-
Hi, thank you for sharing the code!
I am trying to reproduce your results on the local_plane_guidence layer. If I understand correctly, I should just neglect the concatenation step of the output of…
-
Hi, I want to know how do you determine the final best model (End-to-end) in your Table IV at last? There are five indicators in Table IV, How do you balance this 5 indicators to select the final best…
-
尊敬的徐学长,你好!一个小小的问题。您肯定有做过attention机制的消融实验吧!额,请问下您的attention机制能带来多少的精度提升呢?
-
[Issue migrated from SourceForge | ID: 1910557 | Submitted by 'bryan_tripp']
[http://sourceforge.net/support/tracker.php?aid=1910557]
Make it easy to ablate neurons/ensembles. Current method is to se…
-
Have you tried the effect of different loss weights on the distillation results ?
-
Hi
in your ablation experiments where you have used Conv in encoder and decoder, isn‘t that same as UNET?
why then do you have for the Conv variants so less params and GFLOPS compared with UNET…
-
您好,我想问一下在消融实验中移除GA模块,是采用简单的均值聚合替代。
是只是取消分组,但跳数、关系还保留
还是说聚合成一个单独向量,如果是这种方式,后面的可学习编码部分该如何处理,这种方式感觉可学习编码部分也相应失效了
-
I'm running the command to evaluate using openpifpaf.eval on the PoseTrack 2018 dataset. When I run with the command
`python3 -m openpifpaf.eval --checkpoint tshufflenetv2k16 --dataset=posetrack2018 …
-
Hi, Thanks for sharing the code for the interesting paper. May I ask some questions?
1. What is the difference between the ablation study variant Con-GAE-sp and Con-GAE-NonContextDec? (Both variants …
-