训练时是将多个instruction合并起来训练还是一个instruct训练一次 ？

[mengjiexu]

我看训练文本是text only，在训练时好像没有区分instruction和output,是直接concat在一起训练的吗？

[shizhediao]

目前开源的数据是这样的。不过代码里已经支持了text2text的训练。可以将数据转换成text2text的格式，然后进行训练。

---

Currently, the open-source data is like this. However, our code already supports training for text2text. It can be converted into the format of text2text, and then trained.

[shizhediao]

This issue has been marked as stale because it has not had recent activity. If you think this still needs to be addressed please feel free to reopen this issue. Thanks