kornia/kornia
Repository · https://github.com/kornia/kornia
🐍 Geometric Computer Vision Library for Spatial AI
First seen by Oriel on .
AI summary
Kornia is a PyTorch-based differentiable computer vision library offering extensive batched image processing, geometric vision, augmentation, feature matching, and spatial AI capabilities.
- Kornia provides over 1,200 public operators covering image processing, camera models, epipolar geometry, minimal solvers, RANSAC, feature matching, morphology, losses, and metrics.
- Operations run on input devices without NumPy round trips, support automatic differentiation, and accept per-sample parameters in batched calls.
- AugmentationSequential applies and can invert shared transforms across images, masks, boxes, and keypoints, while 3D augmentations transform volumes without annotations or inversion.
- Published 0.9.0 benchmarks report many workloads running 10× to 70× faster than per-image OpenCV or scikit-image on CPU, though OpenCV wins single-image uint8 CPU work.
- CPU half-precision manifests record 243 float16 and 256 bfloat16 failures, while CUDA lacks blocking CI and a checked-in known-failure manifest.
Generated by gpt-5.6-sol. Check the original for details.
Metric trends
Last 30 days · Up to 120 recorded samples per source and metric, with reposts kept separate · First and latest retained
Stars
Total repository stars at each observation.
- GitHub Trending: 12,276, latest observation ; +705 over observed period since
Stars gained (1d)
Daily star gain reported by GitHub Trending. A change compares two daily values; it is not additional stars between observations.
- GitHub Trending: 412, latest observation ; +324 over observed period since
Forks
Total repository forks at each observation.
- GitHub Trending: 1,505, latest observation ; +49 over observed period since
Aliases
| Source | External ID | URL |
|---|---|---|
| GitHub Trending | kornia/kornia | https://github.com/kornia/kornia Read discussion |