kitti viewer about second.pytorch

[traveller59]

1. download result.pkl in model_dir/results/step_xxx
2. add this file to det path in kitti viewer.

[zwyzwy]

but my local PC have no GPU device, now I am using ssh to connect the server cluster ,how can I run the code on the server cluster via ssh and display the qt result on the local screen?

[traveller59]

you can use ssh -XC [email protected] -p xxx, then run python ./kittiviewer/viewer.py in that terminal. this method require high-quality network

[qureshizawar]

try ssh -L localport:localhost:serverport [email protected] -p xxx

then launch kittiviewer at serverport

use the localport when you launch kittiviewer in your browser on the local PC

you'll have to do this for both frontend and backend

[FrancoisMasson1990]

Hello @traveller59 in the kitty web viewer, what is the expected input of checkpointPath and configPath ?

[traveller59]

@FrancoisMasson1990 for example:

checkpointPath: /home/yy/nuscene_test/pp/mida/voxelnet-140670.tckpt
configPath: /home/yy/deeplearning/deeplearning/mypackages/second/configs/car.fhd.config

[HenryJunW]

Did you get it working? the viewer from running on the server GPU @zwyzwy

[xieqi1996]

@FrancoisMasson1990 for example:

checkpointPath: /home/yy/nuscene_test/pp/mida/voxelnet-140670.tckpt
configPath: /home/yy/deeplearning/deeplearning/mypackages/second/configs/car.fhd.config

[xieqi1996]

what is the expected input of "infoPath" and rootPath

[andraspalffy]

what is the expected input of "infoPath" and rootPath

rootPath is the folder above testing and training folders.
infoPath is one of the pkl files in rootPath, e.g. kitti_info_test.pkl

[haoliangjiang]

but my local PC have no GPU device, now I am using ssh to connect the server cluster ,how can I run the code on the server cluster via ssh and display the qt result on the local screen?

Do you mind sharing how you get it work step by step? I am using remote server as well tho not familiar with front end and back end stuffs...

[haoliangjiang]

@zwyzwy