Model description
XLMR-MaCoCu-is is a large pre-trained language model trained on Icelandic texts. It was created by continuing training from the XLM-RoBERTa-large model. It was developed as part of the MaCoCu project and only uses data that was crawled during the project. The main developer is Rik van Noord from the University of Groningen.
XLMR-MaCoCu-is was trained on 4.4GB of Icelandic text, which is equal to 688M tokens. It was trained for 75,000 steps with a batch size of 1,024. It uses the same vocabulary as the original XLMR-large model.
The training and fine-tuning procedures are described in detail on our Github repo.
How to use
from transformers import AutoTokenizer, AutoModel, TFAutoModel
tokenizer = AutoTokenizer.from_pretrained("RVN/XLMR-MaCoCu-is")
model = AutoModel.from_pretrained("RVN/XLMR-MaCoCu-is") # PyTorch
model = TFAutoModel.from_pretrained("RVN/XLMR-MaCoCu-is") # Tensorflow
Data
For training, we used all Icelandic data that was present in the monolingual Icelandic MaCoCu corpus. After de-duplicating the data, we were left with a total of 4.4 GB of text, which equals 688M tokens.
Benchmark performance
We tested the performance of XLMR-MaCoCu-is on benchmarks of XPOS, UPOS, NER and COPA. For UPOS and XPOS, we used the data from the Universal Dependencies project. For NER, we used the data from the MIM-GOLD-NER data set. For COPA, we automatically translated the English data set by using Google Translate. For details please see our Github repo. We compare performance to the strong multi-lingual models XLMR-base and XLMR-large, but also the monolingual IceBERT model. For details regarding the XPOS/UPOS/NER fine-tuning procedure you can checkout our Github.
Scores are averages of three runs, except for COPA, for which we use 10 runs. We use the same hyperparameter settings for all models.
UPOS | UPOS | XPOS | XPOS | NER | NER | COPA | |
---|---|---|---|---|---|---|---|
Dev | Test | Dev | Test | Dev | Test | Test | |
XLM-R-base | 96.8 | 96.5 | 94.6 | 94.3 | 85.3 | 89.7 | 55.2 |
XLM-R-large | 97.0 | 96.7 | 94.9 | 94.7 | 88.5 | 91.7 | 54.3 |
IceBERT | 96.4 | 96.0 | 94.0 | 93.7 | 83.8 | 89.7 | 54.6 |
XLMR-MaCoCu-is | 97.3 | 97.0 | 95.4 | 95.1 | 90.8 | 93.2 | 59.6 |
Acknowledgements
Research supported with Cloud TPUs from Google's TPU Research Cloud (TRC). The authors received funding from the European Union’s Connecting Europe Facility 2014- 2020 - CEF Telecom, under Grant Agreement No.INEA/CEF/ICT/A2020/2278341 (MaCoCu).
Citation
If you use this model, please cite the following paper:
@inproceedings{non-etal-2022-macocu,
title = "{M}a{C}o{C}u: Massive collection and curation of monolingual and bilingual data: focus on under-resourced languages",
author = "Ba{\~n}{\'o}n, Marta and
Espl{\`a}-Gomis, Miquel and
Forcada, Mikel L. and
Garc{\'\i}a-Romero, Cristian and
Kuzman, Taja and
Ljube{\v{s}}i{\'c}, Nikola and
van Noord, Rik and
Sempere, Leopoldo Pla and
Ram{\'\i}rez-S{\'a}nchez, Gema and
Rupnik, Peter and
Suchomel, V{\'\i}t and
Toral, Antonio and
van der Werff, Tobias and
Zaragoza, Jaume",
booktitle = "Proceedings of the 23rd Annual Conference of the European Association for Machine Translation",
month = jun,
year = "2022",
address = "Ghent, Belgium",
publisher = "European Association for Machine Translation",
url = "https://aclanthology.org/2022.eamt-1.41",
pages = "303--304"
}
- Downloads last month
- 15