152 ressourcer fundet

Filtrér resultater
  • Danish multi-task CNN trained on UD Danish DDT and DaNE. Assigns context-specific token vectors, POS tags, dependency parses and named entities. Sources: Danish Universal...
    • Python
    • Source code
  • ML Powered Danish Sentiment Model
    • Python
    • Source code
  • The Danish Universal Dependencies treebank (Johannsen et al., 2015, UD-DDT) is a conversion of the Danish Dependency Treebank (Buch-Kromann et al. 2003) based on texts from...
    • coNLL-U
  • Contents of https://laegemiddelstyrelsen.dk were crawled, aligned on document and sentence level and converted into a parallel corpus. Contains 22699 translation units between...
    • TMX
  • JEX is multi-label classification software that automatically assigns a ranked list of the over six thousand descriptors (classes) from the controlled vocabulary of the EuroVoc...
    • Java
  • Contents of https://www.vikingeskibsmuseet.dk were crawled, aligned on document and sentence level and converted into a parallel corpus. Contains 1939 translation units (EN-DA)....
    • TMX
  • CoREST er et værktøj til sproglige undersøgelser i meget store tekstsamlinger, såkaldte tekstkorpusser. CoREST står for Corpus Retrieval System and Tools. OBS! CoREST kan ikke...
    • Java
  • Et værktøj til at fine-tune til NLP Transformers til sentiment analyse. Udkommer sammen med et sæt modeller fine-tunet til sentiment-analyse på dansk (udgivet på Hugging Face)....
    • Python
  • NERDA' er et værktøj (udgivet som Python-pakke) til at fine-tune NLP transformer-modeller til at identificere personer, organisationer, lokationer m.m. i tekster (=Named-Entity...
    • Python
  • DaCy er et framework til processering af dansk fritekst. Særligt indeholder den tre sprogprocesserings pipelines til dansk fritekst. De pipeline er af 3 forskellige størelser...
    • Python
  • Open-source Python-pakke til dansk talegenkendelse (tale-til-tekst). DanSpeech har arbejdet på at udvikle generelle talegenkendelsesmodeller siden 2018. Projektet har levet som...
    • Python
  • A toolkit for Part-of-Speech tagging and NER in DyNet. It has been tested on Danish, amongst other languages (for the UD POS tags in the UD_Danish-DDT version 1.1 and 2.3)...
    • Python
  • The aligned corpus consists of press releases from the European Commission Press Relase Database (Rapid) harvested in 2009 and 2011 (http://europa.eu/rapid/search.htm). The...
    • TXT
    • TMX
  • Samling af lydoptagelser i 22 kHz 1 kanal (mono). Stammer fra NST (Nordisk Språkteknologi) som gik konkurs i 2003. Er holdt ajour i den norske sprogbank i Nationalbiblioteket....
    • TXT
    • WAV
  • DUDS Jens Bille’s Ballad Book belongs to a corpus of the oldest Danish ballad tradition. The corpus consists of 9 ballad books handed down from Renaissance ballad collectors...
    • XML
  • Elektroniske versioner af størstedelen af Johannes V. Jensens udgivelser. I regi af CLARIN-projektet og i samarbejde med rettighedshaverne, gjorde Jensen Forum i...
    • HTML
    • PDF
  • Gruntvig's Works version 1,12. april 2018 contains N.F.S. Grundtvig's authorship. Corpus folder containing edited texts and OCR texts. Creator: Ravn, Kim Steen License:...
    • XML
  • The DK-CLARIN JRC-Acquis Parallel Corpus (da, en) is a part of the JRC-Acquis mulilingual parallel corpus, containing documents from The Acquis Communautaire (AC) which is the...
    • XML
  • The Danish similarity dataset is a gold standard resource for evaluation of Danish word embedding models. The dataset consists of 99 word pairs rated by 38 human judges...
    • CSV
  • The DK-CLARIN Parallel Financial Corpus comprises 4.3 M Danish and 4.8 M English tokens from translated (parallel) documents, mainly annual reports, of the period 2002-2010 from...
    • XML