vision
https://github.com/pytorch/vision
Python
Datasets, Transforms and Models specific to Computer Vision
Triage Issues!
When you volunteer to triage issues, you'll receive an email each day with a link to an open issue that needs help in this project. You'll also receive instructions on how to triage issues.
Triage Docs!
Receive a documented method or class from your favorite GitHub repos in your inbox every day. If you're really pro, receive undocumented methods or classes and supercharge your commit history.
Python not yet supported17 Subscribers
Add a CodeTriage badge to vision
Help out
- Issues
- Keep complete_box_iou and complete_box_iou_loss finite for zero-heigh…
- Fix NaN gradients in sigmoid_focal_loss for confident predictions when gamma < 1
- RandomPerspective (v1 and v2) crashes with _LinAlgError on images with a one-pixel side; perspective can return garbage coefficients for collinear points
- vision_stable.cpp gets hipified into a duplicate copy under ROCm, causing "multiple definition of vision::cuda_version()"
- Fix duplicate-symbol link error for `vision_stable.cpp` on ROCm
- Fix perspective on float16/bfloat16 bounding boxes
- sigmoid_focal_loss gives NaN gradients for confident predictions when gamma < 1
- complete_box_iou_loss / complete_box_iou: NaN gradients for zero-height boxes, NaN for point boxes
- Fix overflow in XYXY to CXCYWH conversion
- Clip samplers crash on iteration when no video clips are available
- Docs
- Python not yet supported