gkahn13 / gcg Goto Github PK

Hi,

We are in a plan to create an environment which is somewhat similar to the one mentioned in the paper.
Can you suggest me the simulator tool you have used to create the environment.?

Regards,
Ratheesh

hi, I run the script 'python envs/rccar/square_cluttered_env.py',
output 'ImportError: No module named 'sandbox'',
should I do some configuration?

When I run, scattered_env.py, I get this. Any help?
Known pipe types:
wglGraphicsPipe
(all display modules loaded.)
:display(warning): FrameBufferProperties available less than requested.
requested: depth_bits=1 color_bits=3 red_bits=1 green_bits=1 blue_bits=1 alpha
_bits=1 back_buffers=1 force_hardware
got: depth_bits=24 color_bits=24 red_bits=8 green_bits=8 blue_bits=8 alpha_bit
s=8 accum_bits=64 force_hardware

So when you are passing the state input through the convolution layer and into the LSTM, how exactly are you initializing the cell state and the hidden state as the output from the previous convolution layers? I am trying to through the code but it is way too convoluted for me to understand which file is doing what.

In the paper you mention that the policy evaluation is done by sampling K random sample sequences. First, I am just curious as to how big of a K are you using, second with high model horizon of like H=16, how is that you are sampling a sequence that will even have anything close to optimal behavior, so essentially my question is when you say you are sampling K randomly sampled action sequences, are you sampling random actions for each time step because if you are then for large model horizons you need a huge value for K to get to a action sequence that might be close to optimal, or are you doing something else and if so please tell me what that is.

Is there any way to train it on a GPU sever and take the trained weights to a local environment that has only CPU for testing and evaluation?

I set all environment followed the 'Readme.md' step by step.
so I could saw simulation environment and the datafile being created.
($ python run_exp.py --exps ours
Creating your personal config from template...
Personal config created, but you should probably edit it before further experiments are run
Exiting.)

but I have no idea about how I can see the result on simulation environment or other way.
I know it's silly question, but now.. there's no other way except asking to solve this TT
please help me!

Hello, I am trying to test your method but I get this error when creating the anaconda environment.

Pip subprocess error:
parso requires Python '>=3.6' but the running Python is 3.5.2

Do you know of any solution?