search

wangcongrobot / gym-ur5_gripper

13 stars 4 forks source link

ValueError: operands could not be broadcast together with shapes (4,) (11,) #1

Open ChenyangRan opened 5 years ago

ChenyangRan commented 5 years ago

Hi, I'm interested in your work, but not familiar with mujoco. When I clone your pkg and run it as the readme.m, it tells me that: ValueError: operands could not be broadcast together with shapes (4,) (11,)

I think errors happened in gym_ur5_gripper/envs/ur5_gripper_env.py, but I don't know how to correct it. tes

Could you give me some advice? Thanks.

Ubuntu 16.04 python3.6 mujoco200

wangcongrobot commented 5 years ago

Sorry, I haven't accomplished it until now and there may be some bugs. The training environment is based on mujoco200, not ROS. I will try to inspect my code these days. It seems that you use ROS controller in the first picture?

ChenyangRan commented 5 years ago

Sorry, I haven't accomplished it until now and there may be some bugs. The training environment is based on mujoco200, not ROS. I will try to inspect my code these days. It seems that you use ROS controller in the first picture?

Thanks for your reply. I’m sorry that I past the other picture here. This picture is what I got when testing other code. I have deleted the picture.

wangcongrobot commented 5 years ago

There is some useful link you can refer to. Actually, I also refer some code from them. The topic is manipulation, mujoco, arm, gripper.

  1. robotsuite You can choose different arm and gripper to do a lot of manipulation tasks.

  2. Fetch robot in gym It is the most common robot task to do some pick&place task.

  3. DoorGym It is a good example that trains an arm with a gripper to open the door.

For now, if you want to train the robot RL algorithms, you can choose mujoco, pybullet, vrep, gazebo, Unity. I spend some time to learn the API about mujoco.

ChenyangRan commented 5 years ago

There is some useful link you can refer to. Actually, I also refer some code from them. The topic is manipulation, mujoco, arm, gripper.

  1. robotsuite You can choose different arm and gripper to do a lot of manipulation tasks.
  2. Fetch robot in gym It is the most common robot task to do some pick&place task.
  3. DoorGym It is a good example that trains an arm with a gripper to open the door.

For now, if you want to train the robot RL algorithms, you can choose mujoco, pybullet, vrep, gazebo, Unity. I spend some time to learn the API about mujoco.

Thanks for your links. I will read them carefully.

sachin-ag commented 1 year ago

Hello, I am facing the same issue.. have you been able to fix it yet?

I am using mujoco210 as it works on linux arm64.