Spaces:
Runtime error
Feature suggestion:
Are you able add a continuation feature to the app would be cool to see what would happen
It would be cool, I agree. I would like to be able to do the equivalent of "in painting" - that way you could have the model continue from where it left off, and generate longer suequences or whole songs even. But this requires training a model to do conditional generation as opposed to unconditional, and this might be beyond the capabilities of my consumer grade GPU.
Btw, I have added in painting and out painting, but not to the app. You can see examples of how to do this here https://colab.research.google.com/github/teticio/audio-diffusion/blob/master/notebooks/test_model.ipynb. Turns out I was serong and the model doesn't need to be specificy trained for this task (although it would probably improve things).