Train a Neural Network on a Remote System
RemoteConnect can be used to create a connection to a remote machine, and programmatically interact with it. In this case it is used to launch a script, training a neural network on a more capable computer.
Connect to a remote machine.
Create a Wolfram Language script on the remote machine to perform neural network training.
#!wolframscript
data=Flatten@Table[{x,y}->Exp[-Norm[{x,y}]],{x,-3,3,.01},{y,-3,3,.01}];
net=NetChain[{32,Tanh,1}];
trained=NetTrain[net,data,BatchSize->1024, TimeGoal->10,TrainingProgressReporting->"Print"]
Export[$HomeDirectory<>"/trained.wlnet", trained, "WLNet"]
Use WriteLine to send commands to the remote machine without waiting for them to return.
When the evaluation completes, use ReadString to read the results.
Close the connection.