Language modeling: Difference between revisions
No edit summary |
(Marked this version for translation) |
||
| (17 intermediate revisions by 4 users not shown) | |||
| Line 1: | Line 1: | ||
==n-gram modeling== | <languages/> | ||
<translate> | |||
==n-gram modeling== <!--T:1--> | |||
<!--T:11--> | |||
Colibri core is an NLP tool as well as a C++ and Python library for working with basic linguistic constructions such as n-grams and skipgrams (i.e. patterns with one or more gaps, either of fixed or dynamic size) in a quick and memory-efficient way. At the core is the tool colibri-patternmodeller which allows you to build, view, manipulate and query pattern models. | |||
<!--T:2--> | |||
*[http://proycon.github.io/colibri-core/ Github repository] | *[http://proycon.github.io/colibri-core/ Github repository] | ||
== | ==Large Language Models== <!--T:3--> | ||
<!--T:12--> | |||
* [https://huggingface.co/models?search=dutch Hugging Face Dutch Models] | |||
* [https://people.cs.kuleuven.be/~pieter.delobelle/robbert/ RobBERT]: A Dutch RoBERTa-based Language Model | * [https://people.cs.kuleuven.be/~pieter.delobelle/robbert/ RobBERT]: A Dutch RoBERTa-based Language Model | ||
* [https://github.com/wietsedv/bertje BERTje]: A Dutch BERT model | |||
* [https://github.com/Rijgersberg/GEITje GEITje]: A Large Open Language Model | |||
*[https://huggingface.co/Tweeties/tweety-7b-dutch-v24a Tweety] | |||
==Multilingual Language Models including Dutch== <!--T:4--> | |||
<!--T:13--> | |||
* [https://openai.com/ GPT-3] | |||
* [https://huggingface.co/docs/transformers/model_doc/mbart MBart] | |||
==SpaCy== <!--T:5--> | |||
<!--T:14--> | |||
spaCy is a free open-source library for Natural Language Processing in Python. | |||
<!--T:6--> | |||
* [https://spacy.io/models/nl Dutch models] | * [https://spacy.io/models/nl Dutch models] | ||
== Language Modeling Benchmarks == <!--T:7--> | |||
===DUMB=== <!--T:15--> | |||
<!--T:16--> | |||
DUMB is a benchmark for evaluating the quality of language models for Dutch NLP tasks. The set of tasks is designed to be diverse and challenging, to test the limits of current language models. The specific datasets and formats are particularly suitable for fine-tuning encoder models, and applicability for large generative models is yet to be determined. Please read the paper for more details. | |||
<!--T:8--> | |||
* [https://dumbench.nl/ DuMB] | |||
===LLM Leaderboard=== <!--T:9--> | |||
<!--T:17--> | |||
This is a leaderboard for Dutch benchmarks for large language models. | |||
<!--T:10--> | |||
* [https://huggingface.co/spaces/BramVanroy/open_dutch_llm_leaderboard Open Dutch LLM Leaderboard] | |||
=== Scandeval Dutch NLG === <!--T:18--> | |||
<!--T:19--> | |||
*[https://scandeval.com/dutch-nlg/ Scandeval Dutch NLG] | |||
</translate> | |||
Latest revision as of 12:44, 24 June 2024
n-gram modeling
Colibri core is an NLP tool as well as a C++ and Python library for working with basic linguistic constructions such as n-grams and skipgrams (i.e. patterns with one or more gaps, either of fixed or dynamic size) in a quick and memory-efficient way. At the core is the tool colibri-patternmodeller which allows you to build, view, manipulate and query pattern models.
Large Language Models
- Hugging Face Dutch Models
- RobBERT: A Dutch RoBERTa-based Language Model
- BERTje: A Dutch BERT model
- GEITje: A Large Open Language Model
- Tweety
Multilingual Language Models including Dutch
SpaCy
spaCy is a free open-source library for Natural Language Processing in Python.
Language Modeling Benchmarks
DUMB
DUMB is a benchmark for evaluating the quality of language models for Dutch NLP tasks. The set of tasks is designed to be diverse and challenging, to test the limits of current language models. The specific datasets and formats are particularly suitable for fine-tuning encoder models, and applicability for large generative models is yet to be determined. Please read the paper for more details.
LLM Leaderboard
This is a leaderboard for Dutch benchmarks for large language models.