Skip to content

Installation Guide ‐ Parallel Version

Jump to bottom
Maziar Korzani edited this page Mar 16, 2026 · 4 revisions

This page explains how to install and run the serial version of Dynamic CA-ffé on your system.
Two installation methods are provided depending on your operating system.

Windows Installation

Follow the steps below on Windows.

1. Download the Repository

  1. Go to the main GitHub repository.
  2. Click Code → Download ZIP.
  3. Extract the ZIP file to a directory on your computer.

For example:

C:\dyncaffe\

2. Install Visual Studio Build Tools & compile the code

  1. Go to the Visual Studio downloads page and download Visual Studio 2026 Insiders. This will give you a file named VisualStudioSetup.exe.

  2. Run VisualStudioSetup.exe and, in the installer, choose "Visual Studio Build Tools 2026". Keep all the default options selected.

  3. Once installed, open the "x64 Native Tools Command Prompt for VS" from the Start menu.

  4. Navigate to your dyncaffe folder and run the build script:

build_win.bat
  1. If the build is successful, you should see output similar to:
/dll
/implib:.\build\caffe_core_parallel.lib
/out:.\build\caffe_core_parallel.dll
.\build\caffe_core_parallel.obj
Creating library .\build\caffe_core_parallel.lib and object .\build\caffe_core_parallel.exp
  1. Your compiled library files will now be in the build/ folder:
build/
   caffe_core_parallel.dll
   caffe_core_parallel.lib
   caffe_core_parallel.exp

These are the C++ compiled library files used at runtime.

3. Install Miniconda

Dynamic CA-ffé uses a Python environment managed by Conda.

  1. Download Miniconda.

  2. Run the installer and follow the default installation steps.

  3. After installation, open Anaconda Prompt.

4. Create the Python Environment

Navigate to the folder containing the repository and create the environment using the provided YML file.

cd C:\dyncaffe
conda env create -f dyncaffe_conda_env.yml

5. Activate the Environment

Activate the environment before running the code.

conda activate dyncaffe

Once activated, the environment is ready to run the model.

6. Run a test

To verify that the installation was successful, run the following test example:

python run_caffe_parallel.py

Each time you want to use Dynamic CA-ffé, simply activate the environment. Then you can go ahead and run your created model.

Linux Installation

Follow the steps below on a Linux/Debian-based system.

1. Download the Repository

  1. Go to the main GitHub repository.
  2. Click Code → Download ZIP or clone the repository using Git:
git clone https://github.com/MGK-Lab/dyncaffe.git
  1. Navigate to the repository folder:
cd ~/dyncaffe

2. Install Miniconda

Dynamic CA-ffé uses a Python environment managed by Conda.

  1. Download the Linux Miniconda installer:
wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh
  1. Run the installer:
bash Miniconda3-latest-Linux-x86_64.sh

Follow the prompts and accept the default options. After installation, restart your terminal.

3. Create the Python Environment

In the repository folder, create the environment from the provided YML file:

conda env create -f dyncaffe_conda_env.yml

4. Install C++ Dependencies & compile the code

You need to have administrative access to install the package below

sudo apt update
sudo apt install build-essential
bash build_linux.sh

5. Activate the Environment

Activate the environment before running any scripts:

conda activate dyncaffe

Once activated, your environment is ready to run Dynamic CA-ffé.

6. Run a Test

Verify the installation by running a test example:

python run_caffe_parallel.py

Each time you want to use Dynamic CA-ffé, activate the environment first, then run your Python scripts as required.

Clone this wiki locally