MCPcopy Create free account
hub / github.com/InternRobotics/EmbodiedScan / wrapper

Function wrapper

embodiedscan/models/losses/reduce_loss.py:100–123  ·  view source on GitHub ↗

Args: pred (Tensor): The prediction. target (Tensor): Target bboxes. weight (Optional[Tensor], optional): The weight of loss for each prediction. Defaults to None. reduction (str, optional): Options are "none", "mean" and "sum"

(pred: Tensor,
                target: Tensor,
                weight: Optional[Tensor] = None,
                reduction: str = 'mean',
                avg_factor: Optional[int] = None,
                **kwargs)

Source from the content-addressed store, hash-verified

98
99 @functools.wraps(loss_func)
100 def wrapper(pred: Tensor,
101 target: Tensor,
102 weight: Optional[Tensor] = None,
103 reduction: str = 'mean',
104 avg_factor: Optional[int] = None,
105 **kwargs) -> Tensor:
106 """
107 Args:
108 pred (Tensor): The prediction.
109 target (Tensor): Target bboxes.
110 weight (Optional[Tensor], optional): The weight of loss for each
111 prediction. Defaults to None.
112 reduction (str, optional): Options are "none", "mean" and "sum".
113 Defaults to 'mean'.
114 avg_factor (Optional[int], optional): Average factor that is used
115 to average the loss. Defaults to None.
116
117 Returns:
118 Tensor: Loss tensor.
119 """
120 # get element-wise loss
121 loss = loss_func(pred, target, **kwargs)
122 loss = weight_reduce_loss(loss, weight, reduction, avg_factor)
123 return loss
124
125 return wrapper

Callers

nothing calls this directly

Calls 1

weight_reduce_lossFunction · 0.85

Tested by

no test coverage detected