Dongdong Chen

What is the best performance when training by this pytorch repo?

1

@longcw @cory8249 @xuzijian, when training on VOC2007+2012 trainval with current version pytorch implementation, what is the best performance you guys obtained on VOC2007 test dataset now ?

Nan bug

3

In the line 33-35 of accum_layer.cu, the latter number of min should be "src_count -1" but not src_count, I guess this may be the reason why you add some NAN...

Hi @1900zyh yanhong, I suggest you to help sort the paper in a year-descending order, so that it is easier to find the latest progress in the field.

@niuzhiheng ,I want to build the 3rd-party dependencies by myself, could you give me some instructions about building the "leveldb","glog"...with visual studio? Thanks a lot.