# We need to be able to find Python.h and numpy/arrayobject.h. # NOTE: this is required only if you will compile the python interface. Here's which dependencies we'll download using Homebrew:Įxecute the following commands to download and install the dependencies: Here's which dependencies we'll download using Anaconda:

They are listed in the Prerequisites section of the official guide. You need to download all the dependencies of Caffe. Read carefully the guide so that you have a general idea of what needs to be done from the official point of view. Compile and test the Python module for Caffeįirst of all, clone the Caffe repository from GitHub to some directory and head to the official installation guide.Configure the Caffe compilation by editing nfig.Here is a big picture of what we will do: If something is not present on Anaconda, we will use Homebrew to install it. We will use Anaconda as our primary platform, so if something we need is available on Anaconda, we'll tend to use Anaconda to get it. We will compile a very basic working setup without the CUDA support. Homebrew: supplies some dependencies of Caffe.Anaconda: supplies most of the dependencies of Caffe.Here is the setup this guide was tested for: Caffe is a work in progress, so its installation is not that trivial. In this post, I will describe how to install the Caffe neural network framework on Mac OS X as a Python library.

Anaconda mac