luz.metrics module¶
- class Accuracy¶
Bases:
luz.metrics.Metric- compute()¶
Compute metric.
- Return type
float
- reset()¶
Compute on epoch start.
- Return type
None
- update(output, target, **kwargs)¶
Compute on batch end.
- Parameters
output (
Tensor) – Output tensor. Shape: \((N,C)\)target (
Tensor) – Target tensor. One-hot encoded. Shape: \((N,C)\)
- Return type
None
- class DurbinWatson¶
Bases:
luz.metrics.Metric- compute()¶
Compute metric.
- Return type
float
- reset()¶
Reset metric state.
- Return type
None
- update(output, target, **kwargs)¶
Update metric state.
- Return type
None
- class FBeta(beta)¶
Bases:
luz.metrics.Metric- compute()¶
Compute metric.
- Return type
float
- reset(**kwargs)¶
Reset metric state.
- Return type
None
- update(output, target, **kwargs)¶
Compute on batch end.
- Parameters
output (
Tensor) – Output tensor. Shape: \((N,2)\)target (
Tensor) – Target tensor. One-hot encoded. Shape: \((N,2)\)
- Return type
None
- class Max(batch_dim=0)¶
Bases:
luz.metrics.Metric- compute()¶
Compute metric.
- reset()¶
Reset metric state.
- Return type
None
- update(x, **kwargs)¶
Update metric state.
- Return type
None
- class MeanStd(batch_dim=0)¶
Bases:
luz.metrics.Metric- compute()¶
Compute metric.
- reset()¶
Reset metric state.
- Return type
None
- update(x, **kwargs)¶
Update metric state.
- Return type
None
- class Metric¶
Bases:
abc.ABC- abstract compute()¶
Compute metric.
- abstract reset()¶
Reset metric state.
- Return type
None
- abstract update(**kwargs)¶
Update metric state.
- Return type
None
- class Min(batch_dim=0)¶
Bases:
luz.metrics.Metric- compute()¶
Compute metric.
- reset()¶
Reset metric state.
- Return type
None
- update(x, **kwargs)¶
Update metric state.
- Return type
None