search

real-stanford / flingbot

[CoRL 2021 Best System Paper] This repository contains code for training and evaluating FlingBot in both simulation and real-world settings on a dual-UR5 robot arm setup for Ubuntu 18.04
https://flingbot.cs.columbia.edu/
106 stars 25 forks source link

ImportError: /home/hc/dextairity/PyFlex/bindings/build/pyflex.cpython-36m-x86_64-linux-gnu.so: undefined symbol: cudaSetupArgument #6

Open canhe173 opened 1 year ago

canhe173 commented 1 year ago

Hi guys, When I run python test_sim.py, I get the following error:

Traceback (most recent call last): File "test_sim.py", line 2, in <module> from sim_env import SimEnv File "/home/hc/dextairity/sim_env.py", line 7, in <module> import pyflex ImportError: /home/hc/dextairity/PyFlex/bindings/build/pyflex.cpython-36m-x86_64-linux-gnu.so: undefined symbol: cudaSetupArgument

Can anyone help me??

ngkhiem97 commented 1 year ago

I got the same issue. I'm running the program on Ubuntu 20.04

ngkhiem97 commented 1 year ago

I'm thinking maybe CUDA is not installed

ngkhiem97 commented 1 year ago

I have figured out where the error was. I forgot to replace the anaconda3 directory when starting the Nvidia docker container at

nvidia-docker run \
    -v $FLINGBOT_PATH:/workspace/flingbot\
    -v /path/to/your/anaconda3:/path/to/your/anaconda3\
    --gpus all --shm-size=64gb  -d -e DISPLAY=$DISPLAY -e QT_X11_NO_MITSHM=1 -it flingbot
adakan4 commented 1 year ago

I am also having this issue.

adakan4 commented 1 year ago

@canhe173 Running the docker container attached instead of detached (removed the -d tag in the docker run command) worked for me. Make sure to stay in the flingbot conda environment by using conda activate flingbot again after starting the docker container. Then exit the docker container using exit before evaluating or training.

zcswdt commented 9 months ago

@canhe173 Running the docker container attached instead of detached (removed the -d tag in the docker run command) worked for me. Make sure to stay in the flingbot conda environment by using conda activate flingbot again after starting the docker container. Then exit the docker container using exit before evaluating or training.

Have you run through the author's training code yet