Aidge Setup and Installation#

System Requirements#

Python >= 3.10

Additional requirements vary by module. Refer to each module’s documentation for details.

Installation#

Aidge is built on a core library that interfaces with multiple modules, each binded to Python libraries. Each module must be installed independently, aidge_core included.

Tip

We recommend creating a Python environment to work with Aidge:

# Using the virtualenv package, for example:
virtualenv -p python3.10 py_env_aidge

Hint

You can also use Dev Containers if your IDE support it (VScode for example). This is a Docker-based development environment with minimal required tools preinstalled, enabling a separate development environment from your computer. See Dev container configuration section.

1. Install via pip (recommended, Linux only)#

Install the full bundle of default modules:

# Activate your virtual env:
source py_env_aidge/bin/activate

pip install -r default_module_install.txt
# - `aidge_core`
# - `aidge_backend_cpu`
# - `aidge_backend_cuda`
# - `aidge_backend_opencv`
# - `aidge_export_cpp`
# - `aidge_export_arm_cortexm`
# - `aidge_learning`
# - `aidge_onnx`
# - `aidge_quantization`

Alternatively, install only the modules you need:

# Activate your virtual env:
source py_env_aidge/bin/activate

pip install aidge-core aidge-backend-cpu aidge-learning aidge-export-cpp aidge-onnx aidge-quantization

You can install additional modules from the list of available submodules as needed.

2. Build from source#

System Requirements

CMake >= 3.18
Git

Build on Linux & macOS#

Create and activate your virtual environment:

virtualenv -p python3.10 py_env_aidge
source py_env_aidge/bin/activate

Clone the full repository (with submodules), and move to its root directory:

git clone --recursive https://gitlab.eclipse.org/eclipse/aidge/aidge.git
cd aidge

Install the package based on your specific requirements:

For Python development:

# Install all modules
./setup.sh --all --release --python

# Alternatively, install only specific modules
./setup.sh -m core -m backend_cpu -m onnx --release --python

For C++ development:

# Install all modules
./setup.sh --all --release

# Alternatively, install only specific modules, in debug or release mode
./setup.sh -m core -m backend_cpu --release
./setup.sh -m core -m backend_cpu --debug --tests

Tip

Run ./setup.sh -h to view all installation options.

Build on Windows#

Create and activate your virtual environment:

virtualenv -p python3.10 py_env_aidge
source py_env_aidge/bin/activate

Clone the full repository (with submodules), and move to its root directory:

git clone --recursive https://gitlab.eclipse.org/eclipse/aidge/aidge.git
cd aidge

Install the package based on your specific requirements:

Python development:

.\setup.ps1 -All -Release -Python
.\setup.ps1 -Modules core,backend_cpu,onnx -Release -Python

C++ development:

.\setup.ps1 -All -Release
.\setup.ps1 -Modules core,backend_cpu,onnx -Debug -Tests

Tip

Run .\ setup.ps1 -h to view all installation options.

Verify installation#

python -c "import aidge_core; import aidge_backend_cpu; print(aidge_core.Tensor.get_available_backends())"

Expected output:

{'cpu'}

Dev container configuration#

Important

Supported IDE are listed on official website.

To use with Visual Studio Code:

Make sure that the Remote-Containers and Docker extensions are installed in VSCode.
If using podman instead of docker, edit the setting dev.containers.dockerPath for the Docker extension, and set its value to podman, as per below:
```
"dev.containers.dockerPath": "podman",
```
Optionally, you can also modify the dev.containers.socketPath to:
```
/run/podman/podman.sock
```
This will allow the ‘Remote Explorer’ tab to work, providing information about running containers. (Note: the path to the podman socket was found by querying the podman.socket service using systemctl status podman.socket).
Setup sharing git credentials as explained in documentation.
Open the project folder and use the Remote-Containers: Reopen Folder in Container command from the VS Code command palette (or with F1 key).
In the terminal, make sure to source the python virtual env with:
```
source /workspaces/aidge/env_aidge/bin/python
```
In the terminal, make sure all git lfs object has been pulled:
```
git lfs install
git lfs pull
```
Continue Aidge installation as explained in Installation section.