gcunhase/NLPMetrics

Python code for various NLP metrics

44
/ 100
Emerging

This tool helps machine translation researchers and practitioners evaluate the quality of their automated translations. You input a machine-generated translation and one or more human-generated reference translations, and it calculates metrics like BLEU, GLEU, and WER. The output is a score indicating how well the machine translation matches the human references, helping you assess and improve your models.

169 stars. No commits in the last 6 months.

Use this if you need to quantitatively measure the performance of machine translation systems against human-quality translations.

Not ideal if you need metrics for tasks beyond machine translation, such as text summarization or image captioning, which are not yet fully implemented.

Machine Translation NLP Evaluation Translation Quality Assessment Language Model Benchmarking
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 10 / 25
Maturity 16 / 25
Community 18 / 25

How are scores calculated?

Stars

169

Forks

28

Language

Jupyter Notebook

License

MIT

Last pushed

Oct 17, 2019

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/nlp/gcunhase/NLPMetrics"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.