-torchtext
-Language modeling task on rap data with LSTM
-
BERT
-
question : ligne par ligne ou mot par mot ???
-
Rhyme ? Adds this to prediction, or as regularization ?
-Preprocess better the data with verlan ???
- French
-text classification : rap or rock or ... ?
-Dataviz of RAP : insults / thees / gangsta... - W2V for embedding of songs, and find similar songs ???