helpers.py

from tensorflow.keras import backend as K

def recall_m(y_true, y_pred):
        true_positives = K.sum(K.round(K.clip(y_true * y_pred, 0, 1)))
        possible_positives = K.sum(K.round(K.clip(y_true, 0, 1)))
        recall = true_positives / (possible_positives + K.epsilon())
        return recall

def precision_m(y_true, y_pred):
        true_positives = K.sum(K.round(K.clip(y_true * 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(y_true, y_pred):
    precision = precision_m(y_true, y_pred)
    recall = recall_m(y_true, y_pred)
    return 2*((precision*recall)/(precision+recall+K.epsilon()))


def weighted_categorical_crossentropy(weights):
    """
    A weighted version of keras.objectives.categorical_crossentropy
    
    Variables:
        weights: numpy array of shape (C,) where C is the number of classes
    
    Usage:
        weights = np.array([0.5,2,10]) # Class one at 0.5, class 2 twice the normal weights, class 3 10x.
        loss = weighted_categorical_crossentropy(weightsv)
        model.compile(loss=loss,optimizer='adam')
    """
    
    weights = K.variable(weights)
        
    def loss(y_true, y_pred):
        # scale predictions so that the class probas of each sample sum to 1
        y_pred /= K.sum(y_pred, axis=-1, keepdims=True)
        # clip to prevent NaN's and Inf's
        y_pred = K.clip(y_pred, K.epsilon(), 1 - K.epsilon())
        # calc
        loss = y_true * K.log(y_pred) * weights
        loss = -K.sum(loss, -1)
        return loss
    
    return loss

def get_weighted_loss(weights):
    def weighted_loss(y_true, y_pred):
        return K.mean((weights[:,0]**(1-y_true))*(weights[:,1]**(y_true))*K.binary_crossentropy(y_true, y_pred), axis=-1)
    return weighted_loss


def get_weighted_focal_loss(weights, gamma=2):
    def weighted_loss(y_true, y_pred):
        return K.mean((weights[:, 0]**(1-y_true))*(weights[:, 1]**(y_true)) * (((y_pred)**gamma) ** (1-y_true)) * (((1-y_pred)**gamma) ** (y_true)) * K.binary_crossentropy(y_true, y_pred), axis=-1)

    return weighted_loss