wuzhiyang2016

very thanks, i just see your reply, let me take some time to understand your great reply~

hello, i have check the class QuantizeDequantizeG it's backward code, dx is the same with the paper's formula (8), but the scale derivative is not the same with the paper's...

i get, it's really clear ! very thanks!!!

ok, very thanks!

@zfc929 你好，负样本的box标签设置为0之后，训练出来的模型效果怎么样？

//regression box : y x height width regression_box.push_back(cv::Rect(regression_box_temp_[i + count] * width, regression_box_temp_[i] * width, regression_box_temp_[i + count*3] * width, regression_box_temp_[i + count*2] * width)); hi, could you tell me...

At the train time, the label regress as [x,y,w,h],,,at the test time ,,,it must the order [x,y ,w,h],,,but in the code, why save as [y,x,h,w] ??? it is so confused!!!

@foreverYoungGitHub @LittlePeng @GarrickLin

Very thanks,i get it! @GarrickLin

and we set part-face label is 1, positive-face label also 1, i mean these two's number is more than negative-samples,,,it is ok? because i train an 1:3:3，learning rate is 0.0001,,,...