Automatic Speech Recognition using encoder decoder and an attention mechanism between them. It also used a techer forcing method.
Data can be found here -->
ASR Data: https://www.kaggle.com/competitions/11-785-s22-hw4p2/data
The ASR attention model has bidirectional LSTM encoder and attention mechanism which passes the context to the decoder. The model is based on the one described in Listen, Attend and Spell paper.