GitHub - sniperwrb/lstm_note_transcriptor: Using LSTM to estimate pitch, tempo, and transcript notes from pop songs

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? :)

Name		Name	Last commit message	Last commit date
Latest commit History 12 Commits
.gitattributes		.gitattributes
README.md		README.md
README.txt		README.txt
constqm.m		constqm.m
cqgram.m		cqgram.m
draw3d.m		draw3d.m
drawlin.m		drawlin.m
getbias.m		getbias.m
hparams.m		hparams.m
mian.m		mian.m
names.mat		names.mat
novel.m		novel.m
phasetest.m		phasetest.m
pitch2note.m		pitch2note.m
pitch_lstm_444.mat		pitch_lstm_444.mat
pitch_lstm_test.m		pitch_lstm_test.m
pitch_lstm_traindev.m		pitch_lstm_traindev.m
pitchtest.m		pitchtest.m
pv2mat.m		pv2mat.m
tempo.m		tempo.m
tphase.m		tphase.m
wav2res.m		wav2res.m
wav2res_bunch.m		wav2res_bunch.m

Provide feedback