Skip to content

Conversation

@Rickylht
Copy link

Replace sep="" as sep="<s>"
Otherwise, the model will wrongly process the seperation token. It will regard the blank space as stop token and only output one token each time during inferece.
After change, the model outputs correctly.
image

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant