Skip to content →

Tag: Machine Learning

Notes to myself: RTX 2070, Cuda, cudnn, caffe, and faceswap

Published 2019/05/19 by Daoyuan Li

Install NVIDIA driver for RTX 2070: https://www.geforce.com/drivers

sudo bash ~/Downloads/NVIDIA-Linux-x86_64-415.27.run

Install CUDA 10.0: https://developer.nvidia.com/cuda-downloads

sudo bash ~/Downloads/cuda_10.0.130_410.48_linux.run

DO NOT re-install the drivers suggested by the CUDA installer:
Install NVIDIA Accelerated Graphics Driver for Linux-x86_64 410.48?
(y)es/(n)o/(q)uit: n

$ nvidia-smi
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 415.27       Driver Version: 415.27       CUDA Version: 10.0     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  GeForce RTX 2070    Off  | 00000000:1F:00.0 Off |                  N/A |
| 36%   31C    P0    N/A /  N/A |      0MiB /  7949MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+

+-----------------------------------------------------------------------------+
| Processes:                                                       GPU Memory |
|  GPU       PID   Type   Process name                             Usage      |
|=============================================================================|
|  No running processes found                                                 |
+-----------------------------------------------------------------------------+

CuDNN:

cuDNN Runtime Library for Ubuntu18.04 (Deb)

cuDNN Developer Library for Ubuntu18.04 (Deb)

Caffe:

CMake Error: The following variables are used in this project, but they are set to NOTFOUND.
Please set them or make sure they are set and tested correctly in the CMake files:
CUDA_cublas_device_LIBRARY (ADVANCED)
linked by target “caffe” in directory /home/dli/Projects/caffe/src/caffe

Upgrade cmake:

$ ~/cmake-3.13.3-Linux-x86_64/bin/cmake -DCMAKE_INSTALL_PREFIX:PATH=/usr ..
-- Boost version: 1.65.1
-- Found the following Boost libraries:
--   system
--   thread
--   filesystem
--   chrono
--   date_time
--   atomic
-- Found gflags  (include: /usr/include, library: /usr/lib/x86_64-linux-gnu/libgflags.so)
-- Found glog    (include: /usr/include, library: /usr/lib/x86_64-linux-gnu/libglog.so)
-- Found PROTOBUF Compiler: /usr/bin/protoc
-- HDF5: Using hdf5 compiler wrapper to determine C configuration
-- HDF5: Using hdf5 compiler wrapper to determine CXX configuration
-- Found lmdb    (include: /usr/include, library: /usr/lib/x86_64-linux-gnu/liblmdb.so)
-- Found LevelDB (include: /usr/include, library: /usr/lib/x86_64-linux-gnu/libleveldb.so)
-- Found Snappy  (include: /usr/include, library: /usr/lib/x86_64-linux-gnu/libsnappy.so)
-- CUDA detected: 10.0
-- Found cuDNN: ver. 7.4.2 found (include: /usr/include, library: /usr/lib/x86_64-linux-gnu/libcudnn.so)
-- Added CUDA NVCC flags for: sm_75
-- OpenCV found (/usr/share/OpenCV)
-- Found Atlas (include: /usr/include/x86_64-linux-gnu library: /usr/lib/x86_64-linux-gnu/libatlas.so lapack: /usr/lib/x86_64-linux-gnu/liblapack.so
-- NumPy ver. 1.16.0 found (include: /usr/local/lib/python2.7/dist-packages/numpy/core/include)
-- Boost version: 1.65.1
-- Found the following Boost libraries:
--   python
-- Detected Doxygen OUTPUT_DIRECTORY: ./doxygen/
--
-- ******************* Caffe Configuration Summary *******************
-- General:
--   Version           :   1.0.0
--   Git               :   1.0-132-g99bd9979
--   System            :   Linux
--   C++ compiler      :   /usr/bin/c++
--   Release CXX flags :   -O3 -DNDEBUG -fPIC -Wall -Wno-sign-compare -Wno-uninitialized
--   Debug CXX flags   :   -g -fPIC -Wall -Wno-sign-compare -Wno-uninitialized
--   Build type        :   Release
--
--   BUILD_SHARED_LIBS :   ON
--   BUILD_python      :   ON
--   BUILD_matlab      :   OFF
--   BUILD_docs        :   ON
--   CPU_ONLY          :   OFF
--   USE_OPENCV        :   ON
--   USE_LEVELDB       :   ON
--   USE_LMDB          :   ON
--   USE_NCCL          :   OFF
--   ALLOW_LMDB_NOLOCK :   OFF
--   USE_HDF5          :   ON
--
-- Dependencies:
--   BLAS              :   Yes (Atlas)
--   Boost             :   Yes (ver. 1.65)
--   glog              :   Yes
--   gflags            :   Yes
--   protobuf          :   Yes (ver. 3.0.0)
--   lmdb              :   Yes (ver. 0.9.21)
--   LevelDB           :   Yes (ver. 1.20)
--   Snappy            :   Yes (ver. ..)
--   OpenCV            :   Yes (ver. 3.2.0)
--   CUDA              :   Yes (ver. 10.0)
--
-- NVIDIA CUDA:
--   Target GPU(s)     :   Auto
--   GPU arch(s)       :   sm_75
--   cuDNN             :   Yes (ver. 7.4.2)
--
-- Python:
--   Interpreter       :   /usr/bin/python2.7 (ver. 2.7.15)
--   Libraries         :   /usr/lib/x86_64-linux-gnu/libpython2.7.so (ver 2.7.15rc1)
--   NumPy             :   /usr/local/lib/python2.7/dist-packages/numpy/core/include (ver 1.16.0)
--
-- Documentaion:
--   Doxygen           :   /usr/bin/doxygen (1.8.13)
--   config_file       :   /home/dli/Projects/caffe/.Doxyfile
--
-- Install:
--   Install path      :   /usr
--
-- Configuring done
-- Generating done
-- Build files have been written to: /home/dli/Projects/caffe/build
$ make -j16 all
$ make pycaffe
$ sudo make install
[  2%] Built target caffeproto
[ 87%] Built target caffe
[ 89%] Built target upgrade_solver_proto_text
[ 90%] Built target compute_image_mean
[ 90%] Built target caffe.bin
[ 91%] Built target upgrade_net_proto_binary
[ 93%] Built target convert_imageset
[ 94%] Built target extract_features
[ 95%] Built target upgrade_net_proto_text
[ 95%] Built target classification
[ 95%] Built target convert_mnist_data
[ 97%] Built target convert_cifar_data
[ 98%] Built target convert_mnist_siamese_data
[100%] Built target pycaffe
Install the project...
-- Install configuration: "Release"
-- Installing: /usr/share/Caffe/CaffeConfig.cmake
-- Installing: /usr/share/Caffe/CaffeTargets.cmake
-- Installing: /usr/share/Caffe/CaffeTargets-release.cmake
-- Installing: /usr/include/caffe
-- Installing: /usr/include/caffe/test
-- Installing: /usr/include/caffe/test/test_gradient_check_util.hpp
-- Installing: /usr/include/caffe/test/test_caffe_main.hpp
-- Installing: /usr/include/caffe/layers
-- Installing: /usr/include/caffe/layers/cudnn_tanh_layer.hpp
-- Installing: /usr/include/caffe/layers/absval_layer.hpp
-- Installing: /usr/include/caffe/layers/multinomial_logistic_loss_layer.hpp
-- Installing: /usr/include/caffe/layers/dummy_data_layer.hpp
-- Installing: /usr/include/caffe/layers/recurrent_layer.hpp
-- Installing: /usr/include/caffe/layers/scale_layer.hpp
-- Installing: /usr/include/caffe/layers/hdf5_data_layer.hpp
-- Installing: /usr/include/caffe/layers/cudnn_sigmoid_layer.hpp
-- Installing: /usr/include/caffe/layers/clip_layer.hpp
-- Installing: /usr/include/caffe/layers/hinge_loss_layer.hpp
-- Installing: /usr/include/caffe/layers/mvn_layer.hpp
-- Installing: /usr/include/caffe/layers/relu_layer.hpp
-- Installing: /usr/include/caffe/layers/hdf5_output_layer.hpp
-- Installing: /usr/include/caffe/layers/contrastive_loss_layer.hpp
-- Installing: /usr/include/caffe/layers/lrn_layer.hpp
-- Installing: /usr/include/caffe/layers/accuracy_layer.hpp
-- Installing: /usr/include/caffe/layers/conv_layer.hpp
-- Installing: /usr/include/caffe/layers/infogain_loss_layer.hpp
-- Installing: /usr/include/caffe/layers/im2col_layer.hpp
-- Installing: /usr/include/caffe/layers/base_conv_layer.hpp
-- Installing: /usr/include/caffe/layers/euclidean_loss_layer.hpp
-- Installing: /usr/include/caffe/layers/crop_layer.hpp
-- Installing: /usr/include/caffe/layers/window_data_layer.hpp
-- Installing: /usr/include/caffe/layers/bnll_layer.hpp
-- Installing: /usr/include/caffe/layers/eltwise_layer.hpp
-- Installing: /usr/include/caffe/layers/prelu_layer.hpp
-- Installing: /usr/include/caffe/layers/filter_layer.hpp
-- Installing: /usr/include/caffe/layers/cudnn_lcn_layer.hpp
-- Installing: /usr/include/caffe/layers/reduction_layer.hpp
-- Installing: /usr/include/caffe/layers/sigmoid_cross_entropy_loss_layer.hpp
-- Installing: /usr/include/caffe/layers/swish_layer.hpp
-- Installing: /usr/include/caffe/layers/slice_layer.hpp
-- Installing: /usr/include/caffe/layers/cudnn_softmax_layer.hpp
-- Installing: /usr/include/caffe/layers/reshape_layer.hpp
-- Installing: /usr/include/caffe/layers/silence_layer.hpp
-- Installing: /usr/include/caffe/layers/sigmoid_layer.hpp
-- Installing: /usr/include/caffe/layers/power_layer.hpp
-- Installing: /usr/include/caffe/layers/spp_layer.hpp
-- Installing: /usr/include/caffe/layers/exp_layer.hpp
-- Installing: /usr/include/caffe/layers/pooling_layer.hpp
-- Installing: /usr/include/caffe/layers/input_layer.hpp
-- Installing: /usr/include/caffe/layers/data_layer.hpp
-- Installing: /usr/include/caffe/layers/lstm_layer.hpp
-- Installing: /usr/include/caffe/layers/neuron_layer.hpp                                                                                                                                     
-- Installing: /usr/include/caffe/layers/split_layer.hpp
-- Installing: /usr/include/caffe/layers/threshold_layer.hpp
-- Installing: /usr/include/caffe/layers/base_data_layer.hpp
-- Installing: /usr/include/caffe/layers/log_layer.hpp
-- Installing: /usr/include/caffe/layers/loss_layer.hpp
-- Installing: /usr/include/caffe/layers/rnn_layer.hpp
-- Installing: /usr/include/caffe/layers/elu_layer.hpp
-- Installing: /usr/include/caffe/layers/memory_data_layer.hpp
-- Installing: /usr/include/caffe/layers/cudnn_relu_layer.hpp
-- Installing: /usr/include/caffe/layers/tanh_layer.hpp
-- Installing: /usr/include/caffe/layers/flatten_layer.hpp
-- Installing: /usr/include/caffe/layers/dropout_layer.hpp
-- Installing: /usr/include/caffe/layers/bias_layer.hpp
-- Installing: /usr/include/caffe/layers/softmax_loss_layer.hpp
-- Installing: /usr/include/caffe/layers/deconv_layer.hpp
-- Installing: /usr/include/caffe/layers/inner_product_layer.hpp
-- Installing: /usr/include/caffe/layers/batch_norm_layer.hpp
-- Installing: /usr/include/caffe/layers/cudnn_conv_layer.hpp
-- Installing: /usr/include/caffe/layers/parameter_layer.hpp
-- Installing: /usr/include/caffe/layers/tile_layer.hpp
-- Installing: /usr/include/caffe/layers/cudnn_deconv_layer.hpp
-- Installing: /usr/include/caffe/layers/cudnn_pooling_layer.hpp
-- Installing: /usr/include/caffe/layers/python_layer.hpp
-- Installing: /usr/include/caffe/layers/embed_layer.hpp
-- Installing: /usr/include/caffe/layers/image_data_layer.hpp
-- Installing: /usr/include/caffe/layers/batch_reindex_layer.hpp
-- Installing: /usr/include/caffe/layers/softmax_layer.hpp
-- Installing: /usr/include/caffe/layers/concat_layer.hpp
-- Installing: /usr/include/caffe/layers/cudnn_lrn_layer.hpp
-- Installing: /usr/include/caffe/layers/argmax_layer.hpp
-- Installing: /usr/include/caffe/blob.hpp
-- Installing: /usr/include/caffe/util
-- Installing: /usr/include/caffe/util/upgrade_proto.hpp
-- Installing: /usr/include/caffe/util/blocking_queue.hpp
-- Installing: /usr/include/caffe/util/db.hpp
-- Installing: /usr/include/caffe/util/db_lmdb.hpp
-- Installing: /usr/include/caffe/util/insert_splits.hpp
-- Installing: /usr/include/caffe/util/format.hpp
-- Installing: /usr/include/caffe/util/im2col.hpp
-- Installing: /usr/include/caffe/util/gpu_util.cuh
-- Installing: /usr/include/caffe/util/signal_handler.h
-- Installing: /usr/include/caffe/util/cudnn.hpp
-- Installing: /usr/include/caffe/util/hdf5.hpp
-- Installing: /usr/include/caffe/util/io.hpp
-- Installing: /usr/include/caffe/util/nccl.hpp
-- Installing: /usr/include/caffe/util/rng.hpp
-- Installing: /usr/include/caffe/util/db_leveldb.hpp
-- Installing: /usr/include/caffe/util/mkl_alternate.hpp
-- Installing: /usr/include/caffe/util/benchmark.hpp
-- Installing: /usr/include/caffe/util/device_alternate.hpp
-- Installing: /usr/include/caffe/util/math_functions.hpp
-- Installing: /usr/include/caffe/sgd_solvers.hpp
-- Installing: /usr/include/caffe/layer.hpp
-- Installing: /usr/include/caffe/net.hpp
-- Installing: /usr/include/caffe/syncedmem.hpp
-- Installing: /usr/include/caffe/solver.hpp
-- Installing: /usr/include/caffe/filler.hpp
-- Installing: /usr/include/caffe/layer_factory.hpp                                                                                                                                            [0/863]
-- Installing: /usr/include/caffe/data_transformer.hpp
-- Installing: /usr/include/caffe/solver_factory.hpp
-- Installing: /usr/include/caffe/parallel.hpp
-- Installing: /usr/include/caffe/common.hpp
-- Installing: /usr/include/caffe/internal_thread.hpp
-- Installing: /usr/include/caffe/caffe.hpp
-- Installing: /usr/include/caffe/proto/caffe.pb.h
-- Installing: /usr/lib/x86_64-linux-gnu/libcaffe.so.1.0.0
-- Set runtime path of "/usr/lib/x86_64-linux-gnu/libcaffe.so.1.0.0" to "/usr/lib/x86_64-linux-gnu:/usr/lib/x86_64-linux-gnu/hdf5/serial:/usr/local/cuda-10.0/lib64"
-- Installing: /usr/lib/x86_64-linux-gnu/libcaffe.so
-- Installing: /usr/lib/x86_64-linux-gnu/libcaffeproto.a
-- Installing: /usr/python/caffe/proto/caffe_pb2.py
-- Installing: /usr/python/caffe/proto/__init__.py
-- Installing: /usr/bin/caffe
-- Set runtime path of "/usr/bin/caffe" to "/usr/lib/x86_64-linux-gnu:/usr/lib/x86_64-linux-gnu/hdf5/serial:/usr/local/cuda-10.0/lib64"
-- Installing: /usr/bin/compute_image_mean
-- Set runtime path of "/usr/bin/compute_image_mean" to "/usr/lib/x86_64-linux-gnu:/usr/lib/x86_64-linux-gnu/hdf5/serial:/usr/local/cuda-10.0/lib64"
-- Installing: /usr/bin/convert_imageset
-- Set runtime path of "/usr/bin/convert_imageset" to "/usr/lib/x86_64-linux-gnu:/usr/lib/x86_64-linux-gnu/hdf5/serial:/usr/local/cuda-10.0/lib64"
-- Installing: /usr/bin/extract_features
-- Set runtime path of "/usr/bin/extract_features" to "/usr/lib/x86_64-linux-gnu:/usr/lib/x86_64-linux-gnu/hdf5/serial:/usr/local/cuda-10.0/lib64"
-- Installing: /usr/bin/upgrade_net_proto_binary
-- Set runtime path of "/usr/bin/upgrade_net_proto_binary" to "/usr/lib/x86_64-linux-gnu:/usr/lib/x86_64-linux-gnu/hdf5/serial:/usr/local/cuda-10.0/lib64"
-- Installing: /usr/bin/upgrade_net_proto_text
-- Set runtime path of "/usr/bin/upgrade_net_proto_text" to "/usr/lib/x86_64-linux-gnu:/usr/lib/x86_64-linux-gnu/hdf5/serial:/usr/local/cuda-10.0/lib64"
-- Installing: /usr/bin/upgrade_solver_proto_text
-- Set runtime path of "/usr/bin/upgrade_solver_proto_text" to "/usr/lib/x86_64-linux-gnu:/usr/lib/x86_64-linux-gnu/hdf5/serial:/usr/local/cuda-10.0/lib64"
-- Installing: /usr/bin/convert_cifar_data
-- Set runtime path of "/usr/bin/convert_cifar_data" to "/usr/lib/x86_64-linux-gnu:/usr/lib/x86_64-linux-gnu/hdf5/serial:/usr/local/cuda-10.0/lib64"
-- Installing: /usr/bin/classification
-- Set runtime path of "/usr/bin/classification" to "/usr/lib/x86_64-linux-gnu:/usr/lib/x86_64-linux-gnu/hdf5/serial:/usr/local/cuda-10.0/lib64"
-- Installing: /usr/bin/convert_mnist_data
-- Set runtime path of "/usr/bin/convert_mnist_data" to "/usr/lib/x86_64-linux-gnu:/usr/lib/x86_64-linux-gnu/hdf5/serial:/usr/local/cuda-10.0/lib64"
-- Installing: /usr/bin/convert_mnist_siamese_data
-- Set runtime path of "/usr/bin/convert_mnist_siamese_data" to "/usr/lib/x86_64-linux-gnu:/usr/lib/x86_64-linux-gnu/hdf5/serial:/usr/local/cuda-10.0/lib64"
-- Installing: /usr/python/classify.py
-- Installing: /usr/python/detect.py
-- Installing: /usr/python/draw_net.py
-- Installing: /usr/python/requirements.txt
-- Installing: /usr/python/train.py
-- Up-to-date: /usr/python/caffe
-- Installing: /usr/python/caffe/io.py
-- Installing: /usr/python/caffe/pycaffe.py
-- Up-to-date: /usr/python/caffe/proto
-- Installing: /usr/python/caffe/proto/caffe_pb2.py
-- Installing: /usr/python/caffe/proto/__init__.py
-- Installing: /usr/python/caffe/coord_map.py
-- Installing: /usr/python/caffe/classifier.py
-- Installing: /usr/python/caffe/__init__.py
-- Installing: /usr/python/caffe/imagenet
-- Installing: /usr/python/caffe/imagenet/ilsvrc_2012_mean.npy
-- Installing: /usr/python/caffe/net_spec.py
-- Installing: /usr/python/caffe/detector.py
-- Installing: /usr/python/caffe/draw.py
-- Installing: /usr/python/caffe/_caffe.so
-- Set runtime path of "/usr/python/caffe/_caffe.so" to "/usr/lib/x86_64-linux-gnu:/usr/lib/x86_64-linux-gnu/hdf5/serial:/usr/local/cuda-10.0/lib64"

face_swap

$ git diff
diff --git a/CMakeLists.txt b/CMakeLists.txt
index cb34534..762363f 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -69,21 +69,30 @@ find_package(OpenCV REQUIRED highgui imgproc imgcodecs calib3d photo)
 # dlib
 find_package(dlib REQUIRED)

+if(DEFINED CAFFE_DIR)
+    list(APPEND CMAKE_PREFIX_PATH ${CAFFE_DIR})
+endif()
+
 # Caffe
 find_package(Caffe REQUIRED)

+if(DEFINED HDF5_DIR)
+    list(APPEND CMAKE_MODULE_PATH ${HDF5_DIR})
+endif()
+
 # HDF5
 if(MSVC)
   # Find HDF5 using it's hdf5-config.cmake file with MSVC
   if(DEFINED HDF5_DIR)
     list(APPEND CMAKE_MODULE_PATH ${HDF5_DIR})
   endif()
-  find_package(HDF5 COMPONENTS C HL CXX REQUIRED)
   set(HDF5_LIBRARIES hdf5-shared hdf5_cpp-shared)
   set(HDF5_HL_LIBRARIES hdf5_hl-shared)
 else()
   #find_package(HDF5 COMPONENTS HL REQUIRED)
+  # find_package(HDF5 COMPONENTS C CXX HL REQUIRED)
   find_package(HDF5 COMPONENTS C CXX HL REQUIRED)
+  # find_package(HDF5 COMPONENTS C HL CXX REQUIRED PATHS /usr/lib/x86_64-linux-gnu/hdf5/serial)
 endif()
 #find_package(HDF5 REQUIRED CXX)

@@ -132,23 +141,23 @@ endif()
 set(FACE_SWAP_TARGETS face_swap iris_sfs)
 export(TARGETS ${FACE_SWAP_TARGETS}
   FILE "${PROJECT_BINARY_DIR}/face_swap-targets.cmake")
-
+
 # Export the package for use from the build-tree
 # (this registers the build-tree with a global CMake-registry)
 export(PACKAGE face_swap)
-
+
 # Create config files
 configure_file(cmake/face_swap-config.cmake.in
   "${PROJECT_BINARY_DIR}/face_swap-config.cmake" @ONLY)
 configure_file(cmake/face_swap-config-version.cmake.in
   "${PROJECT_BINARY_DIR}/face_swap-config-version.cmake" @ONLY)
-
+
 # Install config files
 install(FILES
   "${PROJECT_BINARY_DIR}/face_swap-config.cmake"
   "${PROJECT_BINARY_DIR}/face_swap-config-version.cmake"
   DESTINATION "cmake" COMPONENT dev)
-
+
 # Install the export set for use with the install-tree
 install(EXPORT face_swap-targets DESTINATION cmake COMPONENT dev)

diff --git a/face_swap/CMakeLists.txt b/face_swap/CMakeLists.txt
index 1fcb187..ca56ecd 100644
--- a/face_swap/CMakeLists.txt
+++ b/face_swap/CMakeLists.txt
@@ -44,6 +44,7 @@ target_include_directories(face_swap PUBLIC
        ${HDF5_INCLUDE_DIRS}
 )
 target_link_libraries(face_swap PUBLIC
+       uuid
        iris_sfs
        ${OpenCV_LIBS}
        dlib::dlib

$ ~/cmake-3.13.3-Linux-x86_64/bin/cmake -DWITH_BOOST_STATIC=OFF -DBUILD_INTERFACE_PYTHON=ON -DBUILD_SHARED_LIBS=OFF -DBUILD_APPS=ON -DBUILD_TESTS=OFF cmake -DCM
AKE_INSTALL_PREFIX=/usr/local/face_swap -DCMAKE_BUILD_TYPE=Release ..
-- The C compiler identification is GNU 7.3.0
-- The CXX compiler identification is GNU 7.3.0
-- Check for working C compiler: /usr/bin/cc
-- Check for working C compiler: /usr/bin/cc -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Detecting C compile features
-- Detecting C compile features - done
-- Check for working CXX compiler: /usr/bin/c++
-- Check for working CXX compiler: /usr/bin/c++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Looking for pthread.h
-- Looking for pthread.h - found
-- Looking for pthread_create
-- Looking for pthread_create - not found
-- Looking for pthread_create in pthreads
-- Looking for pthread_create in pthreads - not found
-- Looking for pthread_create in pthread
-- Looking for pthread_create in pthread - found
-- Found Threads: TRUE
-- Boost version: 1.65.1
-- Found the following Boost libraries:
--   filesystem
--   program_options
--   regex
--   timer
--   thread
--   chrono
--   system
--   date_time
--   atomic
-- Found OpenCV: /usr (found version "3.2.0") found components:  highgui imgproc imgcodecs calib3d photo
-- HDF5: Using hdf5 compiler wrapper to determine C configuration
-- HDF5: Using hdf5 compiler wrapper to determine CXX configuration
-- Found HDF5: /usr/lib/x86_64-linux-gnu/hdf5/serial/libhdf5_cpp.so;/usr/lib/x86_64-linux-gnu/hdf5/serial/libhdf5.so;/usr/lib/x86_64-linux-gnu/libpthread.so;/usr/lib/x86_64-linux-gnu/libsz.so;/usr/lib/x86_64-linux-gnu/libz.so;/usr/lib/x86_64-linux-gnu/libdl.so;/usr/lib/x86_64-linux-gnu/libm.so (found version "1.10.0.1") found components:  C CXX HL
-- Found Eigen3: /usr/include/eigen3 (Required is at least version "2.91.0")
-- Found Protobuf: /usr/lib/x86_64-linux-gnu/libprotobuf.so;-lpthread (found version "3.0.0")
-- Found Doxygen: /usr/bin/doxygen (found version "1.8.13") found components:  doxygen missing components:  dot
-- Found SWIG: /usr/bin/swig3.0 (found version "3.0.12")
-- Performing Test COMPILER_HAS_HIDDEN_VISIBILITY
-- Performing Test COMPILER_HAS_HIDDEN_VISIBILITY - Success
-- Performing Test COMPILER_HAS_HIDDEN_INLINE_VISIBILITY
-- Performing Test COMPILER_HAS_HIDDEN_INLINE_VISIBILITY - Success
-- Performing Test COMPILER_HAS_DEPRECATED_ATTR
-- Performing Test COMPILER_HAS_DEPRECATED_ATTR - Success
-- Found PythonLibs: /usr/lib/x86_64-linux-gnu/libpython3.6m.so (found version "3.6.7")
-- Boost version: 1.65.1
-- Found the following Boost libraries:
--   python
--   numpy
-- Configuring done
-- Generating done
-- Build files have been written to: /home/dli/Projects/face_swap/build
$ make -j16
$ sudo make install
[ 20%] Built target iris_sfs
[ 66%] Built target face_swap
[ 73%] Built target face_swap_image
[ 80%] Built target face_swap_batch
[ 86%] Built target face_swap_single2many
[ 93%] Built target face_swap_image2video
[100%] Built target face_swap_py
Install the project...
-- Install configuration: "Release"
-- Installing: /usr/local/face_swap/cmake/face_swap-config.cmake
-- Installing: /usr/local/face_swap/cmake/face_swap-config-version.cmake
-- Installing: /usr/local/face_swap/cmake/face_swap-targets.cmake
-- Installing: /usr/local/face_swap/cmake/face_swap-targets-release.cmake
-- Installing: /usr/local/face_swap/data/images/brad_pitt_01.jpg
-- Installing: /usr/local/face_swap/data/images/bruce_willis_01.jpg
-- Installing: /usr/local/face_swap/lib/libiris_sfs.a
-- Installing: /usr/local/face_swap/include/face_swap/face_swap_export.h
-- Installing: /usr/local/face_swap/lib/libface_swap.a
-- Installing: /usr/local/face_swap/include/face_swap/basel_3dmm.h
-- Installing: /usr/local/face_swap/include/face_swap/cnn_3dmm.h
-- Installing: /usr/local/face_swap/include/face_swap/cnn_3dmm_expr.h
-- Installing: /usr/local/face_swap/include/face_swap/face_seg.h
-- Installing: /usr/local/face_swap/include/face_swap/face_swap_engine.h
-- Installing: /usr/local/face_swap/include/face_swap/face_swap_engine_impl.h
-- Installing: /usr/local/face_swap/include/face_swap/face_swap_c_interface.h
-- Installing: /usr/local/face_swap/include/face_swap/render_utilities.h
-- Installing: /usr/local/face_swap/include/face_swap/utilities.h
-- Installing: /usr/local/face_swap/include/face_swap/face_detection_landmarks.h
-- Installing: /usr/local/face_swap/include/face_swap/landmarks_utilities.h
-- Installing: /usr/local/face_swap/include/face_swap/segmentation_utilities.h
-- Installing: /usr/local/face_swap/bin/face_swap_image
-- Set runtime path of "/usr/local/face_swap/bin/face_swap_image" to ""
-- Installing: /usr/local/face_swap/bin/face_swap_image.cfg
-- Installing: /usr/local/face_swap/bin/face_swap_batch
-- Set runtime path of "/usr/local/face_swap/bin/face_swap_batch" to ""
-- Installing: /usr/local/face_swap/bin/face_swap_batch.cfg
-- Installing: /usr/local/face_swap/bin/face_swap_single2many
-- Set runtime path of "/usr/local/face_swap/bin/face_swap_single2many" to ""
-- Installing: /usr/local/face_swap/bin/face_swap_single2many.cfg
-- Installing: /usr/local/face_swap/bin/face_swap_image2video
-- Set runtime path of "/usr/local/face_swap/bin/face_swap_image2video" to ""
-- Installing: /usr/local/face_swap/bin/face_swap_image2video.cfg
-- Installing: /usr/local/face_swap/interfaces/python/face_swap_py.so
-- Set runtime path of "/usr/local/face_swap/interfaces/python/face_swap_py.so" to ""
$ ls /face_swap/data/
3dmm_cnn_resnet_101.caffemodel          BaselFace.dat                           face_seg_fcn8s_300_no_aug.zip           images/
3dmm_cnn_resnet_101_deploy.prototxt     BaselFaceModel_mod_wForehead_noEars.h5  face_seg_fcn8s.caffemodel               shape_predictor_68_face_landmarks.dat
3dmm_cnn_resnet_101_mean.binaryproto    face_seg_fcn8s_300.caffemodel           face_seg_fcn8s_deploy.prototxt
3dmm_cnn_resnet_101.zip                 face_seg_fcn8s_300_deploy.prototxt      face_seg_fcn8s.zip
$ sudo cp /face_swap/data/* /usr/local/face_swap/data/ -r
face_swap_image --cfg /usr/local/face_swap/bin/face_swap_image.cfg --input dao.jpg --input min1.jpg --output o2.jpg
Leave a Comment

Installing Theano and CUDA on Mac OS X

Published 2015/10/28 by Daoyuan Li

I started trying Theano today and wanted to use the GPU (NVIDIA GeForce GT 750M 2048 MB) on my Mac. Here’s a brief instruction on how to use the GPU on Mac, largely following the instructions from http://deeplearning.net/software/theano/install.html#mac-os.

Install Theano:

$ pip install Theano

Download and install CUDA: https://developer.nvidia.com/cuda-downloads

Put the following lines into your ~/.bash_profile:

# Theano and CUDA
PATH="/Developer/NVIDIA/CUDA-7.5/bin/:$PATH"
export LD_LIBRARY_PATH=/Developer/NVIDIA/CUDA-7.5/lib/
export CUDA_ROOT=/Developer/NVIDIA/CUDA-7.5/
export THEANO_FLAGS='mode=FAST_RUN,device=gpu,floatX=float32'

Note that the PATH line is necessary. Otherwise you may see the following message:

ERROR (theano.sandbox.cuda): nvcc compiler not found on $PATH. Check your nvcc installation and try again.

Configure Theano:

$ cat .theanorc 
[gcc]
cxxflags = -L/usr/local/lib -L/Developer/NVIDIA/CUDA-7.5/lib/

Test if GPU is used:

$ cat check.py 
from theano import function, config, shared, sandbox
import theano.tensor as T
import numpy
import time

vlen = 10 * 30 * 768  # 10 x #cores x # threads per core
iters = 1000

rng = numpy.random.RandomState(22)
x = shared(numpy.asarray(rng.rand(vlen), config.floatX))
f = function([], T.exp(x))
print(f.maker.fgraph.toposort())
t0 = time.time()
for i in xrange(iters):
    r = f()
t1 = time.time()
print("Looping %d times took %f seconds" % (iters, t1 - t0))
print("Result is %s" % (r,))
if numpy.any([isinstance(x.op, T.Elemwise) for x in f.maker.fgraph.toposort()]):
    print('Used the cpu')
else:
    print('Used the gpu')

$ THEANO_FLAGS=mode=FAST_RUN,device=cpu,floatX=float32 time python check.py 
[Elemwise{exp,no_inplace}(<TensorType(float32, vector)>)]
Looping 1000 times took 1.743682 seconds
Result is [ 1.23178029  1.61879337  1.52278066 ...,  2.20771813  2.29967761
  1.62323284]
Used the cpu
        2.47 real         2.19 user         0.27 sys
$ THEANO_FLAGS=mode=FAST_RUN,device=gpu,floatX=float32 time python check.py 
Using gpu device 0: GeForce GT 750M
[GpuElemwise{exp,no_inplace}(<CudaNdarrayType(float32, vector)>), HostFromGpu(GpuElemwise{exp,no_inplace}.0)]
Looping 1000 times took 1.186971 seconds
Result is [ 1.23178029  1.61879349  1.52278066 ...,  2.20771813  2.29967761
  1.62323296]
Used the gpu
        2.09 real         1.59 user         0.41 sys

A more realistic example:

$ cat lr.py 
import numpy
import theano
import theano.tensor as T
rng = numpy.random

N = 400
feats = 784
D = (rng.randn(N, feats).astype(theano.config.floatX),
rng.randint(size=N,low=0, high=2).astype(theano.config.floatX))
training_steps = 10000

# Declare Theano symbolic variables
x = T.matrix("x")
y = T.vector("y")
w = theano.shared(rng.randn(feats).astype(theano.config.floatX), name="w")
b = theano.shared(numpy.asarray(0., dtype=theano.config.floatX), name="b")
x.tag.test_value = D[0]
y.tag.test_value = D[1]

# Construct Theano expression graph
p_1 = 1 / (1 + T.exp(-T.dot(x, w)-b)) # Probability of having a one
prediction = p_1 > 0.5 # The prediction that is done: 0 or 1
xent = -y*T.log(p_1) - (1-y)*T.log(1-p_1) # Cross-entropy
cost = xent.mean() + 0.01*(w**2).sum() # The cost to optimize
gw,gb = T.grad(cost, [w,b])

# Compile expressions to functions
train = theano.function(
            inputs=[x,y],
            outputs=[prediction, xent],
            updates=[(w, w-0.01*gw), (b, b-0.01*gb)],
            name = "train")
predict = theano.function(inputs=[x], outputs=prediction,
            name = "predict")

if any([x.op.__class__.__name__ in ['Gemv', 'CGemv', 'Gemm', 'CGemm'] for x in
        train.maker.fgraph.toposort()]):
    print('Used the cpu')
elif any([x.op.__class__.__name__ in ['GpuGemm', 'GpuGemv'] for x in
          train.maker.fgraph.toposort()]):
    print('Used the gpu')
else:
    print('ERROR, not able to tell if theano used the cpu or the gpu')
    print(train.maker.fgraph.toposort())

for i in range(training_steps):
    pred, err = train(D[0], D[1])

print("target values for D")
print(D[1])

print("prediction on D")
print(predict(D[0]))
$ THEANO_FLAGS=mode=FAST_RUN,device=cpu,floatX=float32 time python lr.py 
Used the cpu
target values for D
[ 1.  1.  0.  1.  0.  0.  0.  0.  0.  1.  1.  0.  0.  0.  0.  0.  0.  1.
  1.  0.  0.  1.  0.  0.  1.  1.  0.  1.  1.  1.  1.  0.  1.  1.  0.  1.
  0.  0.  0.  0.  0.  1.  0.  0.  0.  1.  1.  0.  1.  1.  1.  0.  1.  0.
  0.  0.  0.  0.  0.  1.  0.  1.  0.  0.  0.  1.  1.  1.  0.  0.  1.  1.
  1.  1.  0.  0.  0.  1.  0.  0.  1.  1.  0.  0.  1.  1.  1.  1.  0.  1.
  0.  0.  0.  0.  1.  0.  0.  1.  1.  1.  0.  0.  1.  1.  1.  1.  1.  1.
  1.  1.  1.  1.  0.  1.  1.  0.  0.  1.  0.  0.  0.  1.  0.  1.  1.  1.
  1.  0.  0.  1.  0.  1.  1.  1.  1.  1.  1.  1.  1.  1.  0.  1.  1.  0.
  1.  0.  1.  1.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  1.  0.  0.
  1.  0.  1.  0.  0.  1.  0.  0.  1.  1.  1.  1.  0.  1.  0.  0.  1.  0.
  0.  0.  1.  1.  1.  1.  1.  1.  1.  0.  1.  1.  1.  0.  1.  0.  1.  0.
  0.  1.  1.  0.  0.  1.  0.  0.  0.  0.  0.  0.  0.  1.  0.  1.  0.  1.
  1.  0.  1.  1.  1.  0.  0.  1.  1.  1.  1.  0.  0.  0.  1.  1.  0.  0.
  1.  0.  0.  0.  0.  1.  1.  1.  0.  1.  1.  1.  0.  1.  0.  0.  0.  0.
  0.  1.  1.  1.  1.  1.  1.  0.  0.  1.  1.  1.  0.  1.  0.  1.  0.  1.
  1.  0.  0.  0.  1.  1.  0.  0.  1.  0.  0.  0.  0.  1.  0.  0.  0.  1.
  0.  1.  0.  1.  1.  0.  1.  1.  0.  0.  0.  0.  1.  0.  0.  0.  0.  1.
  0.  1.  0.  0.  1.  1.  0.  0.  1.  1.  0.  1.  0.  1.  0.  0.  1.  1.
  0.  1.  1.  0.  0.  1.  1.  0.  0.  1.  0.  1.  1.  0.  0.  0.  1.  0.
  0.  0.  1.  0.  0.  0.  0.  1.  1.  0.  1.  1.  1.  0.  1.  1.  1.  1.
  1.  0.  0.  1.  0.  0.  0.  0.  1.  1.  0.  0.  0.  0.  0.  1.  1.  1.
  0.  1.  1.  1.  0.  0.  0.  0.  1.  1.  1.  0.  0.  0.  0.  1.  0.  0.
  1.  1.  0.  1.]
prediction on D
[1 1 0 1 0 0 0 0 0 1 1 0 0 0 0 0 0 1 1 0 0 1 0 0 1 1 0 1 1 1 1 0 1 1 0 1 0
 0 0 0 0 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 0 0 0 1 0 1 0 0 0 1 1 1 0 0 1 1 1 1
 0 0 0 1 0 0 1 1 0 0 1 1 1 1 0 1 0 0 0 0 1 0 0 1 1 1 0 0 1 1 1 1 1 1 1 1 1
 1 0 1 1 0 0 1 0 0 0 1 0 1 1 1 1 0 0 1 0 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1 1
 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 0 1 0 0 1 0 0 1 1 1 1 0 1 0 0 1 0 0 0 1 1 1
 1 1 1 1 0 1 1 1 0 1 0 1 0 0 1 1 0 0 1 0 0 0 0 0 0 0 1 0 1 0 1 1 0 1 1 1 0
 0 1 1 1 1 0 0 0 1 1 0 0 1 0 0 0 0 1 1 1 0 1 1 1 0 1 0 0 0 0 0 1 1 1 1 1 1
 0 0 1 1 1 0 1 0 1 0 1 1 0 0 0 1 1 0 0 1 0 0 0 0 1 0 0 0 1 0 1 0 1 1 0 1 1
 0 0 0 0 1 0 0 0 0 1 0 1 0 0 1 1 0 0 1 1 0 1 0 1 0 0 1 1 0 1 1 0 0 1 1 0 0
 1 0 1 1 0 0 0 1 0 0 0 1 0 0 0 0 1 1 0 1 1 1 0 1 1 1 1 1 0 0 1 0 0 0 0 1 1
 0 0 0 0 0 1 1 1 0 1 1 1 0 0 0 0 1 1 1 0 0 0 0 1 0 0 1 1 0 1]
        8.92 real         8.24 user         1.14 sys
$ THEANO_FLAGS=mode=FAST_RUN,device=gpu,floatX=float32 time python lr.py 
Using gpu device 0: GeForce GT 750M
Used the gpu
target values for D
[ 1.  0.  0.  0.  0.  1.  0.  0.  1.  1.  0.  0.  1.  1.  0.  0.  1.  1.
  0.  0.  0.  1.  1.  0.  1.  1.  1.  0.  0.  1.  1.  1.  1.  1.  1.  0.
  0.  1.  0.  0.  1.  1.  0.  0.  1.  1.  0.  1.  0.  1.  1.  0.  1.  1.
  1.  0.  1.  1.  0.  0.  0.  1.  1.  1.  1.  1.  0.  0.  1.  1.  0.  1.
  1.  1.  1.  0.  1.  1.  0.  1.  1.  1.  0.  0.  0.  1.  1.  0.  0.  0.
  1.  0.  1.  0.  0.  0.  0.  1.  1.  1.  1.  0.  0.  1.  0.  1.  0.  1.
  1.  0.  1.  1.  0.  0.  0.  0.  1.  0.  0.  1.  0.  0.  0.  1.  0.  1.
  1.  1.  0.  0.  0.  1.  0.  1.  0.  1.  0.  1.  1.  1.  1.  1.  0.  1.
  1.  0.  1.  1.  0.  0.  1.  0.  1.  0.  0.  1.  0.  0.  1.  0.  0.  0.
  1.  0.  0.  1.  1.  1.  1.  0.  0.  0.  1.  1.  1.  0.  1.  0.  0.  1.
  1.  1.  1.  1.  1.  1.  1.  1.  0.  0.  0.  0.  0.  1.  1.  1.  0.  1.
  0.  1.  0.  1.  1.  1.  1.  0.  0.  0.  1.  1.  1.  1.  0.  0.  0.  1.
  0.  1.  1.  1.  0.  1.  1.  1.  0.  0.  0.  0.  1.  0.  1.  0.  0.  1.
  0.  0.  1.  1.  0.  1.  0.  1.  1.  1.  0.  0.  1.  1.  0.  0.  0.  0.
  1.  0.  0.  1.  0.  0.  0.  0.  1.  0.  0.  1.  1.  1.  1.  1.  1.  1.
  0.  1.  1.  0.  0.  0.  1.  0.  1.  1.  0.  0.  0.  0.  0.  0.  1.  0.
  1.  1.  1.  0.  0.  1.  0.  1.  0.  0.  1.  0.  1.  0.  0.  1.  0.  0.
  1.  1.  0.  1.  1.  1.  0.  0.  0.  0.  0.  1.  0.  1.  0.  0.  0.  1.
  0.  0.  1.  1.  0.  1.  1.  0.  1.  1.  1.  0.  1.  1.  0.  0.  0.  0.
  0.  0.  1.  1.  1.  1.  1.  1.  1.  1.  0.  1.  1.  1.  0.  1.  0.  1.
  1.  1.  0.  1.  1.  0.  1.  1.  1.  0.  0.  1.  1.  0.  0.  0.  0.  0.
  1.  0.  0.  1.  1.  1.  0.  1.  0.  0.  1.  1.  0.  1.  1.  0.  1.  1.
  0.  0.  1.  0.]
prediction on D
[1 0 0 0 0 1 0 0 1 1 0 0 1 1 0 0 1 1 0 0 0 1 1 0 1 1 1 0 0 1 1 1 1 1 1 0 0
 1 0 0 1 1 0 0 1 1 0 1 0 1 1 0 1 1 1 0 1 1 0 0 0 1 1 1 1 1 0 0 1 1 0 1 1 1
 1 0 1 1 0 1 1 1 0 0 0 1 1 0 0 0 1 0 1 0 0 0 0 1 1 1 1 0 0 1 0 1 0 1 1 0 1
 1 0 0 0 0 1 0 0 1 0 0 0 1 0 1 1 1 0 0 0 1 0 1 0 1 0 1 1 1 1 1 0 1 1 0 1 1
 0 0 1 0 1 0 0 1 0 0 1 0 0 0 1 0 0 1 1 1 1 0 0 0 1 1 1 0 1 0 0 1 1 1 1 1 1
 1 1 1 0 0 0 0 0 1 1 1 0 1 0 1 0 1 1 1 1 0 0 0 1 1 1 1 0 0 0 1 0 1 1 1 0 1
 1 1 0 0 0 0 1 0 1 0 0 1 0 0 1 1 0 1 0 1 1 1 0 0 1 1 0 0 0 0 1 0 0 1 0 0 0
 0 1 0 0 1 1 1 1 1 1 1 0 1 1 0 0 0 1 0 1 1 0 0 0 0 0 0 1 0 1 1 1 0 0 1 0 1
 0 0 1 0 1 0 0 1 0 0 1 1 0 1 1 1 0 0 0 0 0 1 0 1 0 0 0 1 0 0 1 1 0 1 1 0 1
 1 1 0 1 1 0 0 0 0 0 0 1 1 1 1 1 1 1 1 0 1 1 1 0 1 0 1 1 1 0 1 1 0 1 1 1 0
 0 1 1 0 0 0 0 0 1 0 0 1 1 1 0 1 0 0 1 1 0 1 1 0 1 1 0 0 1 0]
       19.78 real        17.61 user         1.24 sys

So it seems this GPU does not outperform the CPU. Well,GT 750M may not be the best GPU you can get… Someone else here has a similar experience.

 

5 Comments