使用cuDNN5編譯py-faster-rcnn錯誤：cudnn.hpp(126): error: argument of type "int" is incompatible ...

前幾天在cuda8+cudnn5.1下編譯py-faster-rcnn/caffe-fast-rcnn時報錯了，

/home/guyadong/caffe/py-faster-rcnn/caffe-fast-rcnn/include/caffe/util/cudnn.hpp(126): error: argument of type “int” is incompatible with parameter of type “cudnnNanPropagation_t”

/home/guyadong/caffe/py-faster-rcnn/caffe-fast-rcnn/include/caffe/util/cudnn.hpp(126): error: too few arguments in function call

2 errors detected in the compilation of “/tmp/tmpxft_000067e2_00000000-5_rmsprop_solver.cpp4.ii”.

CMake Error at cuda_compile_generated_rmsprop_solver.cu.o.cmake:266 (message):

Error generating file

/home/guyadong/caffe/py-faster-rcnn/caffe-fast-rcnn/src/caffe/CMakeFiles/cuda_compile.dir/solvers/./cuda_compile_generated_rmsprop_solver.cu.o

src/caffe/CMakeFiles/caffe.dir/build.make:494: recipe for target ‘src/caffe/CMakeFiles/cuda_compile.dir/solvers/cuda_compile_generated_rmsprop_solver.cu.o’ failed

make2: * [src/caffe/CMakeFiles/cuda_compile.dir/solvers/cuda_compile_generated_rmsprop_solver.cu.o] Error 1

CMakeFiles/Makefile2:273: recipe for target ‘src/caffe/CMakeFiles/caffe.dir/all’ failed

make1: * [src/caffe/CMakeFiles/caffe.dir/all] Error 2

Makefile:116: recipe for target ‘all’ failed

make: * [all] Error 2

不得已改用cudnn4才算編譯通過。

參見 《cuda8+cudnn4 Faster R-CNN安裝塈運作demo》

仔細研究，搞明白了原因：

py-faster-rcnn的作者在做這個項目時從bvlc/caffe(https://github.com/BVLC/caffe.git)項目建立了一個分支(那時還沒有cudnn5),并在這個分支的基礎上添加了自己的代碼(主要是一些layer),形成現在的版本，就是py-faster-rcnn下的caffe-fast-rcnn。

是以py-faster-rcnn/caffe-fast-rcnn還是隻支援cudnn4的代碼，編譯就會報錯。

解決辦法

在google上找了半天，總算有了辦法：

其實BVLC/caffe目前已經支援cudnn5，要做的就是把BVLC/caffe的最新版本代碼合并到py-faster-rcnn/caffe-fast-rcnn：

git版本合并

執行下面的指令實作版本合并

cd caffe-fast-rcnn
# 将BLVC/caffe添加為遠端倉庫，命名為caffe
git remote add caffe https://github.com/BVLC/caffe.git
# 擷取BLVC/caffe代碼
git fetch caffe
# 将BLVC/caffe的master分支合并到目前分支
# -X theirs 指定在合并遇到沖突時以BLVC/caffe版本為主
git merge -X theirs caffe/master
           

執行

git merge -X theirs caffe/master

會出現下面的界面要求輸入注釋資訊：

不必輸入任何内容，直接按Ctrl-X 離開

提示是否存儲更動過的緩沖區，輸入Y ,ENTER

要求寫入的檔案名:$MERGE_MSG,直接ENTER;

就完成了分支合并過程。

修改python_layer.hp

版本合并後，打開caffe-fast-rcnn/include/caffe/layers/python_layer.hp，

self_.attr("phase") = static_cast<int>(this->phase_);

這一行(line29)删除

然後再重新編譯caffe-fast-rcnn,就可以正常編譯通過了。

編譯方式參見 《cuda8+cudnn4 Faster R-CNN安裝塈運作demo》

附：git輸出

執行

git fetch caffe

的輸出

[email protected]:~/tmp/py-faster-rcnn/caffe-fast-rcnn$ git fetch caffe

remote: Counting objects: 13860, done.

remote: Compressing objects: 100% (8/8), done.

remote: Total 13860 (delta 3378), reused 3374 (delta 3374), pack-reused 10478

接收對象中: 100% (13860/13860), 13.76 MiB | 123.00 KiB/s, 完成.

處理 delta 中: 100% (9800/9800), 完成 584 個本地對象.

來自 https://github.com/BVLC/caffe

* [新分支] device-abstraction -> caffe/device-abstraction

* [新分支] gh-pages -> caffe/gh-pages

* [新分支] master -> caffe/master

* [新分支] opencl -> caffe/opencl

* [新分支] parallel -> caffe/parallel

* [新分支] tutorial -> caffe/tutorial

* [新分支] williford-patch-1 -> caffe/williford-patch-1

* [新分支] windows -> caffe/windows

* [新标簽] acm-mm-oss -> acm-mm-oss

* [新标簽] bsd -> bsd

* [新标簽] rc -> rc

* [新标簽] rc2 -> rc2

* [新标簽] rc3 -> rc3

* [新标簽] rcnn-release -> rcnn-release

* [新标簽] v0.1 -> v0.1

* [新标簽] v0.9 -> v0.9

* [新标簽] v0.99 -> v0.99

* [新标簽] v0.999 -> v0.999

* [新标簽] v0.9999 -> v0.9999

* [新标簽] rc4 -> rc4

* [新标簽] rc5 -> rc5

執行

git merge -X theirs caffe/master

的輸出

自動合并 src/caffe/proto/caffe.proto

自動合并 src/caffe/layers/dropout_layer.cpp

删除 src/caffe/data_reader.cpp

删除 scripts/travis/travis_setup_makefile_config.sh

删除 scripts/travis/travis_install.sh

删除 scripts/travis/travis_build_and_test.sh

自動合并 python/caffe/_caffe.cpp

自動合并 python/caffe/init.py

自動合并 include/caffe/layers/python_layer.hpp

自動合并 include/caffe/layer.hpp

删除 include/caffe/data_reader.hpp

删除 examples/hdf5_classification/solver.prototxt

删除 examples/hdf5_classification/nonlinear_solver.prototxt

删除 examples/03-fine-tuning.ipynb

删除 examples/02-brewing-logreg.ipynb

自動合并 LICENSE

Merge made by the ‘recursive’ strategy.

.github/ISSUE_TEMPLATE.md | 19 +

.gitignore | 4 +

.travis.yml | 58 +-

CMakeLists.txt | 32 +-

LICENSE | 15 +-

Makefile | 40 +-

Makefile.config.example | 15 +-

cmake/ConfigGen.cmake | 67 +-

cmake/Cuda.cmake | 35 +-

cmake/Dependencies.cmake | 111 +-

cmake/External/glog.cmake | 1 +

cmake/Modules/FindAtlas.cmake | 6 +-

cmake/Modules/FindMKL.cmake | 2 +-

cmake/Modules/FindNCCL.cmake | 26 +

cmake/Modules/FindvecLib.cmake | 7 +-

cmake/ProtoBuf.cmake | 6 +-

cmake/Summary.cmake | 1 +

cmake/Targets.cmake | 27 +-

cmake/Templates/CaffeConfig.cmake.in | 15 +-

cmake/Templates/caffe_config.h.in | 19 -

data/cifar10/get_cifar10.sh | 2 +-

data/ilsvrc12/get_ilsvrc_aux.sh | 2 +-

data/mnist/get_mnist.sh | 25 +-

docker/README.md | 47 +

docker/cpu/Dockerfile | 45 +

docker/gpu/Dockerfile | 46 +

docs/install_apt.md | 15 +-

docs/install_apt_debian.md | 161 +

docs/install_osx.md | 2 +-

docs/install_yum.md | 4 +-

docs/installation.md | 39 +-

docs/multigpu.md | 2 +-

docs/tutorial/layers.md | 562 +-

docs/tutorial/layers/absval.md | 22 +

docs/tutorial/layers/accuracy.md | 20 +

docs/tutorial/layers/argmax.md | 18 +

docs/tutorial/layers/batchnorm.md | 20 +

docs/tutorial/layers/batchreindex.md | 16 +

docs/tutorial/layers/bias.md | 19 +

docs/tutorial/layers/bnll.md | 25 +

docs/tutorial/layers/concat.md | 40 +

docs/tutorial/layers/contrastiveloss.md | 20 +

docs/tutorial/layers/convolution.md | 63 +

docs/tutorial/layers/crop.md | 20 +

docs/tutorial/layers/data.md | 29 +

docs/tutorial/layers/deconvolution.md | 22 +

docs/tutorial/layers/dropout.md | 20 +

docs/tutorial/layers/dummydata.md | 20 +

docs/tutorial/layers/eltwise.md | 20 +

docs/tutorial/layers/elu.md | 25 +

docs/tutorial/layers/embed.md | 20 +

docs/tutorial/layers/euclideanloss.md | 16 +

docs/tutorial/layers/exp.md | 24 +

docs/tutorial/layers/filter.md | 15 +

docs/tutorial/layers/flatten.md | 21 +

docs/tutorial/layers/hdf5data.md | 20 +

docs/tutorial/layers/hdf5output.md | 25 +

docs/tutorial/layers/hingeloss.md | 19 +

docs/tutorial/layers/im2col.md | 16 +

docs/tutorial/layers/imagedata.md | 27 +

docs/tutorial/layers/infogainloss.md | 23 +

docs/tutorial/layers/innerproduct.md | 59 +

docs/tutorial/layers/input.md | 19 +

docs/tutorial/layers/log.md | 20 +

docs/tutorial/layers/lrn.md | 28 +

docs/tutorial/layers/lstm.md | 21 +

docs/tutorial/layers/memorydata.md | 25 +

docs/tutorial/layers/multinomiallogisticloss.md | 19 +

docs/tutorial/layers/mvn.md | 20 +

docs/tutorial/layers/parameter.md | 21 +

docs/tutorial/layers/pooling.md | 47 +

docs/tutorial/layers/power.md | 46 +

docs/tutorial/layers/prelu.md | 20 +

docs/tutorial/layers/python.md | 27 +

docs/tutorial/layers/recurrent.md | 20 +

docs/tutorial/layers/reduction.md | 20 +

docs/tutorial/layers/relu.md | 32 +

docs/tutorial/layers/reshape.md | 51 +

docs/tutorial/layers/rnn.md | 19 +

docs/tutorial/layers/scale.md | 20 +

docs/tutorial/layers/sigmoid.md | 20 +

docs/tutorial/layers/sigmoidcrossentropyloss.md | 13 +

docs/tutorial/layers/silence.md | 17 +

docs/tutorial/layers/slice.md | 42 +

docs/tutorial/layers/softmax.md | 24 +

docs/tutorial/layers/softmaxwithloss.md | 33 +

docs/tutorial/layers/split.md | 17 +

docs/tutorial/layers/spp.md | 20 +

docs/tutorial/layers/tanh.md | 18 +

docs/tutorial/layers/threshold.md | 18 +

docs/tutorial/layers/tile.md | 20 +

docs/tutorial/layers/windowdata.md | 19 +

docs/tutorial/solver.md | 13 +-

examples/00-classification.ipynb | 13031 +——————

examples/01-learning-lenet.ipynb | 4680 +——

examples/02-brewing-logreg.ipynb | 5771 ——–

examples/02-fine-tuning.ipynb | 1175 ++

examples/03-fine-tuning.ipynb | 947 –

examples/CMakeLists.txt | 2 +-

examples/brewing-logreg.ipynb | 1164 ++

examples/cifar10/cifar10_full.prototxt | 11 +-

…/cifar10/cifar10_full_sigmoid_solver.prototxt | 2 +-

…/cifar10_full_sigmoid_solver_bn.prototxt | 2 +-

examples/cifar10/cifar10_quick.prototxt | 11 +-

examples/cifar10/convert_cifar_data.cpp | 2 +

examples/cifar10/create_cifar10.sh | 1 +

examples/cifar10/train_full.sh | 7 +-

examples/cifar10/train_full_sigmoid.sh | 3 +-

examples/cifar10/train_full_sigmoid_bn.sh | 3 +-

examples/cifar10/train_quick.sh | 5 +-

examples/cpp_classification/classification.cpp | 2 +-

examples/cpp_classification/readme.md | 6 +-

examples/finetune_flickr_style/readme.md | 10 +-

…/hdf5_classification/nonlinear_solver.prototxt | 15 -

examples/hdf5_classification/solver.prototxt | 15 -

examples/imagenet/create_imagenet.sh | 1 +

examples/imagenet/resume_training.sh | 4 +-

examples/imagenet/train_caffenet.sh | 3 +-

examples/images/cat gray.jpg | Bin 0 -> 92726 bytes

examples/mnist/convert_mnist_data.cpp | 89 +-

examples/mnist/create_mnist.sh | 1 +

examples/mnist/lenet.prototxt | 11 +-

examples/mnist/lenet_auto_solver.prototxt | 6 +-

examples/mnist/readme.md | 2 +-

examples/mnist/train_lenet.sh | 3 +-

examples/mnist/train_lenet_adam.sh | 3 +-

examples/mnist/train_lenet_consolidated.sh | 3 +-

examples/mnist/train_lenet_docker.sh | 119 +

examples/mnist/train_lenet_rmsprop.sh | 4 +-

examples/mnist/train_mnist_autoencoder.sh | 3 +-

examples/mnist/train_mnist_autoencoder_adadelta.sh | 3 +-

examples/mnist/train_mnist_autoencoder_adagrad.sh | 3 +-

examples/mnist/train_mnist_autoencoder_nesterov.sh | 3 +-

examples/net_surgery.ipynb | 58 +-

…/net_surgery/bvlc_caffenet_full_conv.prototxt | 15 +-

examples/net_surgery/conv.prototxt | 11 +-

examples/pascal-multilabel-with-datalayer.ipynb | 479 +

…/pycaffe/layers/pascal_multilabel_datalayers.py | 216 +

examples/pycaffe/tools.py | 121 +

examples/siamese/create_mnist_siamese.sh | 1 +

examples/siamese/mnist_siamese.prototxt | 13 +-

examples/siamese/train_mnist_siamese.sh | 3 +-

include/caffe/blob.hpp | 1 +

include/caffe/common.hpp | 19 +-

include/caffe/data_reader.hpp | 82 -

include/caffe/internal_thread.hpp | 4 +-

include/caffe/layer.hpp | 43 +-

include/caffe/layer_factory.hpp | 2 +-

include/caffe/layers/accuracy_layer.hpp | 2 +-

include/caffe/layers/base_data_layer.hpp | 6 +-

include/caffe/layers/batch_norm_layer.hpp | 29 +-

include/caffe/layers/bias_layer.hpp | 10 +-

include/caffe/layers/crop_layer.hpp | 76 +

include/caffe/layers/cudnn_relu_layer.hpp | 1 +

include/caffe/layers/cudnn_sigmoid_layer.hpp | 1 +

include/caffe/layers/cudnn_tanh_layer.hpp | 1 +

include/caffe/layers/data_layer.hpp | 7 +-

include/caffe/layers/hdf5_data_layer.hpp | 6 +-

include/caffe/layers/inner_product_layer.hpp | 1 +

include/caffe/layers/input_layer.hpp | 44 +

include/caffe/layers/lstm_layer.hpp | 154 +

include/caffe/layers/parameter_layer.hpp | 45 +

include/caffe/layers/python_layer.hpp | 5 +-

include/caffe/layers/recurrent_layer.hpp | 187 +

include/caffe/layers/rnn_layer.hpp | 47 +

include/caffe/layers/scale_layer.hpp | 12 +-

…/layers/sigmoid_cross_entropy_loss_layer.hpp | 18 +

include/caffe/net.hpp | 65 +-

include/caffe/parallel.hpp | 93 +-

include/caffe/solver.hpp | 42 +-

include/caffe/solver_factory.hpp | 2 +-

include/caffe/syncedmem.hpp | 26 +-

include/caffe/util/cudnn.hpp | 24 +-

include/caffe/util/db_leveldb.hpp | 5 +-

include/caffe/util/db_lmdb.hpp | 13 +-

include/caffe/util/hdf5.hpp | 4 +-

include/caffe/util/math_functions.hpp | 5 +

include/caffe/util/mkl_alternate.hpp | 5 +

include/caffe/util/nccl.hpp | 37 +

include/caffe/util/upgrade_proto.hpp | 12 +

matlab/+caffe/Net.m | 3 +

matlab/+caffe/Solver.m | 3 +

matlab/+caffe/private/caffe_.cpp | 26 +-

matlab/CMakeLists.txt | 2 +-

matlab/demo/classification_demo.m | 7 +-

models/bvlc_alexnet/deploy.prototxt | 11 +-

models/bvlc_googlenet/deploy.prototxt | 11 +-

models/bvlc_googlenet/train_val.prototxt | 2 +-

models/bvlc_reference_caffenet/deploy.prototxt | 11 +-

…/bvlc_reference_rcnn_ilsvrc13/deploy.prototxt | 11 +-

models/finetune_flickr_style/deploy.prototxt | 11 +-

python/CMakeLists.txt | 28 +-

python/caffe/init.py | 4 +-

python/caffe/_caffe.cpp | 206 +-

python/caffe/classifier.py | 1 +

python/caffe/coord_map.py | 185 +

python/caffe/detector.py | 2 +-

python/caffe/draw.py | 40 +-

python/caffe/io.py | 7 +-

python/caffe/net_spec.py | 8 +-

python/caffe/pycaffe.py | 107 +-

python/caffe/test/test_coord_map.py | 192 +

python/caffe/test/test_io.py | 15 +

python/caffe/test/test_net.py | 273 +-

python/caffe/test/test_python_layer.py | 26 +

python/draw_net.py | 15 +-

python/train.py | 100 +

scripts/build_docs.sh | 3 +

scripts/cpp_lint.py | 8 +-

scripts/download_model_binary.py | 7 +-

scripts/split_caffe_proto.py | 35 +

scripts/travis/build.sh | 13 +

scripts/travis/configure-cmake.sh | 32 +

scripts/travis/configure-make.sh | 36 +

scripts/travis/configure.sh | 11 +

scripts/travis/defaults.sh | 10 +

scripts/travis/install-deps.sh | 110 +

scripts/travis/install-python-deps.sh | 14 +

scripts/travis/setup-venv.sh | 18 +

scripts/travis/test.sh | 19 +

scripts/travis/travis_build_and_test.sh | 54 -

scripts/travis/travis_install.sh | 99 -

scripts/travis/travis_setup_makefile_config.sh | 31 -

src/caffe/CMakeLists.txt | 16 +-

src/caffe/blob.cpp | 22 +-

src/caffe/common.cpp | 47 +-

src/caffe/data_reader.cpp | 119 -

src/caffe/data_transformer.cpp | 2 +-

src/caffe/internal_thread.cpp | 10 +-

src/caffe/layer.cpp | 20 -

src/caffe/layer_factory.cpp | 7 +

src/caffe/layers/base_data_layer.cpp | 44 +-

src/caffe/layers/base_data_layer.cu | 21 +-

src/caffe/layers/batch_norm_layer.cpp | 16 +-

src/caffe/layers/crop_layer.cpp | 141 +

src/caffe/layers/crop_layer.cu | 109 +

src/caffe/layers/cudnn_conv_layer.cu | 12 +-

src/caffe/layers/cudnn_relu_layer.cpp | 1 +

src/caffe/layers/cudnn_relu_layer.cu | 23 +-

src/caffe/layers/cudnn_sigmoid_layer.cpp | 2 +

src/caffe/layers/cudnn_sigmoid_layer.cu | 23 +-

src/caffe/layers/cudnn_tanh_layer.cpp | 1 +

src/caffe/layers/cudnn_tanh_layer.cu | 23 +-

src/caffe/layers/data_layer.cpp | 82 +-

src/caffe/layers/dropout_layer.cpp | 4 +-

src/caffe/layers/exp_layer.cpp | 3 +-

src/caffe/layers/hdf5_data_layer.cpp | 62 +-

src/caffe/layers/hdf5_data_layer.cu | 22 +-

src/caffe/layers/image_data_layer.cpp | 24 +-

src/caffe/layers/inner_product_layer.cpp | 42 +-

src/caffe/layers/inner_product_layer.cu | 31 +-

src/caffe/layers/input_layer.cpp | 27 +

src/caffe/layers/loss_layer.cpp | 4 +-

src/caffe/layers/lstm_layer.cpp | 244 +

src/caffe/layers/lstm_unit_layer.cpp | 131 +

src/caffe/layers/lstm_unit_layer.cu | 154 +

src/caffe/layers/memory_data_layer.cpp | 2 +-

src/caffe/layers/parameter_layer.cpp | 8 +

src/caffe/layers/recurrent_layer.cpp | 295 +

src/caffe/layers/recurrent_layer.cu | 44 +

src/caffe/layers/rnn_layer.cpp | 236 +

src/caffe/layers/scale_layer.cpp | 14 +-

…/layers/sigmoid_cross_entropy_loss_layer.cpp | 79 +-

…/layers/sigmoid_cross_entropy_loss_layer.cu | 80 +-

src/caffe/layers/window_data_layer.cpp | 11 +-

src/caffe/net.cpp | 171 +-

src/caffe/parallel.cpp | 508 +-

src/caffe/proto/caffe.proto | 101 +-

src/caffe/solver.cpp | 53 +-

src/caffe/solvers/adagrad_solver.cpp | 1 -

src/caffe/solvers/nesterov_solver.cpp | 1 -

src/caffe/solvers/sgd_solver.cpp | 4 +-

src/caffe/syncedmem.cpp | 59 +-

src/caffe/test/CMakeLists.txt | 2 +-

src/caffe/test/test_blob.cpp | 8 +

src/caffe/test/test_caffe_main.cpp | 3 -

src/caffe/test/test_crop_layer.cpp | 283 +

src/caffe/test/test_data_layer.cpp | 36 +

src/caffe/test/test_data_transformer.cpp | 1 -

src/caffe/test/test_embed_layer.cpp | 2 +-

src/caffe/test/test_euclidean_loss_layer.cpp | 2 +-

src/caffe/test/test_gradient_based_solver.cpp | 40 +-

src/caffe/test/test_hdf5_output_layer.cpp | 10 +-

src/caffe/test/test_hdf5data_layer.cpp | 32 +-

src/caffe/test/test_image_data_layer.cpp | 44 +-

src/caffe/test/test_inner_product_layer.cpp | 240 +

src/caffe/test/test_lstm_layer.cpp | 288 +

src/caffe/test/test_net.cpp | 308 +-

src/caffe/test/test_neuron_layer.cpp | 20 +

src/caffe/test/test_rnn_layer.cpp | 217 +

…/test/test_sigmoid_cross_entropy_loss_layer.cpp | 28 +

src/caffe/test/test_split_layer.cpp | 61 -

src/caffe/util/benchmark.cpp | 3 +-

src/caffe/util/blocking_queue.cpp | 5 -

src/caffe/util/db_lmdb.cpp | 75 +-

src/caffe/util/hdf5.cpp | 38 +-

src/caffe/util/insert_splits.cpp | 21 +-

src/caffe/util/math_functions.cu | 20 +

src/caffe/util/upgrade_proto.cpp | 102 +-

src/gtest/CMakeLists.txt | 3 +

src/gtest/gtest-all.cpp | 4 +-

src/gtest/gtest.h | 2 +-

tools/caffe.cpp | 57 +-

tools/convert_imageset.cpp | 9 +-

tools/extra/extract_seconds.py | 8 +

tools/extra/parse_log.py | 28 +-

tools/extra/plot_log.gnuplot.example | 2 +-

tools/extra/plot_training_log.py.example | 43 +-

tools/extract_features.cpp | 5 +-

tools/upgrade_net_proto_binary.cpp | 5 +-

tools/upgrade_net_proto_text.cpp | 8 +-

tools/upgrade_solver_proto_text.cpp | 3 +-

312 files changed, 12807 insertions(+), 26312 deletions(-)

create mode 100644 .github/ISSUE_TEMPLATE.md

create mode 100644 cmake/Modules/FindNCCL.cmake

create mode 100644 docker/README.md

create mode 100644 docker/cpu/Dockerfile

create mode 100644 docker/gpu/Dockerfile

create mode 100644 docs/install_apt_debian.md

create mode 100644 docs/tutorial/layers/absval.md

create mode 100644 docs/tutorial/layers/accuracy.md

create mode 100644 docs/tutorial/layers/argmax.md

create mode 100644 docs/tutorial/layers/batchnorm.md

create mode 100644 docs/tutorial/layers/batchreindex.md

create mode 100644 docs/tutorial/layers/bias.md

create mode 100644 docs/tutorial/layers/bnll.md

create mode 100644 docs/tutorial/layers/concat.md

create mode 100644 docs/tutorial/layers/contrastiveloss.md

create mode 100644 docs/tutorial/layers/convolution.md

create mode 100644 docs/tutorial/layers/crop.md

create mode 100644 docs/tutorial/layers/data.md

create mode 100644 docs/tutorial/layers/deconvolution.md

create mode 100644 docs/tutorial/layers/dropout.md

create mode 100644 docs/tutorial/layers/dummydata.md

create mode 100644 docs/tutorial/layers/eltwise.md

create mode 100644 docs/tutorial/layers/elu.md

create mode 100644 docs/tutorial/layers/embed.md

create mode 100644 docs/tutorial/layers/euclideanloss.md

create mode 100644 docs/tutorial/layers/exp.md

create mode 100644 docs/tutorial/layers/filter.md

create mode 100644 docs/tutorial/layers/flatten.md

create mode 100644 docs/tutorial/layers/hdf5data.md

create mode 100644 docs/tutorial/layers/hdf5output.md

create mode 100644 docs/tutorial/layers/hingeloss.md

create mode 100644 docs/tutorial/layers/im2col.md

create mode 100644 docs/tutorial/layers/imagedata.md

create mode 100644 docs/tutorial/layers/infogainloss.md

create mode 100644 docs/tutorial/layers/innerproduct.md

create mode 100644 docs/tutorial/layers/input.md

create mode 100644 docs/tutorial/layers/log.md

create mode 100644 docs/tutorial/layers/lrn.md

create mode 100644 docs/tutorial/layers/lstm.md

create mode 100644 docs/tutorial/layers/memorydata.md

create mode 100644 docs/tutorial/layers/multinomiallogisticloss.md

create mode 100644 docs/tutorial/layers/mvn.md

create mode 100644 docs/tutorial/layers/parameter.md

create mode 100644 docs/tutorial/layers/pooling.md

create mode 100644 docs/tutorial/layers/power.md

create mode 100644 docs/tutorial/layers/prelu.md

create mode 100644 docs/tutorial/layers/python.md

create mode 100644 docs/tutorial/layers/recurrent.md

create mode 100644 docs/tutorial/layers/reduction.md

create mode 100644 docs/tutorial/layers/relu.md

create mode 100644 docs/tutorial/layers/reshape.md

create mode 100644 docs/tutorial/layers/rnn.md

create mode 100644 docs/tutorial/layers/scale.md

create mode 100644 docs/tutorial/layers/sigmoid.md

create mode 100644 docs/tutorial/layers/sigmoidcrossentropyloss.md

create mode 100644 docs/tutorial/layers/silence.md

create mode 100644 docs/tutorial/layers/slice.md

create mode 100644 docs/tutorial/layers/softmax.md

create mode 100644 docs/tutorial/layers/softmaxwithloss.md

create mode 100644 docs/tutorial/layers/split.md

create mode 100644 docs/tutorial/layers/spp.md

create mode 100644 docs/tutorial/layers/tanh.md

create mode 100644 docs/tutorial/layers/threshold.md

create mode 100644 docs/tutorial/layers/tile.md

create mode 100644 docs/tutorial/layers/windowdata.md

delete mode 100644 examples/02-brewing-logreg.ipynb

create mode 100644 examples/02-fine-tuning.ipynb

delete mode 100644 examples/03-fine-tuning.ipynb

create mode 100644 examples/brewing-logreg.ipynb

delete mode 100644 examples/hdf5_classification/nonlinear_solver.prototxt

delete mode 100644 examples/hdf5_classification/solver.prototxt

create mode 100644 examples/images/cat gray.jpg

create mode 100755 examples/mnist/train_lenet_docker.sh

create mode 100644 examples/pascal-multilabel-with-datalayer.ipynb

create mode 100644 examples/pycaffe/layers/pascal_multilabel_datalayers.py

create mode 100644 examples/pycaffe/tools.py

delete mode 100644 include/caffe/data_reader.hpp

create mode 100644 include/caffe/layers/crop_layer.hpp

create mode 100644 include/caffe/layers/input_layer.hpp

create mode 100644 include/caffe/layers/lstm_layer.hpp

create mode 100644 include/caffe/layers/parameter_layer.hpp

create mode 100644 include/caffe/layers/recurrent_layer.hpp

create mode 100644 include/caffe/layers/rnn_layer.hpp

create mode 100644 include/caffe/util/nccl.hpp

mode change 100644 => 100755 models/bvlc_googlenet/train_val.prototxt

create mode 100644 python/caffe/coord_map.py

create mode 100644 python/caffe/test/test_coord_map.py

create mode 100644 python/train.py

create mode 100755 scripts/split_caffe_proto.py

create mode 100755 scripts/travis/build.sh

create mode 100644 scripts/travis/configure-cmake.sh

create mode 100644 scripts/travis/configure-make.sh

create mode 100755 scripts/travis/configure.sh

create mode 100755 scripts/travis/defaults.sh

create mode 100755 scripts/travis/install-deps.sh

create mode 100755 scripts/travis/install-python-deps.sh

create mode 100755 scripts/travis/setup-venv.sh

create mode 100755 scripts/travis/test.sh

delete mode 100755 scripts/travis/travis_build_and_test.sh

delete mode 100755 scripts/travis/travis_install.sh

delete mode 100755 scripts/travis/travis_setup_makefile_config.sh

delete mode 100644 src/caffe/data_reader.cpp

create mode 100644 src/caffe/layers/crop_layer.cpp

create mode 100644 src/caffe/layers/crop_layer.cu

create mode 100644 src/caffe/layers/input_layer.cpp

create mode 100644 src/caffe/layers/lstm_layer.cpp

create mode 100644 src/caffe/layers/lstm_unit_layer.cpp

create mode 100644 src/caffe/layers/lstm_unit_layer.cu

create mode 100644 src/caffe/layers/parameter_layer.cpp

create mode 100644 src/caffe/layers/recurrent_layer.cpp

create mode 100644 src/caffe/layers/recurrent_layer.cu

create mode 100644 src/caffe/layers/rnn_layer.cpp

create mode 100644 src/caffe/test/test_crop_layer.cpp

create mode 100644 src/caffe/test/test_lstm_layer.cpp

create mode 100644 src/caffe/test/test_rnn_layer.cpp

參考資料

《make error with cuda-7.5 and cundnn 5.1》