activeloopai / activeloopai/deeplake
[BUG] Issue with 300-W DataSet
- Vorherrschende Sprache
- C++
- Sterne
- 9.2k
- Forks
- 722
- PR-Merge-Kennzahlen
- Keine gemergten PRs in 30 T.
Beschreibung
### Severity
P1 - Urgent, but non-breaking
### Current Behavior
In the 300w dataset, there are incorrect data points. A bunch of keypoints just have 2 as one of the x,y values causing keypoints to be on the edges:

I originally thought these keypoints may have just been marked with a 2 as a flag that means they aren't useful, but the image on the api website to visualize these keypoints have these points in the correct location:
from: https://app.activeloop.ai/activeloop/300w/firstdbf9474d461a19e9333c2fd19b46115348f

### Steps to Reproduce
This is the code I wrote to show the keypoint locations (image with the green keypoints)
`
ds = deeplake.load("hub://activeloop/300w")
imgs = ds.images
points = ds.keypoints.numpy()
points = points.reshape(599, 102, 2)
img = imgs[0].numpy()
point = points[0]
print(point)
for x, y in point:
cv2.circle(img, (int(x), int(y)), 20, (0, 255, 0), -1)
plt.imshow(img)
plt.axis('off')
plt.show()
`
### Expected/Desired Behavior
I expect the points to not be on the edge.
### Python Version
3.11.7
### OS
Windows 11
### IDE
VScode
### Packages
latest
### Additional Context
_No response_
### Possible Solution
_No response_
### Are you willing to submit a PR?
- [ ] I'm willing to submit a PR (Thank you!)
Beitragsleitfaden
Bewertung
Dieses Issue wurde noch nicht bewertet.