Model cannot produce the human voice?

#26
by sanjeev-bhandari01 - opened

I tried Stable Audio open 1.0(from hugging face space) and stable audio audiosparx 2.0 from stableaudio.com

Both of this model couldnot generate the human voice properly.

Is it limitation of model or it is designed to not generate the human voice?

Is there paper published alongside the model?

Thank you

Here's the paper:
https://arxiv.org/abs/2404.10301v1

Neither Stable Audio Open or Stable Audio 2.0 are trained to produce coherent vocals or speech, by design.

sanjeev-bhandari01 changed discussion status to closed

Sign up or log in to comment