Model card for huBERT (#6893)

* Create README.md

Model card for huBERT.

* Update README.md

lowercase h

* Update model_cards/SZTAKI-HLT/hubert-base-cc/README.md

Co-authored-by: Julien Chaumond <chaumond@gmail.com>
This commit is contained in:
David Mark Nemeskey
2020-09-02 10:50:10 +02:00
committed by GitHub
parent 1889e96c8c
commit e3c55ceb8d

View File

@@ -0,0 +1,43 @@
---
language: hu
license: apache-2.0
datasets:
- common_crawl
- wikipedia
---
# huBERT base model (cased)
## Model description
Cased BERT model for Hungarian, trained on the (filtered, deduplicated) Hungarian subset of the Common Crawl and a snapshot of the Hungarian Wikipedia.
## Intended uses & limitations
The model can be used as any other (cased) BERT model. It has been tested on the chunking and
named entity recognition tasks and set a new state-of-the-art on the former.
## Training
Details of the training data and procedure can be found in the PhD thesis linked below. (With the caveat that it only contains preliminary results
based on the Wikipedia subcorpus. Evaluation of the full model will appear in a future paper.)
## Eval results
When fine-tuned (via `BertForTokenClassification`) on chunking and NER, the model outperforms multilingual BERT, achieves state-of-the-art results on the
former task and comes within 0.5% F1 to the SotA on the latter. The exact scores are
| NER | Minimal NP | Maximal NP |
|-----|------------|------------|
| 97.62% | **97.14%** | **96.97%** |
### BibTeX entry and citation info
```bibtex
@PhDThesis{ Nemeskey:2020,
author = {Nemeskey, Dávid Márk},
title = {Natural Language Processing Methods for Language Modeling},
year = {2020},
school = {E\"otv\"os Lor\'and University}
}
```