Open bei-startdt opened 5 years ago
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!
Thanks a lot!
@richardaecn Hi,have you experiment on detection datasets such as coco, and the results?
Hi @Angzz , we haven't tried it on detection datasets.
@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 ?
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}
Hi @shawnthu, in the formulation, we are using 1 for positive labels and 0 for negative labels.
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^)/~
how to infer the modulator
the code in your repo
for focal loss in tensorflow/models/blob/master/research/object_detection the focal loss form is the same as what is shown in paper
Could you please tell me how to transform the paper form to your form?
Thank you very much!