metrics.py

from tensorflow.keras import backend as K

class Metrics():
    def __init__(self, threshold=0.5):
        self.threshold=threshold
        
    def recall_m(self, y_true, y_pred):
            y_labels = y_true
            y_pred = K.cast(K.greater(K.clip(y_pred, 0, 1), self.threshold), K.floatx())        
            possible_positives = K.sum(K.round(K.clip(y_labels, 0, 1)))
            true_positives = K.sum(K.round(K.clip(y_labels * y_pred, 0, 1)))
            recall = true_positives / (possible_positives + K.epsilon())
            return recall

    def precision_m(self, y_true, y_pred):
            y_labels = y_true
            y_pred = K.cast(K.greater(K.clip(y_pred, 0, 1), self.threshold), K.floatx())        
            true_positives = K.sum(K.round(K.clip(y_labels * y_pred, 0, 1)))
            predicted_positives = K.sum(K.round(K.clip(y_pred, 0, 1)))
            precision = true_positives / (predicted_positives + K.epsilon())
            return precision

    def f1_m(self, y_true, y_pred):
        precision = self.precision_m(y_true, y_pred)
        recall = self.recall_m(y_true, y_pred)
        return 2*((precision*recall)/(precision+recall+K.epsilon()))
    
    def auc2(self, y_true, y_pred):
        auc = tf.metrics.auc(y_true, y_pred)[1]
        return auc
 
def binary_crossentropy_custom(y_true, y_pred):
    y_labels = y_true
    return K.binary_crossentropy(y_labels, 
                                 y_pred)