fighting_liu
fighting_liu
When i import mxnet in 8 processes simultaneously, all cpu resources will be used and the program stagnates for almost 5 minutes. It works well for **mxnet1.1** but failed for...
@BobLiu20 Thanks for your code, but your mAP computing is actually recall at specific confidence threshold, which is totally different from mAP. Here is my test results on your models...
As you note, Inception-v3 512x512 is the fastest model. But from my test, vgg-reduced is around 100FPS, but Inception-v3 512x512 got only 30FPS, is it normal? Any hint? Thanks!
Hi!! I've seen you use another network structure for your new work(pose_deploy_resize), and kernel size for CONV layers is much smaller, is there any difference ? Can you show me...
As the input size is 224*224(ignore the channel), conv1 has kernel_size:7, pad:3, stride:2, as i compute (224+3x2-7) / 2 + 1 = 112.5, so we use floor to get 112?...