Segmentation fault (core dumped)

[zcswdt]

When I was running the training code, the following error occurred，may I ask how to solve this

python train.py

Training with GPU
Created progress directory to: /remote-home/2210975/cloth/dynamic-cloth-folding/trainings/default-run-0
Popen(['git', 'init'], cwd=/remote-home/2210975/cloth/dynamic-cloth-folding/submodules/mujoco-py, universal_newlines=False, shell=None, istream=None)
Popen(['git', 'cat-file', '--batch-check'], cwd=/remote-home/2210975/cloth/dynamic-cloth-folding/submodules/mujoco-py, universal_newlines=False, shell=None, istream=)
Popen(['git', 'cat-file', '--batch'], cwd=/remote-home/2210975/cloth/dynamic-cloth-folding/submodules/mujoco-py, universal_newlines=False, shell=None, istream=)
Popen(['git', 'init'], cwd=/remote-home/2210975/cloth/dynamic-cloth-folding/submodules/rlkit, universal_newlines=False, shell=None, istream=None)
Popen(['git', 'cat-file', '--batch-check'], cwd=/remote-home/2210975/cloth/dynamic-cloth-folding/submodules/rlkit, universal_newlines=False, shell=None, istream=)
Popen(['git', 'cat-file', '--batch'], cwd=/remote-home/2210975/cloth/dynamic-cloth-folding/submodules/rlkit, universal_newlines=False, shell=None, istream=)
Popen(['git', 'init'], cwd=/remote-home/2210975/cloth/dynamic-cloth-folding/submodules/robosuite, universal_newlines=False, shell=None, istream=None)
Popen(['git', 'cat-file', '--batch-check'], cwd=/remote-home/2210975/cloth/dynamic-cloth-folding/submodules/robosuite, universal_newlines=False, shell=None, istream=)
Popen(['git', 'cat-file', '--batch'], cwd=/remote-home/2210975/cloth/dynamic-cloth-folding/submodules/robosuite, universal_newlines=False, shell=None, istream=)
Popen(['git', 'init'], cwd=/remote-home/2210975/cloth/dynamic-cloth-folding, universal_newlines=False, shell=None, istream=None)
Popen(['git', 'cat-file', '--batch-check'], cwd=/remote-home/2210975/cloth/dynamic-cloth-folding, universal_newlines=False, shell=None, istream=)
Popen(['git', 'cat-file', '--batch'], cwd=/remote-home/2210975/cloth/dynamic-cloth-folding, universal_newlines=False, shell=None, istream=)
2023-06-13 02:32:10.319252 UTC | Variant:
2023-06-13 02:32:10.320003 UTC | {
"algorithm": "SAC",
"title": "default-run-0",
"save_folder": "/remote-home/2210975/cloth/dynamic-cloth-folding/trainings/default-run-0",
"random_seed": 0,
"randomization_kwargs": {
"lights_randomization": true,
"materials_randomization": true,
"camera_position_randomization": true,
"lookat_position_randomization": true,
"lookat_position_randomization_radius": 0.03,
"dynamics_randomization": true,
"albumentations_randomization": true,
"cloth_size": 0.2,
"camera_type": "side",
"camera_config": {
"type": "small",
"fovy_range": [
13,
15
],
"height": 100,
"width": 848
},
"position_perturbation_size": 0.2,
"rotation_perturbation_size": 0.75,
"fovy_perturbation_size": 0.05
},
"value_function_kwargs": {
"fc_layer_size": 512,
"fc_layer_depth": 5
},
"policy_kwargs": {
"input_width": 100,
"input_height": 100,
"input_channels": 1,
"kernel_sizes": [
3,
3,
3,
3
],
"n_channels": [
32,
32,
32,
32
],
"strides": [
2,
2,
2,
2
],
"paddings": [
0,
0,
0,
0
],
"hidden_sizes_aux": [
256,
8
],
"hidden_sizes_main": [
256,
256,
256,
256
],
"init_w": 0.0001,
"aux_output_size": 9
},
"env_kwargs": {
"save_folder": "/remote-home/2210975/cloth/dynamic-cloth-folding/trainings/default-run-0",
"timestep": 0.01,
"success_distance": 0.05,
"robot_observation": "ctrl",
"control_frequency": 10,
"ctrl_filter": 0.03,
"kp": 1000.0,
"frame_stack_size": 1,
"damping_ratio": 1,
"success_reward": 0,
"fail_reward": -1,
"extra_reward": 1,
"output_max": 0.03,
"max_close_steps": 10,
"sparse_dense": true,
"goal_noise_range": [
0.0,
0.03
],
"image_obs_noise_mean": 0.5,
"image_obs_noise_std": 0.5,
"model_kwargs_path": "./data/model_params.csv"
},
"eval_kwargs": {
"save_images_every_epoch": 10,
"num_runs": 20,
"max_path_length": 50,
"additional_keys": "['robot_observation']",
"frame_stack_size": 1,
"save_blurred_images": true,
"save_folder": "/remote-home/2210975/cloth/dynamic-cloth-folding/trainings/default-run-0"
},
"algorithm_kwargs": {
"num_epochs": 100,
"num_trains_per_train_loop": 1000,
"num_expl_steps_per_train_loop": 1000,
"num_train_loops_per_epoch": 20,
"max_path_length": 50,
"save_policy_every_epoch": 1,
"batch_size": 256,
"num_demoers": 0,
"save_folder": "/remote-home/2210975/cloth/dynamic-cloth-folding/trainings/default-run-0"
},
"path_collector_kwargs": {
"additional_keys": "['robot_observation']",
"demo_paths": "['./data/demos.csv']",
"demo_divider": 0.03,
"num_processes": 1
},
"replay_buffer_kwargs": {
"max_size": 100000,
"fraction_goals_rollout_goals": 0.19999999999999996,
"internal_keys": "['image', 'robot_observation']"
},
"trainer_kwargs": {
"discount": 0.99,
"soft_target_tau": 0.005,
"target_update_period": 1,
"policy_lr": 0.0003,
"qf_lr": 0.0003,
"reward_scale": 1,
"use_automatic_entropy_tuning": true,
"corner_prediction_loss_coef": 0.001
}
}
Training started
Segmentation fault (core dumped)

[hietalajulius]

Looks like something is segfaulting on your system. Can you try running with e.g. gdb to see further details?

[zcswdt]

Looks like something is segfaulting on your system. Can you try running with e.g. gdb to see further details?

I have already resolved it, thank you

[anh-nn01]

Hi @zcswdt, I am having the same problem. Can you share me what you did to resolve it?

[ZoeQU]

I have the same question, Does anyone know how to solve it?

[zcswdt]

I have the same question, Does anyone know how to solve it?

yes ,i do。

[yuhangz]

I have the same question, Does anyone know how to solve it?

yes ,i do。

Could you share me how to solve it?

[ZoeQU]

I have a new wrong situation here:
after python train.py and I get this error:

UserWarning: Pydantic serializer warnings:
Expected Union[float, tuple[float, float]] but got list - serialized value may not be as expected

could anyone help me?

[hietalajulius]

@zcswdt would you be able to share how you resolved the segfault as others seem to struggle with it as well? Thanks!

[yuhangz]

I have solved this problem. You could solve it by reinstalling mujoco_py with source code.

[zcswdt]

@zcswdt would you be able to share how you resolved the segfault as others seem to struggle with it as well? Thanks!

I'm really sorry. I just saw your email recently. Later, I still couldn't run the real robot arm experiment, so I gave up the project and switched to the flingbot project. I didn't record how I solved it at that time. I'm really sorry. Thank you for your help.