Is it possible to clear the context for the LSTM provided in this library? This would be useful for using multiple sequences for training. For example, if training a network to perform binary addition, for every new set of numbers being added, context should be reset.
Is it possible to clear the context for the LSTM provided in this library? This would be useful for using multiple sequences for training. For example, if training a network to perform binary addition, for every new set of numbers being added, context should be reset.