You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
First of all thanks for this amazing open source work!
Second, I converted eng-pol model to HF format with instruction: https://github.com/huggingface/transformers/tree/master/scripts/tatoeba, but when I use it with transformers library I get sequence of dots ".................." as result. I also tried with pol-eng which is available in HF repo as opus-mt-pl-en but no success again.
Do you have any idea what am I doing wrong?
Thanks
The text was updated successfully, but these errors were encountered:
There was a bug in the conversion script that assumed that special tokens are in the beginning of the vocabulary. That has been fixed recently and it should work with the latest source code. Note, that we also work on updating the conversion procedures to make them more flexible with other variants of marian models. That is not yet ready but hopefully will come soon.
First of all thanks for this amazing open source work!
Second, I converted eng-pol model to HF format with instruction: https://github.com/huggingface/transformers/tree/master/scripts/tatoeba, but when I use it with transformers library I get sequence of dots ".................." as result. I also tried with pol-eng which is available in HF repo as opus-mt-pl-en but no success again.
Do you have any idea what am I doing wrong?
Thanks
The text was updated successfully, but these errors were encountered: