"In order to effectively train and deploy artificial intelligence models, we tend to use large data centers or supercomputers. In order to be able to continuously handle, create and enhance a wide variety of information (images, video, text, and audio) The model we need, the computing power we need should not be underestimated. If we want to deploy these models on your mobile device, then they must be very fast, but this is also very difficult. To overcome these problems, we need one Sustainable, flexible and portable depth learning framework.
Facebook has always created such a framework with other developers of open source communities. Today, Facebook announced the first version of the production available CaffE2, which is a lightweight, modular depth learning framework and emphasizes portability while maintaining scalability and performance.
We are committed to providing high-performance machine learning tools for communities so that everyone can create intelligent applications and services. There are also some related tutorials and cases that are released with Caffe2, including large-scale learning and use one or more GPUs on multiple machines on a machine. Learn in iOS. Android and Raspberry Training and Deployment Models. In addition, you only need to write a few lines of code to call the pre-training model from Caffe2 Model Zoo.
CAFFE2 deploys in Facebook to help R & D staff training large machine learning models and provide a good experience of handling mobile phone users. Now, developers can access a lot of the same tools, allowing them to run large-scale distributed training programs and create machine learning applications for mobile phones. We have worked closely with Ying Weida, Gao Tong, Intel, Amazon and Microsoft to optimize Caffe2 in the cloud and mobile phone. These cooperation will allow the machine to learn the community quickly complete the experimental process using more complex models, and deploy the next generation of artificial intelligence enhanced applications and services.
You can view the Caffe2 documentation and tutorial on caffe2.ai and view the source code in GitHub. If you consider using Caffe2, we are happy to know your specific needs. Please participate in our survey. We will send you information about new versions and special developer activity / webinars.
Home: http://caffe2.ai
GitHub: https://github.com/caffe2/caffe2
Survey: https://www.surveymonkey.com/r/caffe2
The following is the introduction of Caffe2 on the GitHub open source project:
Caffe2 is a deep learning framework for expressiveness, speed, and modularity, is an experimental reconstruction of Caffe, which can be tissue in a more flexible manner.
License
Publish license license for caffe2: https://github.com/yangqing/caffe2/blob/master/license
Caffe2
Detailed build matrix:
Git clone - gene's https://github.com/caffe2/caffe2.git
CD caffe2
Os x
Brew Install Automake Protobuf
Mkdir Build && Cd Build
cmake ..
Make
Ubuntu
Runable version:
Ubuntu 14.04
Ubuntu 16.06
Really dependent package
Sudo Apt-Get Update
Sudo Apt-Get Install -y --NO-Install-Recommends
Build-essential
CMAKE
git
Libgoogle-glog-dev
LIBPROTOBUF-DEV
Protobuf-compiler
Python-dev
Python-PIP
Sudo Pip Install Numpy Protobuf
Can choose GPU support
If you plan to use the GPU, not just using the CPU, you should install NVIDIA CUDA and CUDNN, which is a GPU accelerator for deep neural network. Ying Weida introduces the installation guide in the official blog, or you can try the following quick installation instructions. First of all, be sure to upgrade your graphics driver! Otherwise, you may suffer a great difficulty of being diagnosed.
Install Ubuntu 14.04
Sudo Apt-Get Update && sudo apt-get install wget -y --no-install-recommends
Wget "" http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1404/x86_64/Cuda-repo-ubuntu1404_8.0.61-1_AMD64.DEB ""
Sudo DPKG -I CUDA-REPO-UBUNTU1404_8.0.61-1_AMD64.DEB
Sudo Apt-Get Update
Sudo Apt-Get Install Cuda
Install Ubuntu 16.04
Sudo Apt-Get Update && sudo apt-get install wget -y --no-install-recommends
Wget "" http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1604/x86_64/Cuda-repo-ubuntu1604_8.0.61-1_AMD64.DEB ""
Sudo DPKG -I CUDA-REPO-Ubuntu1604_8.0.61-1_AMD64.DEB
Sudo Apt-Get Update
Sudo Apt-Get Install Cuda
Install Cudnn (all Ubuntu versions)
Cudnn_url = "" http://developer.download.nvidia.com/compute/redist/cudnn/v5.1/cudnn-8.0-linux-x64-v5.1.tgz ""
Wget $ {cudnn_url}
Sudo Tar -XZF CUDNN-8.0-Linux-X64-V5.1.tgz -c / usr / local
RM Cudnn-8.0-linux-x64-v5.1.tgz && sudo ldconfig
Alternative dependency
Note that Ubuntu 14.04 uses libgflags2. Ubuntu 16.04 uses libgflags-dev.
# for ubuntu 14.04
Sudo Apt-Get Install -y --NO-Install-Recommends Libgflags2
# for ubuntu 16.04
Sudo Apt-Get Install -y --NO-Install-Recommends Libgflags-Dev
# for Both Ubuntu 14.04 and 16.04
Sudo Apt-Get Install -y --NO-Install-Recommends
Libgtest-dev
LIBIMP-DEV
LIBLEVELDB-DEV
Liblmdb-dev
Libopencv-dev-dev
Libopenmpi-devi-devi-devi
Libsnappy-dev
OpenMPI-BIN
OpenMPI-DOC
Python-pydot
Check the following Python section and install the optional package before establishing CAFFE2.
Mkdir Build && Cd Build
cmake ..
Make
Android and iOS
We build original binaries using CMake's Android and iOS ports, and then integrate it into the Android or Xcode project. View script /BUILD_ANDROID.SH and /BUILD_IOS.SH gain specific information.
For Android systems, we can build Caffe2 directly with Android Studio. Here is an example item: https: //github.com/bwasti/aicamera. Note that you may need to configure Android Studio so that the SDK and NDK versions you write code will be correct.
Raspberry
For the Raspbian system, you only need to run scripts /BUILD_RASPBIAN.SH on the Raspberry Pie.
Tegra X1
In order to install Caffe2 on the Tieida TEGRA X1 platform, you need to simply install the latest version of the system using the NVIDIA JetPack installer, and then run scripts /BUILD_TEGRA_X1.SH on the Tegra device.
Python support
In order to carry out the following tutorial, the Python environment requires iPython-NoteBooks and MatPlotLib, which can be installed in the OS X system:
Brew Install Matplotlib --with-Python3
Pip Install iPython Notebook
You will find that the following Python libraries are required in the specific tutorials and cases, so you can run the following command line one-time installation all the requirements library:
Sudo PIP Install
Flask
Graphviz
Hypothesis
Jupyter
Matplotlib
Pydot Python-NVD3
Pyyaml
REQUESTS
Scikit-Image
SCIPY
Setuptools
Tornado
Building an environment (known to run)
This article originally address: https://www.eeboard.com/news/facebook-caffe2/
Search "" Love Board "", pay attention, daily update development board, intelligent hardware, open source hardware, activity and other information can make you fully master. Recommended attention!
[WeChat scanning picture can be directly paid]
Technology early know:
Black Technology: "Electronic Fence" technology makes sharing bicycles no longer stop
ASUS sells price of $ 54.99 Tinker development board: Configuring Raspberry Pieces Cannot
Microsoft announces Win 10 ushered in native Linux container
A generation of memories of the memory disappeared on the computer
The public saying that the black technology - four-foot performance monster millet 6 black technology big inventory "
Our other product:
