Frame Level Speech Recognition

In this project we apply feedforward neural network expertise to the task of speech recognition. The given dataset comprises of phoneme state (subphoneme) labels for audio recordings (utterances). The data is taken from Wall Street Journal (WSJ) articles that have been read aloud and labeled using the original language. Finding the phoneme state label for each frame in the test dataset is the task at hand. It's crucial to remember that utterances might vary in length.