n_classes = 4
inp = torch.normal(0, 1, (16, 20, n_classes))
targ = torch.randint(0, n_classes, (16, 20)).to(torch.int8)
_mAP(inp, targ)0.27493315845795063Metrics
Metrics not included in fastai.
MatthewsCorrCoefBinary
def MatthewsCorrCoefBinary(
sample_weight:NoneType=None
):
Matthews correlation coefficient for single-label classification problems
get_task_metrics
def get_task_metrics(
dls, binary_metrics:NoneType=None, multi_class_metrics:NoneType=None, regression_metrics:NoneType=None,
verbose:bool=True
):
Call self as a function.
All metrics applicable to multi classification have been created by Doug Williams (https://github.com/williamsdoug). Thanks a lot Doug!!
F1_multi
def F1_multi(
args:VAR_POSITIONAL, kwargs:VAR_KEYWORD
):
Call self as a function.
Fbeta_multi
def Fbeta_multi(
inp, targ, beta:float=1.0, thresh:float=0.5, sigmoid:bool=True
):
Computes Fbeta when inp and targ are the same size.
balanced_accuracy_multi
def balanced_accuracy_multi(
inp, targ, thresh:float=0.5, sigmoid:bool=True
):
Computes balanced accuracy when inp and targ are the same size.
specificity_multi
def specificity_multi(
inp, targ, thresh:float=0.5, sigmoid:bool=True
):
Computes specificity (true negative rate) when inp and targ are the same size.
recall_multi
def recall_multi(
inp, targ, thresh:float=0.5, sigmoid:bool=True
):
Computes recall when inp and targ are the same size.
precision_multi
def precision_multi(
inp, targ, thresh:float=0.5, sigmoid:bool=True
):
Computes precision when inp and targ are the same size.
metrics_multi_common
def metrics_multi_common(
inp, targ, thresh:float=0.5, sigmoid:bool=True, by_sample:bool=False
):
Computes TP, TN, FP, FN when inp and targ are the same size.
accuracy_multi
def accuracy_multi(
inp, targ, thresh:float=0.5, sigmoid:bool=True, by_sample:bool=False
):
Computes accuracy when inp and targ are the same size.
mae
def mae(
inp, targ
):
Mean absolute error between inp and targ.
mape
def mape(
inp, targ
):
Mean absolute percentage error between inp and targ.