# Input for multilingual EncoderDecoderModel

**URL:** <https://discuss.huggingface.co/t/input-for-multilingual-encoderdecodermodel/22824>\
**Category:** Beginners\
**Created:** [September 9, 2022, 7:23am UTC](https://discuss.huggingface.co/t/input-for-multilingual-encoderdecodermodel/22824 "2022-09-09T07:23:47Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![PandaKata](https://avatars.discourse-cdn.com/v4/letter/p/eb8c5e/32.png) [@PandaKata](https://discuss.huggingface.co/u/PandaKata)\
**Post date:** [September 9, 2022, 7:23am UTC](https://discuss.huggingface.co/t/input-for-multilingual-encoderdecodermodel/22824/1 "2022-09-09T07:23:47Z")

</div>

Hello guys,

I want to train a multilingual EncoderDecoderModel for translation.

```auto
model = EncoderDecoderModel.from_encoder_decoder_pretrained('xlm-roberta-base', 'EleutherAI/gpt-j-6B')

```

Now I’m having a hard time figuring out what input is needed. I have several parallel corpora in different languages - do I need to combine them all into one corpus?

Thank you!