-
Notifications
You must be signed in to change notification settings - Fork 14
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
numeric match and topic tag #72
base: main
Are you sure you want to change the base?
Conversation
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.
Prompt questions
|
||
|
||
class AIME_PIPELINETag(AIME_PIPELINE): | ||
"""This class specifies the config for running AIME benchmark 5 repeated times""" |
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.
update comment so it reflects the functionality of the class
# Each query is tagged with one or more topics from arithmetic, algebra, counting, geometry, number theory, and probability and other. | ||
# These topics follow the description on the official website: https://artofproblemsolving.com/wiki/index.php/American_Invitational_Mathematics_Examination?srsltid=AfmBOooSIQ8ua5aJX00ZtYCKDuOAB4I4c-YE9zr1xYZ86fq8x5RL2sEg. | ||
# In their own words, "The AIME tests mathematical problem solving with arithmetic, algebra, counting, geometry, number theory, and probability and other secondary school math topics" | ||
return pipeline |
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.
Since the class inherits from the original AIME_PIPELINE it will continue to run the rest of the AIME_PIPELINE but with the tagging prompt. For example, this means that it will also try to extract an answer and generate the report. There are two options here: 1) Either to not inherit from AIME_PIPELINE, or 2) Inherit from AIME_PIPELINE but then return only the components you need in the pipeline. For example,
return PipelineConfig(
[
self.data_processing_comp,
self.inference_comp,
self.data_post_processing,
],
self.log_dir,
)
In case 2, also requires changing the answer extractor as the marker is different here.
Add a new metric "numeric match" to compare two numeric values.
Add a pipeline to extract topics from a math question.