Dear Lirui:

I meet some questions when to reproduce your code, could you help to provide some valuable advice?

1. When I run the code bash ./experiments/scripts/train_online_save_buffer.sh bc_save_data.yaml BC for Save Data and Offline Training. I meet the error: ValueError: After taking into account object store and redis memory usage, the amount of memory on this node available for tasks and actors (-11.76 GB) is less than -34% of total. You can adjust these settings with ray.init(memory=<bytes>, object_store_memory=<bytes>).

2. Furthermore, I meet the same error to run the code Online Training and Testing.

3. Finally, I want to known it is right in your code bash ./experiments/scripts/train_online_visdom.sh td3_critic_aux_policy_aux.yaml DDPG of train_online_visdom.sh to be train_online.sh or train_online_continue.sh? Because I cannot find the train_online_visdom.sh in your published code.

Hi Lirui,

Thanks again for this awesome work. Is there any chance that the full shapenet training data can be made available? I assume these are the objects listed in experiments/object_index/filter_shapenet.json.

I tried following the README instructions for processing the shapes data myself but have run into various issues when trying to use OMG-Planner's python -m real_world.process_shape. For example, I'm having a little difficulty getting meshlabserver installed and running properly on my machine.

Hello! When I tried to run the pretrained model demo，I met some problems. I have downloaded the file folder shared_data and demo_model. However, when I run the test procedure, a error called URDF file '/root/GA-DDPG/env/../data/objects/Desktop_94145560501fa53228212dd5e8de73b/model_normalized.urdf' not found. I have checked the whole files contained in the data and there are not such a file.

Dear Liruiw: I have a question about run GA-DDPG in real world. You used the depth heuristic in test_realworld_ros_final.py, but I can't find the function self.graspnet.compute_grasps_score() in the function that calculates the quality of the grasp in termination_heuristics(). Can I use the depth_termination_heuristics(depth_img, mask_img) function in utils to calculate whether the grasp is possible or not?

Hi Lirui,

Thanks for publishing this wonderful work!

I'm interested in behavior cloning in this environment, from images (not pointclouds). It seems like the data from here does not include image observations. Is there a way to regenerate the expert demonstrations, and record image observations?

Thanks, Allan

Dear Lirui: Thanks for sharing so nice work. Could you mind providing more details how to run your GA-PPDG in real world? Specifically, how to do the steps to transfer GA-DDPG in Pybullet to real Franka robot?

Dear Liruiw: I'm confused about how to save video after test.

Actually, in the file folder output_misc, there are some dictionary that are created after running test program.

But when I terminate the program, nothing had been saved. Should I need to type some order in Ipython order line?

Hi Lirui, Thanks for publishing this excellent work. I'm trying to run in real life with realsense similar to your videos, but it hasn't been possible because when I tried to install Pointnet++ i got the error: The detected CUDA version (11.4) mismatches the version that was used to compile PyTorch (10.2). And my GPU supports CUDA 11 and up. Have you thought about migrating it to a newer version? Or do you think that the current version can be run with CUDA 11?

Dear Liruiw: I have a question about model training. I would like to know the idea of model training is to train offline first, using behavioral clones to learn the strategy of omg planning, and then train online, using the trained model to interact with the environment and fine-tune the strategy with hindsight goals

While running bash experiments/scripts/test_demo.sh

I'm getting Segmentation fault error. Could you please suggest the cause of such error.

Below is the ran script for reference:

(gaddpg) akashsingh@MSI:~/GA-DDPG$ bash experiments/scripts/test_demo.sh

• set -e +++ dirname -- experiments/scripts/test_demo.sh ++ cd experiments/scripts ++ pwd

• DIR=/home/akashsingh/GA-DDPG/experiments/scripts

• export PYTHONUNBUFFERED=True

• PYTHONUNBUFFERED=True

• LOG_NAME=agent

• LOG=output//log.txt

• MODEL_NAME=dummy

• RUN_NUM=3

• EPI_NUM=165

• EPOCH=latest ++ date +%Y-%m-%d_%H-%M-%S

• LOG=outputs/dummy/test_log.txt.2021-10-07_16-19-54

• exec ++ tee -a outputs/dummy/test_log.txt.2021-10-07_16-19-54 tee: outputs/dummy/test_log.txt.2021-10-07_16-19-54: No such file or directory

• echo Logging output to outputs/dummy/test_log.txt.2021-10-07_16-19-54 Logging output to outputs/dummy/test_log.txt.2021-10-07_16-19-54

• python -m core.train_test_offline --expert --pretrained output/demo_model --test --render pybullet build time: May 8 2021 05:48:32 Output will be saved to output/demo_model Using config: {'DATA_ROOT_DIR': 'data/scenes', 'EPOCHS': 200, 'EXPERIMENT_OBJ_INDEX_DIR': 'experiments/object_index', 'IMG_SIZE': [112, 112], 'LOG': True, 'MODEL_SPEC_DIR': 'experiments/model_spec', 'OBJECT_DATA_DIR': 'data/objects', 'OFFLINE_BATCH_SIZE': 100, 'OFFLINE_RL_MEMORY_SIZE': 100000, 'ONPOLICY_MEMORY_SIZE': -1, 'OUTPUT_DIR': 'output', 'OUTPUT_MISC_DIR': 'output_misc', 'RL_DATA_ROOT_DIR': 'data/scenes', 'RL_IMG_SIZE': [112, 112], 'RL_MAX_STEP': 20, 'RL_MEMORY_SIZE': 2000000, 'RL_MODEL_SPEC': 'output/demo_model/rl_pointnet_model_spec.yaml', 'RL_SAVE_DATA_NAME': 'data_50k.npz', 'RL_SAVE_DATA_ROOT_DIR': 'data', 'RL_TEST_SCENE': 'data/gaddpg_scenes', 'RL_TRAIN': {'DAGGER_MAX_STEP': 18, 'DAGGER_MIN_STEP': 5, 'DAGGER_RATIO': 0.5, 'DART_MAX_STEP': 18, 'DART_MIN_STEP': 5, 'DART_RATIO': 0.5, 'ENV_FAR': 0.5, 'ENV_NEAR': 0.2, 'ENV_RESET_TRIALS': 7, 'EXPERT_INIT_MAX_STEP': 15, 'EXPERT_INIT_MIN_STEP': 3, 'RL': True, 'SAVE_EPISODE_INTERVAL': 50, 'accumulate_points': True, 'action_noise': 0.01, 'batch_size': 125, 'bc_reward_flag': False, 'buffer_full_size': -1, 'buffer_start_idx': 0, 'change_dynamics': False, 'channel_num': 5, 'clip_grad': 0.5, 'concat_option': 'point_wise', 'critic_aux': True, 'critic_extra_latent': -1, 'critic_goal': False, 'dagger': True, 'dart': True, 'ddpg_coefficients': [0.5, 0.001, 1.0003, 1.0, 0.2], 'domain_randomization': False, 'env_name': 'PandaYCBEnv', 'env_num_objs': 1, 'expert_initial_state': False, 'explore_cap': 1.0, 'explore_ratio': 1.0, 'explore_ratio_list': [0.7, 0.7, 0.7, 0.7, 0.7, 0.7, 0.7], 'feature_input_dim': 512, 'fill_data_step': 10, 'fix_timestep_test': True, 'gamma': 0.95, 'goal_reward_flag': False, 'head_lr': 0.0003, 'hidden_size': 256, 'index_file': 'experiments/object_index/extra_shape.json', 'index_split': 'train', 'init_distance_high': 0.45, 'init_distance_low': 0.15, 'load_buffer': False, 'load_obj_num': 40, 'load_scene_joint': False, 'load_test_scene_new': False, 'log': True, 'lr': 0.0003, 'lr_gamma': 0.5, 'max_epoch': 150000, 'max_num_pts': 20000, 'mix_milestones': [4000, 8000, 16000, 25000, 35000, 45000, 65000, 85000, 100000, 120000], 'mix_policy_ratio_list': [0.1, 0.2, 0.2, 0.2, 0.2, 0.2, 0.2], 'mix_value_ratio_list': [1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0], 'new_scene': True, 'noise_ratio_list': [3.0, 2.5, 2.0, 1.5, 1.2, 1.2, 1, 0.8, 0.5], 'noise_type': 'uniform', 'num_remotes': 8, 'off_policy': True, 'online_buffer_ratio': 0.7, 'onpolicy': True, 'overwrite_feat_milestone': [], 'policy_aux': True, 'policy_extra_latent': -1, 'policy_goal': False, 'policy_milestones': [20000, 40000, 60000, 80000], 'policy_update_gap': 2, 'pt_accumulate_ratio': 0.95, 'refill_buffer': True, 'reinit_factor': 3, 'reinit_lr': 0.0001, 'reinit_optim': False, 'sa_channel_concat': True, 'save_epoch': [3000, 10000, 20000, 40000, 80000, 100000, 140000, 180000, 200000], 'self_supervision': False, 'shared_feature': False, 'shared_objects_across_worker': False, 'target_update_interval': 3000, 'tau': 0.0001, 'train_feature': True, 'train_goal_feature': False, 'train_value_feature': True, 'uniform_num_pts': 1024, 'updates_per_step': 20, 'use_action_limit': True, 'use_expert_plan': False, 'use_image': False, 'use_point_state': True, 'use_time': True, 'value_lr': 0.0003, 'value_lr_gamma': 0.5, 'value_milestones': [20000, 40000, 60000, 80000], 'value_model': True, 'visdom': True}, 'RNG_SEED': 3, 'ROBOT_DATA_DIR': 'data/robots', 'ROOT_DIR': '/home/akashsingh/GA-DDPG/experiments/../', 'SCRIPT_FOLDER': 'experiments/cfgs', 'env_config': {'accumulate_points': True, 'action_space': 'task6d', 'change_dynamics': False, 'data_type': 'RGBDM', 'domain_randomization': False, 'expert_step': 20, 'height': 112, 'img_resize': [112, 112], 'initial_far': 0.5, 'initial_near': 0.2, 'numObjects': 1, 'omg_config': {'allow_collision_point': 0, 'base_obstacle_weight': 1.0, 'clearance': 0.03, 'dynamic_timestep': False, 'extra_smooth_steps': 5, 'goal_idx': -1, 'ik_clearance': 0.07, 'ik_parallel': False, 'ik_seed_num': 13, 'increment_iks': True, 'ol_alg': 'Proj', 'optim_steps': 1, 'pre_terminate': True, 'root_dir': '/home/akashsingh/GA-DDPG/experiments/../', 'scene_file': '', 'silent': True, 'smoothness_base_weight': 3, 'standoff_dist': 0.08, 'target_epsilon': 0.06, 'target_hand_filter_angle': 90, 'target_obj_collision': 1, 'terminate_smooth_loss': 3, 'timesteps': 20, 'traj_delta': 0.05, 'traj_init': 'grasp', 'traj_interpolate': 'linear', 'traj_max_step': 26, 'traj_min_step': 15, 'use_expert_plan': False, 'vis': False}, 'pt_accumulate_ratio': 0.95, 'random_target': True, 'regularize_pc_point_count': True, 'uniform_num_pts': 1024, 'use_hand_finger_point': True, 'width': 112}, 'omg_config': {'allow_collision_point': 0, 'base_obstacle_weight': 1.0, 'clearance': 0.03, 'dynamic_timestep': False, 'extra_smooth_steps': 5, 'goal_idx': -1, 'ik_clearance': 0.07, 'ik_parallel': False, 'ik_seed_num': 13, 'increment_iks': True, 'ol_alg': 'Proj', 'optim_steps': 1, 'pre_terminate': True, 'root_dir': '/home/akashsingh/GA-DDPG/experiments/../', 'scene_file': '', 'silent': True, 'smoothness_base_weight': 3, 'standoff_dist': 0.08, 'target_epsilon': 0.06, 'target_hand_filter_angle': 90, 'target_obj_collision': 1, 'terminate_smooth_loss': 3, 'timesteps': 20, 'traj_delta': 0.05, 'traj_init': 'grasp', 'traj_interpolate': 'linear', 'traj_max_step': 26, 'traj_min_step': 15, 'use_expert_plan': False, 'vis': False}, 'pretrained_time': '', 'script_name': 'td3_critic_aux_policy_aux.yaml'} schedule: [8000, 16000, 30000, 50000, 70000, 90000] schedule: [8000, 16000, 30000, 50000, 70000, 90000] Output will be saved to output/demo_model video output: YCB_td3_critic_aux_policy_aux.yaml stat output: rollout_success.script_td3_critic_aux_policy_aux.yaml.txt load pretrained policy!!!! load pretrained critic!!!! load feat optim load pretrained feature!!!! from: output/demo_model/DDPG_state_feat_PandaYCBEnv_latest step :249921 output_time: demo_model logdir: output/demo_model/PandaYCBEnv_DDPG argv[0]=--opengl2 startThreads creating 1 threads. starting thread 0 started thread 0 argc=3 argv[0] = --unused argv[1] = --opengl2 argv[2] = --start_demo_name=Physics Server ExampleBrowserThreadFunc started X11 functions dynamically loaded using dlopen/dlsym OK! X11 functions dynamically loaded using dlopen/dlsym OK! 4

    visual 0x21 selected
  

GL_VENDOR=Intel GL_RENDERER=Intel(R) UHD Graphics 630 GL_VERSION=1.4 (4.6.0 - Build 30.0.100.9922) GL_SHADING_LANGUAGE_VERSION=(null) pthread_getconcurrency()=0 b3Printf: Selected demo: Physics Server startThreads creating 1 threads. starting thread 0 started thread 0 MotionThreadFunc thread started experiments/scripts/test_demo.sh: line 19: 3313 Segmentation fault python -m core.train_test_offline --expert --pretrained output/demo_model --test --render

Hello, It seems that there's a problem in the state_feature_extractor in the agent.py file. Could you please see the following errors when executing the 'test_demo' script? Best regards, Ronen.

+ echo Logging output to outputs/dummy/test_log.txt.2021-07-20_10-26-24
Logging output to outputs/dummy/test_log.txt.2021-07-20_10-26-24
+ python -m core.train_test_offline --expert --pretrained output/demo_model --test --render
pybullet build time: Jul 14 2021 10:11:29
Output will be saved to `output/demo_model`
Using config:
{'DATA_ROOT_DIR': 'data/scenes',
 'EPOCHS': 200,
 'EXPERIMENT_OBJ_INDEX_DIR': 'experiments/object_index',
 'IMG_SIZE': [112, 112],
 'LOG': True,
 'MODEL_SPEC_DIR': 'experiments/model_spec',
 'OBJECT_DATA_DIR': 'data/objects',
 'OFFLINE_BATCH_SIZE': 100,
 'OFFLINE_RL_MEMORY_SIZE': 100000,
 'ONPOLICY_MEMORY_SIZE': -1,
 'OUTPUT_DIR': 'output',
 'OUTPUT_MISC_DIR': 'output_misc',
 'RL_DATA_ROOT_DIR': 'data/scenes',
 'RL_IMG_SIZE': [112, 112],
 'RL_MAX_STEP': 20,
 'RL_MEMORY_SIZE': 2000000,
 'RL_MODEL_SPEC': 'output/demo_model/rl_pointnet_model_spec.yaml',
 'RL_SAVE_DATA_NAME': 'data_50k.npz',
 'RL_SAVE_DATA_ROOT_DIR': 'data',
 'RL_TEST_SCENE': 'data/gaddpg_scenes',
 'RL_TRAIN': {'DAGGER_MAX_STEP': 18,
              'DAGGER_MIN_STEP': 5,
              'DAGGER_RATIO': 0.5,
              'DART_MAX_STEP': 18,
              'DART_MIN_STEP': 5,
              'DART_RATIO': 0.5,
              'ENV_FAR': 0.5,
              'ENV_NEAR': 0.2,
              'ENV_RESET_TRIALS': 7,
              'EXPERT_INIT_MAX_STEP': 15,
              'EXPERT_INIT_MIN_STEP': 3,
              'RL': True,
              'SAVE_EPISODE_INTERVAL': 50,
              'accumulate_points': True,
              'action_noise': 0.01,
              'batch_size': 125,
              'bc_reward_flag': False,
              'buffer_full_size': -1,
              'buffer_start_idx': 0,
              'change_dynamics': False,
              'channel_num': 5,
              'clip_grad': 0.5,
              'concat_option': 'point_wise',
              'critic_aux': True,
              'critic_extra_latent': -1,
              'critic_goal': False,
              'dagger': True,
              'dart': True,
              'ddpg_coefficients': [0.5, 0.001, 1.0003, 1.0, 0.2],
              'domain_randomization': False,
              'env_name': 'PandaYCBEnv',
              'env_num_objs': 1,
              'expert_initial_state': False,
              'explore_cap': 1.0,
              'explore_ratio': 1.0,
              'explore_ratio_list': [0.7, 0.7, 0.7, 0.7, 0.7, 0.7, 0.7],
              'feature_input_dim': 512,
              'fill_data_step': 10,
              'fix_timestep_test': True,
              'gamma': 0.95,
              'goal_reward_flag': False,
              'head_lr': 0.0003,
              'hidden_size': 256,
              'index_file': 'experiments/object_index/extra_shape.json',
              'index_split': 'train',
              'init_distance_high': 0.45,
              'init_distance_low': 0.15,
              'load_buffer': False,
              'load_obj_num': 40,
              'load_scene_joint': False,
              'load_test_scene_new': False,
              'log': True,
              'lr': 0.0003,
              'lr_gamma': 0.5,
              'max_epoch': 150000,
              'max_num_pts': 20000,
              'mix_milestones': [4000,
                                 8000,
                                 16000,
                                 25000,
                                 35000,
                                 45000,
                                 65000,
                                 85000,
                                 100000,
                                 120000],
              'mix_policy_ratio_list': [0.1, 0.2, 0.2, 0.2, 0.2, 0.2, 0.2],
              'mix_value_ratio_list': [1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0],
              'new_scene': True,
              'noise_ratio_list': [3.0, 2.5, 2.0, 1.5, 1.2, 1.2, 1, 0.8, 0.5],
              'noise_type': 'uniform',
              'num_remotes': 8,
              'off_policy': True,
              'online_buffer_ratio': 0.7,
              'onpolicy': True,
              'overwrite_feat_milestone': [],
              'policy_aux': True,
              'policy_extra_latent': -1,
              'policy_goal': False,
              'policy_milestones': [20000, 40000, 60000, 80000],
              'policy_update_gap': 2,
              'pt_accumulate_ratio': 0.95,
              'refill_buffer': True,
              'reinit_factor': 3,
              'reinit_lr': 0.0001,
              'reinit_optim': False,
              'sa_channel_concat': True,
              'save_epoch': [3000,
                             10000,
                             20000,
                             40000,
                             80000,
                             100000,
                             140000,
                             180000,
                             200000],
              'self_supervision': False,
              'shared_feature': False,
              'shared_objects_across_worker': False,
              'target_update_interval': 3000,
              'tau': 0.0001,
              'train_feature': True,
              'train_goal_feature': False,
              'train_value_feature': True,
              'uniform_num_pts': 1024,
              'updates_per_step': 20,
              'use_action_limit': True,
              'use_expert_plan': False,
              'use_image': False,
              'use_point_state': True,
              'use_time': True,
              'value_lr': 0.0003,
              'value_lr_gamma': 0.5,
              'value_milestones': [20000, 40000, 60000, 80000],
              'value_model': True,
              'visdom': True},
 'RNG_SEED': 3,
 'ROBOT_DATA_DIR': 'data/robots',
 'ROOT_DIR': '/home/nir1tv/projects/GA-DDPG/experiments/../',
 'SCRIPT_FOLDER': 'experiments/cfgs',
 'env_config': {'accumulate_points': True,
                'action_space': 'task6d',
                'change_dynamics': False,
                'data_type': 'RGBDM',
                'domain_randomization': False,
                'expert_step': 20,
                'height': 112,
                'img_resize': [112, 112],
                'initial_far': 0.5,
                'initial_near': 0.2,
                'numObjects': 1,
                'omg_config': {'allow_collision_point': 0,
                               'base_obstacle_weight': 1.0,
                               'clearance': 0.03,
                               'dynamic_timestep': False,
                               'extra_smooth_steps': 5,
                               'goal_idx': -1,
                               'ik_clearance': 0.07,
                               'ik_parallel': False,
                               'ik_seed_num': 13,
                               'increment_iks': True,
                               'ol_alg': 'Proj',
                               'optim_steps': 1,
                               'pre_terminate': True,
                               'root_dir': '/home/nir1tv/projects/GA-DDPG/experiments/../',
                               'scene_file': '',
                               'silent': True,
                               'smoothness_base_weight': 3,
                               'standoff_dist': 0.08,
                               'target_epsilon': 0.06,
                               'target_hand_filter_angle': 90,
                               'target_obj_collision': 1,
                               'terminate_smooth_loss': 3,
                               'timesteps': 20,
                               'traj_delta': 0.05,
                               'traj_init': 'grasp',
                               'traj_interpolate': 'linear',
                               'traj_max_step': 26,
                               'traj_min_step': 15,
                               'use_expert_plan': False,
                               'vis': False},
                'pt_accumulate_ratio': 0.95,
                'random_target': True,
                'regularize_pc_point_count': True,
                'uniform_num_pts': 1024,
                'use_hand_finger_point': True,
                'width': 112},
 'omg_config': {'allow_collision_point': 0,
                'base_obstacle_weight': 1.0,
                'clearance': 0.03,
                'dynamic_timestep': False,
                'extra_smooth_steps': 5,
                'goal_idx': -1,
                'ik_clearance': 0.07,
                'ik_parallel': False,
                'ik_seed_num': 13,
                'increment_iks': True,
                'ol_alg': 'Proj',
                'optim_steps': 1,
                'pre_terminate': True,
                'root_dir': '/home/nir1tv/projects/GA-DDPG/experiments/../',
                'scene_file': '',
                'silent': True,
                'smoothness_base_weight': 3,
                'standoff_dist': 0.08,
                'target_epsilon': 0.06,
                'target_hand_filter_angle': 90,
                'target_obj_collision': 1,
                'terminate_smooth_loss': 3,
                'timesteps': 20,
                'traj_delta': 0.05,
                'traj_init': 'grasp',
                'traj_interpolate': 'linear',
                'traj_max_step': 26,
                'traj_min_step': 15,
                'use_expert_plan': False,
                'vis': False},
 'pretrained_time': '',
 'script_name': 'td3_critic_aux_policy_aux.yaml'}
Let's use 2 GPUs!
schedule: [8000, 16000, 30000, 50000, 70000, 90000]
schedule: [8000, 16000, 30000, 50000, 70000, 90000]
Output will be saved to `output/demo_model`
video output: YCB_td3_critic_aux_policy_aux.yaml stat output: rollout_success.script_td3_critic_aux_policy_aux.yaml.txt
load pretrained policy!!!!
load pretrained critic!!!!
load feat optim
load pretrained feature!!!! from: output/demo_model/DDPG_state_feat_PandaYCBEnv_latest step :249921
output_time: demo_model logdir: output/demo_model/PandaYCBEnv_DDPG
startThreads creating 1 threads.
starting thread 0
started thread 0 
argc=2
argv[0] = --unused
argv[1] = --start_demo_name=Physics Server
ExampleBrowserThreadFunc started
X11 functions dynamically loaded using dlopen/dlsym OK!
X11 functions dynamically loaded using dlopen/dlsym OK!
Creating context
Created GL 3.3 context
Direct GLX rendering context obtained
Making context current
GL_VENDOR=NVIDIA Corporation
GL_RENDERER=NVIDIA GeForce RTX 2080 Ti/PCIe/SSE2
GL_VERSION=3.3.0 NVIDIA 470.42.01
GL_SHADING_LANGUAGE_VERSION=3.30 NVIDIA via Cg compiler
pthread_getconcurrency()=0
Version = 3.3.0 NVIDIA 470.42.01
Vendor = NVIDIA Corporation
Renderer = NVIDIA GeForce RTX 2080 Ti/PCIe/SSE2
b3Printf: Selected demo: Physics Server
startThreads creating 1 threads.
starting thread 0
started thread 0 
MotionThreadFunc thread started
numActiveThreads = 0
stopping threads
destroy semaphore
semaphore destroyed
Thread with taskId 0 exiting
Thread TERMINATED
destroy main semaphore
main semaphore destroyed
finished
numActiveThreads = 0
btShutDownExampleBrowser stopping threads
Thread with taskId 0 exiting
Thread TERMINATED
destroy semaphore
semaphore destroyed
destroy main semaphore
main semaphore destroyed
startThreads creating 1 threads.
starting thread 0
started thread 0 
argc=2
argv[0] = --unused
argv[1] = --start_demo_name=Physics Server
ExampleBrowserThreadFunc started
X11 functions dynamically loaded using dlopen/dlsym OK!
X11 functions dynamically loaded using dlopen/dlsym OK!
Creating context
Created GL 3.3 context
Direct GLX rendering context obtained
Making context current
GL_VENDOR=NVIDIA Corporation
GL_RENDERER=NVIDIA GeForce RTX 2080 Ti/PCIe/SSE2
GL_VERSION=3.3.0 NVIDIA 470.42.01
GL_SHADING_LANGUAGE_VERSION=3.30 NVIDIA via Cg compiler
pthread_getconcurrency()=0
Version = 3.3.0 NVIDIA 470.42.01
Vendor = NVIDIA Corporation
Renderer = NVIDIA GeForce RTX 2080 Ti/PCIe/SSE2
b3Printf: Selected demo: Physics Server
startThreads creating 1 threads.
starting thread 0
started thread 0 
MotionThreadFunc thread started
ven = NVIDIA Corporation
ven = NVIDIA Corporation
/home/nir1tv/projects/GA-DDPG/env/models
>>>> target name: 061_foam_brick
==== loaded scene: scene_0 target: 006_mustard_bottle idx: 5 init joint
Traceback (most recent call last):
  File "/home/nir1tv/miniconda3/envs/gaddpg/lib/python3.6/runpy.py", line 193, in _run_module_as_main
    "__main__", mod_spec)
  File "/home/nir1tv/miniconda3/envs/gaddpg/lib/python3.6/runpy.py", line 85, in _run_code
    exec(code, run_globals)
  File "/home/nir1tv/projects/GA-DDPG/core/train_test_offline.py", line 387, in <module>
    test(run_iter=run_iter)
  File "/home/nir1tv/projects/GA-DDPG/core/train_test_offline.py", line 230, in test
    action, _, _, aux_pred = agent.select_action(state, vis=False, remain_timestep=remain_timestep )
  File "/home/nir1tv/miniconda3/envs/gaddpg/lib/python3.6/site-packages/torch/autograd/grad_mode.py", line 49, in decorate_no_grad
    return func(*args, **kwargs)
  File "/home/nir1tv/projects/GA-DDPG/core/agent.py", line 113, in select_action
    train=False,
  File "/home/nir1tv/projects/GA-DDPG/core/ddpg.py", line 54, in extract_feature
    train=train)
  File "/home/nir1tv/projects/GA-DDPG/core/agent.py", line 77, in unpack_batch
    train=train,
  File "/home/nir1tv/miniconda3/envs/gaddpg/lib/python3.6/site-packages/torch/nn/modules/module.py", line 532, in __call__
    result = self.forward(*input, **kwargs)
  File "/home/nir1tv/miniconda3/envs/gaddpg/lib/python3.6/site-packages/torch/nn/parallel/data_parallel.py", line 152, in forward
    outputs = self.parallel_apply(replicas, inputs, kwargs)
  File "/home/nir1tv/miniconda3/envs/gaddpg/lib/python3.6/site-packages/torch/nn/parallel/data_parallel.py", line 162, in parallel_apply
    return parallel_apply(replicas, inputs, kwargs, self.device_ids[:len(replicas)])
  File "/home/nir1tv/miniconda3/envs/gaddpg/lib/python3.6/site-packages/torch/nn/parallel/parallel_apply.py", line 85, in parallel_apply
    output.reraise()
  File "/home/nir1tv/miniconda3/envs/gaddpg/lib/python3.6/site-packages/torch/_utils.py", line 394, in reraise
    raise self.exc_type(msg)
TypeError: Caught TypeError in replica 1 on device 1.
Original Traceback (most recent call last):
  File "/home/nir1tv/miniconda3/envs/gaddpg/lib/python3.6/site-packages/torch/nn/parallel/parallel_apply.py", line 60, in _worker
    output = module(*input, **kwargs)
  File "/home/nir1tv/miniconda3/envs/gaddpg/lib/python3.6/site-packages/torch/nn/modules/module.py", line 532, in __call__
    result = self.forward(*input, **kwargs)
TypeError: forward() missing 1 required positional argument: 'pc'

numActiveThreads = 0
stopping threads
Thread with taskId 0 exiting
Thread TERMINATED
destroy semaphore
semaphore destroyed
destroy main semaphore
main semaphore destroyed
finished
numActiveThreads = 0
btShutDownExampleBrowser stopping threads
destroy semaphore
semaphore destroyed
Thread with taskId 0 exiting
Thread TERMINATED
destroy main semaphore
main semaphore destroyed