yolo3_grid_loss: Calculates loss for one 'Yolo3' grid.
In maju116/platypus: Tools for Computer Vision in R

Calculates loss for one 'Yolo3' grid.

yolo3_grid_loss(
  y_true,
  y_pred,
  anchors,
  n_class,
  nonobj_threshold,
  bbox_lambda = 1,
  obj_lambda = 1,
  noobj_lambda = 1,
  class_lambda = 1,
  class_weights = rep(1, n_class)
)

`y_true`	Tensor of true coordinates/scores.
`y_pred`	Tensor of predicted coordinates/scores.
`anchors`	Prediction anchors (for one grid). For exact format check `coco_anchors`.
`n_class`	Number of prediction classes.
`nonobj_threshold`	Non-object ignore threshold.
`bbox_lambda`	Bounding box loss lambda.
`obj_lambda`	Object loss lambda.
`noobj_lambda`	Nonobject loss lambda.
`class_lambda`	Class loss lambda.
`class_weights`	Vector of length 'n_class' with class weights.