Closed gaowu9595 closed 5 years ago
Check here:
https://github.com/Eromera/erfnet/blob/master/train/model/loss_general.lua
ERFNets original code was written in lua and then translated to PyTorch. While translating, i assume they just extracted class imbalance weights from cached files in lua and used directly (I assume they did it to save time).
Hi, As said in the code that 'normalization value, as defined in ERFNet paper'. So I read the paper (Efficient Residual Factorized ConvNet for Real-time Semantic) , but I didn't find something about it. Then I went to see the source code of ERFNet, as shown below. The value of weight is fixed...Can you tell me more about the origin of this formula?Thank you