sci-ml/lm-eval - 0.4.13 (stuff)

Search

Install

Install this version:

emerge -a =sci-ml/lm-eval-0.4.13

If this version is masked, you can unmask it using the autounmask tool or standard emerge options:

autounmask =sci-ml/lm-eval-0.4.13

Or alternatively:

emerge --autounmask-write -a =sci-ml/lm-eval-0.4.13

Package Information

Description:
A framework for evaluating language models (lm-evaluation-harness)
Homepage:
https://github.com/EleutherAI/lm-evaluation-harness https://pypi.org/project/lm-eval/
License:
MIT

Ebuild Details

Version EAPI Keywords Slot
0.4.13 8 ~amd64 ~arm64 0
View Raw Ebuild
# 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

# PyPI uses lm_eval in the sdist name but lm-eval for the project.
PYPI_PN="lm_eval"
PYPI_NO_NORMALIZE=1

inherit distutils-r1 pypi

DESCRIPTION="A framework for evaluating language models (lm-evaluation-harness)"
HOMEPAGE="
	https://github.com/EleutherAI/lm-evaluation-harness
	https://pypi.org/project/lm-eval/
"

LICENSE="MIT"
SLOT="0"
KEYWORDS="~amd64 ~arm64"
IUSE="+api ifeval math onnxruntime sentencepiece statsmodels vllm"

# Expose only fully packaged extras; omit lazy ONNX backend's unpackaged
# onnxruntime-genai. rechecked 2026-09-10
# math asserts antlr4 4.11 at task load; ifeval needs nltk>=3.9.1 for punkt fixes
# (#2210, nltk #3266). verified 2026-05-13
# datasets/evaluate are single-impl; wrap the remaining Python dependencies.
RDEPEND="
	>=sci-ml/datasets-2.16.0[${PYTHON_SINGLE_USEDEP}]
	>=sci-ml/evaluate-0.4.0[${PYTHON_SINGLE_USEDEP}]
	onnxruntime? ( sci-ml/transformers[${PYTHON_SINGLE_USEDEP}] )
	vllm? ( >=dev-python/vllm-0.18.0[${PYTHON_SINGLE_USEDEP}] )
	$(python_gen_cond_dep '
		dev-python/dill[${PYTHON_USEDEP}]
		dev-python/jinja2[${PYTHON_USEDEP}]
		dev-python/more-itertools[${PYTHON_USEDEP}]
		dev-python/numpy[${PYTHON_USEDEP}]
		dev-python/pytablewriter[${PYTHON_USEDEP}]
		dev-python/rouge-score[${PYTHON_USEDEP}]
		dev-python/sacrebleu[${PYTHON_USEDEP}]
		dev-python/scikit-learn[${PYTHON_USEDEP}]
		dev-python/sqlitedict[${PYTHON_USEDEP}]
		dev-python/tqdm[${PYTHON_USEDEP}]
		dev-python/typing-extensions[${PYTHON_USEDEP}]
		dev-python/word2number[${PYTHON_USEDEP}]
		api? (
			dev-python/aiohttp[${PYTHON_USEDEP}]
			dev-python/requests[${PYTHON_USEDEP}]
			dev-python/tenacity[${PYTHON_USEDEP}]
			dev-python/tiktoken[${PYTHON_USEDEP}]
			dev-python/tqdm[${PYTHON_USEDEP}]
		)
		ifeval? (
			dev-python/langdetect[${PYTHON_USEDEP}]
			dev-python/immutabledict[${PYTHON_USEDEP}]
			>=dev-python/nltk-3.9.1[${PYTHON_USEDEP}]
		)
		math? (
			=dev-python/antlr4-python3-runtime-4.11*[${PYTHON_USEDEP}]
			>=dev-python/sympy-1.12[${PYTHON_USEDEP}]
			~dev-python/math-verify-0.9.0[${PYTHON_USEDEP}]
		)
		onnxruntime? ( >=sci-libs/onnxruntime-1.23[python,${PYTHON_USEDEP}] )
		sentencepiece? ( >=sci-ml/sentencepiece-0.1.98[${PYTHON_USEDEP}] )
		statsmodels? ( dev-python/statsmodels[${PYTHON_USEDEP}] )
	')
"

USE Flags

Manage flags for this package: euse -i <flag> -p sci-ml/lm-eval | euse -E <flag> -p sci-ml/lm-eval | euse -D <flag> -p sci-ml/lm-eval

Global/Standard Flags

api
Default: Enabled (+)

Inherited Eclasses

Dependencies

RDEPEND

	>=sci-ml/datasets-2.16.0[${PYTHON_SINGLE_USEDEP}]
	>=sci-ml/evaluate-0.4.0[${PYTHON_SINGLE_USEDEP}]
	onnxruntime? ( sci-ml/transformers[${PYTHON_SINGLE_USEDEP}] )
	vllm? ( >=dev-python/vllm-0.18.0[${PYTHON_SINGLE_USEDEP}] )
	$(python_gen_cond_dep '
		dev-python/dill[${PYTHON_USEDEP}]
		dev-python/jinja2[${PYTHON_USEDEP}]
		dev-python/more-itertools[${PYTHON_USEDEP}]
		dev-python/numpy[${PYTHON_USEDEP}]
		dev-python/pytablewriter[${PYTHON_USEDEP}]
		dev-python/rouge-score[${PYTHON_USEDEP}]
		dev-python/sacrebleu[${PYTHON_USEDEP}]
		dev-python/scikit-learn[${PYTHON_USEDEP}]
		dev-python/sqlitedict[${PYTHON_USEDEP}]
		dev-python/tqdm[${PYTHON_USEDEP}]
		dev-python/typing-extensions[${PYTHON_USEDEP}]
		dev-python/word2number[${PYTHON_USEDEP}]
		api? (
			dev-python/aiohttp[${PYTHON_USEDEP}]
			dev-python/requests[${PYTHON_USEDEP}]
			dev-python/tenacity[${PYTHON_USEDEP}]
			dev-python/tiktoken[${PYTHON_USEDEP}]
			dev-python/tqdm[${PYTHON_USEDEP}]
		)
		ifeval? (
			dev-python/langdetect[${PYTHON_USEDEP}]
			dev-python/immutabledict[${PYTHON_USEDEP}]
			>=dev-python/nltk-3.9.1[${PYTHON_USEDEP}]
		)
		math? (
			=dev-python/antlr4-python3-runtime-4.11*[${PYTHON_USEDEP}]
			>=dev-python/sympy-1.12[${PYTHON_USEDEP}]
			~dev-python/math-verify-0.9.0[${PYTHON_USEDEP}]
		)
		onnxruntime? ( >=sci-libs/onnxruntime-1.23[python,${PYTHON_USEDEP}] )
		sentencepiece? ( >=sci-ml/sentencepiece-0.1.98[${PYTHON_USEDEP}] )
		statsmodels? ( dev-python/statsmodels[${PYTHON_USEDEP}] )
	')