Speaker recognition(how to feed data to Lstm model)

Question

My goal is to make a speaker recognition using recurrent network (LSTM , GRE , etc)

My dataset consists of 28 classes (28 people)

filenames = tf.io.gfi