m53297601
/
Paddle
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Merge branch 'develop' of https://github.com/PaddlePaddle/paddle into fix-beam_search-dev

Browse Source
port
guosheng 7 years ago
parent 49ca424d6e 69827f305b
commit f4bdfbc697
210 changed files with 6613 additions and 2464 deletions
Show Stats Download Patch File Download Diff File

1
AUTHORS.md
Unescape View File

@ -22,6 +22,7 @@
| jczaja | Jacek Czaja |
| JiayiFeng | Jia-Yi Feng |
| kbinias | Krzysztof Binias |
| kexinzhao | Ke-Xin Zhao |
| kuke | Yi-Bing Liu |
| lcy-seso | Ying Cao |
| lipeng-unisound | Peng Li |

6
CMakeLists.txt
Unescape View File

@ -61,6 +61,7 @@ option(EIGEN_USE_THREADS "Compile with multi-threaded Eigen" OFF)
option(WITH_ARM_FP16 "Use half precision support on armv8.2-a cpu" OFF)
option(WITH_FAST_BUNDLE_TEST "Bundle tests that can be run in a single process together to reduce launch overhead" OFF)
option(WITH_CONTRIB "Compile the third-party contributation" OFF)
option(WITH_ANAKIN "Compile with Anakin library" OFF)
option(WITH_GRPC "Use grpc as the default rpc framework" ${WITH_DISTRIBUTE})
# CMAKE_BUILD_TYPE
@ -193,7 +194,10 @@ set(EXTERNAL_LIBS
if(WITH_GPU)
include(cuda)
include(tensorrt)
endif(WITH_GPU)
include(external/anakin)
else()
set(WITH_ANAKIN OFF CACHE STRING "Anakin is valid only when GPU is set." FORCE)
endif()
if(WITH_AMD_GPU)
find_package(HIP)

15
benchmark/fluid/models/machine_translation.py
Unescape View File

@ -173,21 +173,6 @@ def seq_to_seq_net(embedding_dim, encoder_size, decoder_size, source_dict_dim,
return avg_cost, feeding_list
def to_lodtensor(data, place):
seq_lens = [len(seq) for seq in data]
cur_len = 0
lod = [cur_len]
for l in seq_lens:
cur_len += l
lod.append(cur_len)
flattened_data = np.concatenate(data, axis=0).astype("int64")
flattened_data = flattened_data.reshape([len(flattened_data), 1])
lod_t = core.LoDTensor()
lod_t.set(flattened_data, place)
lod_t.set_lod([lod])
return lod_t, lod[-1]
def lodtensor_to_ndarray(lod_tensor):
dims = lod_tensor.get_dims()
ndarray = np.zeros(shape=dims).astype('float32')

15
benchmark/fluid/models/stacked_dynamic_lstm.py
Unescape View File

@ -125,18 +125,3 @@ def get_model(args):
batch_size=args.batch_size)
return loss, inference_program, adam, train_reader, test_reader, batch_acc
def to_lodtensor(data, place):
seq_lens = [len(seq) for seq in data]
cur_len = 0
lod = [cur_len]
for l in seq_lens:
cur_len += l
lod.append(cur_len)
flattened_data = numpy.concatenate(data, axis=0).astype("int64")
flattened_data = flattened_data.reshape([len(flattened_data), 1])
res = fluid.LoDTensor()
res.set(flattened_data, place)
res.set_lod([lod])
return res

42
cmake/external/anakin.cmake vendored
Unescape View File

@ -0,0 +1,42 @@
if (NOT WITH_ANAKIN)
return()
endif()
set(ANAKIN_INSTALL_DIR "${THIRD_PARTY_PATH}/install/anakin" CACHE PATH
"Anakin install path." FORCE)
set(ANAKIN_INCLUDE "${ANAKIN_INSTALL_DIR}" CACHE STRING "root of Anakin header files")
set(ANAKIN_LIBRARY "${ANAKIN_INSTALL_DIR}" CACHE STRING "path of Anakin library")
set(ANAKIN_COMPILE_EXTRA_FLAGS -Wno-error=unused-variable -Wno-error=format-extra-args -Wno-error=comment -Wno-error=format -Wno-error=switch -Wno-error=return-type -Wno-error=non-virtual-dtor -Wno-reorder -Wno-error=cpp)
set(ANAKIN_LIBRARY_URL "https://github.com/pangge/Anakin/releases/download/3.0/anakin_release_simple.tar.gz")
# A helper function used in Anakin, currently, to use it, one need to recursively include
# nearly all the header files.
function(fetch_include_recursively root_dir)
if (IS_DIRECTORY ${root_dir})
include_directories(${root_dir})
endif()
file(GLOB ALL_SUB RELATIVE ${root_dir} ${root_dir}/*)
foreach(sub ${ALL_SUB})
if (IS_DIRECTORY ${root_dir}/${sub})
fetch_include_recursively(${root_dir}/${sub})
endif()
endforeach()
endfunction()
# download library
message(STATUS "Download Anakin library from ${ANAKIN_LIBRARY_URL}")
execute_process(COMMAND bash -c "mkdir -p ${ANAKIN_INSTALL_DIR}")
execute_process(COMMAND bash -c "rm -rf ${ANAKIN_INSTALL_DIR}/*")
execute_process(COMMAND bash -c "cd ${ANAKIN_INSTALL_DIR}; wget -q ${ANAKIN_LIBRARY_URL}")
execute_process(COMMAND bash -c "mkdir -p ${ANAKIN_INSTALL_DIR}")
execute_process(COMMAND bash -c "cd ${ANAKIN_INSTALL_DIR}; tar xzf anakin_release_simple.tar.gz")
if (WITH_ANAKIN)
message(STATUS "Anakin for inference is enabled")
message(STATUS "Anakin is set INCLUDE:${ANAKIN_INCLUDE} LIBRARY:${ANAKIN_LIBRARY}")
fetch_include_recursively(${ANAKIN_INCLUDE})
link_directories(${ANAKIN_LIBRARY})
endif()

2
cmake/external/openblas.cmake vendored
Unescape View File

@ -29,6 +29,8 @@ IF(NOT ${CBLAS_FOUND})
"${CBLAS_INSTALL_DIR}/lib/${CMAKE_STATIC_LIBRARY_PREFIX}openblas${CMAKE_STATIC_LIBRARY_SUFFIX}"
CACHE FILEPATH "openblas library." FORCE)
ADD_DEFINITIONS(-DPADDLE_USE_OPENBLAS)
SET(OPENBLAS_CC "${CMAKE_C_COMPILER} -Wno-unused-but-set-variable -Wno-unused-variable")
SET(OPENBLAS_COMMIT "v0.2.20")

11
cmake/inference_lib.cmake
Unescape View File

@ -39,7 +39,7 @@ function(copy TARGET)
message(FATAL_ERROR "${TARGET} source numbers are not equal to destination numbers")
endif()
math(EXPR len "${copy_lib_SRCS_len} - 1")
add_custom_target(${TARGET} DEPENDS ${copy_lib_DEPS})
foreach(index RANGE ${len})
list(GET copy_lib_SRCS ${index} src)
@ -155,6 +155,15 @@ copy(inference_lib DEPS paddle_fluid_shared paddle_fluid
DSTS ${dst_dir}/${module} ${dst_dir}/${module}
)
if(WITH_CONTRIB)
set(contrib_dst_dir "${FLUID_INSTALL_DIR}/contrib/inference")
copy(contrib_inference_lib DEPS paddle_inference_api
SRCS ${PADDLE_SOURCE_DIR}/paddle/contrib/inference/paddle_inference_api.h
${PADDLE_BINARY_DIR}/paddle/contrib/inference/libpaddle_inference_api.*
DSTS ${contrib_dst_dir} ${contrib_dst_dir}
)
endif()
set(module "platform")
copy(platform_lib DEPS profiler_py_proto
SRCS ${src_dir}/${module}/*.h ${src_dir}/${module}/dynload/*.h ${src_dir}/${module}/details/*.h

2
doc/fluid/api/gen_doc.sh
Unescape View File

@ -1,5 +1,5 @@
#!/bin/bash
python gen_doc.py layers --submodules control_flow device io nn ops tensor detection > layers.rst
python gen_doc.py layers --submodules control_flow device io nn ops tensor detection learning_rate_scheduler metric > layers.rst
for module in data_feeder clip metrics executor initializer io nets optimizer param_attr profiler regularizer
do

14
doc/fluid/api/initializer.rst
Unescape View File

@ -33,6 +33,13 @@ Xavier
:members:
:noindex:
Bilinear
--------
.. autoclass:: paddle.fluid.initializer.Bilinear
:members:
:noindex:
force_init_on_cpu
-----------------
@ -73,3 +80,10 @@ XavierInitializer
:members:
:noindex:
BilinearInitializer
-------------------
.. autoclass:: paddle.fluid.initializer.BilinearInitializer
:members:
:noindex:

36
doc/fluid/api/io.rst
Unescape View File

@ -59,3 +59,39 @@ get_inference_program
.. autofunction:: paddle.fluid.io.get_inference_program
:noindex:
save_checkpoint
---------------
.. autofunction:: paddle.fluid.io.save_checkpoint
:noindex:
load_checkpoint
---------------
.. autofunction:: paddle.fluid.io.load_checkpoint
:noindex:
clean_checkpoint
----------------
.. autofunction:: paddle.fluid.io.clean_checkpoint
:noindex:
load_persist_vars_without_grad
------------------------------
.. autofunction:: paddle.fluid.io.load_persist_vars_without_grad
:noindex:
save_persist_vars_without_grad
------------------------------
.. autofunction:: paddle.fluid.io.save_persist_vars_without_grad
:noindex:
get_latest_checkpoint_serial
----------------------------
.. autofunction:: paddle.fluid.io.get_latest_checkpoint_serial
:noindex:

235
doc/fluid/api/layers.rst
View File

File diff suppressed because it is too large Load Diff

7
doc/fluid/api/optimizer.rst
Unescape View File

@ -89,6 +89,13 @@ DecayedAdagradOptimizer
:members:
:noindex:
RMSPropOptimizer
----------------
.. autoclass:: paddle.fluid.optimizer.RMSPropOptimizer
:members:
:noindex:
Adadelta
--------

12
doc/fluid/api/profiler.rst
Unescape View File

@ -23,3 +23,15 @@ profiler
.. autofunction:: paddle.fluid.profiler.profiler
:noindex:
start_profiler
--------------
.. autofunction:: paddle.fluid.profiler.start_profiler
:noindex:
stop_profiler
-------------
.. autofunction:: paddle.fluid.profiler.stop_profiler
:noindex:

2
doc/survey/dynamic_graph.md
Unescape View File

@ -171,7 +171,7 @@ Pytorch chooses immediate evaluation. It avoids ever materializing a "forward gr
## What can fluid learn from them?
TBD
Please refer to `paddle/contrib/dynamic/`.
# Appendix

2
doc/v2/api/config/evaluators.rst
Unescape View File

@ -101,7 +101,7 @@ value_printer
:noindex:
Detection
=====
==========
detection_map
-------------

193
doc/v2/api/config/layer.rst
View File

File diff suppressed because it is too large Load Diff

1
doc/v2/api/index_en.rst
Unescape View File

@ -8,4 +8,3 @@ API
model_configs.rst
data.rst
run_logic.rst
fluid/index.rst

1
doc/v2/build_and_install/pip_install_cn.rst
Unescape View File

@ -60,6 +60,7 @@ paddlepaddle-gpu==0.11.0 使用CUDA 7.5和cuDNN 5编译的0.11.0版
"cpu_noavx_openblas", "`paddlepaddle-latest-cp27-cp27mu-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_CpuNoavxOpenblas/.lastSuccessful/paddlepaddle-latest-cp27-cp27mu-linux_x86_64.whl>`__", "`paddlepaddle-latest-cp27-cp27m-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_CpuNoavxOpenblas/.lastSuccessful/paddlepaddle-latest-cp27-cp27m-linux_x86_64.whl>`_"
"cuda8.0_cudnn5_avx_mkl", "`paddlepaddle_gpu-latest-cp27-cp27mu-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_Cuda80cudnn5cp27cp27mu/.lastSuccessful/paddlepaddle_gpu-latest-cp27-cp27mu-linux_x86_64.whl>`__", "`paddlepaddle_gpu-latest-cp27-cp27m-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_Cuda80cudnn5cp27cp27mu/.lastSuccessful/paddlepaddle_gpu-latest-cp27-cp27m-linux_x86_64.whl>`__"
"cuda8.0_cudnn7_avx_mkl", "`paddlepaddle_gpu-latest-cp27-cp27mu-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_Cuda8cudnn7cp27cp27mu/.lastSuccessful/paddlepaddle_gpu-latest-cp27-cp27mu-linux_x86_64.whl>`__", "`paddlepaddle_gpu-latest-cp27-cp27m-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_Cuda8cudnn7cp27cp27mu/.lastSuccessful/paddlepaddle_gpu-latest-cp27-cp27m-linux_x86_64.whl>`__"
"cuda9.0_cudnn7_avx_mkl", "`paddlepaddle_gpu-latest-cp27-cp27mu-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_Cuda90cudnn7avxMkl/.lastSuccessful/paddlepaddle_gpu-latest-cp27-cp27mu-linux_x86_64.whl>`__", "`paddlepaddle_gpu-latest-cp27-cp27m-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_Cuda90cudnn7avxMkl/.lastSuccessful/paddlepaddle_gpu-latest-cp27-cp27m-linux_x86_64.whl>`__"
.. _pip_dependency:

1
doc/v2/build_and_install/pip_install_en.rst
Unescape View File

@ -63,6 +63,7 @@ If the links below shows up the login form, just click "Log in as guest" to star
"cpu_noavx_openblas", "`paddlepaddle-latest-cp27-cp27mu-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_CpuNoavxOpenblas/.lastSuccessful/paddlepaddle-latest-cp27-cp27mu-linux_x86_64.whl>`__", "`paddlepaddle-latest-cp27-cp27m-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_CpuNoavxOpenblas/.lastSuccessful/paddlepaddle-latest-cp27-cp27m-linux_x86_64.whl>`__"
"cuda8.0_cudnn5_avx_mkl", "`paddlepaddle_gpu-latest-cp27-cp27mu-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_Cuda80cudnn5cp27cp27mu/.lastSuccessful/paddlepaddle_gpu-latest-cp27-cp27mu-linux_x86_64.whl>`__", "`paddlepaddle_gpu-latest-cp27-cp27m-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_Cuda80cudnn5cp27cp27mu/.lastSuccessful/paddlepaddle_gpu-latest-cp27-cp27m-linux_x86_64.whl>`__"
"cuda8.0_cudnn7_avx_mkl", "`paddlepaddle_gpu-latest-cp27-cp27mu-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_Cuda8cudnn7cp27cp27mu/.lastSuccessful/paddlepaddle_gpu-latest-cp27-cp27mu-linux_x86_64.whl>`__", "`paddlepaddle_gpu-latest-cp27-cp27m-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_Cuda8cudnn7cp27cp27mu/.lastSuccessful/paddlepaddle_gpu-latest-cp27-cp27m-linux_x86_64.whl>`__"
"cuda9.0_cudnn7_avx_mkl", "`paddlepaddle_gpu-latest-cp27-cp27mu-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_Cuda90cudnn7avxMkl/.lastSuccessful/paddlepaddle_gpu-latest-cp27-cp27mu-linux_x86_64.whl>`__", "`paddlepaddle_gpu-latest-cp27-cp27m-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_Cuda90cudnn7avxMkl/.lastSuccessful/paddlepaddle_gpu-latest-cp27-cp27m-linux_x86_64.whl>`__"
.. _pip_dependency:

2
doc/v2/dev/contribute_to_paddle_cn.md
Unescape View File

@ -104,7 +104,7 @@ no changes added to commit (use "git add" and/or "git commit -a")
➜ docker run -it -v $(pwd):/paddle paddle:latest-dev bash -c "cd /paddle/build && ctest"
```
关于构建和测试的更多信息，请参见[这篇文档](https://github.com/PaddlePaddle/Paddle/blob/develop/doc/getstarted/build_and_install/docker_install_cn.rst)。
关于构建和测试的更多信息，请参见[使用Docker安装运行](https://github.com/PaddlePaddle/Paddle/blob/develop/doc/v2/build_and_install/docker_install_cn.rst)。
## 提交（commit）

53
paddle/contrib/inference/CMakeLists.txt
Unescape View File

@ -17,48 +17,9 @@ if(APPLE)
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wno-error=pessimizing-move")
endif(APPLE)
set(ANAKIN_INCLUDE "" CACHE STRING "root of Anakin header files")
set(ANAKIN_LIBRARY "" CACHE STRING "path of Anakin library")
set(inference_deps paddle_inference_api paddle_fluid_api)
# if anakin is set enable anakin api implementation
if(ANAKIN_INCLUDE AND ANAKIN_LIBRARY)
set(ANAKIN_FOUND ON)
else()
set(ANAKIN_FOUND OFF)
endif()
function(fetch_include_recursively root_dir)
if (IS_DIRECTORY ${root_dir})
include_directories(${root_dir})
endif()
file(GLOB ALL_SUB RELATIVE ${root_dir} ${root_dir}/*)
foreach(sub ${ALL_SUB})
if (IS_DIRECTORY ${root_dir}/${sub})
fetch_include_recursively(${root_dir}/${sub})
endif()
endforeach()
endfunction()
if (ANAKIN_FOUND)
# Anakin's code style doesn't follow google c style.
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wno-error=unused-variable -Wno-error=format-extra-args -Wno-error=comment -Wno-error=format -Wno-error=switch -Wno-error=return-type -Wno-error=non-virtual-dtor -Wno-reorder -Wno-error=cpp")
message(STATUS "Anakin for inference is enabled")
message(STATUS "Anakin is set INCLUDE:${ANAKIN_INCLUDE} LIBRARY:${ANAKIN_LIBRARY}")
fetch_include_recursively(${ANAKIN_INCLUDE})
link_directories(${ANAKIN_LIBRARY})
nv_library(inference_anakin_api SHARED SRCS paddle_inference_api.cc paddle_inference_api_anakin_engine.cc)
target_link_libraries(inference_anakin_api anakin anakin_saber_common)
list(APPEND inference_deps inference_anakin_api)
endif()
function(inference_api_test TARGET_NAME)
if (WITH_TESTING)
set(options "")
@ -79,7 +40,7 @@ function(inference_api_test TARGET_NAME)
endfunction(inference_api_test)
cc_library(paddle_inference_api
SRCS paddle_inference_api.cc paddle_inference_api_impl.cc
SRCS paddle_inference_api.cc paddle_inference_api_impl.cc
DEPS ${FLUID_CORE_MODULES} ${GLOB_OP_LIB})
cc_test(test_paddle_inference_api
@ -89,9 +50,17 @@ cc_test(test_paddle_inference_api
inference_api_test(test_paddle_inference_api_impl
ARGS test_word2vec test_image_classification)
if (ANAKIN_FOUND)
if (WITH_ANAKIN AND WITH_TESTING) # only needed in CI
# Due to Anakin do not have official library releases and the versions of protobuf and cuda do not match Paddle's,
# so anakin library will not be merged to our official inference library. To use anakin prediction API, one need to
# compile the libinference_anakin_api.a and compile with anakin.so.
nv_library(inference_anakin_api SHARED SRCS paddle_inference_api.cc paddle_inference_api_anakin_engine.cc)
target_compile_options(inference_anakin_api BEFORE PUBLIC ${ANAKIN_COMPILE_EXTRA_FLAGS})
target_link_libraries(inference_anakin_api anakin anakin_saber_common)
cc_test(inference_anakin_test SRCS paddle_inference_api_anakin_engine_tester.cc
DEPS ${inference_deps})
ARGS --model=${ANAKIN_INSTALL_DIR}/mobilenet_v2.anakin.bin
DEPS inference_anakin_api)
target_compile_options(inference_anakin_test BEFORE PUBLIC ${ANAKIN_COMPILE_EXTRA_FLAGS})
endif()
if(WITH_TESTING)

3
paddle/contrib/inference/paddle_inference_api_anakin_engine.cc
Unescape View File

@ -12,9 +12,8 @@
// See the License for the specific language governing permissions and
// limitations under the License.
#include <cuda.h>
#include "paddle/contrib/inference/paddle_inference_api_anakin_engine.h"
#include <cuda.h>
namespace paddle {

3
paddle/contrib/inference/paddle_inference_api_anakin_engine.h
Unescape View File

@ -19,10 +19,9 @@ limitations under the License. */
#pragma once
// NOTE This header file do not have namespace.
//#include <test/framework/net/paddle_api.h>
#include "paddle/contrib/inference/paddle_inference_api.h"
// from anakin
#include "framework/core/net/net.h"
#include "saber/saber_types.h"

6
paddle/contrib/inference/paddle_inference_api_anakin_engine_tester.cc
Unescape View File

@ -12,17 +12,19 @@ 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 <gflags/gflags.h>
#include <glog/logging.h>
#include <gtest/gtest.h>
#include "gflags/gflags.h"
#include "paddle/contrib/inference/paddle_inference_api.h"
DEFINE_string(model, "", "Directory of the inference model.");
namespace paddle {
AnakinConfig GetConfig() {
AnakinConfig config;
config.model_file = "./mobilenet_v2.anakin.bin";
config.model_file = FLAGS_model;
config.device = 0;
config.max_batch_size = 1;
return config;

9
paddle/fluid/framework/executor.cc
Unescape View File

@ -330,8 +330,12 @@ void Executor::RunPreparedContext(ExecutorPrepareContext* ctx, Scope* scope,
}
for (auto& op : ctx->ops_) {
VLOG(3) << place_ << " " << op->DebugStringEx(local_scope);
VLOG(4) << place_ << " " << op->DebugStringEx(local_scope);
op->Run(*local_scope, place_);
// NOTE! Please do not delete this line, it's usefull because the debug
// string before and after op.run are different, after run the output
// will have right shape which is usefull for debug.
VLOG(3) << place_ << " " << op->DebugStringEx(local_scope);
if (FLAGS_benchmark) {
VLOG(2) << "Memory used after operator " + op->Type() + " running: "
@ -402,6 +406,9 @@ void Executor::EnableMKLDNN(const ProgramDesc& program) {
}
}
}
#else
LOG(WARNING)
<< "'MKLDNN' is not supported, Please re-compile with WITH_MKLDNN option";
#endif
}

Some files were not shown because too many files have changed in this diff Show More

Write Preview
Loading…
Cancel
Save