I Can't MLLearn with Standalone build (mlagents-learn with --env with --num-envs={n} )

Unity Engine
, ,
1

Editor communicate with mlagents-learn doing well.
But standalone build cannot communite with mlagents-learn.
I Using ml-agent version. Release 21 (most recent version)

I try to run mlagents-learn like this
‘mlagents-learn ./config/ppo/config.yaml --env=D:\workspace\nemosemo-client\Build\windows\Build --num-envs=2 --run-id=run_1’

I can see the Player log got a error with grpc.
This below log is that.

Unexpected exception when trying to initialize communication: System.NotSupportedException: To marshal a managed method, please add an attribute named ‘MonoPInvokeCallback’ to the method definition. The method we’re attempting to marshal is: Grpc.Core.Internal.NativeLogRedirector::HandleWrite
at Grpc.Core.Internal.NativeLogRedirector.Redirect (Grpc.Core.Internal.NativeMethods native) [0x00000] in <00000000000000000000000000000000>:0
at Grpc.Core.Internal.NativeExtension.Get () [0x00000] in <00000000000000000000000000000000>:0
at Grpc.Core.GrpcEnvironment.GrpcNativeInit () [0x00000] in <00000000000000000000000000000000>:0
at Grpc.Core.GrpcEnvironment…ctor () [0x00000] in <00000000000000000000000000000000>:0
at Grpc.Core.GrpcEnvironment.AddRef () [0x00000] in <00000000000000000000000000000000>:0
at Grpc.Core.Channel…ctor (System.String target, Grpc.Core.ChannelCredentials credentials, System.Collections.Generic.IEnumerable1[T] options) [0x00000] in <00000000000000000000000000000000>:0 at Grpc.Core.Channel..ctor (System.String target, Grpc.Core.ChannelCredentials credentials) [0x00000] in <00000000000000000000000000000000>:0 at Unity.MLAgents.RpcCommunicator.Initialize (System.Int32 port, Unity.MLAgents.CommunicatorObjects.UnityOutputProto unityOutput, Unity.MLAgents.CommunicatorObjects.UnityInputProto& unityInput) [0x00000] in <00000000000000000000000000000000>:0 at Unity.MLAgents.RpcCommunicator.Initialize (Unity.MLAgents.CommunicatorInitParameters initParameters, Unity.MLAgents.UnityRLInitParameters& initParametersOut) [0x00000] in <00000000000000000000000000000000>:0 at Unity.MLAgents.Academy.InitializeEnvironment () [0x00000] in <00000000000000000000000000000000>:0 at Unity.MLAgents.Academy..ctor () [0x00000] in <00000000000000000000000000000000>:0 at Unity.MLAgents.Academy+<>c.<.cctor>b__85_0 () [0x00000] in <00000000000000000000000000000000>:0 at System.Lazy1[T].ViaFactory (System.Threading.LazyThreadSafetyMode mode) [0x00000] in <00000000000000000000000000000000>:0
at System.Lazy1[T].ExecutionAndPublication (System.LazyHelper executionAndPublication, System.Boolean useDefaultConstructor) [0x00000] in <00000000000000000000000000000000>:0 at System.Lazy1[T].get_Value () [0x00000] in <00000000000000000000000000000000>:0
at Unity.MLAgents.Agent.LazyInitialize () [0x00000] in <00000000000000000000000000000000>:0

I want to multiple agent mlagent-learn concurrent.
but this error log stuck my learning. Can’t found solution for this problem in google search or chatgpt.

Help me please.

2

I found below article’s comment
backend must be mono… not il2cpp

when I changed backend il2cpp to mono. It’s doing well!