Toolbox/Guidelines to run for Linemod Dataset?

[ashar6194]

Hi @yuxng ,

Thanks for providing a well structured code package. Could you also provide data pre-processing steps for the LINEMOD object dataset? I am unable to figure out ways to generate train/test index files.

P.S. - I am trying to train your network for benchvise object in linemod.

Thanks,

[2017hack]

@ashar6194 Can you run the the lov_color_2d_train.sh and lov_color_2d_test.sh? Do you happen below problem: "failed to call to cuInit: CUDA_ERROR_NO_DEVICE, No kernel was registered to support Op ‘HoughvotinggpuGrad’ with these attires. registered devices: [cpu], registered kernels: deice = 'gpu'; T in [DT_FLOAT]". How can you solve it, can you share it with us, thanks!

[ashar6194]

Actually, I specifically had to run for linemod, so I did not download LOV before, Now that i have it downloaded with me, I am facing the exact error as you have mentioned above @2017hack .

Do you have any thoughts in the above items @yuxng ?

[yuxng]

Can you make sure you compile the layers in lib correctly?

[huckl3b3rry87]

@yuxng can you please upload the datasetfolder to Google drive i.e.,

/home/yuxiang/Datasets/*

Thank you very much

[huckl3b3rry87]

looks like some data is here

@abdule-mkit do you have the data that would be here (for instance)

/home/yuxiang/Datasets/LINEMOD_SIXD/data/

I found some linemod data sets, but i am not sure which one @yuxng used and if they were processed. https://www.tugraz.at/institute/icg/research/team-lepetit/research-projects/object-detection-and-3d-pose-estimation/

http://campar.in.tum.de/Main/StefanHinterstoisser

[renjiahao0928]

@ashar6194 Can you run the the lov_color_2d_train.sh and lov_color_2d_test.sh? Do you happen below problem: "failed to call to cuInit: CUDA_ERROR_NO_DEVICE, No kernel was registered to support Op ‘HoughvotinggpuGrad’ with these attires. registered devices: [cpu], registered kernels: deice = 'gpu'; T in [DT_FLOAT]". How can you solve it, can you share it with us, thanks!

same errors,have you solved it? @2017hack