kpu/kenlm

Health v2 · maintenance only

Flags: no_releases no_license

Adoption not part of the score

2799 stars · 551 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

KenLM is a C++ toolkit for estimating, filtering, and querying n-gram language models with modified Kneser-Ney smoothing. It is designed for fast, low-memory queries and scalable on-disk estimation, and is commonly embedded in machine translation and speech recognition decoders.

Use cases

• estimate an n-gram language model from a large text corpus
• query language model scores quickly with low memory
• convert ARPA language models to a fast binary format
• filter an ARPA model down to a specific vocabulary
• integrate a language model into a machine translation decoder
• score sentences with a language model for speech recognition rescoring

When to choose

• you need fast, memory-efficient n-gram language model scoring in C++ or Python
• you are training large n-gram LMs that do not fit in memory
• you need a battle-tested LM for MT or ASR decoding pipelines

When to avoid

• you need neural language models rather than n-gram models
• you want a pure-Python solution without C++ compilation
• you need heavy pruning or modern smoothing methods beyond modified Kneser-Ney

Facets

library · maturity stable

nlp machine-learning cli parser machine-learning developer-tools windows cpp python cli language-model n-gram kneser-ney arpa-format speech-recognition-support mt-decoder natural-language-processing linux macos

Member repositories

For agents

markdown · JSON · MCP: product_card(name="kpu/kenlm")