform/place_test.cc index 33e2e5a439ce6801c02daba4bcbd462a74d7a614..c536b59ed8f71bd078bd09c5bd5afeab74c71b28 100644 --- a/paddle/platform/place_test.cc +++ b/paddle/platform/place_test.cc @@ -5,25 +5,37 @@ TEST(Place, Equality) { paddle::platform::CPUPlace cpu; paddle::platform::GPUPlace g0(0), g1(1), gg0(0); + paddle::platform::CUDNNPlace d0(0), d1(1), dd0(0); EXPECT_EQ(cpu, cpu); EXPECT_EQ(g0, g0); EXPECT_EQ(g1, g1); EXPECT_EQ(g0, gg0); + EXPECT_EQ(d0, dd0); EXPECT_NE(g0, g1); + EXPECT_NE(d0, d1); EXPECT_TRUE(paddle::platform::places_are_same_class(g0, gg0)); EXPECT_FALSE(paddle::platform::places_are_same_class(g0, cpu)); + + EXPECT_TRUE(paddle::platform::is_gpu_place(d0)); + EXPECT_FALSE(paddle::platform::places_are_same_class(g0, d0)); } TEST(Place, Default) { EXPECT_TRUE(paddle::platform::is_gpu_place(paddle::platform::get_place())); EXPECT_TRUE(paddle::platform::is_gpu_place(paddle::platform::default_gpu())); EXPECT_TRUE(paddle::platform::is_cpu_place(paddle::platform::default_cpu())); + EXPECT_TRUE( + paddle::platform::is_mkldnn_place(paddle::platform::default_mkldnn())); paddle::platform::set_place(paddle::platform::CPUPlace()); EXPECT_TRUE(paddle::platform::is_cpu_place(paddle::platform::get_place())); + + paddle::platform::set_place(paddle::platform::MKLDNNPlace()); + EXPECT_FALSE(paddle::platform::is_cpu_place(paddle::platform::get_place())); + EXPECT_TRUE(paddle::platform::is_mkldnn_place(paddle::platform::get_place())); } TEST(Place, Print) { diff --git a/paddle/platform/variant.h b/paddle/platform/variant.h index 619897ca19eb2e6f4dbfd9160edf8c4bc58c89a9..284b4c42ac068b23737c12d3f147bbceb135dacc 100644 --- a/paddle/platform/variant.h +++ b/paddle/platform/variant.h @@ -14,6 +14,19 @@ #pragma once +#ifdef __CUDACC__ +#ifdef __CUDACC_VER_MAJOR__ +// CUDA 9 define `__CUDACC_VER__` as a warning message, manually define +// __CUDACC_VER__ instead. +#undef __CUDACC_VER__ + +#define __CUDACC_VER__ \ + (__CUDACC_VER_MAJOR__ * 10000 + __CUDACC_VER_MINOR__ * 100 + \ + __CUDACC_VER_BUILD__) +#endif + +#endif + #include #ifdef PADDLE_WITH_CUDA diff --git a/paddle/pybind/CMakeLists.txt b/paddle/pybind/CMakeLists.txt index fd55f410d3f0fee418e7efffa927e46c38d23a07..1fb69de90d2fb5386dffdd95825c496a8fa559d3 100644 --- a/paddle/pybind/CMakeLists.txt +++ b/paddle/pybind/CMakeLists.txt @@ -1,7 +1,7 @@ if(WITH_PYTHON) cc_library(paddle_pybind SHARED SRCS pybind.cc exception.cc protobuf.cc - DEPS pybind python backward proto_desc paddle_memory executor prune + DEPS pybind python backward proto_desc paddle_memory executor prune init ${GLOB_OP_LIB}) endif(WITH_PYTHON) diff --git a/paddle/pybind/pybind.cc b/paddle/pybind/pybind.cc index c16d3e0cbe01f90a5aa9a5d7a523cd4e282e4771..4248db34c6345bd62e63628c7794b40d8a1adab6 100644 --- a/paddle/pybind/pybind.cc +++ b/paddle/pybind/pybind.cc @@ -16,11 +16,11 @@ limitations under the License. */ #include // for call_once #include -#include "gflags/gflags.h" #include "paddle/framework/backward.h" #include "paddle/framework/executor.h" #include "paddle/framework/feed_fetch_method.h" #include "paddle/framework/framework.pb.h" +#include "paddle/framework/init.h" #include "paddle/framework/lod_rank_table.h" #include "paddle/framework/lod_tensor.h" #include "paddle/framework/lod_tensor_array.h" @@ -51,24 +51,6 @@ static size_t UniqueIntegerGenerator(const std::string &prefix) { return generators[prefix].fetch_add(1); } -std::once_flag gflags_init_flag; - -// TODO(qijun) move init gflags to init.cc -void InitGflags(std::vector &argv) { - std::call_once(gflags_init_flag, [&]() { - int argc = argv.size(); - char **arr = new char *[argv.size()]; - std::string line; - for (size_t i = 0; i < argv.size(); i++) { - arr[i] = &argv[i][0]; - line += argv[i]; - line += ' '; - } - google::ParseCommandLineFlags(&argc, &arr, true); - VLOG(1) << "Init commandline: " << line; - }); -} - bool IsCompileGPU() { #ifndef PADDLE_WITH_CUDA return false; @@ -282,6 +264,23 @@ All parameter, weight, gradient are variables in Paddle. } return ret_values; }); + m.def("get_grad_op_descs", + [](const OpDescBind &op_desc, + const std::unordered_set &no_grad_set, + std::unordered_map &grad_to_var, + const std::vector &grad_sub_block) { + std::vector> grad_op_descs = + framework::OpInfoMap::Instance() + .Get(op_desc.Type()) + .GradOpMaker()(op_desc, no_grad_set, &grad_to_var, + grad_sub_block); + std::vector grad_op_desc_ptrs(grad_op_descs.size()); + std::transform( + grad_op_descs.begin(), grad_op_descs.end(), + grad_op_desc_ptrs.begin(), + [](std::unique_ptr &p) { return p.release(); }); + return grad_op_desc_ptrs; + }); m.def("prune", [](const ProgramDescBind &origin, const std::vector> &targets) { ProgramDescBind prog_with_targets(origin); @@ -421,7 +420,8 @@ All parameter, weight, gradient are variables in Paddle. .def("run", &Executor::Run); m.def("unique_integer", UniqueIntegerGenerator); - m.def("init_gflags", InitGflags); + m.def("init_gflags", framework::InitGflags); + m.def("init_devices", &framework::InitDevices); m.def("is_compile_gpu", IsCompileGPU); m.def("set_feed_variable", framework::SetFeedVariable); diff --git a/paddle/scripts/cluster_train_v2/openmpi/docker_cluster/Dockerfile b/paddle/scripts/cluster_train_v2/openmpi/docker_cluster/Dockerfile index 1a2d19e823541750830fcaa25f65b2f8e1ea2b49..c2f631bdf4ed52a5dfa3fbcf1157d0abbdeadb9b 100644 --- a/paddle/scripts/cluster_train_v2/openmpi/docker_cluster/Dockerfile +++ b/paddle/scripts/cluster_train_v2/openmpi/docker_cluster/Dockerfile @@ -1,7 +1,7 @@ # Build this image: docker build -t mpi . # -FROM paddledev/paddle:0.10.0rc3 +FROM paddlepaddle/paddle:0.10.0rc3 ENV DEBIAN_FRONTEND noninteractive diff --git a/paddle/scripts/tools/build_docs/build_docs.sh b/paddle/scripts/tools/build_docs/build_docs.sh index c6cbbc4eef94fb2e2fc3c1ce71734fbb23fc22d7..f9bc8bf63ae9afdfca1ff660bc83e62e71f03005 100755 --- a/paddle/scripts/tools/build_docs/build_docs.sh +++ b/paddle/scripts/tools/build_docs/build_docs.sh @@ -5,4 +5,4 @@ docker run --rm \ -e "WITH_AVX=ON" \ -e "WITH_DOC=ON" \ -e "WOBOQ=ON" \ - ${1:-"paddledev/paddle:dev"} + ${1:-"paddlepaddle/paddle:latest-dev"} diff --git a/python/.gitignore b/python/.gitignore index cc7d0ece4acaba2a3fa38a89110587fe8dffb992..1ba1d4c9b0301ed920f5303089e75dd3a8e4e3fa 100644 --- a/python/.gitignore +++ b/python/.gitignore @@ -2,6 +2,7 @@ build dist paddle.egg-info +paddlepaddle_gpu.egg-info .idea paddle/proto/*.py paddle/proto/*.pyc diff --git a/python/paddle/trainer_config_helpers/networks.py b/python/paddle/trainer_config_helpers/networks.py index 368fc1e832dc62189d0c5431ecca325a54da8cf8..b5cde7bac779ee1d54395b68941df2693e1ed0f5 100644 --- a/python/paddle/trainer_config_helpers/networks.py +++ b/python/paddle/trainer_config_helpers/networks.py @@ -1198,8 +1198,9 @@ def simple_gru2(input, :param gru_bias_attr: bias parameter attribute of gru layer, False means no bias, None means default bias. :type gru_bias_attr: ParameterAttribute|False|None - :param gru_layer_attr: Extra attribute of the gru layer. - :type gru_layer_attr: ExtraLayerAttribute + :param gru_param_attr: param parameter attribute of gru layer, + None means default param. + :type gru_param_attr: ParameterAttribute|None :return: the gru group. :rtype: LayerOutput """ diff --git a/python/paddle/v2/fluid/evaluator.py b/python/paddle/v2/fluid/evaluator.py index 137c5736226b689340748d5098ca51659d5acff8..e186ee96c387acf24471d4e26ce020c4ecac8d19 100644 --- a/python/paddle/v2/fluid/evaluator.py +++ b/python/paddle/v2/fluid/evaluator.py @@ -1,10 +1,10 @@ import numpy as np import layers -from framework import Program, unique_name, Variable +from framework import Program, unique_name, Variable, program_guard from layer_helper import LayerHelper -__all__ = ['Accuracy'] +__all__ = ['Accuracy', 'ChunkEvaluator'] def _clone_var_(block, var): @@ -49,15 +49,12 @@ class Evaluator(object): if reset_program is None: reset_program = Program() - for var in self.states: - assert isinstance(var, Variable) - g_var = _clone_var_(reset_program.current_block(), var) - layers.fill_constant( - shape=g_var.shape, - value=0.0, - dtype=g_var.dtype, - out=g_var, - main_program=reset_program) + with program_guard(main_program=reset_program): + for var in self.states: + assert isinstance(var, Variable) + g_var = _clone_var_(reset_program.current_block(), var) + layers.fill_constant( + shape=g_var.shape, value=0.0, dtype=g_var.dtype, out=g_var) executor.run(reset_program) @@ -104,20 +101,14 @@ class Accuracy(Evaluator): self.total = self.create_state(dtype='int64', shape=[1], suffix='total') self.correct = self.create_state( dtype='int64', shape=[1], suffix='correct') - kwargs = {'main_program': main_program} total = self.helper.create_tmp_variable(dtype='int') correct = self.helper.create_tmp_variable(dtype='int') acc = layers.accuracy( - input=input, - label=label, - k=k, - total=total, - correct=correct, - **kwargs) - total = layers.cast(x=total, dtype='int64', **kwargs) - correct = layers.cast(x=correct, dtype='int64', **kwargs) - layers.sums(input=[self.total, total], out=self.total, **kwargs) - layers.sums(input=[self.correct, correct], out=self.correct, **kwargs) + input=input, label=label, k=k, total=total, correct=correct) + total = layers.cast(x=total, dtype='int64') + correct = layers.cast(x=correct, dtype='int64') + layers.sums(input=[self.total, total], out=self.total) + layers.sums(input=[self.correct, correct], out=self.correct) self.metrics.append(acc) @@ -125,10 +116,75 @@ class Accuracy(Evaluator): if eval_program is None: eval_program = Program() block = eval_program.current_block() - kwargs = {'main_program': eval_program} - total = _clone_var_(block, self.total) - correct = _clone_var_(block, self.correct) - total = layers.cast(total, dtype='float32', **kwargs) - correct = layers.cast(correct, dtype='float32', **kwargs) - out = layers.elementwise_div(x=correct, y=total, **kwargs) + with program_guard(main_program=eval_program): + total = _clone_var_(block, self.total) + correct = _clone_var_(block, self.correct) + total = layers.cast(total, dtype='float32') + correct = layers.cast(correct, dtype='float32') + out = layers.elementwise_div(x=correct, y=total) return np.array(executor.run(eval_program, fetch_list=[out])[0]) + + +class ChunkEvaluator(Evaluator): + """ + Accumulate counter numbers output by chunk_eval from mini-batches and + compute the precision recall and F1-score using the accumulated counter + numbers. + """ + + def __init__( + self, + input, + label, + chunk_scheme, + num_chunk_types, + excluded_chunk_types=None, ): + super(ChunkEvaluator, self).__init__("chunk_eval") + main_program = self.helper.main_program + if main_program.current_block().idx != 0: + raise ValueError("You can only invoke Evaluator in root block") + + self.num_infer_chunks = self.create_state( + dtype='int64', shape=[1], suffix='num_infer_chunks') + self.num_label_chunks = self.create_state( + dtype='int64', shape=[1], suffix='num_label_chunks') + self.num_correct_chunks = self.create_state( + dtype='int64', shape=[1], suffix='num_correct_chunks') + precision, recall, f1_score, num_infer_chunks, num_label_chunks, num_correct_chunks = layers.chunk_eval( + input=input, + label=label, + chunk_scheme=chunk_scheme, + num_chunk_types=num_chunk_types, + excluded_chunk_types=excluded_chunk_types, ) + layers.sums( + input=[self.num_infer_chunks, num_infer_chunks], + out=self.num_infer_chunks) + layers.sums( + input=[self.num_label_chunks, num_label_chunks], + out=self.num_label_chunks) + layers.sums( + input=[self.num_correct_chunks, num_correct_chunks], + out=self.num_correct_chunks) + + self.metrics.extend([precision, recall, f1_score]) + + def eval(self, executor, eval_program=None): + if eval_program is None: + eval_program = Program() + block = eval_program.current_block() + num_infer_chunks, num_label_chunks, num_correct_chunks = executor.run( + eval_program, + fetch_list=[_clone_var_(block, state) for state in self.states]) + num_infer_chunks = num_infer_chunks[0] + num_label_chunks = num_label_chunks[0] + num_correct_chunks = num_correct_chunks[0] + precision = float( + num_correct_chunks) / num_infer_chunks if num_infer_chunks else 0 + recall = float( + num_correct_chunks) / num_label_chunks if num_label_chunks else 0 + f1_score = float(2 * precision * recall) / ( + precision + recall) if num_correct_chunks else 0 + return np.array( + [precision], dtype='float32'), np.array( + [recall], dtype='float32'), np.array( + [f1_score], dtype='float32') diff --git a/python/paddle/v2/fluid/executor.py b/python/paddle/v2/fluid/executor.py index bdc82eede9d93a7cf904999a6b869ce2d23c90dc..9a99b045dc70a9e4662a6f4da141183ffc8f1846 100644 --- a/python/paddle/v2/fluid/executor.py +++ b/python/paddle/v2/fluid/executor.py @@ -46,6 +46,13 @@ class Executor(object): p.set_place(each) act_places.append(p) + # TODO(dzhwinter) : consider that our fluid tests all written in + # GPUPlace(gpu_id), this will be changed in next PR. + if core.is_compile_gpu(): + core.init_devices(["CPU", "GPU:0"]) + else: + core.init_devices(["CPU"]) + self.executor = core.Executor(act_places) self.places = places diff --git a/python/paddle/v2/fluid/layer_helper.py b/python/paddle/v2/fluid/layer_helper.py index 3963e1322230259230885c097d37b818edda6b13..8df30ad76b0b5ff2140e28935c386bbb603d8bea 100644 --- a/python/paddle/v2/fluid/layer_helper.py +++ b/python/paddle/v2/fluid/layer_helper.py @@ -21,19 +21,11 @@ class LayerHelper(object): @property def main_program(self): - prog = self.kwargs.get('main_program', None) - if prog is None: - return default_main_program() - else: - return prog + return default_main_program() @property def startup_program(self): - prog = self.kwargs.get('startup_program', None) - if prog is None: - return default_startup_program() - else: - return prog + return default_startup_program() def append_op(self, *args, **kwargs): return self.main_program.current_block().append_op(*args, **kwargs) @@ -151,13 +143,6 @@ class LayerHelper(object): persistable=True, initializer=initializer) - @property - def to_kwargs(self): - return { - 'main_program': self.main_program, - 'startup_program': self.startup_program - } - def append_bias_op(self, input_var, dim_start=1, dim_end=None): """ Append bias operator and return its output. If the user does not set diff --git a/python/paddle/v2/fluid/layers.py b/python/paddle/v2/fluid/layers.py deleted file mode 100644 index f67d6d08c7557d939f280d19c5b86914885490bd..0000000000000000000000000000000000000000 --- a/python/paddle/v2/fluid/layers.py +++ /dev/null @@ -1,2025 +0,0 @@ -import contextlib - -import proto.framework_pb2 as framework_pb2 -import core -from framework import OpProtoHolder, Variable, Program, Operator -from initializer import Constant, Normal, Xavier, Initializer -from paddle.v2.fluid.layer_helper import LayerHelper, unique_name -from registry import register_layer -from param_attr import ParamAttr - -__all__ = [ - 'fc', 'data', 'cross_entropy', 'conv2d', 'pool2d', 'embedding', 'concat', - 'StaticRNN', 'cast', 'sequence_conv', 'sequence_pool', 'sums', 'cos_sim', - 'batch_norm', 'accuracy', 'split_lod_tensor', 'While' -] - -_REGISTER_LAYER_FROM_OPS = [ - 'mean', 'mul', 'dropout', 'reshape', 'sigmoid', 'scale', 'transpose', - 'sigmoid_cross_entropy_with_logits', 'elementwise_add', 'elementwise_div', - 'elementwise_sub', 'elementwise_mul', 'clip', 'abs' -] - -for _OP in set(_REGISTER_LAYER_FROM_OPS): - globals()[_OP] = register_layer(_OP) - __all__.append(_OP) - - -def fc(input, - size, - num_flatten_dims=1, - param_attr=None, - bias_attr=None, - act=None, - name=None, - main_program=None, - startup_program=None): - """ - Fully Connected Layer. - - Args: - input: The input tensor to the function - size: The size of the layer - num_flatten_dims: Number of columns in input - param_attr: The parameters/weights to the FC Layer - param_initializer: Initializer used for the weight/parameter. If None, XavierInitializer() is used - bias_attr: The bias parameter for the FC layer - bias_initializer: Initializer used for the bias. If None, then ConstantInitializer() is used - act: Activation to be applied to the output of FC layer - name: Name/alias of the function - main_program: Name of the main program that calls this - startup_program: Name of the startup program - - This function can take in multiple inputs and performs the Fully Connected - function (linear transformation) on top of each of them. - So for input x, the output will be : Wx + b. Where W is the parameter, - b the bias and x is the input. - - The function also applies an activation (non-linearity) on top of the - output, if activation is passed in the input. - - All the input variables of this function are passed in as local variables - to the LayerHelper constructor. - - """ - helper = LayerHelper('fc', **locals()) - - dtype = helper.input_dtype() - - mul_results = [] - for input_var, param_attr in helper.iter_inputs_and_params(): - input_shape = input_var.shape - param_shape = [ - reduce(lambda a, b: a * b, input_shape[num_flatten_dims:], 1) - ] + [size] - w = helper.create_parameter( - attr=param_attr, shape=param_shape, dtype=dtype, is_bias=False) - tmp = helper.create_tmp_variable(dtype) - helper.append_op( - type="mul", - inputs={ - "X": input_var, - "Y": w, - }, - outputs={"Out": tmp}, - attrs={'x_num_col_dims': num_flatten_dims, - 'y_num_col_dims': 1}) - mul_results.append(tmp) - - # sum - if len(mul_results) == 1: - pre_bias = mul_results[0] - else: - pre_bias = helper.create_tmp_variable(dtype) - helper.append_op( - type="sum", inputs={"X": mul_results}, outputs={"Out": pre_bias}) - # add bias - pre_activation = helper.append_bias_op(pre_bias) - # add activation - return helper.append_activation(pre_activation) - - -def embedding(input, - size, - is_sparse=False, - param_attr=None, - dtype='float32', - main_program=None, - startup_program=None): - """ - Embedding Layer. - - Args: - param_initializer: - input: The input to the function - size: The size of the layer - is_sparse: A flag that decleares whether the input is sparse - param_attr: Parameters for this layer - dtype: The type of data : float32, float_16, int etc - main_program: Name of the main program that calls this - startup_program: Name of the startup program - - This function can take in the input (which is a vector of IDs) and - performs a lookup in the lookup_table using these IDs, to result into - the embedding of each ID in the input. - - All the input variables of this function are passed in as local variables - to the LayerHelper constructor. - - """ - - helper = LayerHelper('embedding', **locals()) - w = helper.create_parameter( - attr=helper.param_attr, shape=size, dtype=dtype, is_bias=False) - tmp = helper.create_tmp_variable(dtype) - helper.append_op( - type='lookup_table', - inputs={'Ids': input, - 'W': w}, - outputs={'Out': tmp}, - attrs={'is_sparse': is_sparse}) - return tmp - - -# TODO(qijun): expose H0 and C0 -def dynamic_lstm(input, - size, - param_attr=None, - bias_attr=None, - use_peepholes=True, - is_reverse=False, - gate_activation='sigmoid', - cell_activation='tanh', - candidate_activation='tanh', - dtype='float32', - main_program=None, - startup_program=None): - helper = LayerHelper('lstm', **locals()) - size = size / 4 - weight = helper.create_parameter( - attr=helper.param_attr, shape=[size, 4 * size], dtype=dtype) - bias_size = [1, 7 * size] - if not use_peepholes: - bias_size[1] = 4 * size - bias = helper.create_parameter( - attr=helper.bias_attr, shape=bias_size, dtype=dtype, is_bias=True) - - hidden = helper.create_tmp_variable(dtype) - cell = helper.create_tmp_variable(dtype) - batch_gate = helper.create_tmp_variable(dtype) - batch_cell_pre_act = helper.create_tmp_variable(dtype) - - helper.append_op( - type='lstm', - inputs={'Input': input, - 'Weight': weight, - 'Bias': bias}, - outputs={ - 'Hidden': hidden, - 'Cell': cell, - 'BatchGate': batch_gate, - 'BatchCellPreAct': batch_cell_pre_act - }, - attrs={ - 'use_peepholes': use_peepholes, - 'is_reverse': is_reverse, - 'gate_activation': gate_activation, - 'cell_activation': cell_activation, - 'candidate_activation': candidate_activation - }) - return hidden, cell - - -def gru_unit(input, - hidden, - size, - weight=None, - bias=None, - activation='tanh', - gate_activation='sigmoid', - main_program=None, - startup_program=None): - """ - GRUUnit Operator implements partial calculations of the GRU unit as following: - - $$ - update \ gate: u_t = actGate(xu_t + W_u * h_{t-1} + b_u) \\ - reset \ gate: r_t = actGate(xr_t + W_r * h_{t-1} + b_r) \\ - output \ candidate: {h}_t = actNode(xc_t + W_c * dot(r_t, h_{t-1}) + b_c) \\ - output: h_t = dot((1 - u_t), h_{t-1}) + dot(u_t, {h}_t) - $$ - - which is same as one time step of GRU Operator. - - @note To implement the complete GRU unit, fully-connected operator must be - used before to feed xu, xr and xc as the Input of GRUUnit operator. - - TODO(ChunweiYan) add more document here - """ - activation_dict = dict( - identity=0, - sigmoid=1, - tanh=2, - relu=3, ) - activation = activation_dict[activation] - gate_activation = activation_dict[gate_activation] - - helper = LayerHelper('gru_unit', **locals()) - dtype = helper.input_dtype() - size = size / 3 - - # create weight - if weight is None: - weight = helper.create_parameter( - attr=helper.param_attr, shape=[size, 3 * size], dtype=dtype) - - # create bias - if bias is None: - bias_size = [1, 3 * size] - bias = helper.create_parameter( - attr=helper.bias_attr, shape=bias_size, dtype=dtype, is_bias=True) - - gate = helper.create_tmp_variable(dtype) - reset_hidden_pre = helper.create_tmp_variable(dtype) - updated_hidden = helper.create_tmp_variable(dtype) - - helper.append_op( - type='gru_unit', - inputs={'Input': input, - 'HiddenPrev': hidden, - 'Weight': weight}, - outputs={ - 'Gate': gate, - 'ResetHiddenPrev': reset_hidden_pre, - 'Hidden': updated_hidden, - }, - attrs={ - 'activation': 0, - 'gate_activation': 1, - }) - - return updated_hidden, reset_hidden_pre, gate - - -def data(name, - shape, - append_batch_size=True, - dtype='float32', - lod_level=0, - type=core.VarDesc.VarType.LOD_TENSOR, - main_program=None, - startup_program=None, - stop_gradient=True): - """ - Data Layer. - - Args: - name: The name/alias of the function - shape: Tuple declaring the shape. - append_batch_size: Whether or not to append the data as a batch. - dtype: The type of data : float32, float_16, int etc - type: The output type. By default it is LOD_TENSOR. - lod_level(int): The LoD Level. 0 means the input data is not a sequence. - main_program: Name of the main program that calls this - startup_program: Name of the startup program - stop_gradient: A boolean that mentions whether gradient should flow. - - This function takes in input and based on whether data has - to be returned back as a minibatch, it creates the global variable using - the helper functions. The global variables can be accessed by all the - following operations and layers in the graph. - - All the input variables of this function are passed in as local variables - to the LayerHelper constructor. - - """ - helper = LayerHelper('data', **locals()) - shape = list(shape) - for i in xrange(len(shape)): - if shape[i] is None: - shape[i] = -1 - append_batch_size = False - elif shape[i] < 0: - append_batch_size = False - - if append_batch_size: - shape = [-1] + shape # append batch size as -1 - - return helper.create_global_variable( - name=name, - shape=shape, - dtype=dtype, - type=type, - stop_gradient=stop_gradient, - lod_level=lod_level) - - -def create_tensor(dtype, name=None, main_program=None, startup_program=None): - helper = LayerHelper("create_tensor", **locals()) - return helper.create_variable(name=helper.name, dtype=dtype) - - -def cast(x, dtype, main_program=None): - """ - This function takes in the input with input_dtype - and casts it to the output_dtype as the output. - """ - helper = LayerHelper('cast', **locals()) - out = helper.create_tmp_variable(dtype=dtype) - helper.append_op( - type='cast', - inputs={'X': [x]}, - outputs={'Out': [out]}, - attrs={'in_dtype': x.dtype, - 'out_dtype': out.dtype}) - return out - - -def concat(input, axis, main_program=None, startup_program=None): - """ - This function concats the input along the axis mentioned - and returns that as the output. - """ - helper = LayerHelper('concat', **locals()) - out = helper.create_tmp_variable(dtype=helper.input_dtype()) - helper.append_op( - type='concat', - inputs={'X': input}, - outputs={'Out': [out]}, - attrs={'axis': axis}) - return out - - -def sums(input, out=None, main_program=None, startup_program=None): - """ - This function takes in the input and performs the sum operation on it - and returns that as the output. - """ - helper = LayerHelper('sum', **locals()) - if out is None: - out = helper.create_tmp_variable(dtype=helper.input_dtype()) - helper.append_op(type='sum', inputs={'X': input}, outputs={'Out': out}) - return out - - -def linear_chain_crf(input, - label, - param_attr=None, - main_program=None, - startup_program=None): - helper = LayerHelper('linear_chain_crf', **locals()) - size = input.shape[1] - transition = helper.create_parameter( - attr=helper.param_attr, - shape=[size + 2, size], - dtype=helper.input_dtype()) - alpha = helper.create_tmp_variable(dtype=helper.input_dtype()) - emission_exps = helper.create_tmp_variable(dtype=helper.input_dtype()) - transition_exps = helper.create_tmp_variable(dtype=helper.input_dtype()) - log_likelihood = helper.create_tmp_variable(dtype=helper.input_dtype()) - helper.append_op( - type='linear_chain_crf', - inputs={"Emission": [input], - "Transition": transition, - "Label": label}, - outputs={ - "Alpha": [alpha], - "EmissionExps": [emission_exps], - "TransitionExps": transition_exps, - "LogLikelihood": log_likelihood - }) - - return log_likelihood - - -def crf_decoding(input, - param_attr, - label=None, - main_program=None, - startup_program=None): - helper = LayerHelper('crf_decoding', **locals()) - transition = helper.get_parameter(param_attr.name) - viterbi_path = helper.create_tmp_variable(dtype=helper.input_dtype()) - helper.append_op( - type='crf_decoding', - inputs={"Emission": [input], - "Transition": transition, - "Label": label}, - outputs={"ViterbiPath": [viterbi_path]}) - - return viterbi_path - - -def assign(input, output, main_program=None, startup_program=None): - helper = LayerHelper('assign', **locals()) - helper.append_op( - type='scale', - inputs={'X': [input]}, - outputs={'Out': [output]}, - attrs={'scale': 1.0}) - return output - - -def split_lod_tensor(input, - mask, - level=0, - main_program=None, - startup_program=None): - helper = LayerHelper('split_lod_tensor', **locals()) - out_true = helper.create_tmp_variable(dtype=input.dtype) - out_false = helper.create_tmp_variable(dtype=input.dtype) - helper.append_op( - type='split_lod_tensor', - inputs={ - 'X': input, - 'Mask': mask, - }, - outputs={'OutTrue': out_true, - 'OutFalse': out_false}, - attrs={'level': level}) - return out_true, out_false - - -def merge_lod_tensor(in_true, - in_false, - x, - mask, - level=0, - main_program=None, - startup_program=None): - helper = LayerHelper('merge_lod_tensor', **locals()) - out = helper.create_tmp_variable(dtype=in_true.dtype) - helper.append_op( - type='merge_lod_tensor', - inputs={'X': x, - 'Mask': mask, - 'InTrue': in_true, - 'InFalse': in_false}, - outputs={'Out': out}, - attrs={'level': level}) - return out - - -def cos_sim(X, Y, **kwargs): - """ - This function performs the cosine similarity between two tensors - X and Y and returns that as the output. - """ - helper = LayerHelper('cos_sim', **kwargs) - out = helper.create_tmp_variable(dtype=X.dtype) - xnorm = helper.create_tmp_variable(dtype=X.dtype) - ynorm = helper.create_tmp_variable(dtype=X.dtype) - helper.append_op( - type='cos_sim', - inputs={'X': [X], - 'Y': [Y]}, - outputs={'Out': [out], - 'XNorm': [xnorm], - 'YNorm': [ynorm]}) - return out - - -def cross_entropy(input, label, **kwargs): - """ - This function computes cross_entropy using the input and label. - """ - helper = LayerHelper('cross_entropy', **kwargs) - out = helper.create_tmp_variable(dtype=input.dtype) - helper.append_op( - type='cross_entropy', - inputs={'X': [input], - 'Label': [label]}, - outputs={'Y': [out]}, - attrs=kwargs) - return out - - -def square_error_cost(input, label, **kwargs): - """ - This functions returns the squared error cost using the input and label. - The output is appending the op to do the above. - """ - helper = LayerHelper('square_error_cost', **kwargs) - minus_out = helper.create_tmp_variable(dtype=input.dtype) - helper.append_op( - type='elementwise_sub', - inputs={'X': [input], - 'Y': [label]}, - outputs={'Out': [minus_out]}) - - square_out = helper.create_tmp_variable(dtype=input.dtype) - helper.append_op( - type='square', inputs={'X': [minus_out]}, outputs={'Y': [square_out]}) - return square_out - - -def accuracy(input, label, k=1, correct=None, total=None, **kwargs): - """ - This function computes the accuracy using the input and label. - The output is the top_k inputs and their indices. - """ - helper = LayerHelper("accuracy", **kwargs) - topk_out = helper.create_tmp_variable(dtype=input.dtype) - topk_indices = helper.create_tmp_variable(dtype="int64") - helper.append_op( - type="top_k", - inputs={"X": [input]}, - outputs={"Out": [topk_out], - "Indices": [topk_indices]}, - attrs={"k": k}) - acc_out = helper.create_tmp_variable(dtype="float32") - if correct is None: - correct = helper.create_tmp_variable(dtype="int64") - if total is None: - total = helper.create_tmp_variable(dtype="int64") - helper.append_op( - type="accuracy", - inputs={ - "Out": [topk_out], - "Indices": [topk_indices], - "Label": [label] - }, - outputs={ - "Accuracy": [acc_out], - "Correct": [correct], - "Total": [total], - }) - return acc_out - - -def chunk_eval(input, - label, - chunk_scheme, - num_chunk_types, - excluded_chunk_types=None, - **kwargs): - """ - This function computes the accuracy using the input and label. - The output is the top_k inputs and their indices. - """ - helper = LayerHelper("chunk_eval", **kwargs) - - # prepare output - precision = helper.create_tmp_variable(dtype="float32") - recall = helper.create_tmp_variable(dtype="float32") - f1_score = helper.create_tmp_variable(dtype="float32") - - helper.append_op( - type="chunk_eval", - inputs={"Inference": [input], - "Label": [label]}, - outputs={ - "Precision": [precision], - "Recall": [recall], - "F1-Score": [f1_score] - }, - attrs={ - "num_chunk_types": num_chunk_types, - 'chunk_scheme': chunk_scheme, - 'excluded_chunk_types': excluded_chunk_types or [] - }) - return precision, recall, f1_score - - -def sequence_conv(input, - num_filters, - filter_size=3, - filter_stride=1, - padding=None, - bias_attr=None, - param_attr=None, - act=None, - main_program=None, - startup_program=None): - """ - This function creates the op for sequence_conv, using the inputs and - other convolutional configurations for the filters and stride as given - in the input parameters to the function. - """ - - # FIXME(dzh) : want to unify the argument of python layer - # function. So we ignore some unecessary attributes. - # such as, padding_trainable, context_start. - - helper = LayerHelper('sequence_conv', **locals()) - dtype = helper.input_dtype() - filter_shape = [filter_size * input.shape[1], num_filters] - filter_param = helper.create_parameter( - attr=helper.param_attr, shape=filter_shape, dtype=dtype) - pre_bias = helper.create_tmp_variable(dtype) - - helper.append_op( - type='sequence_conv', - inputs={ - 'X': [input], - 'Filter': [filter_param], - }, - outputs={"Out": pre_bias}, - attrs={ - 'contextStride': filter_stride, - 'contextStart': -int(filter_size / 2), - 'contextLength': filter_size - }) - pre_act = helper.append_bias_op(pre_bias) - return helper.append_activation(pre_act) - - -def conv2d(input, - num_filters, - filter_size, - stride=None, - padding=None, - groups=None, - param_attr=None, - bias_attr=None, - act=None, - name=None, - main_program=None, - startup_program=None): - """ - This function creates the op for a 2-dimensional Convolution. - This is performed using the parameters of filters(size, dimensionality etc) - , stride and other configurations for a Convolution operation. - This funciton can also append an activation on top of the - conv-2d output, if mentioned in the input parameters. - """ - - if stride is None: - stride = [1, 1] - helper = LayerHelper('conv2d', **locals()) - dtype = helper.input_dtype() - - num_channels = input.shape[1] - if groups is None: - num_filter_channels = num_channels - else: - if num_channels % groups != 0: - raise ValueError("num_channels must be divisible by groups.") - num_filter_channels = num_channels / groups - - if isinstance(filter_size, int): - filter_size = [filter_size, filter_size] - if isinstance(stride, int): - stride = [stride, stride] - if isinstance(padding, int): - padding = [padding, padding] - - input_shape = input.shape - filter_shape = [num_filters, num_filter_channels] + filter_size - - def _get_default_param_initializer(): - std = (2.0 / (filter_size[0]**2 * num_channels))**0.5 - return Normal(0.0, std, 0) - - filter_param = helper.create_parameter( - attr=helper.param_attr, - shape=filter_shape, - dtype=dtype, - default_initializer=_get_default_param_initializer()) - - pre_bias = helper.create_tmp_variable(dtype) - - helper.append_op( - type='conv2d_cudnn', - inputs={ - 'Input': input, - 'Filter': filter_param, - }, - outputs={"Output": pre_bias}, - attrs={'strides': stride, - 'paddings': padding, - 'groups': groups}) - - pre_act = helper.append_bias_op(pre_bias, dim_start=1, dim_end=2) - - return helper.append_activation(pre_act) - - -def sequence_pool(input, pool_type, **kwargs): - """ - This function add the operator for sequence pooling. - This is applied on top of the input using pool_type mentioned - in the parameters. - """ - helper = LayerHelper('sequence_pool', input=input, **kwargs) - dtype = helper.input_dtype() - pool_out = helper.create_tmp_variable(dtype) - max_index = helper.create_tmp_variable(dtype) - - helper.append_op( - type="sequence_pool", - inputs={"X": input}, - outputs={"Out": pool_out, - "MaxIndex": max_index}, - attrs={"pooltype": pool_type.upper()}) - - return pool_out - - -def pool2d(input, - pool_size, - pool_type, - pool_stride=None, - pool_padding=None, - global_pooling=False, - main_program=None, - startup_program=None): - """ - This function adds the operator for pooling in 2 dimensions, using the - pooling configurations mentioned in input parameters. - """ - if pool_padding is None: - pool_padding = [0, 0] - if pool_stride is None: - pool_stride = [1, 1] - if pool_type not in ["max", "avg"]: - raise ValueError( - "Unknown pool_type: '%s'. It can only be 'max' or 'avg'.", - str(pool_type)) - if isinstance(pool_size, int): - pool_size = [pool_size, pool_size] - if isinstance(pool_stride, int): - pool_stride = [pool_stride, pool_stride] - if isinstance(pool_padding, int): - pool_padding = [pool_padding, pool_padding] - - helper = LayerHelper('pool2d', **locals()) - dtype = helper.input_dtype() - pool_out = helper.create_tmp_variable(dtype) - - helper.append_op( - type="pool2d", - inputs={"X": input}, - outputs={"Out": pool_out}, - attrs={ - "pooling_type": pool_type, - "ksize": pool_size, - "global_pooling": global_pooling, - "strides": pool_stride, - "paddings": pool_padding - }) - - return pool_out - - -def batch_norm(input, - act=None, - is_test=False, - momentum=0.9, - epsilon=1e-05, - param_attr=None, - bias_attr=None, - data_layout='NCHW', - main_program=None, - startup_program=None): - """ - This function helps create an operator to implement - the BatchNorm layer using the configurations from the input parameters. - """ - helper = LayerHelper('batch_norm', **locals()) - dtype = helper.input_dtype() - - input_shape = input.shape - if data_layout == 'NCHW': - channel_num = input_shape[1] - else: - if data_layout == 'NHWC': - channel_num = input_shape[-1] - else: - raise ValueError("unsupported data layout:" + data_layout) - - param_shape = [channel_num] - - # create parameter - scale = helper.create_parameter( - attr=helper.param_attr, - shape=param_shape, - dtype=dtype, - default_initializer=Constant(1.0)) - - bias = helper.create_parameter( - attr=helper.param_attr, shape=param_shape, dtype=dtype, is_bias=True) - - mean = helper.create_global_variable( - dtype=input.dtype, shape=param_shape, persistable=True) - helper.set_variable_initializer(var=mean, initializer=Constant(0.0)) - - variance = helper.create_global_variable( - dtype=input.dtype, shape=param_shape, persistable=True) - helper.set_variable_initializer(var=variance, initializer=Constant(1.0)) - - # create output - # mean and mean_out share the same memory - mean_out = mean - # variance and variance out share the same memory - variance_out = variance - saved_mean = helper.create_tmp_variable(dtype) - saved_variance = helper.create_tmp_variable(dtype) - - batch_norm_out = helper.create_tmp_variable(dtype) - - helper.append_op( - type="batch_norm", - inputs={ - "X": input, - "Scale": scale, - "Bias": bias, - "Mean": mean, - "Variance": variance - }, - outputs={ - "Y": batch_norm_out, - "MeanOut": mean_out, - "VarianceOut": variance_out, - "SavedMean": saved_mean, - "SavedVariance": saved_variance - }, - attrs={"momentum": momentum, - "epsilon": epsilon, - "is_test": is_test}) - - return helper.append_activation(batch_norm_out) - - -def beam_search_decode(ids, scores, main_program=None, startup_program=None): - helper = LayerHelper('beam_search_decode', **locals()) - sentence_ids = helper.create_tmp_variable(dtype=ids.dtype) - sentence_scores = helper.create_tmp_variable(dtype=ids.dtype) - - helper.append_op( - type="beam_search_decode", - inputs={"Ids": ids, - "Scores": scores}, - outputs={ - "SentenceIds": sentence_ids, - "SentenceScores": sentence_scores - }) - - return sentence_ids, sentence_scores - - -class BlockGuard(object): - """ - BlockGuard class. - - BlockGuard class is used to create a sub-block in a program by - using the Python `with` keyword. - """ - - def __init__(self, main_program): - if not isinstance(main_program, Program): - raise TypeError("BlockGuard takes a program") - self.main_program = main_program - - def __enter__(self): - self.main_program.create_block() - - def __exit__(self, exc_type, exc_val, exc_tb): - self.main_program.rollback() - if exc_type is not None: - return False # re-raise exception - return True - - -class StaticRNNGuard(BlockGuard): - """ - StaticRNNGuard class. - - StaticRNNGuard class is used to create a StaticRNN block in a program. - """ - - def __init__(self, rnn): - if not isinstance(rnn, StaticRNN): - raise TypeError("StaticRNNGuard takes a StaticRNN") - super(StaticRNNGuard, self).__init__(rnn.helper.main_program) - self.rnn = rnn - - def __enter__(self): - self.rnn.status = StaticRNN.IN_RNN_BLOCK - return super(StaticRNNGuard, self).__enter__() - - def __exit__(self, exc_type, exc_val, exc_tb): - if exc_type is not None: - return False - self.rnn.status = StaticRNN.AFTER_RNN_BLOCK - self.rnn.complete_rnn_op() - return super(StaticRNNGuard, self).__exit__(exc_type, exc_val, exc_tb) - - -class StaticRNNMemoryLink(object): - """ - StaticRNNMemoryLink class. - - Args: - init: the initial variable for Memory - init: Variable - pre_mem: the memory variable in previous time step - pre_mem: Variable - mem: the memory variable in current time step - mem: Variable - - StaticRNNMemoryLink class is used to create a link between two - memory cells of a StaticRNN. - """ - - def __init__(self, init, pre_mem, mem=None): - self.init = init - self.pre_mem = pre_mem - self.mem = mem - - -class StaticRNN(object): - """ - StaticRNN class. - - StaticRNN class is used to create a StaticRNN. The RNN will have its - own parameters like inputs, outputs, memories, status and length. - """ - BEFORE_RNN_BLOCK = 0 - IN_RNN_BLOCK = 1 - AFTER_RNN_BLOCK = 2 - - def __init__(self, name=None, main_program=None): - self.helper = LayerHelper( - "static_rnn", name=name, main_program=main_program) - self.memories = {} # memory map, from pre_mem.name --> MemoryLink - self.inputs = [] # input variable list in current block - self.outputs = [] # output variable list in parent block - self.status = StaticRNN.BEFORE_RNN_BLOCK # status flag. - # sequence length, since it is a static RNN, sequence length are fixed. - self.seq_len = None - - def step(self): - return StaticRNNGuard(self) - - def _assert_in_rnn_block_(self, method): - if self.status != StaticRNN.IN_RNN_BLOCK: - raise ValueError("You must invoke {0} in rnn block".format(method)) - - def memory(self, - init=None, - shape=None, - batch_ref=None, - init_value=0.0, - init_batch_dim_idx=0, - ref_batch_dim_idx=1): - """ - Args: - init: boot memory, if not set, a shape, batch_ref must be provided - shape: shape of the boot memory - batch_ref: batch size reference variable - init_value: the init value of boot memory - init_batch_dim_idx: the index of batch size in init's dimension - ref_batch_dim_idx: the index of batch size in batch_ref's dimension - """ - self._assert_in_rnn_block_('memory') - if init is None: - if shape is None or batch_ref is None: - raise ValueError( - "if init is None, memory at least need shape and batch_ref") - parent_block = self.parent_block() - var_name = unique_name("@".join([self.helper.name, "memory_boot"])) - boot_var = parent_block.create_var( - name=var_name, - shape=shape, - dtype=batch_ref.dtype, - persistable=False) - - parent_block.append_op( - type="fill_constant_batch_size_like", - inputs={'Input': [batch_ref]}, - outputs={'Out': [boot_var]}, - attrs={ - 'value': init_value, - 'shape': boot_var.shape, - 'dtype': boot_var.dtype, - 'input_dim_idx': ref_batch_dim_idx, - 'output_dim_idx': init_batch_dim_idx - }) - - return self.memory(init=boot_var) - else: - pre_mem = self.helper.create_variable( - name=unique_name("@".join([self.helper.name, "mem"])), - dtype=init.dtype, - shape=init.shape) - self.memories[pre_mem.name] = StaticRNNMemoryLink( - init=init, pre_mem=pre_mem) - return pre_mem - - def step_input(self, x): - self._assert_in_rnn_block_('step_input') - if not isinstance(x, Variable): - raise TypeError("step input takes a Variable") - if self.seq_len is None: - self.seq_len = x.shape[0] - elif self.seq_len != x.shape[0]: - raise ValueError("Static RNN only take fix seq_len input") - - ipt = self.helper.create_variable( - name=x.name, dtype=x.dtype, shape=list(x.shape[1:]), type=x.type) - self.inputs.append(ipt) - return ipt - - def step_output(self, o): - self._assert_in_rnn_block_('step_output') - if not isinstance(o, Variable): - raise TypeError("step output takes a Variable") - - tmp_o = self.helper.create_tmp_variable(dtype=o.dtype) - self.helper.append_op( - type='rnn_memory_helper', - inputs={'X': [o]}, - outputs={'Out': tmp_o}, - attrs={'dtype': o.dtype}) - - out_var = self.parent_block().create_var( - name=tmp_o.name, - shape=[self.seq_len] + list(tmp_o.shape), - dtype=tmp_o.dtype) - - self.outputs.append(out_var) - - def output(self, *outputs): - for each in outputs: - self.step_output(each) - - def update_memory(self, mem, var): - if not isinstance(mem, Variable) or not isinstance(var, Variable): - raise TypeError("update memory should take variables") - self.memories[mem.name].mem = var - - def parent_block(self): - prog = self.helper.main_program - parent_idx = prog.current_block().parent_idx - assert parent_idx >= 0 - parent_block = prog.block(parent_idx) - return parent_block - - def __call__(self, *args, **kwargs): - if self.status != StaticRNN.AFTER_RNN_BLOCK: - raise ValueError("RNN output can only be retrieved after rnn block") - if len(self.outputs) == 0: - raise ValueError("RNN has no output") - elif len(self.outputs) == 1: - return self.outputs[0] - else: - return self.outputs - - def complete_rnn_op(self): - main_program = self.helper.main_program - rnn_block = main_program.current_block() - parent_block = self.parent_block() - - local_inputs = set() - - for op in rnn_block.ops: - assert isinstance(op, Operator) - for oname in op.output_names: - for out_var_name in op.output(oname): - local_inputs.add(out_var_name) - - for var in self.inputs: - local_inputs.add(var.name) - for m in self.memories: - local_inputs.add(m) - - params = list() - for op in rnn_block.ops: - assert isinstance(op, Operator) - for iname in op.input_names: - for in_var_name in op.input(iname): - if in_var_name not in local_inputs: - params.append(in_var_name) - - parameters = [parent_block.var(name) for name in params] - - step_scope = parent_block.create_var( - type=core.VarDesc.VarType.STEP_SCOPES) - - inlinks = [parent_block.var(i.name) for i in self.inputs] - outlinks = self.outputs - - boot_memories = [] - pre_memories = [] - memories = [] - for _, mem in self.memories.iteritems(): - boot_memories.append(mem.init) - pre_memories.append(mem.pre_mem.name) - mem_var = rnn_block.var(mem.mem.name) - assert isinstance(mem_var, Variable) - new_mem = self.helper.create_tmp_variable(dtype=mem_var.dtype) - - rnn_block.append_op( - type='rnn_memory_helper', - inputs={'X': [mem_var]}, - outputs={'Out': [new_mem]}, - attrs={'dtype': mem_var.dtype}) - - memories.append(new_mem.name) - - parent_block.append_op( - type='recurrent', - inputs={ - 'inputs': inlinks, - 'initial_states': boot_memories, - 'parameters': parameters - }, - outputs={'outputs': outlinks, - 'step_scopes': [step_scope]}, - attrs={ - 'ex_states': pre_memories, - 'states': memories, - 'step_block': rnn_block - }) - - -class WhileGuard(BlockGuard): - def __init__(self, while_op): - if not isinstance(while_op, While): - raise TypeError("WhileGuard takes a while op") - super(WhileGuard, self).__init__(while_op.helper.main_program) - self.while_op = while_op - - def __enter__(self): - self.while_op.status = While.IN_WHILE_BLOCK - return super(WhileGuard, self).__enter__() - - def __exit__(self, exc_type, exc_val, exc_tb): - if exc_type is not None: - return False - self.while_op.status = While.AFTER_WHILE_BLOCK - self.while_op.complete() - return super(WhileGuard, self).__exit__(exc_type, exc_val, exc_tb) - - -class While(object): - BEFORE_WHILE_BLOCK = 0 - IN_WHILE_BLOCK = 1 - AFTER_WHILE_BLOCK = 2 - - def __init__(self, cond, name=None, main_program=None): - self.helper = LayerHelper("while", name=name, main_program=main_program) - self.status = While.BEFORE_WHILE_BLOCK - if not isinstance(cond, Variable): - raise TypeError("condition should be a variable") - assert isinstance(cond, Variable) - if cond.dtype != core.DataType.BOOL: - raise TypeError("condition should be a bool variable") - if reduce(lambda a, b: a * b, cond.shape, 1) != 1: - raise TypeError("condition should be a bool scalar") - self.cond_var = cond - - def block(self): - return WhileGuard(self) - - def complete(self): - main_program = self.helper.main_program - while_block = main_program.current_block() - parent_block = main_program.block(main_program.current_block() - .parent_idx) - - inner_outputs = {self.cond_var.name} - x_name_list = set() - for op in while_block.ops: - for iname in op.input_names: - for in_var_name in op.input(iname): - if in_var_name not in inner_outputs: - x_name_list.add(in_var_name) - - for oname in op.output_names: - for out_var_name in op.output(oname): - inner_outputs.add(out_var_name) - - out_vars = [] - for inner_out_name in inner_outputs: - if inner_out_name in parent_block.vars: - out_vars.append(parent_block.var(inner_out_name)) - - step_scope = parent_block.create_var( - type=core.VarDesc.VarType.STEP_SCOPES) - - parent_block.append_op( - type='while', - inputs={ - 'X': [parent_block.var(x_name) for x_name in x_name_list], - 'Condition': [self.cond_var] - }, - outputs={'Out': out_vars, - 'StepScopes': [step_scope]}, - attrs={'step_block': while_block}) - - -def lstm(x, - c_pre_init, - hidden_dim, - forget_bias=None, - main_program=None, - startup_program=None): - """ - This function helps create an operator for the LSTM (Long Short Term - Memory) cell that can be used inside an RNN. - """ - helper = LayerHelper('lstm_unit', **locals()) - rnn = StaticRNN() - with rnn.step(): - c_pre = rnn.memory(init=c_pre_init) - x_t = rnn.step_input(x) - - before_fc = concat( - input=[x_t, c_pre], - axis=1, - main_program=main_program, - startup_program=startup_program) - after_fc = fc(input=before_fc, - size=hidden_dim * 4, - main_program=main_program, - startup_program=startup_program) - - dtype = x.dtype - c = helper.create_tmp_variable(dtype) - h = helper.create_tmp_variable(dtype) - - helper.append_op( - type='lstm_unit', - inputs={"X": after_fc, - "C_prev": c_pre}, - outputs={"C": c, - "H": h}, - attrs={"forget_bias": forget_bias}) - - rnn.update_memory(c_pre, c) - rnn.output(h) - - return rnn() - - -def lod_rank_table(x, level=0, main_program=None): - """ - This function creates an operator for creating a LOD_RANK_TABLE - using the input x. - """ - helper = LayerHelper("lod_rank_table", **locals()) - table = helper.create_variable( - type=core.VarDesc.VarType.LOD_RANK_TABLE, - name=unique_name("lod_rank_table")) - helper.append_op( - type='lod_rank_table', - inputs={'X': x}, - outputs={'Out': table}, - attrs={'level': level}) - return table - - -def max_sequence_len(rank_table, main_program=None): - """ - This function creates an operator to calculate the length of - max seqence through input rank_table(should be a lod_rank_table) - """ - helper = LayerHelper("max_seqence_len", **locals()) - res = helper.create_tmp_variable(dtype="int64") - helper.append_op( - type="max_sequence_len", - inputs={"RankTable": rank_table}, - outputs={"Out": res}) - return res - - -def topk(input, k, main_program=None, startup_program=None): - helper = LayerHelper('topk', **locals()) - topk_out = helper.create_tmp_variable(dtype=input.data_type) - topk_indices = helper.create_tmp_variable(dtype='int64') - helper.append_op( - type='top_k', - inputs={'X': [input]}, - outputs={'Out': [topk_out], - 'Indices': [topk_indices]}, - attrs={'k': k}) - return topk_out, topk_indices - - -def lod_tensor_to_array(x, table, main_program=None): - """ - This function creates an operator to convert an LOD_Tensor to - an array. - """ - helper = LayerHelper("lod_tensor_to_array", **locals()) - array = helper.create_variable( - name=unique_name("lod_tensor_to_array"), - type=core.VarDesc.VarType.LOD_TENSOR_ARRAY, - dtype=x.dtype) - helper.append_op( - type='lod_tensor_to_array', - inputs={'X': x, - 'RankTable': table}, - outputs={'Out': array}) - return array - - -def array_to_lod_tensor(x, table, main_program=None, startup_program=None): - """ - This function creates an operator to convert an array to a - LOD_Tensor. - """ - helper = LayerHelper("array_to_lod_tensor", **locals()) - tmp = helper.create_tmp_variable(dtype=x.dtype) - helper.append_op( - type="array_to_lod_tensor", - inputs={'X': x, - 'RankTable': table}, - outputs={'Out': tmp}) - return tmp - - -def fill_constant(shape, - dtype, - value, - out=None, - main_program=None, - startup_program=None): - """ - This function creates a tensor , with shape as mentioned in the input and - specified dtype and fills this up with a constant value that - comes in the input. It also sets the stop_gradient to be True. - """ - helper = LayerHelper("fill_constant", **locals()) - if out is None: - out = helper.create_tmp_variable(dtype=dtype) - helper.append_op( - type='fill_constant', - inputs={}, - outputs={'Out': [out]}, - attrs={'shape': shape, - 'dtype': out.dtype, - 'value': float(value)}) - out.stop_gradient = True - return out - - -def fill_constant_batch_size_like(input, - shape, - dtype, - value, - input_dim_idx=0, - output_dim_idx=0, - main_program=None, - startup_program=None): - helper = LayerHelper("fill_constant_batch_size_like", **locals()) - out = helper.create_tmp_variable(dtype=dtype) - helper.append_op( - type='fill_constant_batch_size_like', - inputs={'Input': input}, - outputs={'Out': [out]}, - attrs={ - 'shape': shape, - 'dtype': out.dtype, - 'value': float(value), - 'input_dim_idx': input_dim_idx, - 'output_dim_idx': output_dim_idx - }) - out.stop_gradient = True - return out - - -def ones(shape, dtype, main_program=None): - """ - This function performs the same function as fill_constant() declared above - with the constant value being 1.0. - """ - return fill_constant(value=1.0, **locals()) - - -def zeros(shape, dtype, main_program=None): - """ - This function performs the same function as fill_constant() declared above - with the constant value being 0.0. - """ - return fill_constant(value=0.0, **locals()) - - -def increment(x, - value=1.0, - in_place=True, - main_program=None, - startup_program=None): - """ - This function creates an operator to increment each value in the input - `x` by an amount: `value` as mentioned in the input parameter. This - operation is performed in-place by default. - """ - helper = LayerHelper("increment", **locals()) - if not in_place: - out = helper.create_tmp_variable(dtype=x.dtype) - else: - out = x - helper.append_op( - type='increment', - inputs={'X': [x]}, - outputs={'Out': [out]}, - attrs={'step': float(value)}) - return out - - -def array_write(x, i, array=None, main_program=None, startup_program=None): - """ - This function creates an operator to write the data out as a - LOD_TENSOR_ARRAY. - """ - helper = LayerHelper('array_write', **locals()) - if array is None: - array = helper.create_variable( - name="{0}.out".format(helper.name), - type=core.VarDesc.VarType.LOD_TENSOR_ARRAY, - dtype=x.dtype) - helper.append_op( - type='write_to_array', - inputs={'X': [x], - 'I': [i]}, - outputs={'Out': [array]}) - return array - - -def create_array(dtype, main_program=None): - helper = LayerHelper("array", **locals()) - return helper.create_variable( - name="{0}.out".format(helper.name), - type=core.VarDesc.VarType.LOD_TENSOR_ARRAY, - dtype=dtype) - - -def less_than(x, y, cond=None, main_program=None, **ignored): - helper = LayerHelper("less_than", **locals()) - if cond is None: - cond = helper.create_tmp_variable(dtype='bool') - cond.stop_gradient = True - - helper.append_op( - type='less_than', inputs={'X': [x], - 'Y': [y]}, outputs={'Out': [cond]}) - return cond - - -def array_read(array, i, main_program=None, startup_program=None): - """ - This function creates an operator to read the data in as a - LOD_TENSOR_ARRAY. - """ - helper = LayerHelper('array_read', **locals()) - if not isinstance( - array, - Variable) or array.type != core.VarDesc.VarType.LOD_TENSOR_ARRAY: - raise TypeError("array should be tensor array vairable") - out = helper.create_tmp_variable(dtype=array.dtype) - helper.append_op( - type='read_from_array', - inputs={'X': [array], - 'I': [i]}, - outputs={'Out': [out]}) - return out - - -def shrink_memory(x, i, table, main_program=None, startup_program=None): - """ - This function creates an operator to shrink_rnn_memory using the RankTable - as mentioned in the input parameter. - """ - helper = LayerHelper('shrink_memory', **locals()) - out = helper.create_tmp_variable(dtype=x.dtype) - helper.append_op( - type='shrink_rnn_memory', - inputs={'X': [x], - 'I': [i], - 'RankTable': [table]}, - outputs={'Out': [out]}, - attrs={}) - return out - - -def array_length(array, main_program=None): - """ - This function creates an operator to find the length of the - LOD_TENSOR_ARRAY. - """ - helper = LayerHelper('array_length', **locals()) - tmp = helper.create_tmp_variable(dtype='int64') - tmp.stop_gradient = True - helper.append_op( - type='lod_array_length', inputs={'X': [array]}, outputs={'Out': [tmp]}) - return tmp - - -def conv2d_transpose(input, - num_filters, - output_size=None, - filter_size=None, - padding=None, - stride=None, - param_attr=None, - main_program=None, - startup_program=None): - """ - The transpose of conv2d layer. - - This layer is also known as deconvolution layer. - - Args: - input(Variable): The input image with [N, C, H, W] format. - num_filters(int): The number of filter. It is as same as the output - image channel. - output_size(int|tuple|None): The output image size. If output size is a - tuple, it must contain two integers, (image_H, image_W). This - parameter only works when filter_size is None. - filter_size(int|tuple|None): The filter size. If filter_size is a tuple, - it must contain two integers, (filter_size_H, filter_size_W). - Otherwise, the filter will be a square. None if use output size to - calculate filter_size - padding(int|tuple): The padding size. If padding is a tuple, it must - contain two integers, (padding_H, padding_W). Otherwise, the - padding_H = padding_W = padding. - stride(int|tuple): The stride size. If stride is a tuple, it must - contain two integers, (stride_H, stride_W). Otherwise, the - stride_H = stride_W = stride. - param_attr: Parameter Attribute. - main_program(Program): the main program - startup_program(Program): the startup program - - Returns: - Variable: Output image. - """ - helper = LayerHelper("conv2d_transpose", **locals()) - if not isinstance(input, Variable): - raise TypeError("Input of conv2d_transpose must be Variable") - input_channel = input.shape[1] - - op_attr = dict() - - if isinstance(padding, int): - op_attr['paddings'] = [padding, padding] - elif padding is not None: - op_attr['paddings'] = padding - - if isinstance(stride, int): - op_attr['strides'] = stride - elif stride is not None: - op_attr['strides'] = stride - - if filter_size is None: - if output_size is None: - raise ValueError("output_size must be set when filter_size is None") - if isinstance(output_size, int): - output_size = [output_size, output_size] - - padding = op_attr.get('paddings', [0, 0]) - stride = op_attr.get('strides', [1, 1]) - - h_in = input.shape[2] - w_in = input.shape[3] - filter_size_h = output_size[0] - \ - (h_in - 1) * stride[0] + 2 * padding[0] - filter_size_w = output_size[1] - \ - (w_in - 1) * stride[1] + 2 * padding[1] - filter_size = [filter_size_h, filter_size_w] - elif isinstance(filter_size, int): - filter_size = [filter_size, filter_size] - - filter_shape = [input_channel, num_filters] + filter_size - img_filter = helper.create_parameter( - dtype=input.dtype, shape=filter_shape, attr=helper.param_attr) - - out = helper.create_tmp_variable(dtype=input.dtype) - helper.append_op( - type='conv2d_transpose', - inputs={'Input': [input], - 'Filter': [img_filter]}, - outputs={'Output': out}, - attrs=op_attr) - - return out - - -class ConditionalBlockGuard(BlockGuard): - def __init__(self, block): - if not isinstance(block, ConditionalBlock): - raise TypeError("block should be conditional block") - super(ConditionalBlockGuard, self).__init__(block.helper.main_program) - self.block = block - - def __enter__(self): - return super(ConditionalBlockGuard, self).__enter__() - - def __exit__(self, exc_type, exc_val, exc_tb): - self.block.complete() - return super(ConditionalBlockGuard, self).__exit__(exc_type, exc_val, - exc_tb) - - -class ConditionalBlock(object): - def __init__(self, - inputs, - name=None, - main_program=None, - startup_program=None): - for each_input in inputs: - if not isinstance(each_input, Variable): - raise TypeError("Each input should be variable") - self.inputs = inputs - self.helper = LayerHelper( - 'conditional_block', - name=name, - main_program=main_program, - startup_program=startup_program) - - def block(self): - return ConditionalBlockGuard(self) - - def complete(self): - inside_block = self.helper.main_program.current_block() - parent_block = self.helper.main_program.block(inside_block.parent_idx) - - intermediate = set() - params = set() - - for each_op in inside_block.ops: - assert isinstance(each_op, Operator) - for iname in each_op.input_names: - for in_var_name in each_op.input(iname): - if in_var_name not in intermediate: - params.add(in_var_name) - - for oname in each_op.output_names: - for out_var_name in each_op.output(oname): - intermediate.add(out_var_name) - input_set = set([ipt.name for ipt in self.inputs]) - - param_list = [ - parent_block.var(each_name) for each_name in params - if each_name not in input_set - ] - - out_list = [ - parent_block.var(var_name) for var_name in parent_block.vars - if var_name not in intermediate - ] - - step_scope = parent_block.create_var( - type=core.VarDesc.VarType.STEP_SCOPES) - parent_block.append_op( - type='conditional_block', - inputs={ - 'X': self.inputs, - 'Params': param_list, - }, - outputs={'Out': out_list, - 'Scope': [step_scope]}, - attrs={'block': inside_block}) - - -class IfElseBlockGuard(object): - def __init__(self, is_true, ifelse): - if not isinstance(ifelse, IfElse): - raise TypeError("ifelse must be an instance of IfElse class") - - if ifelse.status != IfElse.OUT_IF_ELSE_BLOCKS: - raise ValueError("You cannot invoke IfElse.block() inside a block") - - self.is_true = is_true - self.ie = ifelse - if is_true: - self.cond_block = ifelse.conditional_true_block - else: - self.cond_block = ifelse.conditional_false_block - - if not isinstance(self.cond_block, ConditionalBlock): - raise TypeError("Unexpected situation") - - self.cond_block = self.cond_block.block() - - def __enter__(self): - self.ie.status = IfElse.IN_IF_ELSE_TRUE_BLOCKS if self.is_true else IfElse.IN_IF_ELSE_FALSE_BLOCKS - self.cond_block.__enter__() - - def __exit__(self, exc_type, exc_val, exc_tb): - if not self.cond_block.__exit__(exc_type, exc_val, exc_tb): - # re-raise inside exception - return False - if len(self.ie.output_table[1 if self.is_true else 0]) == 0: - raise ValueError("Must set output inside block") - self.ie.status = IfElse.OUT_IF_ELSE_BLOCKS - - -class IfElse(object): - OUT_IF_ELSE_BLOCKS = 0 - IN_IF_ELSE_TRUE_BLOCKS = 1 - IN_IF_ELSE_FALSE_BLOCKS = 2 - - def __init__(self, cond, name=None, main_program=None, - startup_program=None): - if not isinstance(cond, Variable): - raise TypeError("cond must be a Variable") - self.helper = LayerHelper( - 'ifelse', - name=name, - main_program=main_program, - startup_program=startup_program) - self.cond = cond - self.input_table = {} - self.status = IfElse.OUT_IF_ELSE_BLOCKS - self.conditional_true_block = ConditionalBlock(inputs=[self.cond]) - self.conditional_false_block = ConditionalBlock(inputs=[self.cond]) - self.output_table = ([], []) # (true_outs, false_outs) - - def input(self, x): - if self.status == IfElse.OUT_IF_ELSE_BLOCKS: - raise ValueError("input must in true/false blocks") - if id(x) not in self.input_table: - parent_block = self.parent_block() - out_true = parent_block.create_var( - name=unique_name('ifelse_input' + self.helper.name), - dtype=x.dtype) - - out_false = parent_block.create_var( - name=unique_name('ifelse_input' + self.helper.name), - dtype=x.dtype) - parent_block.append_op( - type='split_lod_tensor', - inputs={ - 'X': x, - 'Mask': self.cond, - }, - outputs={'OutTrue': out_true, - 'OutFalse': out_false}, - attrs={'level': 0}) - self.input_table[id(x)] = (out_true, out_false) - else: - out_true, out_false = self.input_table[id(x)] - - if self.status == IfElse.IN_IF_ELSE_TRUE_BLOCKS: - return out_true - else: - return out_false - - def parent_block(self): - current_block = self.helper.main_program.current_block() - return self.helper.main_program.block(current_block.parent_idx) - - def true_block(self): - return IfElseBlockGuard(True, self) - - def false_block(self): - return IfElseBlockGuard(False, self) - - def output(self, *outs): - if self.status == self.OUT_IF_ELSE_BLOCKS: - raise ValueError("output can only be invoked in the sub-block") - - out_table = self.output_table[1 if self.status == - self.IN_IF_ELSE_TRUE_BLOCKS else 0] - parent_block = self.parent_block() - for each_out in outs: - if not isinstance(each_out, Variable): - raise TypeError("Each output should be a variable") - # create outside tensor - outside_out = parent_block.create_var( - name=unique_name("_".join([self.helper.name, 'output'])), - dtype=each_out.dtype) - out_table.append(outside_out) - - # assign local var to outside - assign( - input=each_out, - output=outside_out, - main_program=self.helper.main_program, - startup_program=self.helper.startup_program) - - def __call__(self): - if self.status != self.OUT_IF_ELSE_BLOCKS: - raise ValueError("IfElse::__call__ must be out of sub-block") - false_len, true_len = map(len, self.output_table) - if false_len == 0 and true_len == 0: - raise ValueError("Must invoke true_block/false_block before " - "__call__") - elif false_len != true_len and false_len != 0 and true_len != 0: - raise ValueError("The output side must be same") - elif false_len == 0 or true_len == 0: - return self.output_table[0 if false_len != 0 else 1] - - # else none of false_len/true_len is zero - # merge together - rlist = [] - for false_var, true_var in zip(*self.output_table): - rlist.append( - merge_lod_tensor( - in_true=true_var, - in_false=false_var, - mask=self.cond, - x=self.cond, - level=0, - main_program=self.helper.main_program, - startup_program=self.helper.startup_program)) - return rlist - - -class DynamicRNN(object): - BEFORE_RNN = 0 - IN_RNN = 1 - AFTER_RNN = 2 - - def __init__(self, name=None, main_program=None, startup_program=None): - self.helper = LayerHelper( - 'dynamic_rnn', - name=name, - main_program=main_program, - startup_program=startup_program) - self.status = DynamicRNN.BEFORE_RNN - self.lod_rank_table = None - self.max_seq_len = None - self.step_idx = None - self.zero_idx = fill_constant(shape=[1], value=0, dtype='int64') - self.mem_dict = dict() - self.output_array = [] - self.outputs = [] - self.cond = self.helper.create_tmp_variable(dtype='bool') - self.cond.stop_gradient = False - self.while_op = While(self.cond) - self.input_array = [] - self.mem_link = [] - - def step_input(self, x): - self._assert_in_rnn_block_("step_input") - if not isinstance(x, Variable): - raise TypeError( - "step_input() can only take a Variable as its input") - parent_block = self._parent_block_() - if self.lod_rank_table is None: - self.lod_rank_table = parent_block.create_var( - name=unique_name('lod_rank_table'), - type=core.VarDesc.VarType.LOD_RANK_TABLE) - self.lod_rank_table.stop_gradient = True - parent_block.append_op( - type='lod_rank_table', - inputs={"X": x}, - outputs={"Out": self.lod_rank_table}) - self.max_seq_len = parent_block.create_var( - name=unique_name('dynamic_rnn_max_seq_len'), dtype='int64') - self.max_seq_len.stop_gradient = False - parent_block.append_op( - type='max_sequence_len', - inputs={'RankTable': self.lod_rank_table}, - outputs={"Out": self.max_seq_len}) - self.cond.stop_gradient = True - parent_block.append_op( - type='less_than', - inputs={'X': self.step_idx, - 'Y': self.max_seq_len}, - outputs={'Out': self.cond}) - - input_array = parent_block.create_var( - name=unique_name('dynamic_rnn_input_array'), - type=core.VarDesc.VarType.LOD_TENSOR_ARRAY, - dtype=x.dtype) - self.input_array.append((input_array, x.dtype)) - parent_block.append_op( - type='lod_tensor_to_array', - inputs={'X': x, - 'RankTable': self.lod_rank_table}, - outputs={'Out': input_array}) - return array_read( - array=input_array, i=self.step_idx, **self.helper.to_kwargs) - - @contextlib.contextmanager - def block(self): - if self.status != DynamicRNN.BEFORE_RNN: - raise ValueError("rnn.block() can only be invoke once") - self.step_idx = fill_constant(shape=[1], dtype='int64', value=0) - self.step_idx.stop_gradient = False - self.status = DynamicRNN.IN_RNN - with self.while_op.block(): - yield - increment( - x=self.step_idx, - value=1.0, - in_place=True, - **self.helper.to_kwargs) - - for new_mem, mem_array in self.mem_link: - array_write( - x=new_mem, - i=self.step_idx, - array=mem_array, - **self.helper.to_kwargs) - - less_than( - x=self.step_idx, - y=self.max_seq_len, - cond=self.cond, - **self.helper.to_kwargs) - - self.status = DynamicRNN.AFTER_RNN - for each_array in self.output_array: - self.outputs.append( - array_to_lod_tensor( - x=each_array, - table=self.lod_rank_table, - **self.helper.to_kwargs)) - - def __call__(self, *args, **kwargs): - if self.status != DynamicRNN.AFTER_RNN: - raise ValueError( - "Dynamic RNN outputs can only be retrieved after rnn block") - if len(self.outputs) == 1: - return self.outputs[0] - else: - return self.outputs - - def memory(self, init=None, shape=None, value=0.0, dtype='float32'): - self._assert_in_rnn_block_('memory') - if init is not None: - if not isinstance(init, Variable): - raise TypeError( - "The input arg `init` of memory() must be a Variable") - parent_block = self._parent_block_() - mem_array = parent_block.create_var( - name=unique_name('dynamic_rnn_mem_array'), - type=core.VarDesc.VarType.LOD_TENSOR_ARRAY, - dtype=init.dtype) - parent_block.append_op( - type='write_to_array', - inputs={'X': init, - 'I': self.zero_idx}, - outputs={'Out': mem_array}) - retv = array_read( - array=mem_array, i=self.step_idx, **self.helper.to_kwargs) - retv = shrink_memory( - x=retv, - i=self.step_idx, - table=self.lod_rank_table, - **self.helper.to_kwargs) - self.mem_dict[retv.name] = mem_array - return retv - else: - if len(self.input_array) == 0: - raise ValueError( - "step_input should be invoked before memory(shape=..., value=...)" - ) - parent_block = self._parent_block_() - init = parent_block.create_var( - name=unique_name('mem_init'), dtype=dtype) - arr, dtype = self.input_array[0] - in0 = parent_block.create_var(name=unique_name('in0'), dtype=dtype) - parent_block.append_op( - type='read_from_array', - inputs={'X': [arr], - 'I': [self.zero_idx]}, - outputs={'Out': [in0]}) - parent_block.append_op( - type='fill_constant_batch_size_like', - inputs={'Input': [in0]}, - outputs={'Out': [init]}, - attrs={ - 'shape': [-1] + shape, - 'value': float(value), - 'dtype': init.dtype - }) - return self.memory(init=init) - - def update_memory(self, ex_mem, new_mem): - self._assert_in_rnn_block_('update_memory') - if not isinstance(ex_mem, Variable): - raise TypeError("The input arg `ex_mem` of update_memory() must " - "be a Variable") - if not isinstance(new_mem, Variable): - raise TypeError("The input arg `new_mem` of update_memory() must " - "be a Variable") - - mem_array = self.mem_dict.get(ex_mem.name, None) - if mem_array is None: - raise ValueError("Please invoke memory before update_memory") - if self.lod_rank_table is None: - raise ValueError("Please invoke step_input before update_memory") - - self.mem_link.append((new_mem, mem_array)) - - def output(self, *outputs): - self._assert_in_rnn_block_('output') - parent_block = self._parent_block_() - for each in outputs: - outside_array = parent_block.create_var( - name=unique_name("_".join( - [self.helper.name, "output_array", each.name])), - type=core.VarDesc.VarType.LOD_TENSOR_ARRAY, - dtype=each.dtype) - array_write(x=each, i=self.step_idx, array=outside_array) - self.output_array.append(outside_array) - - def _parent_block_(self): - prog = self.helper.main_program - parent_idx = prog.current_block().parent_idx - assert parent_idx >= 0 - parent_block = prog.block(parent_idx) - - return parent_block - - def _assert_in_rnn_block_(self, method): - if self.status != DynamicRNN.IN_RNN: - raise ValueError("{0} can only be invoked inside rnn block.".format( - method)) diff --git a/python/paddle/v2/fluid/layers/__init__.py b/python/paddle/v2/fluid/layers/__init__.py new file mode 100644 index 0000000000000000000000000000000000000000..249f570e13b7a1b50397fb971d1c6f77e0359a5e --- /dev/null +++ b/python/paddle/v2/fluid/layers/__init__.py @@ -0,0 +1,17 @@ +import ops +from ops import * +import nn +from nn import * +import io +from io import * +import tensor +from tensor import * +import control_flow +from control_flow import * + +__all__ = [] +__all__ += nn.__all__ +__all__ += io.__all__ +__all__ += tensor.__all__ +__all__ += control_flow.__all__ +__all__ += ops.__all__ diff --git a/python/paddle/v2/fluid/layers/control_flow.py b/python/paddle/v2/fluid/layers/control_flow.py new file mode 100644 index 0000000000000000000000000000000000000000..dc6c0e7f518ee47b332a501df803a2364e0cffc0 --- /dev/null +++ b/python/paddle/v2/fluid/layers/control_flow.py @@ -0,0 +1,965 @@ +from ..layer_helper import LayerHelper, unique_name +from ..framework import Program, Variable, Operator +from .. import core +from tensor import assign, fill_constant +import contextlib + +__all__ = [ + 'split_lod_tensor', 'merge_lod_tensor', 'BlockGuard', 'StaticRNNGuard', + 'StaticRNNMemoryLink', 'WhileGuard', 'While', 'lod_rank_table', + 'max_sequence_len', 'topk', 'lod_tensor_to_array', 'array_to_lod_tensor', + 'increment', 'array_write', 'create_array', 'less_than', 'array_read', + 'shrink_memory', 'array_length', 'IfElse', 'DynamicRNN', 'ConditionalBlock', + 'StaticRNN' +] + + +def split_lod_tensor(input, mask, level=0): + helper = LayerHelper('split_lod_tensor', **locals()) + out_true = helper.create_tmp_variable(dtype=input.dtype) + out_false = helper.create_tmp_variable(dtype=input.dtype) + helper.append_op( + type='split_lod_tensor', + inputs={ + 'X': input, + 'Mask': mask, + }, + outputs={'OutTrue': out_true, + 'OutFalse': out_false}, + attrs={'level': level}) + return out_true, out_false + + +def merge_lod_tensor(in_true, in_false, x, mask, level=0): + helper = LayerHelper('merge_lod_tensor', **locals()) + out = helper.create_tmp_variable(dtype=in_true.dtype) + helper.append_op( + type='merge_lod_tensor', + inputs={'X': x, + 'Mask': mask, + 'InTrue': in_true, + 'InFalse': in_false}, + outputs={'Out': out}, + attrs={'level': level}) + return out + + +class BlockGuard(object): + """ + BlockGuard class. + + BlockGuard class is used to create a sub-block in a program by + using the Python `with` keyword. + """ + + def __init__(self, main_program): + if not isinstance(main_program, Program): + raise TypeError("BlockGuard takes a program") + self.main_program = main_program + + def __enter__(self): + self.main_program.create_block() + + def __exit__(self, exc_type, exc_val, exc_tb): + self.main_program.rollback() + if exc_type is not None: + return False # re-raise exception + return True + + +class StaticRNNGuard(BlockGuard): + """ + StaticRNNGuard class. + + StaticRNNGuard class is used to create a StaticRNN block in a program. + """ + + def __init__(self, rnn): + if not isinstance(rnn, StaticRNN): + raise TypeError("StaticRNNGuard takes a StaticRNN") + super(StaticRNNGuard, self).__init__(rnn.helper.main_program) + self.rnn = rnn + + def __enter__(self): + self.rnn.status = StaticRNN.IN_RNN_BLOCK + return super(StaticRNNGuard, self).__enter__() + + def __exit__(self, exc_type, exc_val, exc_tb): + if exc_type is not None: + return False + self.rnn.status = StaticRNN.AFTER_RNN_BLOCK + self.rnn.complete_rnn_op() + return super(StaticRNNGuard, self).__exit__(exc_type, exc_val, exc_tb) + + +class StaticRNNMemoryLink(object): + """ + StaticRNNMemoryLink class. + + Args: + init: the initial variable for Memory + init: Variable + pre_mem: the memory variable in previous time step + pre_mem: Variable + mem: the memory variable in current time step + mem: Variable + + StaticRNNMemoryLink class is used to create a link between two + memory cells of a StaticRNN. + """ + + def __init__(self, init, pre_mem, mem=None): + self.init = init + self.pre_mem = pre_mem + self.mem = mem + + +class StaticRNN(object): + """ + StaticRNN class. + + StaticRNN class is used to create a StaticRNN. The RNN will have its + own parameters like inputs, outputs, memories, status and length. + """ + BEFORE_RNN_BLOCK = 0 + IN_RNN_BLOCK = 1 + AFTER_RNN_BLOCK = 2 + + def __init__(self, name=None): + self.helper = LayerHelper("static_rnn", name=name) + self.memories = {} # memory map, from pre_mem.name --> MemoryLink + self.inputs = [] # input variable list in current block + self.outputs = [] # output variable list in parent block + self.status = StaticRNN.BEFORE_RNN_BLOCK # status flag. + # sequence length, since it is a static RNN, sequence length are fixed. + self.seq_len = None + + def step(self): + return StaticRNNGuard(self) + + def _assert_in_rnn_block_(self, method): + if self.status != StaticRNN.IN_RNN_BLOCK: + raise ValueError("You must invoke {0} in rnn block".format(method)) + + def memory(self, + init=None, + shape=None, + batch_ref=None, + init_value=0.0, + init_batch_dim_idx=0, + ref_batch_dim_idx=1): + """ + Args: + init: boot memory, if not set, a shape, batch_ref must be provided + shape: shape of the boot memory + batch_ref: batch size reference variable + init_value: the init value of boot memory + init_batch_dim_idx: the index of batch size in init's dimension + ref_batch_dim_idx: the index of batch size in batch_ref's dimension + """ + self._assert_in_rnn_block_('memory') + if init is None: + if shape is None or batch_ref is None: + raise ValueError( + "if init is None, memory at least need shape and batch_ref") + parent_block = self.parent_block() + var_name = unique_name("@".join([self.helper.name, "memory_boot"])) + boot_var = parent_block.create_var( + name=var_name, + shape=shape, + dtype=batch_ref.dtype, + persistable=False) + + parent_block.append_op( + type="fill_constant_batch_size_like", + inputs={'Input': [batch_ref]}, + outputs={'Out': [boot_var]}, + attrs={ + 'value': init_value, + 'shape': boot_var.shape, + 'dtype': boot_var.dtype, + 'input_dim_idx': ref_batch_dim_idx, + 'output_dim_idx': init_batch_dim_idx + }) + + return self.memory(init=boot_var) + else: + pre_mem = self.helper.create_variable( + name=unique_name("@".join([self.helper.name, "mem"])), + dtype=init.dtype, + shape=init.shape) + self.memories[pre_mem.name] = StaticRNNMemoryLink( + init=init, pre_mem=pre_mem) + return pre_mem + + def step_input(self, x): + self._assert_in_rnn_block_('step_input') + if not isinstance(x, Variable): + raise TypeError("step input takes a Variable") + if self.seq_len is None: + self.seq_len = x.shape[0] + elif self.seq_len != x.shape[0]: + raise ValueError("Static RNN only take fix seq_len input") + + ipt = self.helper.create_variable( + name=x.name, dtype=x.dtype, shape=list(x.shape[1:]), type=x.type) + self.inputs.append(ipt) + return ipt + + def step_output(self, o): + self._assert_in_rnn_block_('step_output') + if not isinstance(o, Variable): + raise TypeError("step output takes a Variable") + + tmp_o = self.helper.create_tmp_variable(dtype=o.dtype) + self.helper.append_op( + type='rnn_memory_helper', + inputs={'X': [o]}, + outputs={'Out': tmp_o}, + attrs={'dtype': o.dtype}) + + out_var = self.parent_block().create_var( + name=tmp_o.name, + shape=[self.seq_len] + list(tmp_o.shape), + dtype=tmp_o.dtype) + + self.outputs.append(out_var) + + def output(self, *outputs): + for each in outputs: + self.step_output(each) + + def update_memory(self, mem, var): + if not isinstance(mem, Variable) or not isinstance(var, Variable): + raise TypeError("update memory should take variables") + self.memories[mem.name].mem = var + + def parent_block(self): + prog = self.helper.main_program + parent_idx = prog.current_block().parent_idx + assert parent_idx >= 0 + parent_block = prog.block(parent_idx) + return parent_block + + def __call__(self, *args, **kwargs): + if self.status != StaticRNN.AFTER_RNN_BLOCK: + raise ValueError("RNN output can only be retrieved after rnn block") + if len(self.outputs) == 0: + raise ValueError("RNN has no output") + elif len(self.outputs) == 1: + return self.outputs[0] + else: + return self.outputs + + def complete_rnn_op(self): + main_program = self.helper.main_program + rnn_block = main_program.current_block() + parent_block = self.parent_block() + + local_inputs = set() + + for op in rnn_block.ops: + assert isinstance(op, Operator) + for oname in op.output_names: + for out_var_name in op.output(oname): + local_inputs.add(out_var_name) + + for var in self.inputs: + local_inputs.add(var.name) + for m in self.memories: + local_inputs.add(m) + + params = list() + for op in rnn_block.ops: + assert isinstance(op, Operator) + for iname in op.input_names: + for in_var_name in op.input(iname): + if in_var_name not in local_inputs: + params.append(in_var_name) + + parameters = [parent_block.var(name) for name in params] + + step_scope = parent_block.create_var( + type=core.VarDesc.VarType.STEP_SCOPES) + + inlinks = [parent_block.var(i.name) for i in self.inputs] + outlinks = self.outputs + + boot_memories = [] + pre_memories = [] + memories = [] + for _, mem in self.memories.iteritems(): + boot_memories.append(mem.init) + pre_memories.append(mem.pre_mem.name) + mem_var = rnn_block.var(mem.mem.name) + assert isinstance(mem_var, Variable) + new_mem = self.helper.create_tmp_variable(dtype=mem_var.dtype) + + rnn_block.append_op( + type='rnn_memory_helper', + inputs={'X': [mem_var]}, + outputs={'Out': [new_mem]}, + attrs={'dtype': mem_var.dtype}) + + memories.append(new_mem.name) + + parent_block.append_op( + type='recurrent', + inputs={ + 'inputs': inlinks, + 'initial_states': boot_memories, + 'parameters': parameters + }, + outputs={'outputs': outlinks, + 'step_scopes': [step_scope]}, + attrs={ + 'ex_states': pre_memories, + 'states': memories, + 'sub_block': rnn_block + }) + + +class WhileGuard(BlockGuard): + def __init__(self, while_op): + if not isinstance(while_op, While): + raise TypeError("WhileGuard takes a while op") + super(WhileGuard, self).__init__(while_op.helper.main_program) + self.while_op = while_op + + def __enter__(self): + self.while_op.status = While.IN_WHILE_BLOCK + return super(WhileGuard, self).__enter__() + + def __exit__(self, exc_type, exc_val, exc_tb): + if exc_type is not None: + return False + self.while_op.status = While.AFTER_WHILE_BLOCK + self.while_op.complete() + return super(WhileGuard, self).__exit__(exc_type, exc_val, exc_tb) + + +class While(object): + BEFORE_WHILE_BLOCK = 0 + IN_WHILE_BLOCK = 1 + AFTER_WHILE_BLOCK = 2 + + def __init__(self, cond, name=None): + self.helper = LayerHelper("while", name=name) + self.status = While.BEFORE_WHILE_BLOCK + if not isinstance(cond, Variable): + raise TypeError("condition should be a variable") + assert isinstance(cond, Variable) + if cond.dtype != core.DataType.BOOL: + raise TypeError("condition should be a bool variable") + if reduce(lambda a, b: a * b, cond.shape, 1) != 1: + raise TypeError("condition should be a bool scalar") + self.cond_var = cond + + def block(self): + return WhileGuard(self) + + def complete(self): + main_program = self.helper.main_program + while_block = main_program.current_block() + parent_block = main_program.block(main_program.current_block() + .parent_idx) + + inner_outputs = {self.cond_var.name} + x_name_list = set() + for op in while_block.ops: + for iname in op.input_names: + for in_var_name in op.input(iname): + if in_var_name not in inner_outputs: + x_name_list.add(in_var_name) + + for oname in op.output_names: + for out_var_name in op.output(oname): + inner_outputs.add(out_var_name) + + out_vars = [] + for inner_out_name in inner_outputs: + if inner_out_name in parent_block.vars: + out_vars.append(parent_block.var(inner_out_name)) + + step_scope = parent_block.create_var( + type=core.VarDesc.VarType.STEP_SCOPES) + + parent_block.append_op( + type='while', + inputs={ + 'X': [parent_block.var(x_name) for x_name in x_name_list], + 'Condition': [self.cond_var] + }, + outputs={'Out': out_vars, + 'StepScopes': [step_scope]}, + attrs={'sub_block': while_block}) + + +def lod_rank_table(x, level=0): + """ + This function creates an operator for creating a LOD_RANK_TABLE + using the input x. + """ + helper = LayerHelper("lod_rank_table", **locals()) + table = helper.create_variable( + type=core.VarDesc.VarType.LOD_RANK_TABLE, + name=unique_name("lod_rank_table")) + helper.append_op( + type='lod_rank_table', + inputs={'X': x}, + outputs={'Out': table}, + attrs={'level': level}) + return table + + +def max_sequence_len(rank_table): + """ + This function creates an operator to calculate the length of + max seqence through input rank_table(should be a lod_rank_table) + """ + helper = LayerHelper("max_seqence_len", **locals()) + res = helper.create_tmp_variable(dtype="int64") + helper.append_op( + type="max_sequence_len", + inputs={"RankTable": rank_table}, + outputs={"Out": res}) + return res + + +def topk(input, k): + helper = LayerHelper('topk', **locals()) + topk_out = helper.create_tmp_variable(dtype=input.data_type) + topk_indices = helper.create_tmp_variable(dtype='int64') + helper.append_op( + type='top_k', + inputs={'X': [input]}, + outputs={'Out': [topk_out], + 'Indices': [topk_indices]}, + attrs={'k': k}) + return topk_out, topk_indices + + +def lod_tensor_to_array(x, table): + """ + This function creates an operator to convert an LOD_Tensor to + an array. + """ + helper = LayerHelper("lod_tensor_to_array", **locals()) + array = helper.create_variable( + name=unique_name("lod_tensor_to_array"), + type=core.VarDesc.VarType.LOD_TENSOR_ARRAY, + dtype=x.dtype) + helper.append_op( + type='lod_tensor_to_array', + inputs={'X': x, + 'RankTable': table}, + outputs={'Out': array}) + return array + + +def array_to_lod_tensor(x, table): + """ + This function creates an operator to convert an array to a + LOD_Tensor. + """ + helper = LayerHelper("array_to_lod_tensor", **locals()) + tmp = helper.create_tmp_variable(dtype=x.dtype) + helper.append_op( + type="array_to_lod_tensor", + inputs={'X': x, + 'RankTable': table}, + outputs={'Out': tmp}) + return tmp + + +def increment(x, value=1.0, in_place=True): + """ + This function creates an operator to increment each value in the input + `x` by an amount: `value` as mentioned in the input parameter. This + operation is performed in-place by default. + """ + helper = LayerHelper("increment", **locals()) + if not in_place: + out = helper.create_tmp_variable(dtype=x.dtype) + else: + out = x + helper.append_op( + type='increment', + inputs={'X': [x]}, + outputs={'Out': [out]}, + attrs={'step': float(value)}) + return out + + +def array_write(x, i, array=None): + """ + This function creates an operator to write the data out as a + LOD_TENSOR_ARRAY. + """ + helper = LayerHelper('array_write', **locals()) + if array is None: + array = helper.create_variable( + name="{0}.out".format(helper.name), + type=core.VarDesc.VarType.LOD_TENSOR_ARRAY, + dtype=x.dtype) + helper.append_op( + type='write_to_array', + inputs={'X': [x], + 'I': [i]}, + outputs={'Out': [array]}) + return array + + +def create_array(dtype): + helper = LayerHelper("array", **locals()) + return helper.create_variable( + name="{0}.out".format(helper.name), + type=core.VarDesc.VarType.LOD_TENSOR_ARRAY, + dtype=dtype) + + +def less_than(x, y, cond=None, **ignored): + helper = LayerHelper("less_than", **locals()) + if cond is None: + cond = helper.create_tmp_variable(dtype='bool') + cond.stop_gradient = True + + helper.append_op( + type='less_than', inputs={'X': [x], + 'Y': [y]}, outputs={'Out': [cond]}) + return cond + + +def array_read(array, i): + """ + This function creates an operator to read the data in as a + LOD_TENSOR_ARRAY. + """ + helper = LayerHelper('array_read', **locals()) + if not isinstance( + array, + Variable) or array.type != core.VarDesc.VarType.LOD_TENSOR_ARRAY: + raise TypeError("array should be tensor array vairable") + out = helper.create_tmp_variable(dtype=array.dtype) + helper.append_op( + type='read_from_array', + inputs={'X': [array], + 'I': [i]}, + outputs={'Out': [out]}) + return out + + +def shrink_memory(x, i, table): + """ + This function creates an operator to shrink_rnn_memory using the RankTable + as mentioned in the input parameter. + """ + helper = LayerHelper('shrink_memory', **locals()) + out = helper.create_tmp_variable(dtype=x.dtype) + helper.append_op( + type='shrink_rnn_memory', + inputs={'X': [x], + 'I': [i], + 'RankTable': [table]}, + outputs={'Out': [out]}, + attrs={}) + return out + + +def array_length(array): + """ + This function creates an operator to find the length of the + LOD_TENSOR_ARRAY. + """ + helper = LayerHelper('array_length', **locals()) + tmp = helper.create_tmp_variable(dtype='int64') + tmp.stop_gradient = True + helper.append_op( + type='lod_array_length', inputs={'X': [array]}, outputs={'Out': [tmp]}) + return tmp + + +class ConditionalBlockGuard(BlockGuard): + def __init__(self, block): + if not isinstance(block, ConditionalBlock): + raise TypeError("block should be conditional block") + super(ConditionalBlockGuard, self).__init__(block.helper.main_program) + self.block = block + + def __enter__(self): + return super(ConditionalBlockGuard, self).__enter__() + + def __exit__(self, exc_type, exc_val, exc_tb): + self.block.complete() + return super(ConditionalBlockGuard, self).__exit__(exc_type, exc_val, + exc_tb) + + +class ConditionalBlock(object): + def __init__(self, inputs, name=None): + for each_input in inputs: + if not isinstance(each_input, Variable): + raise TypeError("Each input should be variable") + self.inputs = inputs + self.helper = LayerHelper('conditional_block', name=name) + + def block(self): + return ConditionalBlockGuard(self) + + def complete(self): + inside_block = self.helper.main_program.current_block() + parent_block = self.helper.main_program.block(inside_block.parent_idx) + + intermediate = set() + params = set() + + for each_op in inside_block.ops: + assert isinstance(each_op, Operator) + for iname in each_op.input_names: + for in_var_name in each_op.input(iname): + if in_var_name not in intermediate: + params.add(in_var_name) + + for oname in each_op.output_names: + for out_var_name in each_op.output(oname): + intermediate.add(out_var_name) + input_set = set([ipt.name for ipt in self.inputs]) + + param_list = [ + parent_block.var(each_name) for each_name in params + if each_name not in input_set + ] + + out_list = [ + parent_block.var(var_name) for var_name in parent_block.vars + if var_name not in intermediate + ] + + step_scope = parent_block.create_var( + type=core.VarDesc.VarType.STEP_SCOPES) + parent_block.append_op( + type='conditional_block', + inputs={ + 'X': self.inputs, + 'Params': param_list, + }, + outputs={'Out': out_list, + 'Scope': [step_scope]}, + attrs={'sub_block': inside_block}) + + +class IfElseBlockGuard(object): + def __init__(self, is_true, ifelse): + if not isinstance(ifelse, IfElse): + raise TypeError("ifelse must be an instance of IfElse class") + + if ifelse.status != IfElse.OUT_IF_ELSE_BLOCKS: + raise ValueError("You cannot invoke IfElse.block() inside a block") + + self.is_true = is_true + self.ie = ifelse + if is_true: + self.cond_block = ifelse.conditional_true_block + else: + self.cond_block = ifelse.conditional_false_block + + if not isinstance(self.cond_block, ConditionalBlock): + raise TypeError("Unexpected situation") + + self.cond_block = self.cond_block.block() + + def __enter__(self): + self.ie.status = IfElse.IN_IF_ELSE_TRUE_BLOCKS if self.is_true else IfElse.IN_IF_ELSE_FALSE_BLOCKS + self.cond_block.__enter__() + + def __exit__(self, exc_type, exc_val, exc_tb): + if not self.cond_block.__exit__(exc_type, exc_val, exc_tb): + # re-raise inside exception + return False + if len(self.ie.output_table[1 if self.is_true else 0]) == 0: + raise ValueError("Must set output inside block") + self.ie.status = IfElse.OUT_IF_ELSE_BLOCKS + + +class IfElse(object): + OUT_IF_ELSE_BLOCKS = 0 + IN_IF_ELSE_TRUE_BLOCKS = 1 + IN_IF_ELSE_FALSE_BLOCKS = 2 + + def __init__(self, cond, name=None): + if not isinstance(cond, Variable): + raise TypeError("cond must be a Variable") + self.helper = LayerHelper('ifelse', name=name) + self.cond = cond + self.input_table = {} + self.status = IfElse.OUT_IF_ELSE_BLOCKS + self.conditional_true_block = ConditionalBlock(inputs=[self.cond]) + self.conditional_false_block = ConditionalBlock(inputs=[self.cond]) + self.output_table = ([], []) # (true_outs, false_outs) + + def input(self, x): + if self.status == IfElse.OUT_IF_ELSE_BLOCKS: + raise ValueError("input must in true/false blocks") + if id(x) not in self.input_table: + parent_block = self.parent_block() + out_true = parent_block.create_var( + name=unique_name('ifelse_input' + self.helper.name), + dtype=x.dtype) + + out_false = parent_block.create_var( + name=unique_name('ifelse_input' + self.helper.name), + dtype=x.dtype) + parent_block.append_op( + type='split_lod_tensor', + inputs={ + 'X': x, + 'Mask': self.cond, + }, + outputs={'OutTrue': out_true, + 'OutFalse': out_false}, + attrs={'level': 0}) + self.input_table[id(x)] = (out_true, out_false) + else: + out_true, out_false = self.input_table[id(x)] + + if self.status == IfElse.IN_IF_ELSE_TRUE_BLOCKS: + return out_true + else: + return out_false + + def parent_block(self): + current_block = self.helper.main_program.current_block() + return self.helper.main_program.block(current_block.parent_idx) + + def true_block(self): + return IfElseBlockGuard(True, self) + + def false_block(self): + return IfElseBlockGuard(False, self) + + def output(self, *outs): + if self.status == self.OUT_IF_ELSE_BLOCKS: + raise ValueError("output can only be invoked in the sub-block") + + out_table = self.output_table[1 if self.status == + self.IN_IF_ELSE_TRUE_BLOCKS else 0] + parent_block = self.parent_block() + for each_out in outs: + if not isinstance(each_out, Variable): + raise TypeError("Each output should be a variable") + # create outside tensor + outside_out = parent_block.create_var( + name=unique_name("_".join([self.helper.name, 'output'])), + dtype=each_out.dtype) + out_table.append(outside_out) + + # assign local var to outside + assign(input=each_out, output=outside_out) + + def __call__(self): + if self.status != self.OUT_IF_ELSE_BLOCKS: + raise ValueError("IfElse::__call__ must be out of sub-block") + false_len, true_len = map(len, self.output_table) + if false_len == 0 and true_len == 0: + raise ValueError("Must invoke true_block/false_block before " + "__call__") + elif false_len != true_len and false_len != 0 and true_len != 0: + raise ValueError("The output side must be same") + elif false_len == 0 or true_len == 0: + return self.output_table[0 if false_len != 0 else 1] + + # else none of false_len/true_len is zero + # merge together + rlist = [] + for false_var, true_var in zip(*self.output_table): + rlist.append( + merge_lod_tensor( + in_true=true_var, + in_false=false_var, + mask=self.cond, + x=self.cond, + level=0)) + return rlist + + +class DynamicRNN(object): + BEFORE_RNN = 0 + IN_RNN = 1 + AFTER_RNN = 2 + + def __init__(self, name=None): + self.helper = LayerHelper('dynamic_rnn', name=name) + self.status = DynamicRNN.BEFORE_RNN + self.lod_rank_table = None + self.max_seq_len = None + self.step_idx = None + self.zero_idx = fill_constant(shape=[1], value=0, dtype='int64') + self.mem_dict = dict() + self.output_array = [] + self.outputs = [] + self.cond = self.helper.create_tmp_variable(dtype='bool') + self.cond.stop_gradient = False + self.while_op = While(self.cond) + self.input_array = [] + self.mem_link = [] + + def step_input(self, x): + self._assert_in_rnn_block_("step_input") + if not isinstance(x, Variable): + raise TypeError( + "step_input() can only take a Variable as its input") + parent_block = self._parent_block_() + if self.lod_rank_table is None: + self.lod_rank_table = parent_block.create_var( + name=unique_name('lod_rank_table'), + type=core.VarDesc.VarType.LOD_RANK_TABLE) + self.lod_rank_table.stop_gradient = True + parent_block.append_op( + type='lod_rank_table', + inputs={"X": x}, + outputs={"Out": self.lod_rank_table}) + self.max_seq_len = parent_block.create_var( + name=unique_name('dynamic_rnn_max_seq_len'), dtype='int64') + self.max_seq_len.stop_gradient = False + parent_block.append_op( + type='max_sequence_len', + inputs={'RankTable': self.lod_rank_table}, + outputs={"Out": self.max_seq_len}) + self.cond.stop_gradient = True + parent_block.append_op( + type='less_than', + inputs={'X': self.step_idx, + 'Y': self.max_seq_len}, + outputs={'Out': self.cond}) + + input_array = parent_block.create_var( + name=unique_name('dynamic_rnn_input_array'), + type=core.VarDesc.VarType.LOD_TENSOR_ARRAY, + dtype=x.dtype) + self.input_array.append((input_array, x.dtype)) + parent_block.append_op( + type='lod_tensor_to_array', + inputs={'X': x, + 'RankTable': self.lod_rank_table}, + outputs={'Out': input_array}) + return array_read(array=input_array, i=self.step_idx) + + @contextlib.contextmanager + def block(self): + if self.status != DynamicRNN.BEFORE_RNN: + raise ValueError("rnn.block() can only be invoke once") + self.step_idx = fill_constant(shape=[1], dtype='int64', value=0) + self.step_idx.stop_gradient = False + self.status = DynamicRNN.IN_RNN + with self.while_op.block(): + yield + increment(x=self.step_idx, value=1.0, in_place=True) + + for new_mem, mem_array in self.mem_link: + array_write(x=new_mem, i=self.step_idx, array=mem_array) + + less_than(x=self.step_idx, y=self.max_seq_len, cond=self.cond) + + self.status = DynamicRNN.AFTER_RNN + for each_array in self.output_array: + self.outputs.append( + array_to_lod_tensor( + x=each_array, table=self.lod_rank_table)) + + def __call__(self, *args, **kwargs): + if self.status != DynamicRNN.AFTER_RNN: + raise ValueError( + "Dynamic RNN outputs can only be retrieved after rnn block") + if len(self.outputs) == 1: + return self.outputs[0] + else: + return self.outputs + + def memory(self, init=None, shape=None, value=0.0, dtype='float32'): + self._assert_in_rnn_block_('memory') + if init is not None: + if not isinstance(init, Variable): + raise TypeError( + "The input arg `init` of memory() must be a Variable") + parent_block = self._parent_block_() + mem_array = parent_block.create_var( + name=unique_name('dynamic_rnn_mem_array'), + type=core.VarDesc.VarType.LOD_TENSOR_ARRAY, + dtype=init.dtype) + parent_block.append_op( + type='write_to_array', + inputs={'X': init, + 'I': self.zero_idx}, + outputs={'Out': mem_array}) + retv = array_read(array=mem_array, i=self.step_idx) + retv = shrink_memory( + x=retv, i=self.step_idx, table=self.lod_rank_table) + self.mem_dict[retv.name] = mem_array + return retv + else: + if len(self.input_array) == 0: + raise ValueError( + "step_input should be invoked before memory(shape=..., value=...)" + ) + parent_block = self._parent_block_() + init = parent_block.create_var( + name=unique_name('mem_init'), dtype=dtype) + arr, dtype = self.input_array[0] + in0 = parent_block.create_var(name=unique_name('in0'), dtype=dtype) + parent_block.append_op( + type='read_from_array', + inputs={'X': [arr], + 'I': [self.zero_idx]}, + outputs={'Out': [in0]}) + parent_block.append_op( + type='fill_constant_batch_size_like', + inputs={'Input': [in0]}, + outputs={'Out': [init]}, + attrs={ + 'shape': [-1] + shape, + 'value': float(value), + 'dtype': init.dtype + }) + return self.memory(init=init) + + def update_memory(self, ex_mem, new_mem): + self._assert_in_rnn_block_('update_memory') + if not isinstance(ex_mem, Variable): + raise TypeError("The input arg `ex_mem` of update_memory() must " + "be a Variable") + if not isinstance(new_mem, Variable): + raise TypeError("The input arg `new_mem` of update_memory() must " + "be a Variable") + + mem_array = self.mem_dict.get(ex_mem.name, None) + if mem_array is None: + raise ValueError("Please invoke memory before update_memory") + if self.lod_rank_table is None: + raise ValueError("Please invoke step_input before update_memory") + + self.mem_link.append((new_mem, mem_array)) + + def output(self, *outputs): + self._assert_in_rnn_block_('output') + parent_block = self._parent_block_() + for each in outputs: + outside_array = parent_block.create_var( + name=unique_name("_".join( + [self.helper.name, "output_array", each.name])), + type=core.VarDesc.VarType.LOD_TENSOR_ARRAY, + dtype=each.dtype) + array_write(x=each, i=self.step_idx, array=outside_array) + self.output_array.append(outside_array) + + def _parent_block_(self): + prog = self.helper.main_program + parent_idx = prog.current_block().parent_idx + assert parent_idx >= 0 + parent_block = prog.block(parent_idx) + + return parent_block + + def _assert_in_rnn_block_(self, method): + if self.status != DynamicRNN.IN_RNN: + raise ValueError("{0} can only be invoked inside rnn block.".format( + method)) diff --git a/python/paddle/v2/fluid/layers/io.py b/python/paddle/v2/fluid/layers/io.py new file mode 100644 index 0000000000000000000000000000000000000000..f4c5907f48b46ee5d9bcaba48370e5baf036c615 --- /dev/null +++ b/python/paddle/v2/fluid/layers/io.py @@ -0,0 +1,55 @@ +from .. import core +from ..layer_helper import LayerHelper + +__all__ = ['data'] + + +def data(name, + shape, + append_batch_size=True, + dtype='float32', + lod_level=0, + type=core.VarDesc.VarType.LOD_TENSOR, + stop_gradient=True): + """ + Data Layer. + + Args: + name: The name/alias of the function + shape: Tuple declaring the shape. + append_batch_size: Whether or not to append the data as a batch. + dtype: The type of data : float32, float_16, int etc + type: The output type. By default it is LOD_TENSOR. + lod_level(int): The LoD Level. 0 means the input data is not a sequence. + main_program: Name of the main program that calls this + startup_program: Name of the startup program + stop_gradient: A boolean that mentions whether gradient should flow. + + This function takes in input and based on whether data has + to be returned back as a minibatch, it creates the global variable using + the helper functions. The global variables can be accessed by all the + following operations and layers in the graph. + + All the input variables of this function are passed in as local variables + to the LayerHelper constructor. + + """ + helper = LayerHelper('data', **locals()) + shape = list(shape) + for i in xrange(len(shape)): + if shape[i] is None: + shape[i] = -1 + append_batch_size = False + elif shape[i] < 0: + append_batch_size = False + + if append_batch_size: + shape = [-1] + shape # append batch size as -1 + + return helper.create_global_variable( + name=name, + shape=shape, + dtype=dtype, + type=type, + stop_gradient=stop_gradient, + lod_level=lod_level) diff --git a/python/paddle/v2/fluid/layers/nn.py b/python/paddle/v2/fluid/layers/nn.py new file mode 100644 index 0000000000000000000000000000000000000000..5863957c5fb6f65ae299e2203bae324283c850e7 --- /dev/null +++ b/python/paddle/v2/fluid/layers/nn.py @@ -0,0 +1,828 @@ +""" +All layers just related to the neural network. +""" + +from ..layer_helper import LayerHelper +from ..initializer import Normal, Constant +from ..framework import Variable + +__all__ = [ + 'fc', 'embedding', 'dynamic_lstm', 'gru_unit', 'linear_chain_crf', + 'crf_decoding', 'cos_sim', 'cross_entropy', 'square_error_cost', 'accuracy', + 'chunk_eval', 'sequence_conv', 'conv2d', 'sequence_pool', 'pool2d', + 'batch_norm', 'beam_search_decode', 'conv2d_transpose', 'sequence_expand' +] + + +def fc(input, + size, + num_flatten_dims=1, + param_attr=None, + bias_attr=None, + act=None, + name=None): + """ + Fully Connected Layer. + + Args: + input: The input tensor to the function + size: The size of the layer + num_flatten_dims: Number of columns in input + param_attr: The parameters/weights to the FC Layer + param_initializer: Initializer used for the weight/parameter. If None, XavierInitializer() is used + bias_attr: The bias parameter for the FC layer + bias_initializer: Initializer used for the bias. If None, then ConstantInitializer() is used + act: Activation to be applied to the output of FC layer + name: Name/alias of the function + main_program: Name of the main program that calls this + startup_program: Name of the startup program + + This function can take in multiple inputs and performs the Fully Connected + function (linear transformation) on top of each of them. + So for input x, the output will be : Wx + b. Where W is the parameter, + b the bias and x is the input. + + The function also applies an activation (non-linearity) on top of the + output, if activation is passed in the input. + + All the input variables of this function are passed in as local variables + to the LayerHelper constructor. + + """ + helper = LayerHelper('fc', **locals()) + + dtype = helper.input_dtype() + + mul_results = [] + for input_var, param_attr in helper.iter_inputs_and_params(): + input_shape = input_var.shape + param_shape = [ + reduce(lambda a, b: a * b, input_shape[num_flatten_dims:], 1) + ] + [size] + w = helper.create_parameter( + attr=param_attr, shape=param_shape, dtype=dtype, is_bias=False) + tmp = helper.create_tmp_variable(dtype) + helper.append_op( + type="mul", + inputs={ + "X": input_var, + "Y": w, + }, + outputs={"Out": tmp}, + attrs={'x_num_col_dims': num_flatten_dims, + 'y_num_col_dims': 1}) + mul_results.append(tmp) + + # sum + if len(mul_results) == 1: + pre_bias = mul_results[0] + else: + pre_bias = helper.create_tmp_variable(dtype) + helper.append_op( + type="sum", inputs={"X": mul_results}, outputs={"Out": pre_bias}) + # add bias + pre_activation = helper.append_bias_op(pre_bias) + # add activation + return helper.append_activation(pre_activation) + + +def embedding(input, size, is_sparse=False, param_attr=None, dtype='float32'): + """ + Embedding Layer. + + Args: + param_initializer: + input: The input to the function + size: The size of the layer + is_sparse: A flag that decleares whether the input is sparse + param_attr: Parameters for this layer + dtype: The type of data : float32, float_16, int etc + main_program: Name of the main program that calls this + startup_program: Name of the startup program + + This function can take in the input (which is a vector of IDs) and + performs a lookup in the lookup_table using these IDs, to result into + the embedding of each ID in the input. + + All the input variables of this function are passed in as local variables + to the LayerHelper constructor. + + """ + + helper = LayerHelper('embedding', **locals()) + w = helper.create_parameter( + attr=helper.param_attr, shape=size, dtype=dtype, is_bias=False) + tmp = helper.create_tmp_variable(dtype) + helper.append_op( + type='lookup_table', + inputs={'Ids': input, + 'W': w}, + outputs={'Out': tmp}, + attrs={'is_sparse': is_sparse}) + return tmp + + +# TODO(qijun): expose H0 and C0 +def dynamic_lstm(input, + size, + param_attr=None, + bias_attr=None, + use_peepholes=True, + is_reverse=False, + gate_activation='sigmoid', + cell_activation='tanh', + candidate_activation='tanh', + dtype='float32'): + helper = LayerHelper('lstm', **locals()) + size = size / 4 + weight = helper.create_parameter( + attr=helper.param_attr, shape=[size, 4 * size], dtype=dtype) + bias_size = [1, 7 * size] + if not use_peepholes: + bias_size[1] = 4 * size + bias = helper.create_parameter( + attr=helper.bias_attr, shape=bias_size, dtype=dtype, is_bias=True) + + hidden = helper.create_tmp_variable(dtype) + cell = helper.create_tmp_variable(dtype) + batch_gate = helper.create_tmp_variable(dtype) + batch_cell_pre_act = helper.create_tmp_variable(dtype) + + helper.append_op( + type='lstm', + inputs={'Input': input, + 'Weight': weight, + 'Bias': bias}, + outputs={ + 'Hidden': hidden, + 'Cell': cell, + 'BatchGate': batch_gate, + 'BatchCellPreAct': batch_cell_pre_act + }, + attrs={ + 'use_peepholes': use_peepholes, + 'is_reverse': is_reverse, + 'gate_activation': gate_activation, + 'cell_activation': cell_activation, + 'candidate_activation': candidate_activation + }) + return hidden, cell + + +def gru_unit(input, + hidden, + size, + weight=None, + bias=None, + activation='tanh', + gate_activation='sigmoid'): + """ + GRUUnit Operator implements partial calculations of the GRU unit as following: + + $$ + update \ gate: u_t = actGate(xu_t + W_u * h_{t-1} + b_u) \\ + reset \ gate: r_t = actGate(xr_t + W_r * h_{t-1} + b_r) \\ + output \ candidate: {h}_t = actNode(xc_t + W_c * dot(r_t, h_{t-1}) + b_c) \\ + output: h_t = dot((1 - u_t), h_{t-1}) + dot(u_t, {h}_t) + $$ + + which is same as one time step of GRU Operator. + + @note To implement the complete GRU unit, fully-connected operator must be + used before to feed xu, xr and xc as the Input of GRUUnit operator. + + TODO(ChunweiYan) add more document here + """ + activation_dict = dict( + identity=0, + sigmoid=1, + tanh=2, + relu=3, ) + activation = activation_dict[activation] + gate_activation = activation_dict[gate_activation] + + helper = LayerHelper('gru_unit', **locals()) + dtype = helper.input_dtype() + size = size / 3 + + # create weight + if weight is None: + weight = helper.create_parameter( + attr=helper.param_attr, shape=[size, 3 * size], dtype=dtype) + + # create bias + if bias is None: + bias_size = [1, 3 * size] + bias = helper.create_parameter( + attr=helper.bias_attr, shape=bias_size, dtype=dtype, is_bias=True) + + gate = helper.create_tmp_variable(dtype) + reset_hidden_pre = helper.create_tmp_variable(dtype) + updated_hidden = helper.create_tmp_variable(dtype) + + helper.append_op( + type='gru_unit', + inputs={'Input': input, + 'HiddenPrev': hidden, + 'Weight': weight}, + outputs={ + 'Gate': gate, + 'ResetHiddenPrev': reset_hidden_pre, + 'Hidden': updated_hidden, + }, + attrs={ + 'activation': 0, + 'gate_activation': 1, + }) + + return updated_hidden, reset_hidden_pre, gate + + +def linear_chain_crf(input, label, param_attr=None): + helper = LayerHelper('linear_chain_crf', **locals()) + size = input.shape[1] + transition = helper.create_parameter( + attr=helper.param_attr, + shape=[size + 2, size], + dtype=helper.input_dtype()) + alpha = helper.create_tmp_variable(dtype=helper.input_dtype()) + emission_exps = helper.create_tmp_variable(dtype=helper.input_dtype()) + transition_exps = helper.create_tmp_variable(dtype=helper.input_dtype()) + log_likelihood = helper.create_tmp_variable(dtype=helper.input_dtype()) + helper.append_op( + type='linear_chain_crf', + inputs={"Emission": [input], + "Transition": transition, + "Label": label}, + outputs={ + "Alpha": [alpha], + "EmissionExps": [emission_exps], + "TransitionExps": transition_exps, + "LogLikelihood": log_likelihood + }) + + return log_likelihood + + +def crf_decoding(input, param_attr, label=None): + helper = LayerHelper('crf_decoding', **locals()) + transition = helper.get_parameter(param_attr.name) + viterbi_path = helper.create_tmp_variable(dtype=helper.input_dtype()) + helper.append_op( + type='crf_decoding', + inputs={"Emission": [input], + "Transition": transition, + "Label": label}, + outputs={"ViterbiPath": [viterbi_path]}) + + return viterbi_path + + +def cos_sim(X, Y, **kwargs): + """ + This function performs the cosine similarity between two tensors + X and Y and returns that as the output. + """ + helper = LayerHelper('cos_sim', **kwargs) + out = helper.create_tmp_variable(dtype=X.dtype) + xnorm = helper.create_tmp_variable(dtype=X.dtype) + ynorm = helper.create_tmp_variable(dtype=X.dtype) + helper.append_op( + type='cos_sim', + inputs={'X': [X], + 'Y': [Y]}, + outputs={'Out': [out], + 'XNorm': [xnorm], + 'YNorm': [ynorm]}) + return out + + +def cross_entropy(input, label, **kwargs): + """ + This function computes cross_entropy using the input and label. + """ + helper = LayerHelper('cross_entropy', **kwargs) + out = helper.create_tmp_variable(dtype=input.dtype) + helper.append_op( + type='cross_entropy', + inputs={'X': [input], + 'Label': [label]}, + outputs={'Y': [out]}, + attrs=kwargs) + return out + + +def square_error_cost(input, label, **kwargs): + """ + This functions returns the squared error cost using the input and label. + The output is appending the op to do the above. + """ + helper = LayerHelper('square_error_cost', **kwargs) + minus_out = helper.create_tmp_variable(dtype=input.dtype) + helper.append_op( + type='elementwise_sub', + inputs={'X': [input], + 'Y': [label]}, + outputs={'Out': [minus_out]}) + + square_out = helper.create_tmp_variable(dtype=input.dtype) + helper.append_op( + type='square', inputs={'X': [minus_out]}, outputs={'Y': [square_out]}) + return square_out + + +def accuracy(input, label, k=1, correct=None, total=None, **kwargs): + """ + This function computes the accuracy using the input and label. + The output is the top_k inputs and their indices. + """ + helper = LayerHelper("accuracy", **kwargs) + topk_out = helper.create_tmp_variable(dtype=input.dtype) + topk_indices = helper.create_tmp_variable(dtype="int64") + helper.append_op( + type="top_k", + inputs={"X": [input]}, + outputs={"Out": [topk_out], + "Indices": [topk_indices]}, + attrs={"k": k}) + acc_out = helper.create_tmp_variable(dtype="float32") + if correct is None: + correct = helper.create_tmp_variable(dtype="int64") + if total is None: + total = helper.create_tmp_variable(dtype="int64") + helper.append_op( + type="accuracy", + inputs={ + "Out": [topk_out], + "Indices": [topk_indices], + "Label": [label] + }, + outputs={ + "Accuracy": [acc_out], + "Correct": [correct], + "Total": [total], + }) + return acc_out + + +def chunk_eval(input, + label, + chunk_scheme, + num_chunk_types, + excluded_chunk_types=None, + **kwargs): + """ + This function computes and outputs the precision, recall and + F1-score of chunk detection. + """ + helper = LayerHelper("chunk_eval", **kwargs) + + # prepare output + precision = helper.create_tmp_variable(dtype="float32") + recall = helper.create_tmp_variable(dtype="float32") + f1_score = helper.create_tmp_variable(dtype="float32") + num_infer_chunks = helper.create_tmp_variable(dtype="int64") + num_label_chunks = helper.create_tmp_variable(dtype="int64") + num_correct_chunks = helper.create_tmp_variable(dtype="int64") + + helper.append_op( + type="chunk_eval", + inputs={"Inference": [input], + "Label": [label]}, + outputs={ + "Precision": [precision], + "Recall": [recall], + "F1-Score": [f1_score], + "NumInferChunks": [num_infer_chunks], + "NumLabelChunks": [num_label_chunks], + "NumCorrectChunks": [num_correct_chunks] + }, + attrs={ + "num_chunk_types": num_chunk_types, + 'chunk_scheme': chunk_scheme, + 'excluded_chunk_types': excluded_chunk_types or [] + }) + return precision, recall, f1_score, num_infer_chunks, num_label_chunks, num_correct_chunks + + +def sequence_conv(input, + num_filters, + filter_size=3, + filter_stride=1, + padding=None, + bias_attr=None, + param_attr=None, + act=None): + """ + This function creates the op for sequence_conv, using the inputs and + other convolutional configurations for the filters and stride as given + in the input parameters to the function. + """ + + # FIXME(dzh) : want to unify the argument of python layer + # function. So we ignore some unecessary attributes. + # such as, padding_trainable, context_start. + + helper = LayerHelper('sequence_conv', **locals()) + dtype = helper.input_dtype() + filter_shape = [filter_size * input.shape[1], num_filters] + filter_param = helper.create_parameter( + attr=helper.param_attr, shape=filter_shape, dtype=dtype) + pre_bias = helper.create_tmp_variable(dtype) + + helper.append_op( + type='sequence_conv', + inputs={ + 'X': [input], + 'Filter': [filter_param], + }, + outputs={"Out": pre_bias}, + attrs={ + 'contextStride': filter_stride, + 'contextStart': -int(filter_size / 2), + 'contextLength': filter_size + }) + pre_act = helper.append_bias_op(pre_bias) + return helper.append_activation(pre_act) + + +def conv2d(input, + num_filters, + filter_size, + stride=None, + padding=None, + groups=None, + param_attr=None, + bias_attr=None, + act=None, + name=None): + """ + This function creates the op for a 2-dimensional Convolution. + This is performed using the parameters of filters(size, dimensionality etc) + , stride and other configurations for a Convolution operation. + This funciton can also append an activation on top of the + conv-2d output, if mentioned in the input parameters. + """ + + if stride is None: + stride = [1, 1] + helper = LayerHelper('conv2d', **locals()) + dtype = helper.input_dtype() + + num_channels = input.shape[1] + if groups is None: + num_filter_channels = num_channels + else: + if num_channels % groups != 0: + raise ValueError("num_channels must be divisible by groups.") + num_filter_channels = num_channels / groups + + if isinstance(filter_size, int): + filter_size = [filter_size, filter_size] + if isinstance(stride, int): + stride = [stride, stride] + if isinstance(padding, int): + padding = [padding, padding] + + input_shape = input.shape + filter_shape = [num_filters, num_filter_channels] + filter_size + + def _get_default_param_initializer(): + std = (2.0 / (filter_size[0]**2 * num_channels))**0.5 + return Normal(0.0, std, 0) + + filter_param = helper.create_parameter( + attr=helper.param_attr, + shape=filter_shape, + dtype=dtype, + default_initializer=_get_default_param_initializer()) + + pre_bias = helper.create_tmp_variable(dtype) + + helper.append_op( + type='conv2d_cudnn', + inputs={ + 'Input': input, + 'Filter': filter_param, + }, + outputs={"Output": pre_bias}, + attrs={'strides': stride, + 'paddings': padding, + 'groups': groups}) + + pre_act = helper.append_bias_op(pre_bias, dim_start=1, dim_end=2) + + return helper.append_activation(pre_act) + + +def sequence_pool(input, pool_type, **kwargs): + """ + This function add the operator for sequence pooling. + This is applied on top of the input using pool_type mentioned + in the parameters. + """ + helper = LayerHelper('sequence_pool', input=input, **kwargs) + dtype = helper.input_dtype() + pool_out = helper.create_tmp_variable(dtype) + max_index = helper.create_tmp_variable(dtype) + + helper.append_op( + type="sequence_pool", + inputs={"X": input}, + outputs={"Out": pool_out, + "MaxIndex": max_index}, + attrs={"pooltype": pool_type.upper()}) + + return pool_out + + +def pool2d(input, + pool_size, + pool_type, + pool_stride=None, + pool_padding=None, + global_pooling=False): + """ + This function adds the operator for pooling in 2 dimensions, using the + pooling configurations mentioned in input parameters. + """ + if pool_padding is None: + pool_padding = [0, 0] + if pool_stride is None: + pool_stride = [1, 1] + if pool_type not in ["max", "avg"]: + raise ValueError( + "Unknown pool_type: '%s'. It can only be 'max' or 'avg'.", + str(pool_type)) + if isinstance(pool_size, int): + pool_size = [pool_size, pool_size] + if isinstance(pool_stride, int): + pool_stride = [pool_stride, pool_stride] + if isinstance(pool_padding, int): + pool_padding = [pool_padding, pool_padding] + + helper = LayerHelper('pool2d', **locals()) + dtype = helper.input_dtype() + pool_out = helper.create_tmp_variable(dtype) + + helper.append_op( + type="pool2d", + inputs={"X": input}, + outputs={"Out": pool_out}, + attrs={ + "pooling_type": pool_type, + "ksize": pool_size, + "global_pooling": global_pooling, + "strides": pool_stride, + "paddings": pool_padding + }) + + return pool_out + + +def batch_norm(input, + act=None, + is_test=False, + momentum=0.9, + epsilon=1e-05, + param_attr=None, + bias_attr=None, + data_layout='NCHW'): + """ + This function helps create an operator to implement + the BatchNorm layer using the configurations from the input parameters. + """ + helper = LayerHelper('batch_norm', **locals()) + dtype = helper.input_dtype() + + input_shape = input.shape + if data_layout == 'NCHW': + channel_num = input_shape[1] + else: + if data_layout == 'NHWC': + channel_num = input_shape[-1] + else: + raise ValueError("unsupported data layout:" + data_layout) + + param_shape = [channel_num] + + # create parameter + scale = helper.create_parameter( + attr=helper.param_attr, + shape=param_shape, + dtype=dtype, + default_initializer=Constant(1.0)) + + bias = helper.create_parameter( + attr=helper.param_attr, shape=param_shape, dtype=dtype, is_bias=True) + + mean = helper.create_global_variable( + dtype=input.dtype, shape=param_shape, persistable=True) + helper.set_variable_initializer(var=mean, initializer=Constant(0.0)) + + variance = helper.create_global_variable( + dtype=input.dtype, shape=param_shape, persistable=True) + helper.set_variable_initializer(var=variance, initializer=Constant(1.0)) + + # create output + # mean and mean_out share the same memory + mean_out = mean + # variance and variance out share the same memory + variance_out = variance + saved_mean = helper.create_tmp_variable(dtype) + saved_variance = helper.create_tmp_variable(dtype) + + batch_norm_out = helper.create_tmp_variable(dtype) + + helper.append_op( + type="batch_norm", + inputs={ + "X": input, + "Scale": scale, + "Bias": bias, + "Mean": mean, + "Variance": variance + }, + outputs={ + "Y": batch_norm_out, + "MeanOut": mean_out, + "VarianceOut": variance_out, + "SavedMean": saved_mean, + "SavedVariance": saved_variance + }, + attrs={"momentum": momentum, + "epsilon": epsilon, + "is_test": is_test}) + + return helper.append_activation(batch_norm_out) + + +def beam_search_decode(ids, scores): + helper = LayerHelper('beam_search_decode', **locals()) + sentence_ids = helper.create_tmp_variable(dtype=ids.dtype) + sentence_scores = helper.create_tmp_variable(dtype=ids.dtype) + + helper.append_op( + type="beam_search_decode", + inputs={"Ids": ids, + "Scores": scores}, + outputs={ + "SentenceIds": sentence_ids, + "SentenceScores": sentence_scores + }) + + return sentence_ids, sentence_scores + + +def conv2d_transpose(input, + num_filters, + output_size=None, + filter_size=None, + padding=None, + stride=None, + param_attr=None): + """ + The transpose of conv2d layer. + + This layer is also known as deconvolution layer. + + Args: + input(Variable): The input image with [N, C, H, W] format. + num_filters(int): The number of filter. It is as same as the output + image channel. + output_size(int|tuple|None): The output image size. If output size is a + tuple, it must contain two integers, (image_H, image_W). This + parameter only works when filter_size is None. + filter_size(int|tuple|None): The filter size. If filter_size is a tuple, + it must contain two integers, (filter_size_H, filter_size_W). + Otherwise, the filter will be a square. None if use output size to + calculate filter_size + padding(int|tuple): The padding size. If padding is a tuple, it must + contain two integers, (padding_H, padding_W). Otherwise, the + padding_H = padding_W = padding. + stride(int|tuple): The stride size. If stride is a tuple, it must + contain two integers, (stride_H, stride_W). Otherwise, the + stride_H = stride_W = stride. + param_attr: Parameter Attribute. + main_program(Program): the main program + startup_program(Program): the startup program + + Returns: + Variable: Output image. + """ + helper = LayerHelper("conv2d_transpose", **locals()) + if not isinstance(input, Variable): + raise TypeError("Input of conv2d_transpose must be Variable") + input_channel = input.shape[1] + + op_attr = dict() + + if isinstance(padding, int): + op_attr['paddings'] = [padding, padding] + elif padding is not None: + op_attr['paddings'] = padding + + if isinstance(stride, int): + op_attr['strides'] = stride + elif stride is not None: + op_attr['strides'] = stride + + if filter_size is None: + if output_size is None: + raise ValueError("output_size must be set when filter_size is None") + if isinstance(output_size, int): + output_size = [output_size, output_size] + + padding = op_attr.get('paddings', [0, 0]) + stride = op_attr.get('strides', [1, 1]) + + h_in = input.shape[2] + w_in = input.shape[3] + filter_size_h = output_size[0] - \ + (h_in - 1) * stride[0] + 2 * padding[0] + filter_size_w = output_size[1] - \ + (w_in - 1) * stride[1] + 2 * padding[1] + filter_size = [filter_size_h, filter_size_w] + elif isinstance(filter_size, int): + filter_size = [filter_size, filter_size] + + filter_shape = [input_channel, num_filters] + filter_size + img_filter = helper.create_parameter( + dtype=input.dtype, shape=filter_shape, attr=helper.param_attr) + + out = helper.create_tmp_variable(dtype=input.dtype) + helper.append_op( + type='conv2d_transpose', + inputs={'Input': [input], + 'Filter': [img_filter]}, + outputs={'Output': out}, + attrs=op_attr) + + return out + + +def sequence_expand(x, y, main_program=None, startup_program=None): + """Sequence Expand Layer. This layer will expand the input variable **x** + according to LoD information of **y**. And the following examples will + explain how sequence_expand works: + + .. code-block:: text + + * Case 1 + x is a LoDTensor: + x.lod = [[0, 2, 3], + [0, 1, 3, 4]] + x.data = [a, b, c, d] + x.dims = [4, 1] + + y is a LoDTensor: + y.lod = [[0, 2, 4], + [0, 3, 6, 7, 8]] + + with condition len(y.lod[-1]) - 1 == x.dims[0] + + then output is a 2-level LoDTensor: + out.lod = [[0, 2, 4], + [0, 3, 6, 7, 8]] + out.data = [a, a, a, b, b, b, c, d] + out.dims = [8, 1] + + * Case 2 + x is a Tensor: + x.data = [a, b, c] + x.dims = [3, 1] + + y is a LoDTensor: + y.lod = [[0, 2, 3, 6]] + + with condition len(y.lod[-1]) - 1 == x.dims[0] + + then output is a 1-level LoDTensor: + out.lod = [[0, 2, 3, 6]] + out.data = [a, a, b, c, c, c] + out.dims = [6, 1] + + Args: + x (Variable): The input variable which is a Tensor or LoDTensor. + y (Variable): The input variable which is a LoDTensor. + main_program (Program): The main program. + startup_program (Program): The startup program. + + Returns: + Variable: The expanded variable which is a LoDTensor. + + Examples: + .. code-block:: python + + x = fluid.layers.data(name='x', shape=[10], dtype='float32') + y = fluid.layers.data(name='y', shape=[10, 20], + dtype='float32', lod_level=1) + out = layers.sequence_expand(x=x, y=y) + """ + helper = LayerHelper('sequence_expand', input=x, **locals()) + dtype = helper.input_dtype() + tmp = helper.create_tmp_variable(dtype) + helper.append_op( + type='sequence_expand', inputs={'X': x, + 'Y': y}, outputs={'Out': tmp}) + return tmp diff --git a/python/paddle/v2/fluid/layers/ops.py b/python/paddle/v2/fluid/layers/ops.py new file mode 100644 index 0000000000000000000000000000000000000000..fa312ace60390e5fdd9637dccc71ccf8b247ca47 --- /dev/null +++ b/python/paddle/v2/fluid/layers/ops.py @@ -0,0 +1,9 @@ +from ..registry import register_layer +__all__ = [ + 'mean', 'mul', 'dropout', 'reshape', 'sigmoid', 'scale', 'transpose', + 'sigmoid_cross_entropy_with_logits', 'elementwise_add', 'elementwise_div', + 'elementwise_sub', 'elementwise_mul', 'clip', 'abs' +] + +for _OP in set(__all__): + globals()[_OP] = register_layer(_OP) diff --git a/python/paddle/v2/fluid/layers/tensor.py b/python/paddle/v2/fluid/layers/tensor.py new file mode 100644 index 0000000000000000000000000000000000000000..bda017b141dcba5ac268c34388742c433a533337 --- /dev/null +++ b/python/paddle/v2/fluid/layers/tensor.py @@ -0,0 +1,123 @@ +from ..layer_helper import LayerHelper + +__all__ = [ + 'create_tensor', 'cast', 'concat', 'sums', 'assign', + 'fill_constant_batch_size_like', 'fill_constant', 'ones', 'zeros' +] + + +def create_tensor(dtype, name=None): + helper = LayerHelper("create_tensor", **locals()) + return helper.create_variable(name=helper.name, dtype=dtype) + + +def cast(x, dtype): + """ + This function takes in the input with input_dtype + and casts it to the output_dtype as the output. + """ + helper = LayerHelper('cast', **locals()) + out = helper.create_tmp_variable(dtype=dtype) + helper.append_op( + type='cast', + inputs={'X': [x]}, + outputs={'Out': [out]}, + attrs={'in_dtype': x.dtype, + 'out_dtype': out.dtype}) + return out + + +def concat(input, axis): + """ + This function concats the input along the axis mentioned + and returns that as the output. + """ + helper = LayerHelper('concat', **locals()) + out = helper.create_tmp_variable(dtype=helper.input_dtype()) + helper.append_op( + type='concat', + inputs={'X': input}, + outputs={'Out': [out]}, + attrs={'axis': axis}) + return out + + +def sums(input, out=None): + """ + This function takes in the input and performs the sum operation on it + and returns that as the output. + """ + helper = LayerHelper('sum', **locals()) + if out is None: + out = helper.create_tmp_variable(dtype=helper.input_dtype()) + helper.append_op(type='sum', inputs={'X': input}, outputs={'Out': out}) + return out + + +def assign(input, output): + helper = LayerHelper('assign', **locals()) + helper.append_op( + type='scale', + inputs={'X': [input]}, + outputs={'Out': [output]}, + attrs={'scale': 1.0}) + return output + + +def fill_constant(shape, dtype, value, out=None): + """ + This function creates a tensor , with shape as mentioned in the input and + specified dtype and fills this up with a constant value that + comes in the input. It also sets the stop_gradient to be True. + """ + helper = LayerHelper("fill_constant", **locals()) + if out is None: + out = helper.create_tmp_variable(dtype=dtype) + helper.append_op( + type='fill_constant', + inputs={}, + outputs={'Out': [out]}, + attrs={'shape': shape, + 'dtype': out.dtype, + 'value': float(value)}) + out.stop_gradient = True + return out + + +def fill_constant_batch_size_like(input, + shape, + dtype, + value, + input_dim_idx=0, + output_dim_idx=0): + helper = LayerHelper("fill_constant_batch_size_like", **locals()) + out = helper.create_tmp_variable(dtype=dtype) + helper.append_op( + type='fill_constant_batch_size_like', + inputs={'Input': input}, + outputs={'Out': [out]}, + attrs={ + 'shape': shape, + 'dtype': out.dtype, + 'value': float(value), + 'input_dim_idx': input_dim_idx, + 'output_dim_idx': output_dim_idx + }) + out.stop_gradient = True + return out + + +def ones(shape, dtype): + """ + This function performs the same function as fill_constant() declared above + with the constant value being 1.0. + """ + return fill_constant(value=1.0, **locals()) + + +def zeros(shape, dtype): + """ + This function performs the same function as fill_constant() declared above + with the constant value being 0.0. + """ + return fill_constant(value=0.0, **locals()) diff --git a/python/paddle/v2/fluid/nets.py b/python/paddle/v2/fluid/nets.py index 7ef524318e637604cc22ba9d8d7cafe1b7505261..54886a8f2cc63474fe82290c0a12771b4cbdba72 100644 --- a/python/paddle/v2/fluid/nets.py +++ b/python/paddle/v2/fluid/nets.py @@ -10,25 +10,19 @@ def simple_img_conv_pool(input, pool_stride, act, param_attr=None, - pool_type='max', - main_program=None, - startup_program=None): + pool_type='max'): conv_out = layers.conv2d( input=input, num_filters=num_filters, filter_size=filter_size, param_attr=param_attr, - act=act, - main_program=main_program, - startup_program=startup_program) + act=act) pool_out = layers.pool2d( input=conv_out, pool_size=pool_size, pool_type=pool_type, - pool_stride=pool_stride, - main_program=main_program, - startup_program=startup_program) + pool_stride=pool_stride) return pool_out @@ -42,9 +36,7 @@ def img_conv_group(input, conv_with_batchnorm=False, conv_batchnorm_drop_rate=None, pool_stride=1, - pool_type=None, - main_program=None, - startup_program=None): + pool_type=None): """ Image Convolution Group, Used for vgg net. """ @@ -75,31 +67,19 @@ def img_conv_group(input, filter_size=conv_filter_size[i], padding=conv_padding[i], param_attr=param_attr[i], - act=local_conv_act, - main_program=main_program, - startup_program=startup_program) + act=local_conv_act) if conv_with_batchnorm[i]: - tmp = layers.batch_norm( - input=tmp, - act=conv_act, - main_program=main_program, - startup_program=startup_program) + tmp = layers.batch_norm(input=tmp, act=conv_act) drop_rate = conv_batchnorm_drop_rate[i] if abs(drop_rate) > 1e-5: - tmp = layers.dropout( - x=tmp, - dropout_prob=drop_rate, - main_program=main_program, - startup_program=startup_program) + tmp = layers.dropout(x=tmp, dropout_prob=drop_rate) pool_out = layers.pool2d( input=tmp, pool_size=pool_size, pool_type=pool_type, - pool_stride=pool_stride, - main_program=main_program, - startup_program=startup_program) + pool_stride=pool_stride) return pool_out @@ -108,21 +88,13 @@ def sequence_conv_pool(input, filter_size, param_attr=None, act="sigmoid", - pool_type="max", - main_program=None, - startup_program=None): + pool_type="max"): conv_out = layers.sequence_conv( input=input, num_filters=num_filters, filter_size=filter_size, param_attr=param_attr, - act=act, - main_program=main_program, - startup_program=startup_program) + act=act) - pool_out = layers.sequence_pool( - input=conv_out, - pool_type=pool_type, - main_program=main_program, - startup_program=startup_program) + pool_out = layers.sequence_pool(input=conv_out, pool_type=pool_type) return pool_out diff --git a/python/paddle/v2/fluid/optimizer.py b/python/paddle/v2/fluid/optimizer.py index bbdfab2df9519b77e5df184c00aadf703ec765e0..9f03eeea83e6d212da5fbe3d090d82028fa378ac 100644 --- a/python/paddle/v2/fluid/optimizer.py +++ b/python/paddle/v2/fluid/optimizer.py @@ -2,7 +2,7 @@ from collections import defaultdict import framework from backward import append_backward_ops -from framework import unique_name +from framework import unique_name, program_guard from initializer import Constant from layer_helper import LayerHelper from regularizer import append_regularization_ops @@ -159,34 +159,32 @@ class Optimizer(object): # Create any accumulators program = loss.block.program - self.helper = LayerHelper( - self.__class__.__name__, - main_program=program, - startup_program=startup_program) - self._create_accumulators(loss.block, - [p[0] for p in parameters_and_grads]) - - optimize_ops = [] - for param_and_grad in parameters_and_grads: - if param_and_grad[0].trainable is True and param_and_grad[ - 1] is not None: - optimize_op = self._append_optimize_op(loss.block, - param_and_grad) - optimize_ops.append(optimize_op) - - # Returned list of ops can include more ops in addition - # to optimization ops - return_ops = optimize_ops - - # Get custom finish ops for subclasses - # FIXME: Need to fix this once we figure out how to handle dependencies - finish_ops = self._finish_update(loss.block) - if finish_ops is not None: - return_ops += finish_ops - - if self._global_step is not None: - return_ops.append(self._increment_global_step(loss.block)) - return return_ops + with program_guard(program, startup_program): + self.helper = LayerHelper(self.__class__.__name__) + self._create_accumulators(loss.block, + [p[0] for p in parameters_and_grads]) + + optimize_ops = [] + for param_and_grad in parameters_and_grads: + if param_and_grad[0].trainable is True and param_and_grad[ + 1] is not None: + optimize_op = self._append_optimize_op(loss.block, + param_and_grad) + optimize_ops.append(optimize_op) + + # Returned list of ops can include more ops in addition + # to optimization ops + return_ops = optimize_ops + + # Get custom finish ops for subclasses + # FIXME: Need to fix this once we figure out how to handle dependencies + finish_ops = self._finish_update(loss.block) + if finish_ops is not None: + return_ops += finish_ops + + if self._global_step is not None: + return_ops.append(self._increment_global_step(loss.block)) + return return_ops def minimize(self, loss, diff --git a/python/paddle/v2/fluid/tests/.gitignore b/python/paddle/v2/fluid/tests/.gitignore index a648f2b387c2c7b9422eea6749e43e7b8871f60f..62f82151eb42342cd90657b1e4dfc93410950e62 100644 --- a/python/paddle/v2/fluid/tests/.gitignore +++ b/python/paddle/v2/fluid/tests/.gitignore @@ -1,3 +1,4 @@ image/ fit_a_line.model/ tmp +cuda_profiler.txt diff --git a/python/paddle/v2/fluid/tests/book/test_image_classification_train.py b/python/paddle/v2/fluid/tests/book/test_image_classification_train.py index 4e71b6f345ea7a1e6d29bc4ad810bc5b5f99d456..3d336ffe9582ddd9a2031e7aa7e2c26a772820f8 100644 --- a/python/paddle/v2/fluid/tests/book/test_image_classification_train.py +++ b/python/paddle/v2/fluid/tests/book/test_image_classification_train.py @@ -1,9 +1,9 @@ from __future__ import print_function -import numpy as np +import sys + import paddle.v2 as paddle import paddle.v2.fluid as fluid -import sys def resnet_cifar10(input, depth=32): diff --git a/python/paddle/v2/fluid/tests/book/test_label_semantic_roles.py b/python/paddle/v2/fluid/tests/book/test_label_semantic_roles.py index d2693b602ea5de9d2d60fbe114820b25119bfa3f..c3591a613acafb268a5bd70618cd4555450bac29 100644 --- a/python/paddle/v2/fluid/tests/book/test_label_semantic_roles.py +++ b/python/paddle/v2/fluid/tests/book/test_label_semantic_roles.py @@ -150,7 +150,7 @@ def main(): crf_decode = fluid.layers.crf_decoding( input=feature_out, param_attr=fluid.ParamAttr(name='crfw')) - precision, recall, f1_score = fluid.layers.chunk_eval( + chunk_evaluator = fluid.evaluator.ChunkEvaluator( input=crf_decode, label=target, chunk_scheme="IOB", @@ -176,20 +176,21 @@ def main(): batch_id = 0 for pass_id in xrange(PASS_NUM): + chunk_evaluator.reset(exe) for data in train_data(): - outs = exe.run(fluid.default_main_program(), - feed=feeder.feed(data), - fetch_list=[avg_cost, precision, recall, f1_score]) - avg_cost_val = np.array(outs[0]) - precision_val = np.array(outs[1]) - recall_val = np.array(outs[2]) - f1_score_val = np.array(outs[3]) + cost, precision, recall, f1_score = exe.run( + fluid.default_main_program(), + feed=feeder.feed(data), + fetch_list=[avg_cost] + chunk_evaluator.metrics) + pass_precision, pass_recall, pass_f1_score = chunk_evaluator.eval( + exe) if batch_id % 10 == 0: - print("avg_cost=" + str(avg_cost_val)) - print("precision_val=" + str(precision_val)) - print("recall_val:" + str(recall_val)) - print("f1_score_val:" + str(f1_score_val)) + print("avg_cost:" + str(cost) + " precision:" + str( + precision) + " recall:" + str(recall) + " f1_score:" + str( + f1_score) + " pass_precision:" + str( + pass_precision) + " pass_recall:" + str(pass_recall) + + " pass_f1_score:" + str(pass_f1_score)) # exit early for CI exit(0) diff --git a/python/paddle/v2/fluid/tests/book/test_recognize_digits_mlp.py b/python/paddle/v2/fluid/tests/book/test_recognize_digits_mlp.py index 4dc2c50e1c963a189b727f0a7edcb6886abd9038..d77f19660ebcd470837e8b4e63509683de4a7a82 100644 --- a/python/paddle/v2/fluid/tests/book/test_recognize_digits_mlp.py +++ b/python/paddle/v2/fluid/tests/book/test_recognize_digits_mlp.py @@ -33,11 +33,10 @@ opts = optimizer.minimize(avg_cost) accuracy = fluid.evaluator.Accuracy(input=predict, label=label) inference_program = fluid.default_main_program().clone() -test_accuracy = fluid.evaluator.Accuracy( - input=predict, label=label, main_program=inference_program) -test_target = [avg_cost] + test_accuracy.metrics + test_accuracy.states -inference_program = fluid.io.get_inference_program( - test_target, main_program=inference_program) +with fluid.program_guard(inference_program): + test_accuracy = fluid.evaluator.Accuracy(input=predict, label=label) + test_target = [avg_cost] + test_accuracy.metrics + test_accuracy.states + inference_program = fluid.io.get_inference_program(test_target) train_reader = paddle.batch( paddle.reader.shuffle( diff --git a/python/paddle/v2/fluid/tests/book/test_understand_sentiment_lstm.py b/python/paddle/v2/fluid/tests/book/test_understand_sentiment_lstm.py index 80f859967979ec07536d652d4ea620fd4ddb2daa..633de66bea2af7404ab0d325b425e7b9e63d3e43 100644 --- a/python/paddle/v2/fluid/tests/book/test_understand_sentiment_lstm.py +++ b/python/paddle/v2/fluid/tests/book/test_understand_sentiment_lstm.py @@ -1,6 +1,39 @@ import numpy as np import paddle.v2 as paddle import paddle.v2.fluid as fluid +from paddle.v2.fluid.layer_helper import LayerHelper + + +def lstm(x, c_pre_init, hidden_dim, forget_bias=None): + """ + This function helps create an operator for the LSTM (Long Short Term + Memory) cell that can be used inside an RNN. + """ + helper = LayerHelper('lstm_unit', **locals()) + rnn = fluid.layers.StaticRNN() + with rnn.step(): + c_pre = rnn.memory(init=c_pre_init) + x_t = rnn.step_input(x) + + before_fc = fluid.layers.concat(input=[x_t, c_pre], axis=1) + after_fc = fluid.layers.fc(input=before_fc, size=hidden_dim * 4) + + dtype = x.dtype + c = helper.create_tmp_variable(dtype) + h = helper.create_tmp_variable(dtype) + + helper.append_op( + type='lstm_unit', + inputs={"X": after_fc, + "C_prev": c_pre}, + outputs={"C": c, + "H": h}, + attrs={"forget_bias": forget_bias}) + + rnn.update_memory(c_pre, c) + rnn.output(h) + + return rnn() def lstm_net(dict_dim, class_dim=2, emb_dim=32, seq_len=80, batch_size=50): @@ -23,8 +56,7 @@ def lstm_net(dict_dim, class_dim=2, emb_dim=32, seq_len=80, batch_size=50): c_pre_init = fluid.layers.fill_constant( dtype=emb.dtype, shape=[batch_size, emb_dim], value=0.0) c_pre_init.stop_gradient = False - layer_1_out = fluid.layers.lstm( - emb, c_pre_init=c_pre_init, hidden_dim=emb_dim) + layer_1_out = lstm(emb, c_pre_init=c_pre_init, hidden_dim=emb_dim) layer_1_out = fluid.layers.transpose(x=layer_1_out, axis=[1, 0, 2]) prediction = fluid.layers.fc(input=layer_1_out, diff --git a/python/paddle/v2/fluid/tests/test_chunk_eval_op.py b/python/paddle/v2/fluid/tests/test_chunk_eval_op.py index 819e65a653437f0c34e14403f76317ff3b7f37f4..53bf6f815b8c7baf4c92d9fd488b69722ab0bef5 100644 --- a/python/paddle/v2/fluid/tests/test_chunk_eval_op.py +++ b/python/paddle/v2/fluid/tests/test_chunk_eval_op.py @@ -147,7 +147,13 @@ class TestChunkEvalOp(OpTest): 'Recall': np.asarray( [recall], dtype='float32'), 'F1-Score': np.asarray( - [f1], dtype='float32') + [f1], dtype='float32'), + 'NumInferChunks': np.asarray( + [self.num_infer_chunks], dtype='int64'), + 'NumLabelChunks': np.asarray( + [self.num_label_chunks], dtype='int64'), + 'NumCorrectChunks': np.asarray( + [self.num_correct_chunks], dtype='int64') } def setUp(self): diff --git a/python/paddle/v2/fluid/tests/test_image_classification_layer.py b/python/paddle/v2/fluid/tests/test_image_classification_layer.py index 2fd609d4474e97ecd96adcd146f2f550e0772740..b621d1525e33693869e24e2bb233bc8e257b077f 100644 --- a/python/paddle/v2/fluid/tests/test_image_classification_layer.py +++ b/python/paddle/v2/fluid/tests/test_image_classification_layer.py @@ -5,12 +5,7 @@ import paddle.v2.fluid.nets as nets from paddle.v2.fluid.framework import Program -def conv_block(input, - num_filter, - groups, - dropouts, - main_program=None, - startup_program=None): +def conv_block(input, num_filter, groups, dropouts): return nets.img_conv_group( input=input, pool_size=2, @@ -20,90 +15,54 @@ def conv_block(input, conv_act='relu', conv_with_batchnorm=True, conv_batchnorm_drop_rate=dropouts, - pool_type='max', - main_program=main_program, - startup_program=startup_program) + pool_type='max') class TestLayer(unittest.TestCase): def test_batch_norm_layer(self): main_program = Program() startup_program = Program() - images = fluid.layers.data( - name='pixel', - shape=[3, 48, 48], - dtype='float32', - main_program=main_program) - hidden1 = fluid.layers.batch_norm( - input=images, - main_program=main_program, - startup_program=startup_program) - hidden2 = fluid.layers.fc(input=hidden1, - size=128, - act='relu', - main_program=main_program) - hidden3 = fluid.layers.batch_norm( - input=hidden2, - main_program=main_program, - startup_program=startup_program) + with fluid.program_guard(main_program, startup_program): + images = fluid.layers.data( + name='pixel', shape=[3, 48, 48], dtype='float32') + hidden1 = fluid.layers.batch_norm(input=images) + hidden2 = fluid.layers.fc(input=hidden1, size=128, act='relu') + fluid.layers.batch_norm(input=hidden2) print str(main_program) def test_dropout_layer(self): main_program = Program() startup_program = Program() - images = fluid.layers.data( - name='pixel', - shape=[3, 48, 48], - dtype='float32', - main_program=main_program) - fluid.layers.dropout( - x=images, - dropout_prob=0.5, - main_program=main_program, - startup_program=startup_program) + with fluid.program_guard(main_program, startup_program): + images = fluid.layers.data( + name='pixel', shape=[3, 48, 48], dtype='float32') + fluid.layers.dropout(x=images, dropout_prob=0.5) - # print str(main_program) + print str(main_program) def test_img_conv_group(self): main_program = Program() startup_program = Program() - images = fluid.layers.data( - name='pixel', - shape=[3, 48, 48], - dtype='float32', - main_program=main_program, - startup_program=startup_program) - conv1 = conv_block(images, 64, 2, [0.3, 0], main_program, - startup_program) - conv2 = conv_block(conv1, 256, 3, [0.4, 0.4, 0], main_program, - startup_program) + with fluid.program_guard(main_program, startup_program): + images = fluid.layers.data( + name='pixel', shape=[3, 48, 48], dtype='float32') + conv1 = conv_block(images, 64, 2, [0.3, 0]) + conv_block(conv1, 256, 3, [0.4, 0.4, 0]) - # print str(main_program) + print str(main_program) def test_elementwise_add_with_act(self): main_program = Program() startup_program = Program() - image1 = fluid.layers.data( - name='pixel1', - shape=[3, 48, 48], - dtype='float32', - main_program=main_program, - startup_program=startup_program) - image2 = fluid.layers.data( - name='pixel2', - shape=[3, 48, 48], - dtype='float32', - main_program=main_program, - startup_program=startup_program) - out = fluid.layers.elementwise_add( - x=image1, - y=image2, - act='relu', - main_program=main_program, - startup_program=startup_program) - # print(main_program) + with fluid.program_guard(main_program, startup_program): + image1 = fluid.layers.data( + name='pixel1', shape=[3, 48, 48], dtype='float32') + image2 = fluid.layers.data( + name='pixel2', shape=[3, 48, 48], dtype='float32') + fluid.layers.elementwise_add(x=image1, y=image2, act='relu') + print(main_program) if __name__ == '__main__': diff --git a/python/paddle/v2/fluid/tests/test_inference_model_io.py b/python/paddle/v2/fluid/tests/test_inference_model_io.py index 60aed62ead83dedbeb9438c431ec292558d88ce5..71ca3e6c105c4437470f8e9f596e723d879b65e4 100644 --- a/python/paddle/v2/fluid/tests/test_inference_model_io.py +++ b/python/paddle/v2/fluid/tests/test_inference_model_io.py @@ -6,7 +6,7 @@ import paddle.v2.fluid.core as core import paddle.v2.fluid.executor as executor import paddle.v2.fluid.layers as layers import paddle.v2.fluid.optimizer as optimizer -from paddle.v2.fluid.framework import Program +from paddle.v2.fluid.framework import Program, program_guard from paddle.v2.fluid.io import save_inference_model, load_inference_model @@ -16,35 +16,18 @@ class TestBook(unittest.TestCase): init_program = Program() program = Program() - x = layers.data( - name='x', - shape=[2], - dtype='float32', - main_program=program, - startup_program=init_program) - y = layers.data( - name='y', - shape=[1], - dtype='float32', - main_program=program, - startup_program=init_program) - - y_predict = layers.fc(input=x, - size=1, - act=None, - main_program=program, - startup_program=init_program) - - cost = layers.square_error_cost( - input=y_predict, - label=y, - main_program=program, - startup_program=init_program) - avg_cost = layers.mean( - x=cost, main_program=program, startup_program=init_program) - - sgd_optimizer = optimizer.SGDOptimizer(learning_rate=0.001) - sgd_optimizer.minimize(avg_cost, init_program) + + with program_guard(program, init_program): + x = layers.data(name='x', shape=[2], dtype='float32') + y = layers.data(name='y', shape=[1], dtype='float32') + + y_predict = layers.fc(input=x, size=1, act=None) + + cost = layers.square_error_cost(input=y_predict, label=y) + avg_cost = layers.mean(x=cost) + + sgd_optimizer = optimizer.SGDOptimizer(learning_rate=0.001) + sgd_optimizer.minimize(avg_cost, init_program) place = core.CPUPlace() exe = executor.Executor(place) diff --git a/python/paddle/v2/fluid/tests/test_layers.py b/python/paddle/v2/fluid/tests/test_layers.py index 9b88080158139f267e253c598e60a4d92a0eff68..2286e94a90a4810dfb170ba6e929a7c4f3edaba1 100644 --- a/python/paddle/v2/fluid/tests/test_layers.py +++ b/python/paddle/v2/fluid/tests/test_layers.py @@ -161,6 +161,15 @@ class TestBook(unittest.TestCase): x=dat, label=lbl)) print(str(program)) + def test_seq_expand(self): + program = Program() + with program_guard(program): + x = layers.data(name='x', shape=[10], dtype='float32') + y = layers.data( + name='y', shape=[10, 20], dtype='float32', lod_level=1) + self.assertIsNotNone(layers.sequence_expand(x=x, y=y)) + print(str(program)) + if __name__ == '__main__': unittest.main() diff --git a/python/paddle/v2/fluid/tests/test_lod_tensor_array_ops.py b/python/paddle/v2/fluid/tests/test_lod_tensor_array_ops.py index 0a916a55bc3d097e17fb504b0d6b2f2818f030c9..5fdabbcf889448114ac4e55e7944cb6c57ba5f3c 100644 --- a/python/paddle/v2/fluid/tests/test_lod_tensor_array_ops.py +++ b/python/paddle/v2/fluid/tests/test_lod_tensor_array_ops.py @@ -2,7 +2,7 @@ import unittest import paddle.v2.fluid.core as core import numpy import paddle.v2.fluid.layers as layers -from paddle.v2.fluid.framework import Program +from paddle.v2.fluid.framework import Program, program_guard from paddle.v2.fluid.executor import Executor from paddle.v2.fluid.backward import append_backward_ops @@ -118,16 +118,17 @@ class TestCPULoDTensorArrayOps(unittest.TestCase): def main(self, tensor, expect_array, expect_lod, expect_max_len, level=0): place = self.place() program = Program() - x = layers.data(name='x', shape=[10], main_program=program) - x.persistable = True - table = layers.lod_rank_table(x, level=level, main_program=program) - max_len = layers.max_sequence_len(table, main_program=program) - max_len.persistable = True - array = layers.lod_tensor_to_array(x, table, main_program=program) - array.persistable = True - - result = layers.array_to_lod_tensor(array, table, main_program=program) - result.persistable = True + with program_guard(program): + x = layers.data(name='x', shape=[10]) + x.persistable = True + table = layers.lod_rank_table(x, level=level) + max_len = layers.max_sequence_len(table) + max_len.persistable = True + array = layers.lod_tensor_to_array(x, table) + array.persistable = True + + result = layers.array_to_lod_tensor(array, table) + result.persistable = True exe = Executor(place) scope = core.Scope() exe.run(program, feed={'x': tensor}, scope=scope) @@ -160,19 +161,16 @@ class TestCPULoDTensorArrayOpGrad(unittest.TestCase): place = core.CPUPlace() program = Program() - x = layers.data( - name='x', - shape=[1], - dtype='float32', - main_program=program, - stop_gradient=False) - table = layers.lod_rank_table(x, level=0, main_program=program) - array = layers.lod_tensor_to_array(x, table, main_program=program) - result = layers.array_to_lod_tensor(array, table, main_program=program) + with program_guard(program): + x = layers.data( + name='x', shape=[1], dtype='float32', stop_gradient=False) + table = layers.lod_rank_table(x, level=0) + array = layers.lod_tensor_to_array(x, table) + result = layers.array_to_lod_tensor(array, table) - mean = layers.mean(x=result, main_program=program) + mean = layers.mean(x=result) - append_backward_ops(mean) + append_backward_ops(mean) tensor = core.LoDTensor() tensor.set(numpy.arange(10).reshape(10, 1).astype('float32'), place) diff --git a/python/paddle/v2/fluid/tests/test_mnist_if_else_op.py b/python/paddle/v2/fluid/tests/test_mnist_if_else_op.py index 50fcc4a72ddbd6d7a3d3b73434c6ac8de5a006e2..33558c6105442b169b2e26abc7f39e15b7fe7322 100644 --- a/python/paddle/v2/fluid/tests/test_mnist_if_else_op.py +++ b/python/paddle/v2/fluid/tests/test_mnist_if_else_op.py @@ -1,5 +1,5 @@ import paddle.v2.fluid.layers as layers -from paddle.v2.fluid.framework import Program +from paddle.v2.fluid.framework import Program, program_guard, default_main_program, default_startup_program from paddle.v2.fluid.executor import Executor from paddle.v2.fluid.optimizer import MomentumOptimizer import paddle.v2.fluid.core as core @@ -10,44 +10,42 @@ import numpy as np class TestMNISTIfElseOp(unittest.TestCase): def test_raw_api(self): - kwargs = {'startup_program': Program(), 'main_program': Program()} - image = layers.data(name='x', shape=[784], dtype='float32', **kwargs) + prog = Program() + startup_prog = Program() + with program_guard(prog, startup_prog): + image = layers.data(name='x', shape=[784], dtype='float32') - label = layers.data(name='y', shape=[1], dtype='int64', **kwargs) + label = layers.data(name='y', shape=[1], dtype='int64') - limit = layers.fill_constant_batch_size_like( - input=label, dtype='int64', shape=[1], value=5.0, **kwargs) + limit = layers.fill_constant_batch_size_like( + input=label, dtype='int64', shape=[1], value=5.0) + cond = layers.less_than(x=label, y=limit) + true_image, false_image = layers.split_lod_tensor( + input=image, mask=cond) - cond = layers.less_than(x=label, y=limit, **kwargs) - true_image, false_image = layers.split_lod_tensor( - input=image, mask=cond, **kwargs) + true_out = layers.create_tensor(dtype='float32') + true_cond = layers.ConditionalBlock([true_image]) - true_out = layers.create_tensor(dtype='float32', **kwargs) - true_cond = layers.ConditionalBlock([true_image], **kwargs) + with true_cond.block(): + hidden = layers.fc(input=true_image, size=100, act='tanh') + prob = layers.fc(input=hidden, size=10, act='softmax') + layers.assign(input=prob, output=true_out) - with true_cond.block(): - hidden = layers.fc(input=true_image, size=100, act='tanh', **kwargs) - prob = layers.fc(input=hidden, size=10, act='softmax', **kwargs) - layers.assign(input=prob, output=true_out, **kwargs) + false_out = layers.create_tensor(dtype='float32') + false_cond = layers.ConditionalBlock([false_image]) - false_out = layers.create_tensor(dtype='float32', **kwargs) - false_cond = layers.ConditionalBlock([false_image], **kwargs) + with false_cond.block(): + hidden = layers.fc(input=false_image, size=200, act='tanh') + prob = layers.fc(input=hidden, size=10, act='softmax') + layers.assign(input=prob, output=false_out) - with false_cond.block(): - hidden = layers.fc(input=false_image, - size=200, - act='tanh', - **kwargs) - prob = layers.fc(input=hidden, size=10, act='softmax', **kwargs) - layers.assign(input=prob, output=false_out, **kwargs) + prob = layers.merge_lod_tensor( + in_true=true_out, in_false=false_out, mask=cond, x=image) + loss = layers.cross_entropy(input=prob, label=label) + avg_loss = layers.mean(x=loss) - prob = layers.merge_lod_tensor( - in_true=true_out, in_false=false_out, mask=cond, x=image, **kwargs) - loss = layers.cross_entropy(input=prob, label=label, **kwargs) - avg_loss = layers.mean(x=loss, **kwargs) - - optimizer = MomentumOptimizer(learning_rate=0.001, momentum=0.9) - optimizer.minimize(avg_loss, kwargs['startup_program']) + optimizer = MomentumOptimizer(learning_rate=0.001, momentum=0.9) + optimizer.minimize(avg_loss, startup_prog) train_reader = paddle.batch( paddle.reader.shuffle( @@ -57,7 +55,7 @@ class TestMNISTIfElseOp(unittest.TestCase): place = core.CPUPlace() exe = Executor(place) - exe.run(kwargs['startup_program']) + exe.run(startup_prog) PASS_NUM = 100 for pass_id in range(PASS_NUM): for data in train_reader(): @@ -65,7 +63,7 @@ class TestMNISTIfElseOp(unittest.TestCase): y_data = np.array(map(lambda x: x[1], data)).astype("int64") y_data = np.expand_dims(y_data, axis=1) - outs = exe.run(kwargs['main_program'], + outs = exe.run(prog, feed={'x': x_data, 'y': y_data}, fetch_list=[avg_loss]) @@ -75,39 +73,36 @@ class TestMNISTIfElseOp(unittest.TestCase): self.assertFalse(True) def test_ifelse(self): - kwargs = {'startup_program': Program(), 'main_program': Program()} - image = layers.data(name='x', shape=[784], dtype='float32', **kwargs) - - label = layers.data(name='y', shape=[1], dtype='int64', **kwargs) - - limit = layers.fill_constant_batch_size_like( - input=label, dtype='int64', shape=[1], value=5.0, **kwargs) - - cond = layers.less_than(x=label, y=limit, **kwargs) - - ie = layers.IfElse(cond, **kwargs) - - with ie.true_block(): - true_image = ie.input(image) - hidden = layers.fc(input=true_image, size=100, act='tanh', **kwargs) - prob = layers.fc(input=hidden, size=10, act='softmax', **kwargs) - ie.output(prob) - - with ie.false_block(): - false_image = ie.input(image) - hidden = layers.fc(input=false_image, - size=200, - act='tanh', - **kwargs) - prob = layers.fc(input=hidden, size=10, act='softmax', **kwargs) - ie.output(prob) - - prob = ie() - loss = layers.cross_entropy(input=prob[0], label=label, **kwargs) - avg_loss = layers.mean(x=loss, **kwargs) - - optimizer = MomentumOptimizer(learning_rate=0.001, momentum=0.9) - optimizer.minimize(avg_loss, kwargs['startup_program']) + prog = Program() + startup_prog = Program() + with program_guard(prog, startup_prog): + image = layers.data(name='x', shape=[784], dtype='float32') + + label = layers.data(name='y', shape=[1], dtype='int64') + + limit = layers.fill_constant_batch_size_like( + input=label, dtype='int64', shape=[1], value=5.0) + cond = layers.less_than(x=label, y=limit) + ie = layers.IfElse(cond) + + with ie.true_block(): + true_image = ie.input(image) + hidden = layers.fc(input=true_image, size=100, act='tanh') + prob = layers.fc(input=hidden, size=10, act='softmax') + ie.output(prob) + + with ie.false_block(): + false_image = ie.input(image) + hidden = layers.fc(input=false_image, size=200, act='tanh') + prob = layers.fc(input=hidden, size=10, act='softmax') + ie.output(prob) + + prob = ie() + loss = layers.cross_entropy(input=prob[0], label=label) + avg_loss = layers.mean(x=loss) + + optimizer = MomentumOptimizer(learning_rate=0.001, momentum=0.9) + optimizer.minimize(avg_loss, startup_prog) train_reader = paddle.batch( paddle.reader.shuffle( paddle.dataset.mnist.train(), buf_size=8192), @@ -135,4 +130,5 @@ class TestMNISTIfElseOp(unittest.TestCase): if __name__ == '__main__': - unittest.main() + # temp disable if else unittest since it could be buggy. + exit(0) diff --git a/python/paddle/v2/fluid/tests/test_program.py b/python/paddle/v2/fluid/tests/test_program.py index 1a9313c68aab165d85ae29051faeacb4927ac2c9..e6da0b2be77533811c98751de1067dbbeac11309 100644 --- a/python/paddle/v2/fluid/tests/test_program.py +++ b/python/paddle/v2/fluid/tests/test_program.py @@ -1,7 +1,7 @@ from __future__ import print_function import unittest -from paddle.v2.fluid.framework import Program, default_main_program +from paddle.v2.fluid.framework import Program, default_main_program, program_guard import paddle.v2.fluid.layers as layers main_program = default_main_program() @@ -129,13 +129,10 @@ class TestProgram(unittest.TestCase): def test_program_clone_with_parameter(self): main_program = Program() startup_program = Program() - kwargs = { - 'main_program': main_program, - 'startup_program': startup_program - } - d = layers.data(name='x', shape=[784], dtype='float32', **kwargs) - hidden = layers.fc(input=d, size=100, **kwargs) - layers.fc(input=hidden, size=100, **kwargs) + with program_guard(main_program, startup_program): + d = layers.data(name='x', shape=[784], dtype='float32') + hidden = layers.fc(input=d, size=100) + layers.fc(input=hidden, size=100) new_program = main_program.clone() self.assertNotEqual(0, len(new_program.blocks[0].all_parameters())) diff --git a/python/paddle/v2/fluid/tests/test_reduce_op.py b/python/paddle/v2/fluid/tests/test_reduce_op.py index 70359d60cbe656150877673c63e81eae92d8ab9a..a021d4dd91bb9cc1e5d85411b3813b966ef5b296 100644 --- a/python/paddle/v2/fluid/tests/test_reduce_op.py +++ b/python/paddle/v2/fluid/tests/test_reduce_op.py @@ -85,5 +85,19 @@ class Test1DReduce(OpTest): self.check_grad(['X'], 'Out') +class TestReduceAll(OpTest): + def setUp(self): + self.op_type = "reduce_sum" + self.inputs = {'X': np.random.random((5, 6, 2, 10)).astype("float32")} + self.attrs = {'reduce_all': True} + self.outputs = {'Out': self.inputs['X'].sum()} + + def test_check_output(self): + self.check_output() + + def test_check_grad(self): + self.check_grad(['X'], 'Out') + + if __name__ == '__main__': unittest.main() diff --git a/python/paddle/v2/fluid/tests/test_reshape_op.py b/python/paddle/v2/fluid/tests/test_reshape_op.py index 16bb6bb2af67f7d32a2fafc1cb37412084ec0829..18ee3aece656276fec9671df9baf298b7fd3c9b1 100644 --- a/python/paddle/v2/fluid/tests/test_reshape_op.py +++ b/python/paddle/v2/fluid/tests/test_reshape_op.py @@ -17,5 +17,19 @@ class TestReshapeOp(OpTest): self.check_grad(["X"], "Out") +class TestReshapeOpDimInfer(OpTest): + def setUp(self): + self.op_type = "reshape" + self.inputs = {'X': np.random.random((10, 20)).astype("float32")} + self.attrs = {'shape': [4, -1, 5]} + self.outputs = {'Out': self.inputs['X'].reshape(self.attrs['shape'])} + + def test_check_output(self): + self.check_output() + + def test_check_grad(self): + self.check_grad(["X"], "Out") + + if __name__ == '__main__': unittest.main() diff --git a/python/paddle/v2/fluid/tests/test_seq_expand.py b/python/paddle/v2/fluid/tests/test_sequence_expand.py similarity index 89% rename from python/paddle/v2/fluid/tests/test_seq_expand.py rename to python/paddle/v2/fluid/tests/test_sequence_expand.py index ff17edd04bfd34ab8449a0ae05aacf66632dabc8..0f22612d3dbe483e4d5a8638636e44e172160156 100644 --- a/python/paddle/v2/fluid/tests/test_seq_expand.py +++ b/python/paddle/v2/fluid/tests/test_sequence_expand.py @@ -3,7 +3,7 @@ import numpy as np from op_test import OpTest -class TestSeqExpand(OpTest): +class TestSequenceExpand(OpTest): def set_data(self): x_data = np.random.uniform(0.1, 1, [3, 1]).astype('float32') y_data = np.random.uniform(0.1, 1, [8, 1]).astype('float32') @@ -21,7 +21,7 @@ class TestSeqExpand(OpTest): self.outputs = {'Out': out} def setUp(self): - self.op_type = 'seq_expand' + self.op_type = 'sequence_expand' self.set_data() self.compute() @@ -32,7 +32,7 @@ class TestSeqExpand(OpTest): self.check_grad(["X"], "Out") -class TestSeqExpandCase1(TestSeqExpand): +class TestSequenceExpandCase1(TestSequenceExpand): def set_data(self): x_data = np.random.uniform(0.1, 1, [5, 1]).astype('float32') x_lod = [[0, 2, 5]] @@ -41,7 +41,7 @@ class TestSeqExpandCase1(TestSeqExpand): self.inputs = {'X': (x_data, x_lod), 'Y': (y_data, y_lod)} -class TestSeqExpandCase2(TestSeqExpand): +class TestSequenceExpandCase2(TestSequenceExpand): def set_data(self): x_data = np.random.uniform(0.1, 1, [1, 2, 2]).astype('float32') x_lod = [[0, 1]] @@ -50,7 +50,7 @@ class TestSeqExpandCase2(TestSeqExpand): self.inputs = {'X': (x_data, x_lod), 'Y': (y_data, y_lod)} -class TestSeqExpandCase3(TestSeqExpand): +class TestSequenceExpandCase3(TestSequenceExpand): def set_data(self): x_data = np.random.uniform(0.1, 1, [4, 1]).astype('float32') x_lod = [[0, 1, 2, 3, 4]] diff --git a/python/paddle/v2/fluid/tests/test_split_and_merge_lod_tensor_op.py b/python/paddle/v2/fluid/tests/test_split_and_merge_lod_tensor_op.py index f5da4e408f0a83dbf6da530b478e91bbf9cd5ab2..8cdd59ff3cc7deb57252fc5218d239f86016cb9c 100644 --- a/python/paddle/v2/fluid/tests/test_split_and_merge_lod_tensor_op.py +++ b/python/paddle/v2/fluid/tests/test_split_and_merge_lod_tensor_op.py @@ -2,7 +2,7 @@ import unittest import paddle.v2.fluid.core as core import numpy as np import paddle.v2.fluid.layers as layers -from paddle.v2.fluid.framework import Program +from paddle.v2.fluid.framework import Program, program_guard from paddle.v2.fluid.executor import Executor from paddle.v2.fluid.backward import append_backward_ops @@ -75,26 +75,22 @@ class TestCPULoDTensorArrayOps(unittest.TestCase): level=0): place = self.place() program = Program() - x = layers.data(name='x', shape=[1], main_program=program) - x.persistable = True + with program_guard(program): + x = layers.data(name='x', shape=[1]) + x.persistable = True - y = layers.data(name='y', shape=[1], main_program=program) - y.persistable = True + y = layers.data(name='y', shape=[1]) + y.persistable = True - out_true, out_false = layers.split_lod_tensor( - input=x, mask=y, level=level, main_program=program) - out_true.persistable = True - out_false.persistable = True + out_true, out_false = layers.split_lod_tensor( + input=x, mask=y, level=level) + out_true.persistable = True + out_false.persistable = True - out = layers.merge_lod_tensor( - in_true=out_true, - in_false=out_false, - mask=y, - x=x, - level=level, - main_program=program) + out = layers.merge_lod_tensor( + in_true=out_true, in_false=out_false, mask=y, x=x, level=level) - out.persistable = True + out.persistable = True exe = Executor(place) scope = core.Scope() @@ -123,34 +119,21 @@ class TestCPUSplitMergeLoDTensorGrad(unittest.TestCase): def test_grad(self): place = core.CPUPlace() program = Program() + with program_guard(program): + x = layers.data( + name='x', shape=[1], dtype='float32', stop_gradient=False) + y = layers.data( + name='y', shape=[1], dtype='bool', stop_gradient=False) - x = layers.data( - name='x', - shape=[1], - dtype='float32', - main_program=program, - stop_gradient=False) - y = layers.data( - name='y', - shape=[1], - dtype='bool', - main_program=program, - stop_gradient=False) - - level = 0 - - out_true, out_false = layers.split_lod_tensor( - input=x, mask=y, level=level, main_program=program) - out = layers.merge_lod_tensor( - in_true=out_true, - in_false=out_false, - mask=y, - x=x, - level=level, - main_program=program) - mean = layers.mean(x=out, main_program=program) - - append_backward_ops(mean) + level = 0 + + out_true, out_false = layers.split_lod_tensor( + input=x, mask=y, level=level) + out = layers.merge_lod_tensor( + in_true=out_true, in_false=out_false, mask=y, x=x, level=level) + mean = layers.mean(x=out) + + append_backward_ops(mean) tensor = core.LoDTensor() tensor.set(np.arange(10).reshape(10, 1).astype('float32'), place) diff --git a/python/paddle/v2/fluid/tests/test_spp_op.py b/python/paddle/v2/fluid/tests/test_spp_op.py new file mode 100644 index 0000000000000000000000000000000000000000..007723f0e35ad194c427401337bc9b13756576de --- /dev/null +++ b/python/paddle/v2/fluid/tests/test_spp_op.py @@ -0,0 +1,68 @@ +import unittest +import numpy as np +from op_test import OpTest +from test_pool2d_op import max_pool2D_forward_naive +from test_pool2d_op import avg_pool2D_forward_naive + + +class TestSppOp(OpTest): + def setUp(self): + self.op_type = "spp" + self.init_test_case() + input = np.random.random(self.shape).astype("float32") + nsize, csize, hsize, wsize = input.shape + out_level_flatten = [] + for i in xrange(self.pyramid_height): + bins = np.power(2, i) + kernel_size = [0, 0] + padding = [0, 0] + kernel_size[0] = np.ceil(hsize / + bins.astype("double")).astype("int32") + padding[0] = ( + (kernel_size[0] * bins - hsize + 1) / 2).astype("int32") + + kernel_size[1] = np.ceil(wsize / + bins.astype("double")).astype("int32") + padding[1] = ( + (kernel_size[1] * bins - wsize + 1) / 2).astype("int32") + out_level = self.pool2D_forward_naive(input, kernel_size, + kernel_size, padding) + out_level_flatten.append( + out_level.reshape(nsize, bins * bins * csize)) + if i == 0: + output = out_level_flatten[i] + else: + output = np.concatenate((output, out_level_flatten[i]), 1) + # output = np.concatenate(out_level_flatten.tolist(), 0); + self.inputs = {'X': input.astype('float32'), } + self.attrs = { + 'pyramid_height': self.pyramid_height, + 'pooling_type': self.pool_type + } + + self.outputs = {'Out': output.astype('float32')} + + def test_check_output(self): + self.check_output() + + def test_check_grad(self): + if self.pool_type != "avg": + self.check_grad(['X'], 'Out', max_relative_error=0.05) + + def init_test_case(self): + self.shape = [3, 2, 4, 4] + self.pyramid_height = 3 + self.pool2D_forward_naive = max_pool2D_forward_naive + self.pool_type = "max" + + +class TestCase2(TestSppOp): + def init_test_case(self): + self.shape = [3, 2, 4, 4] + self.pyramid_height = 3 + self.pool2D_forward_naive = avg_pool2D_forward_naive + self.pool_type = "avg" + + +if __name__ == '__main__': + unittest.main() diff --git a/python/paddle/v2/parameters.py b/python/paddle/v2/parameters.py index bd97dc1199fedc8ac91c1c6086957e8cce88bdc4..7b7d1a1d1672802e0e91a857100604758683224e 100644 --- a/python/paddle/v2/parameters.py +++ b/python/paddle/v2/parameters.py @@ -383,19 +383,22 @@ class Parameters(object): params.deserialize(param_name, f) return params - def init_from_tar(self, f): + def init_from_tar(self, f, exclude_params=[]): """ Different from `from_tar`, this interface can be used to init partial network parameters from another saved model. :param f: the initialized model file. :type f: tar file + :param exclude_params: the names of parameters that should + not be initialized from the model file. + :type exclude_params: list of strings :return: Nothing. """ tar_param = Parameters.from_tar(f) for pname in tar_param.names(): - if pname in self.names(): + if pname in self.names() and pname not in exclude_params: self.set(pname, tar_param.get(pname)) diff --git a/python/paddle/v2/reader/decorator.py b/python/paddle/v2/reader/decorator.py index 7e457f987d36e52e9e7c7727b4f996ad31c6bf08..27c82c95f79e0a3e3129627bfa33d85e0d3cd862 100644 --- a/python/paddle/v2/reader/decorator.py +++ b/python/paddle/v2/reader/decorator.py @@ -390,8 +390,6 @@ def pipe_reader(left_cmd, if not callable(parser): raise TypeError("parser must be a callable object") - process = subprocess.Popen( - left_cmd.split(" "), bufsize=bufsize, stdout=subprocess.PIPE) # TODO(typhoonzero): add a thread to read stderr # Always init a decompress object is better than @@ -400,6 +398,8 @@ def pipe_reader(left_cmd, 32 + zlib.MAX_WBITS) # offset 32 to skip the header def reader(): + process = subprocess.Popen( + left_cmd.split(" "), bufsize=bufsize, stdout=subprocess.PIPE) remained = "" while True: buff = process.stdout.read(bufsize) diff --git a/python/paddle/v2/reader/tests/decorator_test.py b/python/paddle/v2/reader/tests/decorator_test.py index 5a92951b100fa51ab6df7039d9c6b54d1f9d963e..06e14796daf27812b9aeb1e4b024f294c7609991 100644 --- a/python/paddle/v2/reader/tests/decorator_test.py +++ b/python/paddle/v2/reader/tests/decorator_test.py @@ -145,5 +145,35 @@ class TestXmap(unittest.TestCase): self.assertEqual(e, mapper(idx)) +class TestPipeReader(unittest.TestCase): + def test_pipe_reader(self): + def simple_parser(lines): + return lines + + import tempfile + + records = [str(i) for i in xrange(5)] + temp = tempfile.NamedTemporaryFile() + try: + with open(temp.name, 'w') as f: + for r in records: + f.write('%s\n' % r) + + cmd = "cat %s" % temp.name + reader = paddle.v2.reader.pipe_reader( + cmd, simple_parser, bufsize=128) + for i in xrange(4): + result = [] + for r in reader(): + result.append(r) + + for idx, e in enumerate(records): + print e, result[idx] + self.assertEqual(e, result[idx]) + finally: + # delete the temporary file + temp.close() + + if __name__ == '__main__': unittest.main() diff --git a/python/setup.py.in b/python/setup.py.in index 9ccb4dc1762ac761212347fa7c7c94b223d75e24..8396fb44cfcee28211b5d3db7684a4adce1fb1f6 100644 --- a/python/setup.py.in +++ b/python/setup.py.in @@ -68,6 +68,7 @@ packages=['paddle', 'paddle.v2.plot', 'paddle.v2.fluid', 'paddle.v2.fluid.proto', + 'paddle.v2.fluid.layers', 'py_paddle'] with open('@PADDLE_SOURCE_DIR@/python/requirements.txt') as f: