gps
gps copied to clipboard
cbfinn
Guided Policy Search
I hope this is helpful to you. [https://blog.csdn.net/qq_41009742/article/details/103131546](https://blog.csdn.net/qq_41009742/article/details/103131546)
when i run python python/gps/gps_main.py box2d_pointmass_pi2_example, the following error occurs. ```bash ('__file__:', 'python/gps/gps_main.py') ('add path:', 'python') pygame 1.9.6 Hello from the pygame community. https://www.pygame.org/contribute.html Initializing pygame framework... DEBUG:tm._add: /joy, sensor_msgs/Joy,...
MPC GPS
This PR implement MPC Guided Policy Search, which described in [1]. Main contribution: 1. `gps_main.py` is modified to make the agent able to run the MPC trajectory optimizer during sampling....
I have a problems, when run python .../gps_main.py pr2_tensorflow_example -p 9 to test the policy, follows: **File "/home/eric/PycharmProjects/gps/python/gps/algorithm/policy_opt/policy_opt_tf.py", line 249, in restore_model self.saver.restore(self.sess, fname)** ..... ...... **DataLossError: 2 root error(s)...
To run the V-REP agent: 1. Download V-REP. http://www.coppeliarobotics.com/downloads.html 2. The V-REP agent communicates with V-REP using its remote API: http://www.coppeliarobotics.com/helpFiles/en/remoteApiOverview.htm http://www.coppeliarobotics.com/helpFiles/en/remoteApiClientSide.htm Help Python find the vrep module: $ export...
When running "python python/gps/gps_main.py mjc_badmm_example‘， I got "ImportError: cannot import name 'QtCore'". Please help me.
In this repo, it offers two methods to set the init_traj_distr, which are init_pd, init_lqr. when dX < 2 * dU, errors occur. how to fix this bug? ```python #...
when i use `python python/gps/gps_main.py box2d_pointmass_pigps_example --resume 5`, some errors occur. when i check the implementation, i found the reason in `policy_opt_tf.py`. ```python def save_model(self, fname): print "self.sess:", self.sess print...
I am installing gps on ubuntu 16.04. Everything works fine with box2d examples. The compilation for mjcpy is also successful (I have generated mjcpy.so). However, when I execute the example...
It seems the following line computes new action without any use of nominal trajectory: https://github.com/cbfinn/gps/blob/82fa6cc930c4392d55d2525f6b792089f1d2ccfe/python/gps/algorithm/traj_opt/traj_opt_lqr_python.py#L257 But in the following papers, nominal trajectory seems to be essential in the whole iLQR...
