ERROR: Shadow framebuffer is not complete, error 0x8cd7

[dmzsyl]

(psl) yl@yl:~/planseqlearn$ python planseqlearn/launch_scripts/psl_mopa.py [2024-07-13 17:23:44,778][absl][INFO] - MUJOCO_GL=egl, attempting to import specified OpenGL backend. libEGL warning: DRI2: failed to create dri screen libEGL warning: Not allowed to force software rendering when API explicitly selects a hardware device. libEGL warning: DRI2: failed to create dri screen libEGL warning: DRI2: failed to create dri screen libEGL warning: Not allowed to force software rendering when API explicitly selects a hardware device. libEGL warning: DRI2: failed to create dri screen [2024-07-13 17:23:44,854][absl][INFO] - MuJoCo library version is: 2.3.5 [robosuite WARNING] No private macro file found! (init.py:7) [2024-07-13 17:23:45,409][robosuite_logs][WARNING] - No private macro file found! [robosuite WARNING] It is recommended to use a private macro file (init.py:8) [2024-07-13 17:23:45,409][robosuite_logs][WARNING] - It is recommended to use a private macro file [robosuite WARNING] To setup, run: python /home/yl/yes/envs/psl/lib/python3.8/site-packages/robosuite/scripts/setup_macros.py (init.py:9) [2024-07-13 17:23:45,409][robosuite_logs][WARNING] - To setup, run: python /home/yl/yes/envs/psl/lib/python3.8/site-packages/robosuite/scripts/setup_macros.py /home/yl/yes/envs/psl/lib/python3.8/site-packages/gym/spaces/box.py:73: UserWarning: WARN: Box bound precision lowered by casting to float32 logger.warn( workspace: /home/yl/planseqlearn/exp_local/mopa_psl_final_mopa_SawyerLiftObstacle-v0_wrist_67384919_2024.07.13_17:23:43 [07/13 17:23] INFO: initial qpos: [ 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.55 -0.25 0.86 -0.87958388 0. 0. 0.47574384] [07/13 17:23] INFO: initial qvel: [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.] [07/13 17:23] INFO: is_limited: [ True True True True True True True True True] [07/13 17:23] INFO: control_range: [[-3.04 -3.8 -3.04 -3.04 -2.98 -2.98 -4.71 -0.0115 -0.0115 ] [ 3.04 3.8 3.04 3.04 2.98 2.98 4.71 0.020833 0.020833]] /home/yl/yes/envs/psl/lib/python3.8/site-packages/gym/spaces/box.py:73: UserWarning: WARN: Box bound precision lowered by casting to float32 logger.warn( [07/13 17:23] INFO: initial qpos: [ 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.55 -0.25 0.86 -0.87958388 0. 0. 0.47574384] [07/13 17:23] INFO: initial qvel: [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.] [07/13 17:23] INFO: is_limited: [ True True True True True True True True True] [07/13 17:23] INFO: control_range: [[-3.04 -3.8 -3.04 -3.04 -2.98 -2.98 -4.71 -0.0115 -0.0115 ] [ 3.04 3.8 3.04 3.04 2.98 2.98 4.71 0.020833 0.020833]] Text plan: [] ERROR: Shadow framebuffer is not complete, error 0x8cd7

Press Enter to exit ...

[mihdalal]

Can you please provide additional details?

[dmzsyl]

Thank you for your reply. Now I want to know how I can evaluate the results, can you tell me which instruction to use

[mihdalal]

We don't have a specific evaluation script, but in the training loop evaluation is done here: https://github.com/mihdalal/planseqlearn/blob/master/planseqlearn/train.py#L230

[dmzsyl]

Thank you very much for your reply, I now have two more questions that I would like to ask, and would be grateful if you could respond to them

1. The training result can only be in wandb, can it show pybullet or MUJOCO visualisation interface?
2. The success rate in wandb is only displayed to one decimal place, I read that the success rate of robosuite_NutAssemblyRound in your paper is as high as 96%, and the following is the result that I ran.

[mihdalal]

1. there should also be videos saved out
2. probably let it train for longer + run multiple seeds, it looks like it is still improving

[YeJX0108]

How did you fix this ERROR, 0x8cd7? @dmzsyl

[mihdalal]

closing due to inactivity