google/mt5-base

Google's mT5 is pretrained on the mC4 corpus, covering 101 languages . The model has to be fine-tuned before it is useable on a downstream task . It is pre-trained on a new Common Crawl-based dataset . The code and model checkpoints used in this work are publicly available . For more information, visit the paper here: mT5: A massively multilingual pre-training text-to-text transformer. The work is published by Linting Xue, Noah Constant, Adam Roberts, Mihir Kale, Rami Al-Rfou, Aditya Siddhant and Colin Raffel at Google.com,