Peizhen Zhang
Peizhen Zhang
I attempted to train with larger base lr that bigger than 0.001 ( specifically, 0.002~0.005) but all of these bring about NaN CE and smoothL1 loss values. This verifies that...
Oh ! Of course, I 've already tried the same bash command as I said in previous comments -- "For strong ablation, I also strictly execute the official training comman...
+1. Could release the vgg16 pre-trained model on ImageNet too. @foreverYoungGitHub @ShuangXieIrene
Hello ! Have you ever manged to modify the project codes to support multiple-gpu training and testing ? @isalirezag
> I use the labelenc in a custom model. In training step1 the autoencoder loss are not going down。in my opinion，the autoencoder loss should be in sync with detection loss...
> I'm really sorry to bother you. Recently, I have also been learning this project, but I have encountered some problems. The source code did not work in my own...
> > > I'm really sorry to bother you. Recently, I have also been learning this project, but I have encountered some problems. The source code did not work in...
Any update here ?
看报错的log来看, 你的meta_arch是None，请问您是跑什么backbone呢。可能需要在你的config文件里指定好这个变量名。
24G I guess, since I run a RTX TITAN 24G with bsz=1, it is full. OOM occurs after setting bsz>=2. By the way, I notice in the paper the authors...