File size: 749 Bytes
346a35c c433bb3 346a35c c433bb3 346a35c |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 |
---
language:
- de
- en
pipeline_tag: translation
---
# mt5-small-nc16-250k-deen
This model is released as part of the work from [Are Character-level Translations Worth the Wait? Comparing Character- and Subword-level Models for Machine Translation](https://arxiv.org/abs/2302.14220).
It is an mT5 model finetuned on German-->English translation using 250k sentence pairs from the WMT NewsCommentary v16 dataset.
To use the model correctly, you must prepend the prompt with "translate X to Y: ", where X and Y are your source and target languages (e.g. German, English).
NOTE: The decoder_start_token_id is 259 for byt5 models and 250099 for mt5 models, which is different from the default token from google's byt5 and mt5 models (which is 0). |