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
I'd like to find a clean way to decouple loading the Frames of a video Sample from loading the sample itself. Currently we have this behavior:
importfiftyone.zooasfozdataset=foz.load_zoo_dataset("quickstart-video")
# takes a non-trivial amount of time since the frames are always # attached, even if I don't want themprint(dataset.first())
I'd rather make frame loading JIT if possible (noting that, when DatasetViews are involved, the aggregation pipeline may have internally attached and then detached frames in order to compute which samples are in the view):
forsampleinvideo_dataset:
# `sample` loads fast because `frames` haven't been loaded yetprint(sample.id)
# now frames are loadedprint(len(sample.frames))
The text was updated successfully, but these errors were encountered:
Originally posted by @brimoor in #858 (comment)
I'd like to find a clean way to decouple loading the
Frames
of a videoSample
from loading the sample itself. Currently we have this behavior:I'd rather make frame loading JIT if possible (noting that, when
DatasetView
s are involved, the aggregation pipeline may have internally attached and then detached frames in order to compute which samples are in the view):The text was updated successfully, but these errors were encountered: