Closed osma closed 1 year ago
Kudos, SonarCloud Quality Gate passed!
0 Bugs
0 Vulnerabilities
0 Security Hotspots
0 Code Smells
No Coverage information
0.0% Duplication
Base: 99.56% // Head: 99.56% // Increases project coverage by +0.00%
:tada:
Coverage data is based on head (
f4b55cd
) compared to base (a7e3b4b
). Patch coverage: 100.00% of modified lines in pull request are covered.
:umbrella: View full report at Codecov.
:loudspeaker: Do you have feedback about the report comment? Let us know in this issue.
This PR clarifies that backends only have to implement either one of
_suggest
and_suggest_batch
, then implements batched suggest in the Omikuji backend. In practice, only the text vectorization is performed on the whole batch at once; the Omikuji implementation only supports a predict method for a single document at a time so it has to be done within afor
loop.There seems to be a small performance benefit. I tested this using
annif eval
the Finto AI yso-parabel-fi project/model, with the kirjaesittelyt2021/fin/test corpus. The evaluation results were unchanged, only the amount of time spent was slightly different. Memory usage remained pretty much the same.With 1 job
With 4 jobs
Fixes #665