It’s also possible that you didn’t need 300 epochs, but 200 would work as well… if you’re interested in optimising this, you can save a model checkpoint (checkpoint_model flag), but bear in mind that setting aside some data for validation means there will be less data for training. So you should have plenty of data if you want to try this! You can also just try different numbers of epochs and see what works, but I wouldn’t spend too much time on this. Important thing is that it works!
Good luck with rules, I think you will like them! 