(huge) weirdab@flexo:~/goalrelabel_locobot_fullgcsl$ python launch_main.py --env_name bridge_finetune --method huge
/home/weirdab/miniconda3/envs/huge/lib/python3.8/site-packages/scipy/__init__.py:146: UserWarning: A NumPy version >=1.16.5 and <1.23.0 is required for this version of SciPy (detected version 1.23.5
warnings.warn(f"A NumPy version >={np_minversion} and <{np_maxversion}"
wandb: Currently logged in as: yuancarrieyjy. Use `wandb login --relogin` to force relogin
wandb: Tracking run with wandb version 0.15.10
wandb: Run data is saved locally in /home/weirdab/goalrelabel_locobot_fullgcsl/wandb/run-20230910_212016-uf22c09t
wandb: Run `wandb offline` to turn off syncing.
wandb: Syncing run bridge_finetune_huge_1_
wandb: ⭐️ View project at https://wandb.ai/yuancarrieyjy/bridge_finetune_huge
wandb: 🚀 View run at https://wandb.ai/yuancarrieyjy/bridge_finetune_huge/runs/uf22c09t
params before run {'lr': 0.0005, 'save_buffer': False, 'load_buffer': False, 'load_goal_selector': False, 'batch_size': 100, 'goal_selector_name': '', 'select_best_sample_size': 1000, 'explore_episodes': 10, 'render': False, 'display_plots': True, 'goal_selector_num_samples': 1000, 'remove_last_steps_when_stopped': True, 'train_goal_selector_freq': 5, 'exploration_when_stopped': True, 'distance_noise_std': 0, 'eval_episodes': 5, 'save_videos': True, 'eval_freq': 50, 'task_config': 'slide_cabinet,microwave', 'epsilon_greedy_exploration': 0, 'epsilon_greedy_rollout': 1, 'select_last_k_steps': 21, 'remove_last_k_steps': 8, 'network_layers': '400,600,600,300', 'reward_layers': '400,600,600,300', 'weighted_sl': False, 'num_blocks': 3, 'random_goal': False, 'maze_type': 3, 'buffer_size': 200, 'use_horizon': False, 'sample_new_goal_freq': 10, 'k_goal': 1, 'throw_trajectories_not_reaching_goal': False, 'fourier': True, 'fourier_goal_selector': False, 'command_goal_if_too_close': False, 'display_trajectories_freq': 50, 'contrastive': False, 'deterministic_rollout': False, 'expl_noise_std': 1, 'normalize': False, 'goal_threshold': -1, 'goal_selector_epochs': 400, 'goal_selector_batch_size': 64, 'check_if_stopped': True, 'use_wrong_oracle': False, 'human_data_file': '', 'stop_training_goal_selector_after': -1, 'policy_updates_per_step': 100, 'pretrain': False, 'num_demos': 0, 'desired_goal_sampling_freq': 0, 'img_width': 256, 'img_height': 256, 'human_input': False, 'max_timesteps': 1000000, 'explore_length': 20, 'max_path_length': 50, 'stopped_thresh': 0.05, 'frontier_expansion_freq': 20, 'frontier_expansion_rate': 5, 'start_frontier': 30, 'repeat_previous_action_prob': 0, 'select_goal_from_last_k_trajectories': 20, 'pretrain_policy': False, 'pretrain_goal_selector': False, 'train_with_preferences': True, 'use_oracle': False, 'sample_softmax': True, 'continuous_action_space': True, 'use_images_in_reward_model': True, 'input_image_size': 128, 'gpu': 0, 'seed': 1, 'env_name': 'bridge_finetune', 'comment': '', 'method': 'huge', 'use_images': False, 'data_folder': 'bridge_finetune__use_oracle_1'}
2023-09-10 21:20:25.260265: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
No GPU/TPU found, falling back to CPU. (Set TF_CPP_MIN_LOG_LEVEL=0 and rerun for more info.)
Not using GPU. Will be slow.
hello
waiting for widowx to be set up...
Robot Name: wx250s
Robot Model: wx250s
Initialized InterbotixRobotXSCore!
Arm Group Name: arm
Moving Time: 2.00 seconds
Acceleration Time: 0.30 seconds
Drive Mode: Time-Based-Profile
Initialized InterbotixArmXSInterface!
Gripper Name: gripper
Gripper Pressure: 50%
Initialized InterbotixGripperXSInterface!
Trying to read camera info from /cam0/camera_info
Successfully read camera info
moving to neutral..
moving to neutral failed!
bridgeEnv: {'_hp': {'robot_name': 'wx250s', 'robot_controller': <class 'widowx_envs.widowx.src.widowx_controller.WidowX_Controller'>, 'gripper_attached': 'custom', 'camera_topics': [<robonetv3.widowx_envs.widowx_envs.utils.multicam_server_rospkg.src.topic_utils.IMTopic object at 0x7f7dc5d647c0>], 'start_at_neutral': False, 'start_at_current_pos': False, 'OFFSET_TOL': 0.1, 'lower_bound_delta': [0.0, 0.0, 0.0, 0.0, 0.0], 'upper_bound_delta': [0.0, 0.0, 0.0, 0.0, 0.0], 'print_debug': False, 'move_duration': 0.2, 'action_clipping': 'xyz', 'override_workspace_boundaries': [[0.1, -0.15, -0.1, -1.57, 0], [0.45, 0.25, 0.18, 1.57, 0]], 'resetqpos_after_every_step': False, 'absolute_grasp_action': True, 'continuous_gripper': True, 'action_mode': '3trans3rot', 'start_state': None, 'wait_time': 0, 'adaptive_wait': True, 'workspace_rotation_angle_z': 0, 'wait_until_gripper_pose_reached': False, 'catch_environment_except': False, 'gripper_params': {'des_pos_max': 1, 'des_pos_min': 0}, 'randomize_initpos': 'full_area', 'mode_rel': [True, True, True, True, True], 'start_transform': None, 'skip_move_to_neutral': True, 'move_to_rand_start_freq': 1, 'fix_zangle': 0.1, 'image_crop_xywh': None, 'return_full_image': False}, 'savedir': None, '_robot_name': 'wx250s', '_low_bound': array([ 0.1 , -0.15, -0.1 , -1.57, 0. ]), '_high_bound': array([0.45, 0.25, 0.18, 1.57, 0. ]), '_obs_tol': 0.1, '_controller': <widowx_envs.widowx.src.widowx_controller.WidowX_Controller object at 0x7f7d43dc56a0>, '_cameras': [<widowx_envs.utils.multicam_server_rospkg.src.camera_recorder.CameraRecorder object at 0x7f7d43e53520>], '_camera_info': [header:
seq: 86795
stamp:
secs: 1694405858
nsecs: 802920478
frame_id: "world"
height: 480
width: 640
distortion_model: "plumb_bob"
D: [0.0, 0.0, 0.0, 0.0, 0.0]
K: [320.0, 0.0, 320.0, 0.0, 240.0, 240.0, 0.0, 0.0, 1.0]
R: [1.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 1.0]
P: [320.0, 0.0, 320.0, 0.0, 0.0, 240.0, 240.0, 0.0, 0.0, 0.0, 1.0, 0.0]
binning_x: 0
binning_y: 0
roi:
x_offset: 0
y_offset: 0
height: 0
width: 0
do_rectify: False], '_depth_cameras': [], '_depth_camera_info': [], '_height': 480, '_width': 640, '_cam_names': ['front'], '_reset_counter': 0, '_previous_target_qpos': None, 'action_space': Box([-0.05 -0.05 -0.05 -0.25 -0.25 -0.25 0. ], [0.05 0.05 0.05 0.25 0.25 0.25 1. ], (7,), float32), '_adim': 7, '_base_adim': 7, '_sdim': 7, '_base_sdim': 7, 'time_for_get_obs': 0.08, 'image_size': 128, 'task_id': None, 'num_tasks': None, 'observation_space': Dict('desired_goal': Box(-inf, inf, (3,), float64), 'image': Box(0, 255, (49152,), uint8), 'state': Box(-inf, inf, (7,), float64)), 'move_except': False, 'reward_function': None, 'control_viewpoint': 0}
{'eval_freq': 10000, 'eval_episodes': 50, 'max_timesteps': 1000000.0, 'goal_threshold': 0.05, 'max_trajectory_length': 50, 'network_layers': '400,600,600,300', 'reward_layers': '400,600,600,300', 'buffer_size': 200, 'use_horizon': False, 'fourier': False, 'fourier_goal_selector': False, 'normalize': False, 'env_name': 'bridge_finetune', 'goal_selector_buffer_size': 8, 'input_image_size': 128, 'img_width': 256, 'img_height': 256, 'use_images_in_policy': False, 'use_images_in_reward_model': True, 'use_images_in_stopping_criteria': False, 'close_frames': 2, 'far_frames': 10}
INIT RW MODEL False [400, 600, 600, 300]
2704 1 False [400, 600, 600, 300] <class 'torch.nn.modules.activation.ReLU'>
<class 'huge.envs.env_utils.DummyWrappedEnv'>
buffer_size = 200, max_trajectory_length = 50,input_image_size = 128
*env.action_space.shape 7
*env.state_space.shape 147456
<class 'huge.envs.env_utils.DummyWrappedEnv'>
buffer_size = 200, max_trajectory_length = 50,input_image_size = 128
*env.action_space.shape 7
*env.state_space.shape 147456
violation [ 344.06005859 -1378.94999695 -800.47000122 -725.1499939
-897.30999994 -907.42999268]
motor number: 0
max effort reached: [1614. 126.43000031 94.15000153 174.8500061 13.45000076
142.57000732]
max effort allowed [1200. 1500. 900. 900. 900. 1050.]
moving to neutral..
moving to neutral failed!
{'max_path_length': 50, 'goal_threshold': 0.05, 'explore_episodes': 10, 'eval_freq': 50, 'eval_episodes': 5, 'save_every_iteration': False, 'max_timesteps': 1000000, 'expl_noise': 0.0, 'batch_size': 100, 'n_accumulations': 1, 'policy_updates_per_step': 100, 'lr': 0.0005, 'goal_selector_buffer_validation': <huge.algo.buffer.RewardModelBuffer object at 0x7f7cefeccfd0>, 'validation_buffer': <huge.algo.buffer.ReplayBuffer object at 0x7f7d43dfa760>, 'remove_last_k_steps': 8, 'select_last_k_steps': 21, 'continuous_action_space': True, 'expl_noise_std': 1, 'check_if_stopped': True, 'num_demos': 0, 'demo_epochs': 100000, 'demo_goal_selector_epochs': 1000, 'input_image_size': 64, 'use_images_in_policy': False, 'use_images_in_reward_model': True, 'classifier_model': None, 'use_images_in_stopping_criteria': False}
Use images in reward model
stop training goal selector after 1000000
Select goal from last k trajectories 20
Curr frontier beginning 30
Train policy freq is, 1
cuda device cuda:0
[WARN] [1694406065.378961]: Would exceed position limits on joint 2.
[WARN] [1694406065.380197]: Limits are [-2.146755, 1.605703], value was -3.440000.
[WARN] [1694406065.381487]: No valid pose could be found. Returned theta_list variable may be nonsense.
no IK solution found
moving to neutral..
moving to neutral failed!
Goal selector batch size 64
Evaluate policy
0%| | 0/5 [00:00<?, ?it/s][WARN] [1694406071.024276]: Would exceed position limits on joint 2.
[WARN] [1694406071.027939]: Limits are [-2.146755, 1.605703], value was -2.283000.
[WARN] [1694406071.029195]: No valid pose could be found. Returned theta_list variable may be nonsense.
no IK solution found
moving to neutral..
moving to neutral failed!
time to get goal to rollout 0.0001659393310546875
> /home/weirdab/goalrelabel_locobot_fullgcsl/huge/algo/huge.py(1050)sample_trajectory()
1048 from IPython.core.debugger import set_trace
1049 set_trace()
-> 1050 while t < curr_max: #self.curr_frontier: #self.max_path_length:
1051 if (curr_max - t == self.explore_length) and not stopped:
1052 stopped = True
ipdb> c
samples before normalized [ 0.97025514 1.3716675 -1.7952583 -0.48900414 -0.01262892 -0.14722599
-0.40280282] <class 'jaxlib.xla_extension.ArrayImpl'> (7,)
Added action is [0.00931571 0.01290708 0.01255678 0.02615689 0.02875672 0.07803939
0.2465638 ]
action clipped!
samples before normalized [ 0.93283796 1.4770036 -2.0304914 -0.23736706 0.1958166 -0.13575329
-0.23227417] <class 'jaxlib.xla_extension.ArrayImpl'> (7,)
Added action is [0.00930849 0.01292147 0.01259108 0.02611061 0.0286925 0.07804254
0.34839708]
action clipped!
samples before normalized [ 0.7924475 1.4374168 -1.8981501 -0.23285323 0.05775591 -0.16565613
-0.75542134] <class 'jaxlib.xla_extension.ArrayImpl'> (7,)
Added action is [0.0092814 0.01291606 0.01257178 0.02610978 0.02873504 0.07803433
0.03599337]
action clipped!
repated images for get_image method!
repated images for get_image method!
samples before normalized [ 0.73545426 1.445658 -1.8924967 -0.22196919 0.0497895 -0.16653901
-1.0642703 ] <class 'jaxlib.xla_extension.ArrayImpl'> (7,)
Added action is [ 0.0092704 0.01291719 0.01257096 0.02610778 0.02873749 0.0780341
-0.14843956]
action clipped!
samples before normalized [ 0.72231317 1.4503988 -1.8732325 -0.20745376 0.04127735 -0.16549581
-1.0117553 ] <class 'jaxlib.xla_extension.ArrayImpl'> (7,)
Added action is [ 0.00926786 0.01291784 0.01256815 0.02610511 0.02874011 0.07803438
-0.11707965]
action clipped!
samples before normalized [ 0.7278633 1.4368619 -1.8756726 -0.22618791 0.06283711 -0.16449644
-1.0342791 ] <class 'jaxlib.xla_extension.ArrayImpl'> (7,)
Added action is [ 0.00926893 0.01291599 0.0125685 0.02610856 0.02873347 0.07803465
-0.13052997]
action clipped!
samples before normalized [ 0.735414 1.4452916 -1.8781239 -0.20955886 0.05863629 -0.17685837
-1.0222832 ] <class 'jaxlib.xla_extension.ArrayImpl'> (7,)
Added action is [ 0.00927039 0.01291714 0.01256886 0.0261055 0.02873476 0.07803126
-0.12336645]
action clipped!
samples before normalized [ 0.72697735 1.4368008 -1.8524523 -0.24289031 0.06032449 -0.19015732
-1.3830544 ] <class 'jaxlib.xla_extension.ArrayImpl'> (7,)
Added action is [ 0.00926876 0.01291598 0.01256512 0.02611163 0.02873424 0.07802761
-0.33880535]
action clipped!
samples before normalized [ 0.7147742 1.4438474 -1.8367518 -0.22444993 0.07499987 -0.21300814
-1.3749485 ] <class 'jaxlib.xla_extension.ArrayImpl'> (7,)
Added action is [ 0.00926641 0.01291694 0.01256283 0.02610824 0.02872972 0.07802135
-0.33396485]
action clipped!
samples before normalized [ 0.70229685 1.4473615 -1.8244907 -0.21400252 0.08964357 -0.22391719
-1.2654428 ] <class 'jaxlib.xla_extension.ArrayImpl'> (7,)
Added action is [ 0.009264 0.01291742 0.01256104 0.02610632 0.02872521 0.07801836
-0.26857218]
action clipped!
samples before normalized [ 0.7340973 1.473994 -1.8508734 -0.19784576 0.0530136 -0.23718667
-1.3954504 ] <class 'jaxlib.xla_extension.ArrayImpl'> (7,)
Added action is [ 0.00927014 0.01292106 0.01256489 0.02610335 0.0287365 0.07801472
-0.34620777]
action clipped!
samples before normalized [ 0.7316608 1.4378705 -1.8220754 -0.29516345 0.04273079 -0.21541089
-1.7023227 ] <class 'jaxlib.xla_extension.ArrayImpl'> (7,)
Added action is [ 0.00926967 0.01291613 0.01256069 0.02612124 0.02873966 0.07802069
-0.5294603 ]
action clipped!
samples before normalized [ 0.8037339 1.4601629 -1.8548878 -0.30685765 0.05872571 -0.19143265
-1.6176925 ] <class 'jaxlib.xla_extension.ArrayImpl'> (7,)
Added action is [ 0.00928358 0.01291917 0.01256547 0.02612339 0.02873474 0.07802726
-0.47892234]
action clipped!
samples before normalized [ 0.8643358 1.4982544 -1.9606819 -0.2422189 0.10006149 -0.19918488
-1.2781966 ] <class 'jaxlib.xla_extension.ArrayImpl'> (7,)
Added action is [ 0.00929527 0.01292438 0.0125809 0.02611151 0.028722 0.07802514
-0.27618822]
action clipped!
samples before normalized [ 0.786553 1.4527746 -1.9388633 -0.26058972 0.09892879 -0.19647223
-1.369046 ] <class 'jaxlib.xla_extension.ArrayImpl'> (7,)
Added action is [ 0.00928026 0.01291816 0.01257772 0.02611488 0.02872235 0.07802588
-0.33044007]
action clipped!
samples before normalized [ 0.8048543 1.441401 -1.9533207 -0.26221204 0.09413055 -0.1946082
-1.3175772 ] <class 'jaxlib.xla_extension.ArrayImpl'> (7,)
Added action is [ 0.00928379 0.01291661 0.01257983 0.02611518 0.02872383 0.0780264
-0.29970488]
action clipped!
samples before normalized [ 0.75758004 1.4279449 -1.9691696 -0.28349072 0.1290725 -0.2034349
-1.309882 ] <class 'jaxlib.xla_extension.ArrayImpl'> (7,)
Added action is [ 0.00927467 0.01291477 0.01258214 0.0261191 0.02871306 0.07802398
-0.2951096 ]
action clipped!
samples before normalized [ 0.758535 1.4169228 -1.9726912 -0.31229183 0.1488091 -0.18123063
-1.169313 ] <class 'jaxlib.xla_extension.ArrayImpl'> (7,)
Added action is [ 0.00927485 0.01291326 0.01258265 0.02612439 0.02870698 0.07803006
-0.21116707]
action clipped!
samples before normalized [ 0.8334554 1.4201878 -1.9645644 -0.34066227 0.14435545 -0.18844792
-1.2291652 ] <class 'jaxlib.xla_extension.ArrayImpl'> (7,)
Added action is [ 0.00928931 0.01291371 0.01258147 0.02612961 0.02870836 0.07802808
-0.24690858]
action clipped!
samples before normalized [ 0.7883356 1.3972486 -1.9425926 -0.3545811 0.15943909 -0.18690675
-1.2435839 ] <class 'jaxlib.xla_extension.ArrayImpl'> (7,)
Added action is [ 0.0092806 0.01291057 0.01257826 0.02613217 0.02870371 0.07802851
-0.25551888]
action clipped!
samples before normalized [ 0.9248098 1.4382524 -2.0810895 -0.3233588 0.21680126 -0.19203143
-0.9877667 ] <class 'jaxlib.xla_extension.ArrayImpl'> (7,)
Added action is [ 0.00930694 0.01291618 0.01259846 0.02612643 0.02868604 0.0780271
-0.1027545 ]
action clipped!
samples before normalized [ 1.0160712 1.4608828 -2.0165439 -0.27167726 0.15307921 -0.19598076
-1.0541446 ] <class 'jaxlib.xla_extension.ArrayImpl'> (7,)
Added action is [ 0.00932455 0.01291927 0.01258905 0.02611692 0.02870567 0.07802602
-0.1423929 ]
action clipped!
samples before normalized [ 0.90875304 1.4339551 -1.953509 -0.37012368 0.19297053 -0.1880789
-0.66839504] <class 'jaxlib.xla_extension.ArrayImpl'> (7,)
Added action is [0.00930384 0.01291559 0.01257985 0.02613503 0.02869338 0.07802819
0.08796218]
action clipped!
samples before normalized [ 0.88828635 1.4096627 -1.9044924 -0.40591693 0.17726931 -0.17186946
-0.5442407 ] <class 'jaxlib.xla_extension.ArrayImpl'> (7,)
Added action is [0.00929989 0.01291227 0.0125727 0.02614161 0.02869822 0.07803264
0.16210243]
action clipped!
repated images for get_image method!
samples before normalized [ 0.84730136 1.3898685 -1.8352371 -0.42736462 0.18377748 -0.17858896
-0.60812604] <class 'jaxlib.xla_extension.ArrayImpl'> (7,)
Added action is [0.00929198 0.01290956 0.01256261 0.02614556 0.02869621 0.07803079
0.12395254]
action clipped!
samples before normalized [ 0.86698526 1.4043908 -1.9024822 -0.38717565 0.22943318 -0.1789239
-0.40881008] <class 'jaxlib.xla_extension.ArrayImpl'> (7,)
Added action is [0.00929578 0.01291155 0.01257241 0.02613816 0.02868214 0.0780307
0.2429765 ]
action clipped!
samples before normalized [ 1.0275571 1.3581132 -1.8302563 -0.47271284 0.1817059 -0.18089122
-0.44631827] <class 'jaxlib.xla_extension.ArrayImpl'> (7,)
Added action is [0.00932677 0.01290522 0.01256188 0.02615389 0.02869685 0.07803015
0.22057801]
action clipped!
samples before normalized [ 0.80170894 1.3675776 -1.8259588 -0.406202 0.2325322 -0.17103955
-0.50945836] <class 'jaxlib.xla_extension.ArrayImpl'> (7,)
Added action is [0.00928319 0.01290652 0.01256125 0.02614166 0.02868119 0.07803286
0.18287316]
action clipped!
samples before normalized [ 0.79734176 1.3993565 -1.9319774 -0.3239236 0.35101235 -0.15463872
-0.4119243 ] <class 'jaxlib.xla_extension.ArrayImpl'> (7,)
Added action is [0.00928234 0.01291086 0.01257671 0.02612653 0.02864469 0.07803736
0.2411168 ]
action clipped!
[WARN] [1694406120.374322]: No valid pose could be found. Returned theta_list variable may be nonsense.
no IK solution found
moving to neutral..
moving to neutral failed!
samples before normalized [ 0.73561954 1.4322803 -1.8646414 -0.32850367 0.1455136 -0.14187118
-0.58181715] <class 'jaxlib.xla_extension.ArrayImpl'> (7,)
Added action is [0.00927043 0.01291536 0.01256689 0.02612737 0.028708 0.07804086
0.13966322]
action clipped!
[WARN] [1694406125.047158]: No valid pose could be found. Returned theta_list variable may be nonsense.
no IK solution found
moving to neutral..
moving to neutral failed!
Added action is [ 0.53381091 -0.0919733 1.91382039 0.33079713 1.14194252 -1.12959516
-0.85005238]
action clipped!
0%| | 0/5 [00:59<?, ?it/s]
Traceback (most recent call last):
File "launch_main.py", line 286, in <module>
run(**params)
File "launch_main.py", line 210, in run
algo.train()
File "/home/weirdab/goalrelabel_locobot_fullgcsl/huge/algo/huge.py", line 1513, in train
self.evaluate_policy(self.eval_episodes, greedy=True, prefix='Eval')
File "/home/weirdab/goalrelabel_locobot_fullgcsl/huge/algo/huge.py", line 1866, in evaluate_policy
states, actions, goal_state, _, _ , img_states= self.sample_trajectory(goal=goal, greedy=greedy, save_video_trajectory=index==0, video_filename=video_filename)
File "/home/weirdab/goalrelabel_locobot_fullgcsl/huge/algo/huge.py", line 1050, in sample_trajectory
while t < curr_max: #self.curr_frontier: #self.max_path_length:
File "/home/weirdab/goalrelabel_locobot_fullgcsl/huge/algo/huge.py", line 885, in create_video
skvideo.io.vwrite(f"{self.trajectories_videos_folder}/{video_filename}.mp4", images)
File "/home/weirdab/miniconda3/envs/huge/lib/python3.8/site-packages/skvideo/io/io.py", line 53, in vwrite
videodata = vshape(videodata)
File "/home/weirdab/miniconda3/envs/huge/lib/python3.8/site-packages/skvideo/utils/__init__.py", line 318, in vshape
raise ValueError("Improper data input")
ValueError: Improper data input
Killed