This is useless >30 secs
#1
by
qm9
- opened
ibid
Hey @dnato - the model is only trained up to context lengths of 30s. For workarounds to generate longer audio samples, you can check out the repository https://github.com/camenduru/MusicGen-colab