amolchanov86 / quad_sim2multireal Goto Github PK

i want to try the model in the code

Hi @amolchanov86 ,

Thanks a lot for your great work! It's amazing performance simply using model-free RL policy transfer.

I'm running your code right now and wondering if there is any way to visualize the simulation and print the training information, like reward, episode number, etc.

Thank you!

Best,
Xubo

Can you provide the pre-trained weights for the basic tasks such as hovering or takeoff?

The issue has been solved.

Hello, I have trained a good policy to control the quadrotor, but when I test in the RotorS gazebo environment, the action always fluctuated when the quadrotor was hovering, why does this happen? The quadrotor could hover at a position with low bias.

Hi @whoenig @pushpreet @amolchanov86 @TaoChenOSU @JialouWang

Thanks for your work on sim2multireal project!

I am trying to use the script "test_controller.py". However, it uses a package called gym_art which I cannot find from current repo.
But I also find it similar to the "quad_sim" package. Do they share the same code? Can I use "quad_sim" instead of "gym_art".

Edit: also no module named "simulators_investigation". Where to find this?

Thanks!