phate09 / safedrl Goto Github PK

I am a graduate student who has just started researching the use of PRISM for some validation tasks. Instead of using a PRISM desktop application, I wanted to do this directly by building the PRISM model in Python and using the program interface.

I notice that your work uses PRISM through Py4j, for example "MDP = gateway.entry_point.getmdpsimple ()". I'm curious if "getMdpSimple()" is in a Java program you wrote yourself, and how can I use PRISM's Java API?

I would be honored if you could take note of my problem！

Hi I am a graduate student, and I am trying to run this code and getting into errors of "Illegal instruction (core dumped)".
Here is my situation:
My host machine is windows. I am running the code on a ubuntu virtual machine on Virtualbox.
Right now, Prism is installed and the prism binary works. Gradle is installed and works.
I am able to run "train_pendulum.py" and obtain the dqn model.
But when running the "domain_analysis_sym.py", it runs into error "Illegal instruction (core dumped)".

By checking the diagnosis message, it reports "python domain_a[23732] trap invalid opcode ip:7f650a3094d0 sp:7ffc9c1a7ba8 error:0 in _pywrap_tensorflow_internal.so[7f64fa800000+2a2d8000]", attached below.

I searched for the error code online, and it might due to that the windows virtualization Hyper-V needs to be turned off in Virutalbox. I haven't succeeded with that yet. I would like to know what os you have run these code with. Is it a linux machine without virtualization?

I was wondering if you could come up with any hint on this problem? Your help would be much appreciated.

when i try to run runnables/symbolic/dqn/domain_analysis_sym.py
That error occurred：

Resetting the StateStorage
Loading complete
Building the tree
Resetting the tree
Finished building the tree
Iteration 0
Traceback (most recent call last):
  File "runnables/symbolic/dqn/domain_analysis_sym.py", line 46, in <module>
    allow_assign_actions=True)
  File "/usr/local/SafeDRL-master/symbolic/unroll_methods.py", line 599, in probability_iteration
    storage.recreate_prism(horizon * 2)
  File "/usr/local/SafeDRL-master/prism/state_storage.py", line 102, in recreate_prism
    mdp = gateway.entry_point.reset_mdp()
  File "/root/miniconda3/envs/safedrl/lib/python3.7/site-packages/py4j/java_gateway.py", line 1310, in __call__
    answer, self.gateway_client, self.target_id, self.name)
  File "/root/miniconda3/envs/safedrl/lib/python3.7/site-packages/py4j/protocol.py", line 332, in get_return_value
    format(target_id, ".", name, value))
py4j.protocol.Py4JError: An error occurred while calling t.reset_mdp. Trace:
py4j.Py4JException: Method reset_mdp([]) does not exist
	at py4j.reflection.ReflectionEngine.getMethod(ReflectionEngine.java:318)
	at py4j.reflection.ReflectionEngine.getMethod(ReflectionEngine.java:326)
	at py4j.Gateway.invoke(Gateway.java:274)
	at py4j.commands.AbstractCommand.invokeMethod(AbstractCommand.java:132)
	at py4j.commands.CallCommand.execute(CallCommand.java:79)
	at py4j.GatewayConnection.run(GatewayConnection.java:238)
	at java.lang.Thread.run(Thread.java:748)

why come to that
Is it due to spark version?