> Specifically, the sLSTM has still recurrence (memory mixing) in it, i.e. you cannot fully parallelize the computation.
If you mean that you cannot fully parallelize inference, this might be true but also not quite relevant since the computational demands of inference are low. And you can always "parallelize" training to some extent, just by training larger batches.
If you mean that you cannot fully parallelize inference, this might be true but also not quite relevant since the computational demands of inference are low. And you can always "parallelize" training to some extent, just by training larger batches.