gaolinorange / SDMNET_stamp

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Can not reproduce chair correctly by following the steps in README.md

susan31213 opened this issue · comments

This is a great work! I am interested in it very much.
I tried to reproduce by following the steps in README.md. I got the right result of plane but wrong result with chair data.
The armrests' positions are wrong(like the picture) and they are different from the Fig.1 in the paper.
Screenshot from 2021-03-26 17-14-55
How can I get the correct result?

Thanks a lot if any help provided.

Thanks & Regards,
Susan

Update:
Here is the output when I ran

cd ../chair_reproduce
CUDA_VISIBLE_DEVICES='' python ./test_stacknewvae.py --output_dir ./05060123_6863bin_1-joint_1-l0_100.0-l2_10.0-l3_1.0-l4_0.001-model_chair-trcet_1.0 --interids '4e664dae1bafe49f19fb4103277a6b93' '1c17cc67b8c747c3febad4f49b26ec52' '2bbf00f0c583fd8a4b3c42e318f3affc'

log.txt

It seems that you did not run the optimization process. You should run post-process after interpolation.

GetOptimizedObj('../python/chair_reproduce/05060123_6863bin_1-joint_1-l0_100.0-l2_10.0-l3_1.0-l4_0.001-model_chair-trcet_1.0/interpolation100000', 'chair', 2, 0, 0)

Hi shinxg,

Thanks for your reply! I tried to run post-process(GetOptimizedObj('../python/chair_reproduce/05060123_6863bin_1-joint_1-l0_100.0-l2_10.0-l3_1.0-l4_0.001-model_chair-trcet_1.0/interpolation100000', 'chair', 2, 0, 0)), but there are some problem in matlab codes.
I also tried to fix them and create a pull request, but I still got a different result from Fig.1:
image
Next, I tried to set use struct = 1(e.g. GetOptimizedObj('../python/chair_reproduce/05060123_6863bin_1-joint_1-l0_100.0-l2_10.0-l3_1.0-l4_0.001-model_chair-trcet_1.0/interpolation100000', 'chair', 2, 1, 0)) and got this:
image
The armrests are shifted but still in a wrong position. 😢
I also check the origin model and it looks like this:
image
Do you have any idea about the reason why cannot get the correct armrests' positions?

Thanks for your time to read this.

Regards,
Susan

Sorry for the late reply. We updated a new version. Please update your local repo and run again.

It works! Thanks for the update!