Faster R-cnn和Cudnn5.0衝突解決過程記錄
阿新 • • 發佈：2019-01-02
執行Faster R-cnn編譯時遇到了如下問題
In file included from ./include/caffe/util/device_alternate.hpp:40:0, from ./include/caffe/common.hpp:19, from src/caffe/syncedmem.cpp:1: ./include/caffe/util/cudnn.hpp: In function ‘void caffe::cudnn::createPoolingDesc(cudnnPoolingStruct**, caffe::PoolingParameter_PoolMethod, cudnnPoolingMode_t*, int, int, int, int, int, int)’: ./include/caffe/util/cudnn.hpp:127:41: error: too few arguments to function ‘cudnnStatus_t cudnnSetPooling2dDescriptor(cudnnPoolingDescriptor_t, cudnnPoolingMode_t, cudnnNanPropagation_t, int, int, int, int, int, int)’ pad_h, pad_w, stride_h, stride_w)); ^ ./include/caffe/util/cudnn.hpp:15:28: note: in definition of macro ‘CUDNN_CHECK’ cudnnStatus_t status = condition; \ ^ In file included from ./include/caffe/util/cudnn.hpp:5:0, from ./include/caffe/util/device_alternate.hpp:40, from ./include/caffe/common.hpp:19, from src/caffe/syncedmem.cpp:1: /usr/include/cudnn.h:803:27: note: declared here cudnnStatus_t CUDNNWINAPI cudnnSetPooling2dDescriptor( ^ Makefile:563: recipe for target '.build_release/src/caffe/syncedmem.o' failed make: *** [.build_release/src/caffe/syncedmem.o] Error 1 make: *** Waiting for unfinished jobs.... In file included from ./include/caffe/util/device_alternate.hpp:40:0, from ./include/caffe/common.hpp:19, from ./include/caffe/blob.hpp:8, from src/caffe/blob.cpp:4: ./include/caffe/util/cudnn.hpp: In function ‘void caffe::cudnn::createPoolingDesc(cudnnPoolingStruct**, caffe::PoolingParameter_PoolMethod, cudnnPoolingMode_t*, int, int, int, int, int, int)’: ./include/caffe/util/cudnn.hpp:127:41: error: too few arguments to function ‘cudnnStatus_t cudnnSetPooling2dDescriptor(cudnnPoolingDescriptor_t, cudnnPoolingMode_t, cudnnNanPropagation_t, int, int, int, int, int, int)’ pad_h, pad_w, stride_h, stride_w));
出錯原因應該為Faster R-cnn 和cudnn5.0不相容，解決辦法一是降低更換cudnn的版本，另一種是提高caffe的版本。在下還搜到了大神的一個解決辦法。
http://blog.csdn.net/rzjmpb/article/details/52373012，程式碼貼在下面
cd py-faster-rcnn/caffe-fast-rcnn
Git remote add caffe https://github.com/BVLC/caffe.git
git fetch caffe
git merge caffe/master鄙人愚鈍，對於git合併命令不是很熟悉，一週折騰merge命令還是沒有成功，我把過程附到最後，希望大牛們看到此文章能指點一二。
最後解決辦法還是在csdn上一番苦尋找到的，參考了卜居和阿·甘的部落格，感謝兩位的分享。直接貼連結比較合適，後期我還會給出我的caffe-fast-rcnn版本，裡面改了一些其它的資料，由於沒有詳細記錄，直接貼結果方便大家使用，大家可以直接下載替換faster-rcnn中的caffe
http://blog.csdn.net/kexinmcu/article/details/53178428（阿甘）
http://blog.csdn.net/kkk584520/article/details/51163564（卜居）
附：
但在執行最後一步（合併）的時候，又遇到了棘手的問題。
error: Your local changes to the following files would be overwritten by merge:
Makefile
Please, commit your changes or stash them before you can merge.
error: The following untracked working tree files would be overwritten by merge:
.github/ISSUE_TEMPLATE.md
cmake/Modules/FindNCCL.cmake
docker/README.md
docker/cpu/Dockerfile
docker/gpu/Dockerfile
docs/install_apt_debian.md
docs/tutorial/layers/absval.md
docs/tutorial/layers/accuracy.md
docs/tutorial/layers/argmax.md
docs/tutorial/layers/batchnorm.md
docs/tutorial/layers/batchreindex.md
docs/tutorial/layers/bias.md
docs/tutorial/layers/bnll.md
docs/tutorial/layers/concat.md
docs/tutorial/layers/contrastiveloss.md
docs/tutorial/layers/convolution.md
docs/tutorial/layers/crop.md
docs/tutorial/layers/data.md
docs/tutorial/layers/deconvolution.md
docs/tutorial/layers/dropout.md
docs/tutorial/layers/dummydata.md
docs/tutorial/layers/eltwise.md
docs/tutorial/layers/elu.md
docs/tutorial/layers/embed.md
docs/tutorial/layers/euclideanloss.md
docs/tutorial/layers/exp.md
docs/tutorial/layers/filter.md
docs/tutorial/layers/flatten.md
docs/tutorial/layers/hdf5data.md
docs/tutorial/layers/hdf5output.md
docs/tutorial/layers/hingeloss.md
docs/tutorial/layers/im2col.md
docs/tutorial/layers/imagedata.md
docs/tutorial/layers/infogainloss.md
docs/tutorial/layers/innerproduct.md
docs/tutorial/layers/input.md
docs/tutorial/layers/log.md
docs/tutorial/layers/lrn.md
docs/tutorial/layers/lstm.md
docs/tutorial/layers/memorydata.md
docs/tutorial/layers/multinomiallogisticloss.md
docs/tutorial/layers/mvn.md
docs/tutorial/layers/parameter.md
docs/tutorial/layers/pooling.md
docs/tutorial/layers/power.md
docs/tutorial/layers/prelu.md
docs/tutorial/layers/python.md
docs/tutorial/layers/recurrent.md
docs/tutorial/layers/reduction.md
docs/tutorial/layers/relu.md
docs/tutorial/layers/reshape.md
docs/tutorial/layers/rnn.md
docs/tutorial/layers/scale.md
docs/tutorial/layers/sigmoid.md
docs/tutorial/layers/sigmoidcrossentropyloss.md
docs/tutorial/layers/silence.md
docs/tutorial/layers/slice.md
docs/tutorial/layers/softmax.md
docs/tutorial/layers/softmaxwithloss.md
docs/tutorial/layers/split.md
docs/tutorial/layers/spp.md
docs/tutorial/layers/tanh.md
docs/tutorial/layers/threshold.md
docs/tutorial/layers/tile.md
docs/tutorial/layers/windowdata.md
examples/02-fine-tuning.ipynb
examples/brewing-logreg.ipynb
examples/images/cat gray.jpg
examples/mnist/train_lenet_docker.sh
examples/pascal-multilabel-with-datalayer.ipynb
examples/pycaffe/layers/pascal_multilabel_datalayers.py
examples/pycaffe/tools.py
include/caffe/layers/crop_layer.hpp
include/caffe/layers/input_layer.hpp
include/caffe/layers/lstm_layer.hpp
include/caffe/layers/parameter_layer.hpp
include/caffe/layers/recurrent_layer.hpp
include/caffe/layers/rnn_layer.hpp
include/caffe/util/nccl.hpp
python/caffe/coord_map.py
python/caffe/test/test_coord_map.py
python/caffe/test/test_draw.py
python/train.py
scripts/split_caffe_proto.py
scripts/travis/build.sh
scripts/travis/configure-cmake.sh
scripts/travis/configure-make.sh
scripts/travis/configure.sh
scripts/travis/defaults.sh
scripts/travis/install-deps.sh
scripts/travis/install-python-deps.sh
scripts/travis/setup-venv.sh
scripts/travis/test.sh
src/caffe/layers/crop_layer.cpp
src/caffe/layers/crop_layer.cu
src/caffe/layers/input_layer.cpp
src/caffe/layers/lstm_layer.cpp
src/caffe/layers/lstm_unit_layer.cpp
src/caffe/layers/lstm_unit_layer.cu
src/caffe/layers/parameter_layer.cpp
src/caffe/layers/recurrent_layer.cpp
src/caffe/layers/recurrent_layer.cu
src/caffe/layers/rnn_layer.cpp
src/caffe/test/test_crop_layer.cpp
src/caffe/test/test_lstm_layer.cpp
src/caffe/test/test_rnn_layer.cpp
Please move or remove them before you can merge.
Aborting sudo git clean -d -fx成功移除合併多餘的部分,移除後的部分為
sudo git clean -d -fx
Removing .github/
Removing docker/
Removing docs/tutorial/layers/再次合併
sudo git merge caffe/master
Auto-merging src/caffe/proto/caffe.proto
Auto-merging src/caffe/layers/dropout_layer.cpp
Removing src/caffe/data_reader.cpp
Removing scripts/travis/travis_setup_makefile_config.sh
Removing scripts/travis/travis_install.sh
Removing scripts/travis/travis_build_and_test.sh
Auto-merging python/caffe/_caffe.cpp
CONFLICT (content): Merge conflict in python/caffe/_caffe.cpp
Auto-merging python/caffe/__init__.py
CONFLICT (content): Merge conflict in python/caffe/__init__.py
Auto-merging include/caffe/layers/python_layer.hpp
Auto-merging include/caffe/layer.hpp
Removing include/caffe/data_reader.hpp
Removing examples/hdf5_classification/solver.prototxt
Removing examples/hdf5_classification/nonlinear_solver.prototxt
Removing examples/03-fine-tuning.ipynb
Removing examples/02-brewing-logreg.ipynb
Auto-merging LICENSE
CONFLICT (content): Merge conflict in LICENSE
Automatic merge failed; fix conflicts and then commit the result.此次合併後雖仍有部分檔案沒有合併成功，但已經不影響編譯和使用