I would like to start an NLP project, analysing financial data in Hebrew. To do so, I would like to train a Bert model on the Financial PhraseBank. Currently, the sentences are in English and was wondering if anyone would like to/know anyone that could assist me in translating this.
Hi The Udster,
You could take a look at https://towardsdatascience.com/going-global-how-to-multi-task-in-multiple-languages-with-the-mt5-transformer-892617cd890c
Somewhere in the article the author describes how to use the simpletransformers library (based on Huggingface Transformers) to translate datasets into several language using the MarianMT models. (opus_xyz)
The translation quality is not so high as the one from google cloud, but it still beats manual translation