Installation

System Requirements

PyNAS requires the following system dependencies:

Python 3.9 or higher
CUDA-capable GPU (recommended for training)
At least 8GB RAM
Linux, macOS, or Windows

Python Dependencies

PyNAS depends on several key Python packages:

PyTorch 2.0.1
PyTorch Lightning
NumPy < 2.0.0
Pandas
OpenCV Python
Matplotlib
Seaborn
Scikit-learn
Scikit-image
TQDM

Installation Methods

From Source (Recommended)

Clone the repository:

git clone https://github.com/sirbastiano/pynas.git
cd pynas

Create a virtual environment:

# Using conda
conda create -n pynas python=3.9
conda activate pynas

# OR using venv
python -m venv pynas_env
source pynas_env/bin/activate  # On Windows: pynas_env\\Scripts\\activate

Install the package in development mode:

pip install -e .

Using PDM (Recommended)

PDM is the recommended package manager for PyNAS development:

# Clone the repository
git clone https://github.com/esa-philab/PyNAS.git
cd PyNAS

# Install using PDM
pdm install

Using pip

Alternatively, you can use pip:

pip install pynas

Verify Installation

To verify that PyNAS is installed correctly, run the following Python code:

import pynas
from pynas.core.individual import Individual
from pynas.core.population import Population

print(f"PyNAS version: {pynas.__version__ if hasattr(pynas, '__version__') else '0.1.0'}")
print("PyNAS installed successfully!")

GPU Setup

For optimal performance, ensure CUDA is properly installed:

Install CUDA Toolkit (version 11.0 or higher recommended)
Verify PyTorch can access GPU:

import torch
print(f"CUDA available: {torch.cuda.is_available()}")
print(f"CUDA devices: {torch.cuda.device_count()}")

Troubleshooting

Common Issues

ImportError: No module named ‘pynas’

Ensure you’ve activated the correct virtual environment and installed the package.

CUDA out of memory

Reduce batch size or use a smaller population size in your genetic algorithm configuration.

Missing dependencies

If you encounter missing dependencies, install them manually:

pip install torch==2.0.1 pytorch-lightning numpy pandas opencv-python

Development Installation

For development purposes, install additional dependencies:

pip install -e ".[dev]"

This includes testing and documentation dependencies.