nlp benchmarking

Mukayese

Mukayese is an extensive benchmarking platform for Turkish NLP, providing datasets, benchmarks, and leaderboards to evaluate LLMs and other NLP tools.

Mukayese screenshot

About Mukayese

Mukayese (Turkish for "comparison/benchmarking") is a platform designed to address the lack of organized benchmarks and baselines in Turkish Natural Language Processing. It provides an extensive set of datasets and benchmarks for various Turkish NLP tasks, aiming to advance state-of-the-art systems. The platform includes the MukayeseLLM Leaderboard, specifically for ranking and evaluating the performance of large language models on Turkish tasks. Its NLP benchmarks cover a wide range of functionalities, including Language Modeling, Machine Translation, Named Entity Recognition, Sentence Segmentation, Spellchecking and Correction, Summarization, and Text Classification. Each benchmark on the Mukayese platform features a leaderboard, includes one or more datasets, and provides two or more baseline models for comprehensive evaluation.

Mukayese features

  • text classification benchmarks
  • spellchecking and correction benchmarks
  • named entity recognition benchmarks
  • machine translation benchmarks
  • language modeling benchmarks
  • mukayesellm leaderboard for turkish llms
  • benchmarking platform for turkish nlp tasks
  • comprehensive turkish nlp datasets

Ratings & reviews

No reviews yet. Be the first to share how Mukayese worked for you.