Skip to content

preliminary analysis of quantitative evaluation of translation #61

@rlskoeser

Description

@rlskoeser

output is a google document or marimo notebook (whichever is easier)

  • report on attempted correlation, do the scores align with quality of translation

initial / light effort - run some stats; are the numbers useful at all or are they trash? (or bad for our data because the translations are not literal?) any value in reporting these scores? Does comet-kiwi do anything better / more useful?

Metadata

Metadata

Assignees

Labels

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions