How to use HMM for state estimation with multiple observations per state

Hi, I am interested in using the HMM toolbox for state estimation. My training data is a sequence of N states with a corresponding MxN sequence of observations, where at each state a fixed number of observations is recorded. I can calculate the state transition matrix and the state emission matrix using MLE, but I don't know if I can use the Viterbi algorithm to find the most probable state sequence given the observation sequence. Thank you