I can't find a decent pytorch implementation of softmax focal loss for Semantic Segmentation. So I code one on my own.
type python loss.py to test the loss.
An intuitive pytorch implementation of softmax focal loss which can be used to train classification and segmentation tasks.
I can't find a decent pytorch implementation of softmax focal loss for Semantic Segmentation. So I code one on my own.
type python loss.py to test the loss.