Metadata-Version: 2.4
Name: inferencebench-mt
Version: 0.1.0
Summary: Machine-translation plugin for InferenceBench Suite (chrF/BLEU/exact-match on bundled fixtures)
Project-URL: Homepage, https://github.com/yobitelcomm/bench
Author-email: Yobitel Communications <bench@yobitel.com>
License: Apache-2.0
Keywords: ai,benchmark,bleu,chrf,llm,ml,translation
Classifier: Development Status :: 2 - Pre-Alpha
Classifier: Intended Audience :: Developers
Classifier: Intended Audience :: Science/Research
Classifier: License :: OSI Approved :: Apache Software License
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.12
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Requires-Python: >=3.12
Requires-Dist: inferencebench-envelope
Requires-Dist: inferencebench-harness
Requires-Dist: pydantic~=2.9
Requires-Dist: pyyaml~=6.0
Description-Content-Type: text/markdown

# inferencebench-mt

Machine-translation plugin for the InferenceBench Suite.

Scores model translations against bundled reference fixtures using chrF (character
n-gram F-score), token-level BLEU, or exact match. Mirrors the contract of the
other plugins (`list_benchmarks` / `get_benchmark` / `validate` / `run`) and
emits the canonical signed envelope.

Two bundled benchmarks ship out of the box:

- `llm.mt.flores-200-mini-en-fr` — FLORES-200-style English to French, chrF.
- `llm.mt.flores-200-mini-en-de` — FLORES-200-style English to German, chrF.

The fixtures are tiny (eight rows each, mixed across greeting / news / technical
/ conversational domains) — intended for skeleton verification, not headline
numbers.
