update

[jtn191]

change this in line 211, in model.py:
tf.nn.seq2seq.rnn_decoder

to:
tf.contrib.legacy_seq2seq.rnn_decoder

[ghostPath]

@jtn191 Have you successflly run thiis code?? I change the code follow your instruction. But got another error!

[xunxdd]

https://github.com/xunxdd/music-generator

i updated code for tensorflow 1.8. It works. @ghostPath, hopefully it helps

Though I am disappointed that it only works with a very limited and basic set of midi files. I tried with Jazz and classical pianos, none work yet. Rigid ragtimemusic does work.