Creating symlink /home/nvidia/caffe/python/caffe/_caffe.so → /home/nvidia/caffe/lib/_caffe.so
[100%] Built target pycaffe
The imagenet-camera program uses the pre-trained network snapshot from Caffe/DIGITS to perform classification inference at runtime. It’s not using online learning. Labelling is the process of annotating the ground truth training dataset. For complex DNN’s like GoogLeNet and ResNet and the training data size, a larger PC platform is typically used to significantly save time. Caffe isn’t especially geared for online learning, but TensorFlow or pyTorch frameworks could be (I have a separate jetson-reinforcement repo which learns on the Jetson), although the networks are typically less complex there.
I agree with you.
I have labeled an initial data set manually.
I want to feed it to an algorithm for I can somehow use the retrained model with devboard video camera for it labels[recognizes and attaches label in the top] images on-the-fly based on the retrained model.
I tried tests:
the outcome is as follows.
I haven’t tried the exact the link above way. Jetsonhacks seems to define just Caffe, not nvCaffe installation, as it seem to me.
If the listed errors are fatal.I shall try the second link.
[----------] Global test environment tear-down
[==========] 2122 tests from 287 test cases ran. (2551732 ms total)
[ PASSED ] 2116 tests.
[ FAILED ] 6 tests, listed below:
[ FAILED ] DetectNetTransformationLayerTest/3.TestScaleDown, where TypeParam = caffe::GPUDevice
[ FAILED ] DetectNetTransformationLayerTest/3.TestDesaturation, where TypeParam = caffe::GPUDevice
[ FAILED ] DetectNetTransformationLayerTest/3.TestNoAugmentation, where TypeParam = caffe::GPUDevice
[ FAILED ] DetectNetTransformationLayerTest/3.TestScaleUp, where TypeParam = caffe::GPUDevice
[ FAILED ] DetectNetTransformationLayerTest/3.TestRotation, where TypeParam = caffe::GPUDevice
[ FAILED ] DetectNetTransformationLayerTest/3.TestFlip, where TypeParam = caffe::GPUDevice
6 FAILED TESTS
src/caffe/test/CMakeFiles/runtest.dir/build.make:57: recipe for target ‘src/caffe/test/CMakeFiles/runtest’ failed
make[3]: *** [src/caffe/test/CMakeFiles/runtest] Error 1
CMakeFiles/Makefile2:328: recipe for target ‘src/caffe/test/CMakeFiles/runtest.dir/all’ failed
make[2]: *** [src/caffe/test/CMakeFiles/runtest.dir/all] Error 2
CMakeFiles/Makefile2:335: recipe for target ‘src/caffe/test/CMakeFiles/runtest.dir/rule’ failed
make[1]: *** [src/caffe/test/CMakeFiles/runtest.dir/rule] Error 2
Makefile:240: recipe for target ‘runtest’ failed
make: *** [runtest] Error 2
For the link [url]https://github.com/dusty-nv/jetson-inference/blob/master/docs/building-nvcaffe.md[/url],
experimental/fp16
branch seems not to exist.
I did:
$ sudo apt-get update -y
$ sudo apt-get install cmake -y
$ sudo apt-get install libprotobuf-dev libleveldb-dev libsnappy-dev libhdf5-serial-dev protobuf-compiler -y
$ sudo apt-get install --no-install-recommends libboost-all-dev -y
$ sudo apt-get install libatlas-base-dev -y
$ sudo apt-get install libopenblas-dev
$ sudo apt-get install libgflags-dev libgoogle-glog-dev liblmdb-dev -y
$ sudo apt-get install python-dev python-numpy python-skimage python-protobuf -y
$ sudo ln -s /usr/lib/libsnappy.so.1 /usr/lib/libsnappy.so
$ sudo ldconfig
$ cd ~
$ git clone -b caffe-0.15 GitHub - NVIDIA/caffe: Caffe: a fast open framework for deep learning.
$ cd caffe
$ cp Makefile.config.example Makefile.config
$ rm -rf cmake/ CMakeLists.txt
$ sed -i ‘s/# NATIVE_FP16/NATIVE_FP16/g’ Makefile.config
$ sed -i ‘s/# USE_CUDNN/USE_CUDNN/g’ Makefile.config
$ sed -i ‘s/-gencode arch=compute_50,code=compute_50/-gencode arch=compute_53,code=sm_53 -gencode arch=compute_53,code=compute_53/g’ Makefile.config
$ emacs -nw Makefile.config
±-------------------------------------------------+
|—INCLUDE_DIRS := $(PYTHON_INCLUDE) /usr/local/include
|—LIBRARY_DIRS := $(PYTHON_LIB) /usr/local/lib /usr/lib
|+++INCLUDE_DIRS := $(PYTHON_INCLUDE) /usr/local/include /usr/include/hdf5/serial/
|+++LIBRARY_DIRS := $(PYTHON_LIB) /usr/local/lib /usr/lib /usr/lib/aarch64-linux-gnu/hdf5/serial/
|
|—# USE_PKG_CONFIG := 1
|+++USE_PKG_CONFIG := 1
±-------------------------------------------------+
$ emacs -nw 3rdparty/cub/host/mutex.cuh
±-------------------------------------------------+
|— #ifndef arm
|+++ #if !defined(arm) && !defined(aarch64)
±-------------------------------------------------+
$ sudo ln -s /usr/lib/aarch64-linux-gnu/libhdf5_serial.so.10 /usr/lib/aarch64-linux-gnu/libhdf5.so
$ sudo ln -s /usr/lib/aarch64-linux-gnu/libhdf5_serial_hl.so.10 /usr/lib/aarch64-linux-gnu/libhdf5_hl.so
$ make -j4
$ make pycaffe
$ make distribute
$ emacs -nw ~/.bashrc
±-------------------------------------------------+
|—export PATH=/usr/local/cuda-8.0/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/snap/bin:
|+++export PATH=/usr/local/cuda-8.0/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/snap/bin:/home/nvidia/caffe/build/tools
|+++export PYTHONPATH=/home/nvidia/caffe/python:$PYTHONPATH
±-------------------------------------------------+
$ make runtest
The last make runtest
is optional and caffe-0.15
branch is compatible with current TensorRT according to @dusty_nv.
([url]https://devtalk.nvidia.com/default/topic/1022718/create-object-detection-model-without-digits-/[/url])