[solved] "Found no NVIDIA driver on your system" and "cannot connect to X server"

[MartinaRuocco]

Hi @nlakshmanan!

I was trying to follow your README (run from Docker) but unfortunately, the execution failed and the processes were killed.
Here are the errors and how I managed to solve them:

1. NVIDIA driver

UserWarning: CUDA initialization: Found no NVIDIA driver on your system. Please check that you have an NVIDIA GPU and installed a driver from http://www.nvidia.com/Download/index.aspx (Triggered internally at  /pytorch/c10/cuda/CUDAFunctions.cpp:100.)

which was strange as I do have installed both the nvidia drivers and nvidia-docker.
I solved it by modifying the "docker run" command (solution taken from here):
docker run --runtime=nvidia -e NVIDIA_DRIVER_CAPABILITIES=compute,utility -e NVIDIA_VISIBLE_DEVICES=all -it nnlnachu/surreal_pybullet bash
an alternative would be to build your image on top of the nvidia/cuda image.

2. X server

X11 functions dynamically loaded using dlopen/dlsym OK!
    cannot connect to X server                                   

which I solved by installing xvbf and running:
xvfb-run python3 surreal_subproc.py -al ppo --env gym:HalfCheetahPyBulletEnv-v0 exp1

3. SDF files
   pybullet.error: Cannot load SDF file.
   I solved it by installing pybullet with pip install -e . instead of python setup.py install as explained here.

To summarize, I think your Step 3 should be the following:

cd ./surreal
python3 -m pip install -r requirements.txt
python3 -m pip install https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow_cpu-2.1.0-cp37-cp37m-manylinux2010_x86_64.whl
python3 setup.py install
cd ../pybullet-gym
pip install -e .
apt-get update -y
apt-get install -y xvfb

After that, in order to not repeat these commands every time we want to launch a training, we can save this container state as a new image, using the command:
docker commit container-name new-image-name
(solution taken from here)

I hope this helps!