Please describe how to use BERT or ELECTRA Models for Longformer training.
- Linguagem predominante
- Python
- Estrelas
- 2.2k
- Forks
- 286
- Métricas de merge de PRs
- Nenhum PR com merge em 30d
Descrição
Hi,
I am interested in traing a german longformer model. But there is no german RoBERTa. I have seen that it is possible to use BERT as a basis but you only have a description to use RoBERTa. Since there seems to be some demand:
1. https://github.com/allenai/longformer/issues/85
2. https://github.com/allenai/longformer/issues/21
3. https://github.com/allenai/longformer/issues/82
4. https://github.com/allenai/longformer/issues/77
5. https://github.com/allenai/longformer/issues/72
6. https://github.com/allenai/longformer/issues/65
7. https://github.com/allenai/longformer/issues/61
Could you please add a 2nd jupyter notebook on how to use BERT as basis or describe in detail what to change on the existing notebook for RoBERTa? That would realy be awsome.
Thanks
Philip
Guia de contribuição
Nenhum guia de contribuição indexado para este repositório
Avaliação
Esta issue ainda não foi avaliada.