bruceyo / gla-gcn Goto Github PK

I read your paper, but I can't find the global loss and shared individually connected layer in your code.
Can you tell me where can I use these module?

RuntimeError: one of the variables needed for gradient computation has been modified by an inplace operation ,may i ask why,thank you!

Thanks for your work!

• indicates reconstructing an intermediate 3D pose sequence.
  use python run_s-agcn.py -tta --evaluate 96_gt_243_supervised.bin this command can not reproduce 21.0
  how can I reproduce this result.

Great Work! Thank you for sharing for such a great project!

I have a question though..

Is it possible to change T (frames) value into 81 or lower instead of 243 to test in low frame rate?

Thank you very much!

hello again!
I was wondering why you substracted z min value in this line when you added trajectory when GT is available?

thnx

Hi, I am testing 3dhp dataset with gla-gcn model.
I have followed your instruction which you have wrote..

I've found trans_model checkpoint but I couldn't find refine_model checkpoint.

Can you help through this issue?

Thanks

The code in run_s-agcn.py says
“
if args.causal:
from common.causal_model import *
else:
from common.s_agcn import *
”
But there is no causal_model in the common folder. Could you tell me what the code is?

I found only 13 points in ‘data_2d_humaneva15_detectron_pt_coco.npz’ that matched the label . Could you tell me which joints you used?

Hi, Great Job. I want to ask why my 3DHP dataset doesn't have a valid_frame metric. Is it because I didn't process the dataset, or because my downloaded dataset has an error. Looking forward to your reply, thank you very much!

Hi， As in your case，My model is great at gt but worse at cpn，Adding some module can make gt better but cpn gets worse.What does the author think about this situation? ls there any way to solve it.
Looking forward to your reply!
Thanks！