When switching translation providers, the switching doesn't stop the translateLocally instance running in the background.
Same thing occurs in the benchmark: any translateLocally instances used for any of the runs will remain active (albeit dormant, but still using memory) while the benchmark page is open.
When switching translation providers, the switching doesn't stop the translateLocally instance running in the background.
Same thing occurs in the benchmark: any translateLocally instances used for any of the runs will remain active (albeit dormant, but still using memory) while the benchmark page is open.