search

mivp-agent / moos-ivp-agent

Model agnostic ML tooling for MOOS-IvP
https://mivp-agent.github.io/
MIT License
8 stars 3 forks source link

Program results #16

Closed En-lightener closed 2 months ago

En-lightener commented 2 months ago

Hello, could you please tell me the results of AgentAquaticusDemo.gif and DebuggingDemo.gif and DiscretizerDemo.gif and QTableDemo.gif and TestingDemo.gif to input which commands can be run? Can you provide the running instructions?

CarterFendley commented 2 months ago

Hi @En-lightener, for all of these you should be running the commands specified in the docker container. You can get a terminal on the docker container by ./docker.sh connect (after previously running ./docker.sh run). All the UI elements are forwarded to the host machine via X11 (you won't see anything at all if that is not properly setup) there are some MacOS specific notes in the main README about that. Below is information on each of those.

The AgentAquaticusDemo.gif was produced by running the examples/QTable example through the ./test.sh script. There are a few modifications required to get the simulation to show up graphically, namely here and here, uncomment the below and comment the lines above so that it looks like the following.

# TIME_WARP="10"
TIME_WARP="1"
#   ./launch_shoreside.sh --no_gui $LOGGING $TIME_WARP >& /dev/null &
  ./launch_shoreside.sh $LOGGING $TIME_WARP >& /dev/null &

The DebuggingDemo.gif was produced by running the examples/pLearn example which allows for backwards compatibility with previous ML tooling. There are details in that README, to get the graphs shown in the GIF to appear you need to add --debug For example you could use:

./run.sh --model model/trained/topModel/model/ --debug

The DiscretizerDemo.gif was produced by running the examples/DiscreteField demo. There is a README.md in that folder with extensive information on it.

The AgentAquaticusDemo.gif was produced by running the examples/QTable example through the ./run.sh script. No additional modifications should be needed to run that one.

The TestingDemo.gif was produced by running examples/pLearn examples. You will have to run the commands listed below.

NOTE: That GIF is extremely speed up.

./model/test.py
# AND in a separate terminal
./model/scripts/launch_sim.sh
CarterFendley commented 2 months ago

Good questions, closing since it seems there was no followup questions.