License question: Mel-Band RoFormer Karaoke model (aufr33/viperx) — use in a free Android app

Hi, and first of all thank you for the amazing separation models — they are genuinely impressive work.

Question for @aufr33 (and viperx): I'd like to ask about the license / usage terms of the Mel-Band RoFormer Karaoke model (mel_band_roformer_karaoke_aufr33_viperx_sdr_10.1956.ckpt).

Context: I'm an amateur musician in Japan building a small stem-separation Android app for band practice. It splits a song into stems (lead vocals / backing vocals / guitar L-R / drums / bass / piano) so you can mute or solo each part while practicing. For the lead/backing vocal split it uses the karaoke model above via python-audio-separator, with inference running server-side on my own GPU server — the model file itself is not bundled into or redistributed with the app.

I'd like to publish the app on Google Play (free to start with; possibly paid plans in the future), but I could not find any license or usage terms for this model — neither on the Hugging Face mirror (jarredou/aufr33-viperx-karaoke-melroformer-model) nor in the release announcement.

So my questions are:

1. May I use this model in a publicly distributed app as described above?
2. If yes, under what conditions? (attribution required / free apps only / non-commercial only / etc.)
3. If this isn't the right place to ask, could you point me to where I should reach you or viperx?

I will of course follow whatever terms you set, including credit in the app and on the store listing — or removing the model entirely if you'd prefer. Thank you!