kornia.metrics

Metrics to monitor and evaluate vision models during training and validation.

Classification	Top-k accuracy.
Segmentation	Confusion matrix and mean IoU.
Detection	Mean average precision and box IoU.
Image quality	PSNR and SSIM in 2D and 3D.
Optical flow	Average end-point error.
Stereo	Disparity error metrics.
Pose	Rotation and translation errors and AUC.
Monitoring	Running averages for training loops.