bokbokbok.loss_functions.classification
WeightedCrossEntropyLoss(alpha=0.5)
¶
Calculates the Weighted Cross-Entropy Loss, which applies a factor alpha, allowing one to trade off recall and precision by up- or down-weighting the cost of a positive error relative to a negative error.
A value alpha > 1 decreases the false negative count, hence increasing the recall. Conversely, setting alpha < 1 decreases the false positive count and increases the precision.
Source code in bokbokbok/loss_functions/classification/classification_loss_functions.py
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 |
|
WeightedFocalLoss(alpha=1.0, gamma=2.0)
¶
Calculates the Weighted Focal Loss.
Note that if using alpha = 1 and gamma = 0, this is the same as using regular Cross Entropy.
The more gamma is increased, the more the model is focussed on the hard, misclassified examples.
A value alpha > 1 decreases the false negative count, hence increasing the recall. Conversely, setting alpha < 1 decreases the false positive count and increases the precision.
Source code in bokbokbok/loss_functions/classification/classification_loss_functions.py
85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 |
|