-
I suppose you know what LSTM is. If you don't, this project is a good example of LSTM, and you can make your own LSTM from my files.
-
You will need MIR-1K dataset (which is open source) if you want to retrain the model.
-
Run "pv2mat.m" then "pitch_lstm_traindev.m" to retrain.
-
Run "mian.m" to transcript notes. (Don't ask me why it is not "main.m".)
-
I will add comments if I have time. (I will not do this if I do not have time.)
-
If you publish a paper based on this, can you please list me as an author? :)
sniperwrb / lstm_note_transcriptor Goto Github PK
View Code? Open in Web Editor NEWUsing LSTM to estimate pitch, tempo, and transcript notes from pop songs