metadata
language: eu
license: cc-by-sa-4.0
datasets:
- cc100
widget:
- text: Euria egingo <mask> gaur ?
- text: <mask> umeari liburua eman dio.
- text: Zein da zure <mask> ?
RoBERTa Basque x-small model (Uncased)
Prerequisites
transformers==4.19.2
Model architecture
This model uses half the size of RoBERTa base setttings.
Tokenizer
Using BPE tokenizer with vocabulary size 50,000.
Training Data
- Subset of CC-100/eu : Monolingual Datasets from Web Crawl Data
Usage
from transformers import pipeline
unmasker = pipeline('fill-mask', model='ClassCat/roberta-xsmall-basque')
unmasker("Zein da zure <mask> ?")