In both the LSTM and CNN driver files, we should save models after training. This will involve adjusting the code in driver's main and the command line parser to give the user the option of choosing to save models and the the directory where they should be saved to. I've literally been stalking following the parsing setup that cilia uses, but one can also look up tutorials for configarparse.
Maybe for now we can just create a temporary "out" folder that holds the models, though we might want to make sure we save them somewhere accessible on Logan once we start seriously comparing them. This is not something I'm super familiar with.
In both the LSTM and CNN driver files, we should save models after training. This will involve adjusting the code in driver's main and the command line parser to give the user the option of choosing to save models and the the directory where they should be saved to. I've literally been
stalkingfollowing the parsing setup that cilia uses, but one can also look up tutorials for configarparse.Maybe for now we can just create a temporary "out" folder that holds the models, though we might want to make sure we save them somewhere accessible on Logan once we start seriously comparing them. This is not something I'm super familiar with.