Open jichilen opened 5 years ago
Excellent question, that exp was from an experiment (since this is used in darknet), it should be removed or behind a flag. I will remove it for now.
Unfortunately I don't have a good answer for the weights, it really depends on the network. You might try even weighting then doing some search for better params
https://github.com/generalized-iou/g-tensorflow-models/blob/d0ff754286a3b9b106eb3a0b55b4cbedbfb24269/research/object_detection/core/losses.py#L215
should i use the exp function for faster r-cnn training?
https://github.com/generalized-iou/g-tensorflow-models/blob/d0ff754286a3b9b106eb3a0b55b4cbedbfb24269/research/object_detection/core/losses.py#L217 And what weigths should i choose? 0.25 in g-darknet or 10 as paper metioned about?