focal loss modulator about class-balanced-loss HOT 8 OPEN

Hi @bei-startdt

Thanks for pointing this out! The implementation you mentioned is not very numerically stable (same for the implementation in https://github.com/tensorflow/tpu/blob/master/models/official/retinanet/retinanet_model.py#L130-L162). When gamma is small (< 1), there might be NaN occurs during back-propagation.

The full derivation can be found in the figure below. Hope this will help!

from class-balanced-loss.

Thanks a lot!

from class-balanced-loss.

@richardaecn Hi,have you experiment on detection datasets such as coco, and the results?

from class-balanced-loss.

Hi @Angzz , we haven't tried it on detection datasets.

from class-balanced-loss.

@richardaecn Hi , have you compared the class balanced focal loss with the orignal focal loss using resnet 50 or 101 ? When did such comparsion , you used resnet 32 in your paper. Will stronger networks weaken the framework you proposed ?

from class-balanced-loss.

modulator = tf.exp(-gamma * labels * logits - gamma * tf.log1p( tf.exp(-1.0 * logits)))
should be
modulator = tf.exp(-gamma * labels * logits - gamma * tf.log1p( tf.exp(-1.0 * labels * logits)))
labels in {-1, 1}

from class-balanced-loss.

Hi @shawnthu, in the formulation, we are using 1 for positive labels and 0 for negative labels.

from class-balanced-loss.

Hi @shawnthu, in the formulation, we are using 1 for positive labels and 0 for negative labels.

in fact we are both right, but your solution more concise (^o^)/~

from class-balanced-loss.