Fix Kinetics dataset docstring #8121
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/vision/8121
Note: Links to docs will display an error until the docs builds have been completed. ✅ You can merge normally! (2 Unrelated Failures)As of commit 365823c with merge base 4433680 ( FLAKY - The following jobs failed but were likely due to flakiness present on trunk:
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
NicolasHug
added
enhancement
module: datasets
module: documentation
and removed
cla signed
labels
| @@ -152,13 +152,13 @@ def _compute_frame_pts(self) -> None: | |||
| with tqdm(total=len(dl)) as pbar: | |||
| for batch in dl: | |||
| pbar.update(1) | |||
| clips, fps = list(zip(*batch)) | |||
| batch_pts, batch_fps = list(zip(*batch)) | |||
There was a problem hiding this comment.
This is just renaming of local variables. The main thing is: the variable that used to be called clips really really should not have been called clips. These are pts (presentation time stamps) for each frame in the videos. Nothing to do with a clip in this VideoClips context.
There was a problem hiding this comment.
I don't know how, but this is somehow causing segfaults: https://github.com/pytorch/vision/actions/runs/6920303074/attempts/1#summary-18824649629
|
It's a pre-existing failure: https://hud.pytorch.org/pr/pytorch/vision/7990 I saw it earlier last week, but couldn't match it to a change in torch core. I was hoping it would just go away but we might need look into it more seriously. In any case, it's not related to this PR. |
Reviewed By: vmoens Differential Revision: D52539009 fbshipit-source-id: 7631bf6d4de0c144d581eb324944f14163f4d5e5
The docstring is using the term "clip" instead of "video".
In the context of the video datasets / io / samplers:
cc @pmeier