daming5432

关于paddle resnet50模型动态图concat reduce_mean、reduce_mean_grad、c_broadcast 的作用

1

你好，在打印模型训练log时发现resnet50会执行到concat、reduce_mean、reduce_mean_grad、c_braodcast这些op。我用 netron打开保存的模型并没有发现这些op。能帮忙解释一下为什么会用到这些op，以及这些op在训练过程中的作用吗？非常感谢~

standard normalised landmarks mean error

1

How can I caculate the standard normalised landmarks mean error? I caculate use mean error = (mean_nor / distance). Is the "distance" = "eye_distance" right?

load model

3

How can I load several models simultaneously?

the input is ./build/bin/ocr_test -ds=/home/daming/Work/Pytorch/LPR/TaiWanData/1 -mnd=/home/daming/Work/Pytorch/LPR/LPRMTCNN/src/ocr_test/model_platecar -pmd=/home/daming/Work/Pytorch/LPR/LPRMTCNN/src/ocr_test/plateCard_test 0 /home/daming/Work/Pytorch/LPR/TaiWanData 0384159482759-87_93-367&427_654&550-656&527_369&554_379&441_666&414-0_0_13_19_33_27_24-95-81.jpg detector point 224.399368,363.916168 671.017761,263.650604 667.478333,568.336975 229.654694,683.795410 cost: 17.7248 (ms) detect plate region (0, 0) (720 x 1160) with confidence 0.000000,...

How can I load several models simultaneously?