Namespace: Models

Accuracy, defined by probability of correct classification.

Functions and values

Function or valueDescription
correctlyClassified trgt pred
Signature: trgt:Tensor<'T> -> pred:Tensor<'T> -> float<MeasureOne>
Type parameters: 'T

Calculates the number of correctly classified samples. Target must be in one-hot encoding. Shapes: pred[smpl, class], target[smpl, class]

ofClassifier (...)
Signature: dataset:TrnValTst<'S> -> batchSize:int64 -> trgtFn:('S -> Tensor<'T>) -> predFn:('S -> Tensor<'T>) -> float<MeasureOne> * float<MeasureOne> * float<MeasureOne>
Type parameters: 'S, 'T

Calculates the accuracies of a classifier on the training, validation and test sets.

Fork me on GitHub