Installation

This guide provides detailed steps to install and configure the AIMinify package, a Python library available via the AIminify PyPi server.

Prerequisites

Before proceeding with the installation, ensure you have the following:
  • Python installed on your system
  • An AIminify username and password for the PyPi server
  • An AIminify password for using the code

System requirements

AIminify requires:
Python ≥ 3.8
PyTorch ≥ 2.0
TensorFlow ≥ 2.0

Basic installation

# Pip
pip install aiminify -i https://USERNAME:PASSWORD@pypi.aiminify.com/ --extra-index-url https://pypi.org/simple/
 
# Conda
conda run -n your_env_name pip install aiminify -i https://USERNAME:PASSWORD@pypi.aiminify.com/  --extra-index-url https://pypi.org/simple/
 
# Pipx
pipx install aiminify --pip-args="--index-url https://USERNAME:PASSWORD@pypi.aiminify.com/  --extra-index-url https://pypi.org/simple/"
 
# Poetry
poetry source add --priority=primary private https://USERNAME:PASSWORD@pypi.aiminify.com/
poetry source add --priority=supplemental pypi https://upload.pypi.org/simple/
poetry add aiminify

# Pipenv
PIPENV_EXTRA_INDEX_URL="https://USERNAME:PASSWORD@pypi.aiminify.com/ https://pypi.org/simple/" pipenv install aiminify --skip-lock
After the installation process is complete, AIminify should be successfully installed in your Python environment.

TensorRT quantization support (optional)

# Pip
pip install aiminify[tensorrt] -i https://USERNAME:PASSWORD@pypi.aiminify.com/ --extra-index-url https://pypi.org/simple/
 
# Conda
conda run -n your_env_name pip install aiminify[tensorrt] -i https://USERNAME:PASSWORD@pypi.aiminify.com/  --extra-index-url https://pypi.org/simple/
 
# Pipx
pipx install aiminify[tensorrt] --pip-args="--index-url https://USERNAME:PASSWORD@pypi.aiminify.com/  --extra-index-url https://pypi.org/simple/"
 
# Poetry
poetry source add --priority=primary private https://USERNAME:PASSWORD@pypi.aiminify.com/
poetry source add --priority=supplemental pypi https://upload.pypi.org/simple/
poetry add aiminify[tensorrt]

# Pipenv
PIPENV_EXTRA_INDEX_URL="https://USERNAME:PASSWORD@pypi.aiminify.com/ https://pypi.org/simple/" pipenv install aiminify --skip-lock
TensorRT dependencies included:
  • tensorrt==10.13.0.35
  • pycuda>=2025.1.0
  • onnx>=1.17.0

After the installation process is complete, AIminify should be successfully installed in your Python environment.

Configuration

Before using AIminify, configure the library with your username using one of the following commands based on your package manager:
# Pip, pip-tools and Pipx
aiminify configure USERNAME
 
# Conda
conda run -n your_env_name aiminify configure USERNAME
 
# Poetry
poetry run aiminify configure USERNAME
 
# Pipenv
pipenv run aiminify configure USERNAME

Validation

To verify that AIminify is correctly installed and configured, attempt to import the library using the appropriate command for your environment:
# Pip, pip-tools and Pipx
python <<< "from aiminify import minify"
 
# Conda
conda run -n your_env_name python -c "from aiminify import minify"
 
# Poetry
poetry run python -c "from aiminify import minify"
 
# Pipenv
pipenv run python -c "from aiminify import minify"
To verify TensorRT support is available:
python -c "from aiminify.minify_pytorch.quantization.static_quantization import CUDA_AVAILABLE; print('TensorRT available:', CUDA_AVAILABLE)"

Now you have installed, configured and verified that AIminify is working in your environment. Now head over to  the usage manual  to continue!