You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository has been archived by the owner on Dec 21, 2023. It is now read-only.
1 - Each time tc.one_shot_object_detector.create(...) is called the tar file is always extracted, even if those files have already been extracted. The extraction takes a little over 30 seconds. So we shouldn't be doing it if we don't need to do it.
2 - The files are extracted to the user's current directory (in an SArray named one_shot_backgrounds.sarray). We should be extracting these files to somewhere in the Turi Create's data cache (i.e. somewhere under the value return by tc.toolkits._pre_trained_models._get_cache_dir(_type="data")).
The text was updated successfully, but these errors were encountered:
There are two related issues here:
1 - Each time
tc.one_shot_object_detector.create(...)
is called the tar file is always extracted, even if those files have already been extracted. The extraction takes a little over 30 seconds. So we shouldn't be doing it if we don't need to do it.2 - The files are extracted to the user's current directory (in an SArray named
one_shot_backgrounds.sarray
). We should be extracting these files to somewhere in the Turi Create's data cache (i.e. somewhere under the value return bytc.toolkits._pre_trained_models._get_cache_dir(_type="data")
).The text was updated successfully, but these errors were encountered: