Skip to content
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.

Sign up for GitHub

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

Jump to bottom

Support excerpts with multiple degradations #32

Open
apmcleod opened this issue Aug 30, 2019 · 4 comments
Open

Support excerpts with multiple degradations #32

apmcleod opened this issue Aug 30, 2019 · 4 comments
Assignees
@apmcleod
Labels
enhancement New feature or request requires_discussion Either a large change or changes functionality - discuss/debate before implementation
Milestone
Later

Comments

@apmcleod
Copy link
Collaborator

This includes no degradations (necessary), as well as multiple degradations (would be nice).
@apmcleod apmcleod mentioned this issue Aug 30, 2019
Closed
@apmcleod apmcleod added the enhancement New feature or request label Oct 1, 2019
@apmcleod
Copy link
Collaborator Author

apmcleod commented Oct 7, 2019

This would make the tasks very difficult, and we should leave this until future work, if anything.

@apmcleod apmcleod changed the title Make sure to support excerpts with variable number of degradations Support excerpts with multiple degradations Oct 16, 2019
@apmcleod
Copy link
Collaborator Author

No degradations is implemented. >1 is not yet (though it is theoretically possible to do so manually).

@JamesOwers JamesOwers added the requires_discussion Either a large change or changes functionality - discuss/debate before implementation label Nov 13, 2019
@apmcleod apmcleod self-assigned this May 8, 2020
@apmcleod
Copy link
Collaborator Author

A suggestion for how to do this: Currently, make_dataset gets sum(deg_proportions) > 1, it implicitly normalizes.

I suggest, for sum(deg_proportions) <= 1, it still normalizes, but for > 1, it no longer does, and instead samples degradations such that, given a degraded excerpt, the given values are expected counts of each degradation, rather than proportions. This implementation would allow the measure_errors.py script to work with no changes.

@apmcleod
Copy link
Collaborator Author

(Still not advocating for implementing this immediately)

@apmcleod apmcleod mentioned this issue Jul 17, 2020
Closed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@apmcleod apmcleod

Labels
enhancement New feature or request requires_discussion Either a large change or changes functionality - discuss/debate before implementation
Projects
None yet
Milestone
Later
Development

No branches or pull requests
2 participants
@JamesOwers @apmcleod