Bertrade Contexual Embeddings Models for Old French
Creators
- 1. Modyco, Université Paris Nanterre and CNRS, Nanterre France
- 2. Institut für Linguistik/Romanistik (ILR), Universität Stuttgart, Deutschland
- 3. Inria, Paris, France
- 4. LLF, CNRS and Université Paris Cité, Paris, France
Description
These are the contextual embeddings models for Old French developed for the article “BERTrade: Using Contextual Embeddings to Parse Old French” (Grobol et al., 2022). They are meant to be used with the 🤗 Transformers library and are known to work with versions from 4.12.0 to 4.18.0 of this library (other versions might be compatible, but we offer no guarantee).
See the article for a list of the resources and the settings used to develop these models as well as assessment of their suitability for syntactic dependency parsing of Old French.
References:
Grobol, Loïc, Mathilde Regnault, Pedro Javier Ortiz Suárez, Benoît Sagot, Laurent Romary, and Benoit Crabbé. 2022. “BERTrade: Using Contextual Embeddings to Parse Old French.” In Proceedings of the 13th International Conference on Language Resources and Evaluation. European Language Resource Association.
Files
Files
(3.1 GB)
Name | Size | Download all |
---|---|---|
md5:4fd3f1e4348f0f8a86844ee5f13a91d4
|
451.8 MB | Download |
md5:4934017a214e3be73c4ee009d0083272
|
644.3 MB | Download |
md5:0f81f5362690c2cfec666df7a56d9244
|
334.5 MB | Download |
md5:bdbb77d5efcaa64a5553e2e743eeebd5
|
334.1 MB | Download |
md5:6ff3679f87753be20585883fc9563b06
|
154.2 MB | Download |
md5:d30be53a1d600a9fcba21d49c07fc64a
|
43.2 MB | Download |
md5:cf47f3b3f76cf267ce6d6312b5a4bd50
|
43.2 MB | Download |
md5:40e6caccca3aa056cfa03abfc1e4736e
|
154.2 MB | Download |
md5:acea51a7245f92a0f6e9a76d70d704ee
|
5.7 MB | Download |
md5:a5e8309f27363a50c1d3bafa998db730
|
402.6 MB | Download |
md5:49d238e0ad8eeb5cd370cfc60cab416a
|
502.7 MB | Download |
Additional details
Funding
- PROFITEROLE – PRocessing Old French Instrumented TExts for the Representation Of Language Evolution ANR-16-CE38-0010
- Agence Nationale de la Recherche