Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into fix_api_reference_docs

Author: reyoung

AUTHORS.md

@@ -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 |

CMakeLists.txt

@@ -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)

benchmark/fluid/models/machine_translation.py

@@ -173,21 +173,6 @@ def simple_attention(encoder_vec, encoder_proj, decoder_state):
         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')

benchmark/fluid/models/stacked_dynamic_lstm.py

@@ -125,18 +125,3 @@ def gate_common(
         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

cmake/external/anakin.cmake

@@ -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()

cmake/external/openblas.cmake

@@ -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")
 

cmake/inference_lib.cmake

@@ -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

doc/fluid/api/gen_doc.sh

@@ -1,5 +1,5 @@
 #!/bin/bash
-python gen_doc.py layers --submodules control_flow device io nn ops tensor detection learning_rate_scheduler > 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

doc/fluid/api/initializer.rst

@@ -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:
+

doc/fluid/api/io.rst

@@ -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:
+