-
Notifications
You must be signed in to change notification settings - Fork 136
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Fixes wikitext prompts + some patches on tg models #64
Conversation
…s, fix small bug in tgi models
I'll need to come back to this to inspect a bit more following some comments in the original issues about disrepancies in score |
@@ -261,7 +261,10 @@ def _get_docs_from_split(self, splits: list[str], few_shots=False) -> list[Doc]: | |||
# vs when it's used for the actual prompt. That's why we store whether we are currently using the | |||
# doc for a fewshot sample (few_shots=True) or not, which then leads to the creation of a different Doc. | |||
item["__few_shots"] = few_shots | |||
docs.extend(as_list(self.formatter(item, self.name))) | |||
cur_docs = self.formatter(item, self.name) | |||
if cur_docs is None: |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
why would it be empty ? Not sure that's expected behaviour
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I did some experiments with the original wikitext
, and it contains empty rows, which we want to skip - it should not happen a lot though
Patches #57
Might need to edit test suite