"You should find 3 or 4 different NLP tasks for fine-tuning e.g. Masked Language Modelling (e.g. on Wikipedia pages) , Natural Language Inference (XNLI dataset), Paraphrasing (PAWS-X dataset), and Review Rating (MARC dataset)"
And determine which tasks contains the all languages that we want to investigate on.
Go through following suggested tasks
"You should find 3 or 4 different NLP tasks for fine-tuning e.g. Masked Language Modelling (e.g. on Wikipedia pages) , Natural Language Inference (XNLI dataset), Paraphrasing (PAWS-X dataset), and Review Rating (MARC dataset)"
And determine which tasks contains the all languages that we want to investigate on.