Isaaclab simulation crashed when training, only report warnings

Hi. I meet a problem when I’m running the tutorial example of Unitree H1 biped robot, following the Isaaclab documentation:
python source/standalone/workflows/rl_games/train.py --task Isaac-H1-Direct-v0 –headless
The simulation crashed and reported these warnings.

[INFO][AppLauncher]: Loading experience file: D:\projects\IsaacLab\source\apps\isaaclab.python.headless.kit
[Warning] [omni.isaac.kit.simulation_app] Modules: [‘omni.kit_app’] were loaded before SimulationApp was started and might not be loaded correctly.
[Warning] [omni.isaac.kit.simulation_app] Please check to make sure no extra omniverse or pxr modules are imported before the call to SimulationApp(…)
Loading user config located at: ‘c:/users/17985/.conda/envs/isaaclab/lib/site-packages/omni/data/Kit/Isaac-Sim/4.2/user.config.json’
[Info] [carb] Logging to file: c:/users/17985/.conda/envs/isaaclab/lib/site-packages/omni/logs/Kit/Isaac-Sim/4.2/kit_20241121_201412.log
2024-11-21 12:14:12 [0ms] [Warning] [omni.kit.app.plugin] No crash reporter present, dumps uploading isn’t available.
2024-11-21 12:14:15 [3,055ms] [Warning] [gpu.foundation.plugin] Skipping unsupported non-NVIDIA GPU: Intel(R) Iris(R) Xe Graphics
2024-11-21 12:14:15 [3,055ms] [Warning] [gpu.foundation.plugin] Skipping unsupported non-NVIDIA GPU: Intel(R) Iris(R) Xe Graphics

[INFO]: Parsing configuration from: <class ‘omni.isaac.lab_tasks.direct.humanoid.h1_env.H1EnvCfg’>
[INFO]: Parsing configuration from: D:\projects\IsaacLab\source\extensions\omni.isaac.lab_tasks\omni\isaac\lab_tasks\direct\humanoid\agents\rl_games_ppo_cfg.yaml
[INFO] Logging experiment in directory: D:\projects\IsaacLab\logs\rl_games\humanoid_direct
Setting seed: 42
[INFO]: Base environment:
Environment device : cuda:0
Environment seed : 42
Physics step-size : 0.008333333333333333
Rendering step-size : 0.016666666666666666
Environment step-size : 0.016666666666666666
[INFO]: Time taken for scene creation : 2.677555 seconds
[INFO]: Scene manager:
Number of environments: 2
Environment spacing : 4.0
Source prim name : /World/envs/env_0
Global prim paths : [‘/World/ground’]
Replicate physics : True
[INFO]: Starting the simulation. This may take a few seconds. Please wait…
[INFO]: Time taken for simulation start : 0.540319 seconds
[INFO]: Completed setting up the environment…
self.seed = 42
Started to train
Exact experiment name requested from command line: 2024-11-21_20-14-19
seq_length: 4
current training device: cuda:0
Error executing job with overrides:
Traceback (most recent call last):
File “D:\projects\IsaacLab\source\standalone\workflows\rl_games\train.py”, line 178, in
main()
File “D:\projects\IsaacLab\source\extensions\omni.isaac.lab_tasks\omni\isaac\lab_tasks\utils\hydra.py”, line 102, in wrapper
hydra_main()
File “C:\Users\17985.conda\envs\isaaclab\lib\site-packages\hydra\main.py”, line 94, in decorated_main
_run_hydra(
File “C:\Users\17985.conda\envs\isaaclab\lib\site-packages\hydra_internal\utils.py”, line 394, in _run_hydra
_run_app(
File “C:\Users\17985.conda\envs\isaaclab\lib\site-packages\hydra_internal\utils.py”, line 457, in _run_app
run_and_report(
File “C:\Users\17985.conda\envs\isaaclab\lib\site-packages\hydra_internal\utils.py”, line 223, in run_and_report
raise ex
File “C:\Users\17985.conda\envs\isaaclab\lib\site-packages\hydra_internal\utils.py”, line 220, in run_and_report
return func()
File “C:\Users\17985.conda\envs\isaaclab\lib\site-packages\hydra_internal\utils.py”, line 458, in
lambda: hydra.run(
File “C:\Users\17985.conda\envs\isaaclab\lib\site-packages\hydra_internal\hydra.py”, line 132, in run
_ = ret.return_value
File “C:\Users\17985.conda\envs\isaaclab\lib\site-packages\hydra\core\utils.py”, line 260, in return_value
raise self._return_value
File “C:\Users\17985.conda\envs\isaaclab\lib\site-packages\hydra\core\utils.py”, line 186, in run_job
ret.return_value = task_function(task_cfg)
File “D:\projects\IsaacLab\source\extensions\omni.isaac.lab_tasks\omni\isaac\lab_tasks\utils\hydra.py”, line 99, in hydra_main
func(env_cfg, agent_cfg, *args, **kwargs)
File “D:\projects\IsaacLab\source\standalone\workflows\rl_games\train.py”, line 170, in main
runner.run({“train”: True, “play”: False, “sigma”: train_sigma})
File “C:\Users\17985.conda\envs\isaaclab\lib\site-packages\rl_games\torch_runner.py”, line 133, in run
self.run_train(args)
File “C:\Users\17985.conda\envs\isaaclab\lib\site-packages\rl_games\torch_runner.py”, line 113, in run_train
agent = self.algo_factory.create(self.algo_name, base_name=‘run’, params=self.params)
File “C:\Users\17985.conda\envs\isaaclab\lib\site-packages\rl_games\common\object_factory.py”, line 15, in create
return builder(**kwargs)
File “C:\Users\17985.conda\envs\isaaclab\lib\site-packages\rl_games\torch_runner.py”, line 37, in
self.algo_factory.register_builder(‘a2c_continuous’, lambda **kwargs : a2c_continuous.A2CAgent(**kwargs))
File “C:\Users\17985.conda\envs\isaaclab\lib\site-packages\rl_games\algos_torch\a2c_continuous.py”, line 15, in init
a2c_common.ContinuousA2CBase.init(self, base_name, params)
File “C:\Users\17985.conda\envs\isaaclab\lib\site-packages\rl_games\common\a2c_common.py”, line 1143, in init
A2CBase.init(self, base_name, params)
File “C:\Users\17985.conda\envs\isaaclab\lib\site-packages\rl_games\common\a2c_common.py”, line 249, in init
assert(self.batch_size % self.minibatch_size == 0)
AssertionError
2024-11-21 12:14:22 [10,280ms] [Warning] [carb] Recursive unloadAllPlugins() detected!

I have transfered the assets to local server and the Unitree h1.usd model can be successfully loaded in demo script.
So I think it may not be caused by model file broken or bad network.
But the simulation runs normally when switch to another task with the same train.py script:
python source/standalone/workflows/rl_games/train.py --task Isaac-Humanoid-Direct-v0.
I want to know where the error happens.

My laptop is running on windows 11, mounted with NVIDIA GeForce RTX 3060 GPU (VRAM 6G), CPU 12th Gen Intel(R) Core™ i9-12900H x 14 (RAM 64G), Nvidia driver version 560.94, and Cuda version 12.6.
I appreciate much your help.

Thank you.

Are you following Modifying an existing Direct RL Environment when you encounter the issue? Why don’t you run with the isaaclab.sh script?

Which CUDA 12.6 update number are you using? You may need to upgrade your driver. Please review this release doc. If you still face this problem, please post an issue in Isaac Lab’s repository. Thanks.

Yes, I follow that tutorial. It is running on windows, so I use isaaclab.bat:
./isaaclab.bat -p source/standalone/workflows/rl_games/train.py --task Isaac-H1-Direct-v0 --headless
But it still doesn’t work.

I download the cuda 12.6 update 3 toolkit from CUDA Toolkit 12.6 Update 3 Downloads | NVIDIA Developer, and the driver version is now 561.17. The problem still comes. Thank you after all.

Thank you for trying this. This issue has been addressed here.