Installation

Requirements

The current package requires Python 3.9 or later. The core installation includes the dependencies for text, image, static, and dynamic inference. Video file input/output and hyperparameter tuning use optional dependencies, as described below.

From PyPI (Version 0.2.1)

gpi_pack 0.2.1 is available from PyPI. Install or upgrade the core package with:

python -m pip install --upgrade gpi-pack

To include video support, Optuna tuning, or both, install the corresponding extras:

# Video reading and H.264 writing
python -m pip install --upgrade "gpi-pack[video]"

# Hyperparameter tuning
python -m pip install --upgrade "gpi-pack[tune]"

# Both optional feature sets
python -m pip install --upgrade "gpi-pack[video,tune]"

Version 0.2.1 includes the static interfaces, video processing, scalar and repeated-outcome dynamic estimators, and both optional dependency groups.

From Source

For an editable development installation, clone the source repository and install it from the repository root:

git clone https://github.com/gpi-pack/gpi_pack.git
cd gpi_pack
python -m pip install -e .

Use -e ".[video]", -e ".[tune]", or -e ".[video,tune]" to add the optional dependencies. You can confirm the installed package version with:

import gpi_pack

print(gpi_pack.__version__)  # 0.2.1