Focal_Loss_Keras
Focal_Loss_Keras copied to clipboard
Multi-class classification with focal loss for imbalanced datasets
Considering the dataset you use, you only have two classes, so this is a binary classification problem. Unlike cifar10 for example, where the are 10 classes, hence multilabel classification
I noticed that the value for alpha in keras_focal_loss.ipynb is set to 1. Could you please tell me how did you find this value? Thank you for your help!
In the following lines, the computation of `weight` multiplies `y_true` one more time. ```Python ce = tf.multiply(y_true, -tf.log(model_out)) weight = tf.multiply(y_true, tf.pow(tf.subtract(1., model_out), gamma)) ``` I find that in current...
Tough its not an issue but more of a confusion. in the paper alpha is set to .25 you have set it to 4 When i try other available implementations,...