ProcTHOR static object poses are all zero

[BlGene]

I've found a workaround for this issue, so it's not urgent.

Hi @StoneT2000

I was trying to get the poses of items in ProcTHOR, however these were returning only zeros.
(via procthor_scene.scene_objects[object_name].pose.get_p() )

I tried to fix it by setting a value for builder.initial_pose = sapien.Pose() however, this messed the scene up.

It looks like the debug viewer is also giving zero values for the poses. (see below)

The workaround for me was to add a self._default_scene_objects_poses.append((actor, pose)) to the scene builder. Could provide a PR if it's helpful.

[StoneT2000]

Is this on the GPU sim or CPU sim?

[BlGene]

Hi @StoneT2000,

Yes sure! The output of env.print_sim_details() is :

# -------------------------------------------------------------------------- #
Task ID: ClevrMove-v1, 1 parallel environments, sim_backend=cpu
obs_mode=rgb+depth+segmentation, control_mode=pd_joint_pos
render_mode=rgb_array, sensor_details=RGBD(448x448), RGBD(128x128), RGBD(128x128)
sim_freq=100, control_freq=20
observation space: Dict('agent': Dict('qpos': Box(-inf, inf, (1, 15), float32), 'qvel': Box(-inf, inf, (1, 15), float32)), 'extra': Dict('tcp_pose': Box(-inf, inf, (1, 7), float32)), 'sensor_param': Dict('render_camera': Dict('extrinsic_cv': Box(-inf, inf, (1, 3, 4), float32), 'cam2world_gl': Box(-inf, inf, (1, 4, 4), float32), 'intrinsic_cv': Box(-inf, inf, (1, 3, 3), float32)), 'fetch_head': Dict('extrinsic_cv': Box(-inf, inf, (1, 3, 4), float32), 'cam2world_gl': Box(-inf, inf, (1, 4, 4), float32), 'intrinsic_cv': Box(-inf, inf, (1, 3, 3), float32)), 'fetch_hand': Dict('extrinsic_cv': Box(-inf, inf, (1, 3, 4), float32), 'cam2world_gl': Box(-inf, inf, (1, 4, 4), float32), 'intrinsic_cv': Box(-inf, inf, (1, 3, 3), float32))), 'sensor_data': Dict('render_camera': Dict('depth': Box(-32768, 32767, (1, 448, 448, 1), int16), 'rgb': Box(0, 255, (1, 448, 448, 3), uint8), 'segmentation': Box(-9223372036854775808, 9223372036854775807, (1, 448, 448, 1), int64)), 'fetch_head': Dict('depth': Box(-32768, 32767, (1, 128, 128, 1), int16), 'rgb': Box(0, 255, (1, 128, 128, 3), uint8), 'segmentation': Box(-9223372036854775808, 9223372036854775807, (1, 128, 128, 1), int64)), 'fetch_hand': Dict('depth': Box(-32768, 32767, (1, 128, 128, 1), int16), 'rgb': Box(0, 255, (1, 128, 128, 3), uint8), 'segmentation': Box(-9223372036854775808, 9223372036854775807, (1, 128, 128, 1), int64))))
(single) action space: Box([-1.606 -1.221   -inf -2.251   -inf -2.16    -inf -1.    -1.    -1.
 -1.    -1.    -1.   ], [1.606 1.518   inf 2.251   inf 2.16    inf 1.    1.    1.    1.    1.
 1.   ], (13,), float32)
# -------------------------------------------------------------------------- #