2022 Developer Survey is open! Take survey.

New answers tagged

1 vote
Accepted

When exactly am I overfitting -- contradicting metrics

From the loss graph I would conclude, that at approx 2k steps overfitting starts, so using the model at approx 2k steps would be the best choice. But looking at the precision graph, training e.g. ...
user avatar
  • 23.3k
0 votes

Role of confidence or classification score in object detection mAP metrics

Nice question got me thinking, hope these help you understand better Source MS COCO — uses 101-Recall-points on PR-chart mAP (mean Average Precision) for Object Detection PascalVOC2007 — uses 11-...
user avatar
1 vote

Is there way to segment an image without labeling/classification, as well as supervised learning?

You could run your images through a pretrained CNN for feature extraction, then run an unsupervised algorithm to see if there are any obvious groups of "dogs" and "cups". Don't ...
user avatar
1 vote
Accepted

Extract person silhouette from photo or video

Check out pytorch/pyvision. This is wonderful in that it has many great models that are pretrained that you can use to train your model on and then go further and use what is trained to train further....
user avatar

Top 50 recent answers are included