Install this version:
emerge -a =dev-python/stop-sequencer-1.2.3-r1If this version is masked, you can unmask it using the autounmask tool or standard emerge options:
autounmask =dev-python/stop-sequencer-1.2.3-r1Or alternatively:
emerge --autounmask-write -a =dev-python/stop-sequencer-1.2.3-r1# Copyright 1999-2026 Gentoo Authors
# Distributed under the terms of the GNU General Public License v2
EAPI=8
DISTUTILS_USE_PEP517=setuptools
PYTHON_COMPAT=( python3_{12..14} )
DISTUTILS_SINGLE_IMPL=1
# upstream sdist filename uses dash, eclass would normalise to underscore
PYPI_NO_NORMALIZE=1
inherit distutils-r1 pypi
DESCRIPTION="Stop-sequence support for Hugging Face Transformers generation"
HOMEPAGE="
https://github.com/hyunwoongko/stop-sequencer
https://pypi.org/project/stop-sequencer/
"
LICENSE="Apache-2.0"
SLOT="0"
KEYWORDS="~amd64 ~arm64"
# upstream's setup.py declares only transformers, but both modules open with
# a bare `import torch`, so the package cannot be imported without it. The
# transformers dependency does not cover that: sci-ml/transformers has
# IUSE="torch" with torch off by default, so a default build of it pulls no
# pytorch at all. verified 2026-07-27
RDEPEND="
sci-ml/pytorch[${PYTHON_SINGLE_USEDEP}]
>=sci-ml/transformers-4.3.0[${PYTHON_SINGLE_USEDEP}]
"
src_prepare() {
# upstream find_packages() picks up tests/ which lands as a stray
# top-level dir in site-packages
sed -i 's|find_packages()|find_packages(exclude=["tests*"])|' setup.py || die
distutils-r1_src_prepare
}