Found 0 GPUs for rendering. Using device 0.

[ggggfff1]

Hi,

I have encountered some problems with MetaWorld evaluation on a headless server. I suspect that MuJoCo is failing to find a GPU for rendering. However, I have confirmed that my GPUs work properly during training. Have you encountered this issue when deploying DP3 on a remote server? Looking forward to your reply!

Here is the error message:

Found 0 GPUs for rendering. Using device 0.
Device id outside of range of available devices.
Traceback (most recent call last):
File "test.py", line 27, in
img = sim.render(12, 12, camera_name='corner2', depth=False, device_id=-1,mode='offscreen',)
File "mjsim.pyx", line 156, in mujoco_py.cymj.MjSim.render
File "mjsim.pyx", line 158, in mujoco_py.cymj.MjSim.render
File "mjrendercontext.pyx", line 46, in mujoco_py.cymj.MjRenderContext.init
File "mjrendercontext.pyx", line 114, in mujoco_py.cymj.MjRenderContext._setup_opengl_context
File "opengl_context.pyx", line 130, in mujoco_py.cymj.OffscreenOpenGLContext.init
RuntimeError: Failed to initialize OpenGL

Additionally, I referred to another issue, but it did not solve my problem.

Thank you for your help!

[YanjieZe]

This seems to be a problem in metaworld/mujoco/opengl. like this?
openai/mujoco-py#187

[wut19]

@ggggfff1 Have you handled this problem? I've met the same problem. Any suggestions?
Thank you.

[ggggfff1]

@ggggfff1 Have you handled this problem? I've met the same problem. Any suggestions? Thank you.

Hi, I found a solution, is to use

self.viewer = mujoco_py.MjRenderContextOffscreen(self.env.sim, None, device_id=self.device_id) 
self.viewer.render(self.image_size, self.image_size, self.env.sim.model.camera_name2id("corner2"))
img = self.viewer.read_pixels(self.image_size, self.image_size, depth=False)

instead of
img = self.env.sim.render(width=self.image_size, height=self.image_size, camera_name="corner2", device_id=self.device_id)
in 3D-Diffusion-Policy/diffusion_policy_3d/env/metaworld/metaworld_wrapper.py line130

but it will occur new problems. The eval speed will become slower and the result will become bad. @wut19

[ggggfff1]

This seems to be a problem in metaworld/mujoco/opengl. like this? openai/mujoco-py#187

I have tried this method, but it doesn't work. @YanjieZe