fastwer and werx
These are **competitors**, as both are Python packages designed for fast Word Error Rate (WER) calculation, making them alternatives for the same core task in ASR evaluation.
About fastwer
kahne/fastwer
A PyPI package for fast word/character error rate (WER/CER) calculation
Leverages C++ bindings via pybind11 for high-performance computation of both word and character-level error rates. Supports dual granularity evaluation through corpus-level aggregation and sentence-level scoring, enabling detailed error analysis across different scopes.
About werx
analyticsinmotion/werx
🐍📦 Easy-to-use Python package for lightning-fast Word Error Rate (WER) analysis
Powered by a Rust core, WERx delivers exceptional performance—processing ~907,000 utterances per second on LibriSpeech benchmarks with minimal memory overhead. Beyond standard WER metrics, it provides detailed word-level error breakdowns (insertions, deletions, substitutions) with customizable error weights, and integrates seamlessly with Pandas and Polars for downstream analysis. Designed for production use with minimal dependencies, it handles edge cases gracefully while supporting both single-sentence and corpus-level evaluations.
Related comparisons
Scores updated daily from GitHub, PyPI, and npm data. How scores work