File "/home/robot/3D-Diffusion-Policy/third_party/pytorch3d_simplified/pytorch3d/ops/sample_farthest_points.py", line 139, in sample_farthest_points idx = _C.sample_farthest_points(points, lengths, K, start_idxs) RuntimeError: Not compiled with GPU support.

[dbdxnuliba]

(dp3) robot@robot-LEGION-REN7000K-26IRB:~/3D-Diffusion-Policy$ bash scripts/train_policy.sh dp3 adroit_hammer 0112 0 0 gpu id (to use): 0 Train mode /home/robot/anaconda3/envs/dp3/lib/python3.8/site-packages/hydra/_internal/defaults_list.py:251: UserWarning: In 'dp3.yaml': Defaults list is missing _self_. See https://hydra.cc/docs/upgrades/1.0_to_1.1/default_composition_order for more information warnings.warn(msg, UserWarning) [DP3Encoder] point cloud shape: [512, 3] [DP3Encoder] state shape: [24] [DP3Encoder] imagination point shape: None [PointNetEncoderXYZ] use_layernorm: True [PointNetEncoderXYZ] use_final_norm: layernorm [DP3Encoder] output dim: 128 [DiffusionUnetHybridPointcloudPolicy] use_pc_color: False [DiffusionUnetHybridPointcloudPolicy] pointnet_type: pointnet [2024-05-24 10:30:10,598][diffusion_policy_3d.model.diffusion.conditional_unet1d][INFO] - number of parameters: 2.551533e+08

Class name: DP3 Number of parameters: 255.2181M _dummy_variable: 0.0000M (0.00%) obs_encoder: 0.0648M (0.03%) model: 255.1533M (99.97%) mask_genrobottor: 0.0000M (0.00%)

Replay Buffer: state, shape (1000, 24), dtype float32, range -1.20~1.60 Replay Buffer: action, shape (1000, 26), dtype float32, range -1.00~1.00 Replay Buffer: point_cloud, shape (1000, 512, 6), dtype float32, range -0.25~255.00 Replay Buffer: img, shape (1000, 84, 84, 3), dtype uint8, range 0.00~255.00

/home/robot/3D-Diffusion-Policy/third_party/gym-0.21.0/gym/logger.py:34: UserWarning: WARN: Box bound precision lowered by casting to float32 warnings.warn(colorize("%s: %s" % ("WARN", msg % args), "yellow")) [MujocoPointcloudWrapper] use_point_crop: True [MujocoPointcloudWrapper] sampling 512 points from point cloud using fps

[WandB] group: adroit_hammer-dp3-0112 [WandB] name: 0

wandb: Currently logged in as: robot_nuli. Use wandb login --relogin to force relogin wandb: Tracking run with wandb version 0.17.0 wandb: Run data is saved locally in /home/robot/3D-Diffusion-Policy/3D-Diffusion-Policy/data/outputs/adroit_hammer-dp3-0112_seed0/wandb/run-20240524_103013-zg67wyim wandb: Run wandb offline to turn off syncing. wandb: Syncing run 0 wandb: ⭐️ View project at https://wandb.ai/robot_nuli/dp3 wandb: 🚀 View run at https://wandb.ai/robot_nuli/dp3/runs/zg67wyim Eval in Adroit hammer Pointcloud Env: 0%| | 0/20 [00:00<?, ?it/s]Found 1 GPUs for rendering. Using device 0. test_mean_score: 0.0
test_mean_score: 1.0
test_mean_score: 1.0
test_mean_score: 1.0
Error executing job with overrides: ['task=adroit_hammer', 'training.debug=False', 'training.seed=0', 'training.device=cuda:0', 'exp_name=adroit_hammer-dp3-0112', 'logging.mode=online', 'checkpoint.save_ckpt=True'] Traceback (most recent call last): File "train.py", line 506, in main() File "/home/robot/anaconda3/envs/dp3/lib/python3.8/site-packages/hydra/main.py", line 90, in decorated_main _run_hydra( File "/home/robot/anaconda3/envs/dp3/lib/python3.8/site-packages/hydra/_internal/utils.py", line 389, in _run_hydra _run_app( File "/home/robot/anaconda3/envs/dp3/lib/python3.8/site-packages/hydra/_internal/utils.py", line 452, in _run_app run_and_report( File "/home/robot/anaconda3/envs/dp3/lib/python3.8/site-packages/hydra/_internal/utils.py", line 216, in run_and_report raise ex File "/home/robot/anaconda3/envs/dp3/lib/python3.8/site-packages/hydra/_internal/utils.py", line 213, in run_and_report return func() File "/home/robot/anaconda3/envs/dp3/lib/python3.8/site-packages/hydra/_internal/utils.py", line 453, in lambda: hydra.run( File "/home/robot/anaconda3/envs/dp3/lib/python3.8/site-packages/hydra/internal/hydra.py", line 132, in run = ret.return_value File "/home/robot/anaconda3/envs/dp3/lib/python3.8/site-packages/hydra/core/utils.py", line 260, in return_value raise self._return_value File "/home/robot/anaconda3/envs/dp3/lib/python3.8/site-packages/hydra/core/utils.py", line 186, in run_job ret.return_value = task_function(task_cfg) File "train.py", line 503, in main workspace.run() File "train.py", line 257, in run runner_log = env_runner.run(policy) File "/home/robot/3D-Diffusion-Policy/3D-Diffusion-Policy/diffusion_policy_3d/env_runner/adroit_runner.py", line 102, in run obs, reward, done, info = env.step(action) File "/home/robot/3D-Diffusion-Policy/3D-Diffusion-Policy/diffusion_policy_3d/gym_util/multistep_wrapper.py", line 145, in step observation, reward, done, info = super().step(act) File "/home/robot/3D-Diffusion-Policy/third_party/gym-0.21.0/gym/core.py", line 289, in step return self.env.step(action) File "/home/robot/3D-Diffusion-Policy/3D-Diffusion-Policy/diffusion_policy_3d/gym_util/video_recording_wrapper.py", line 34, in step result = super().step(action) File "/home/robot/3D-Diffusion-Policy/third_party/gym-0.21.0/gym/core.py", line 289, in step return self.env.step(action) File "/home/robot/3D-Diffusion-Policy/3D-Diffusion-Policy/diffusion_policy_3d/gym_util/mjpc_diffusion_wrapper.py", line 177, in step point_cloud, depth = self.get_point_cloud() File "/home/robot/3D-Diffusion-Policy/3D-Diffusion-Policy/diffusion_policy_3d/gym_util/mjpc_diffusion_wrapper.py", line 166, in get_point_cloud point_cloud = point_cloud_sampling(point_cloud=point_cloud, File "/home/robot/3D-Diffusion-Policy/3D-Diffusion-Policy/diffusion_policy_3d/gym_util/mjpc_diffusion_wrapper.py", line 80, in point_cloudsampling , sampled_indices = torch3d_ops.sample_farthest_points(points=point_cloud[...,:3], K=num_points) File "/home/robot/3D-Diffusion-Policy/third_party/pytorch3d_simplified/pytorch3d/ops/sample_farthest_points.py", line 139, in sample_farthest_points idx = _C.sample_farthest_points(points, lengths, K, start_idxs) RuntimeError: Not compiled with GPU support. wandb: \ 0.125 MB of 0.125 MB uploaded wandb: Run history: wandb: SR_test_L3 ▁▅▆█ wandb: SR_test_L5 ▁▆▇█ wandb: bc_loss █▂▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁ wandb: epoch ▁▁▁▁▂▂▂▂▂▃▃▃▃▃▃▄▄▄▄▄▅▅▅▅▅▅▆▆▆▆▆▇▇▇▇▇▇███ wandb: global_step ▁▁▁▁▂▂▂▂▂▃▃▃▃▃▃▄▄▄▄▄▅▅▅▅▅▅▆▆▆▆▆▇▇▇▇▇▇███ wandb: lr ▁▃▆███████████████████████▇▇▇▇▇▇▇▇▇▇▇▇▇▇ wandb: mean_n_goal_achieved ▁███ wandb: mean_success_rates ▁███ wandb: test_mean_score ▁███ wandb: train_action_mse_error █▂▂▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁ wandb: train_loss █▂▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁ wandb: wandb: Run summary: wandb: SR_test_L3 1.0 wandb: SR_test_L5 0.75 wandb: bc_loss 0.0002 wandb: epoch 800 wandb: global_step 6406 wandb: lr 9e-05 wandb: mean_n_goal_achieved 63.65 wandb: mean_success_rates 1.0 wandb: test_mean_score 1.0 wandb: train_action_mse_error 8e-05 wandb: train_loss 0.0002 wandb: wandb: 🚀 View run 0 at: https://wandb.ai/robot_nuli/dp3/runs/zg67wyim wandb: ⭐️ View project at: https://wandb.ai/robot_nuli/dp3 wandb: Synced 6 W&B file(s), 4 media file(s), 2 artifact file(s) and 0 other file(s) wandb: Find logs at: ./data/outputs/adroit_hammer-dp3-0112_seed0/wandb/run-20240524_103013-zg67wyim/logs (dp3) robot@robot-LEGION-REN7000K-26IRB:~/3D-Diffusion-Policy$

[YanjieZe]

Hi, does your computer have Nvidia GPU? If you have, then maybe re-install pytorch3d under your current conda env.