Hey lemmy!
I was wondering if anyone was interested in checking out an AI radio app that I’ve put together.
Really just looking for feedback and overall impressions. Obviously if it’s against the rules (I didnt see a rule this would violate), I wont post the link. If you guys are interested, I’ll edit the post to include the link.
Thanks guys!
EDIT: as @db2 mentions, I didnt describe what the app actually does… It’s an AI radio app with various stations that all play back AI generated music. All different genres. Everything but lyrics is generated locally (I need to experiment with some local models to see which produce lyrics well)
You must log in or register to comment.
naw, man.
No.
Nobody’s going to know, you haven’t described what it actually does.
Sorry for the oversight! It’s basically an AI radio app (with stations playing different music styles). There’s even an AI DJ feature that’s audience aware (think weather callouts for listener locations, audience polls etc)
Don’t mind the downvotes, they’re from people who don’t even use this community.
Yeah I’m not disheartened, my mom said it was cool /s
In all seriousness, my wife and I think it’s cool and I literally use it (all day) during the week while I’m working
If you’re going to do lyrics don’t do all the stations, leave some just to play the music.
Yup, I have instrumental only stations (and they’re tagged as such so you can filter by those only)
You might not even know how to answer this, but can it use accelerators to do the generation?
so i’m using an open source model which leverages a 3090. I tried to make the app itself as agnostic as possible so I can just plug any API compatible server into each component (track metadata generation, actual song generation, AI DJ script, DJ script and so on). I figured making this thing as flexible as possible would be best overall
Edit: model is ace step 1.5
You know before i read the post i was thinking of news or podcasts created in AI. But yeah music would be cool too.
I’d love to hear what local model you settle on for lyrics, I’ve been having a lot of fun with ACE-Step 1.5 but the lyric generator it’s bundled with produces semi-nonsense lyrics that have nothing to do with what I prompt it with. Which is actually kind of fun in its own way, I literally never know what the song’s going to be about, but I’d like a little control sometimes too. :)
Yeah I completely agree on the lyrics it can generate! That being said I actually haven’t tried using the 4b lm model version just yet. I should probably give that a shot…
It’s the one I use, so don’t expect miraculous improvement. :)
I’m curious, do you have ADG enabled at all? How many steps do you use generally?
ADG == Audio-Driven Guidance? I haven’t played around with that part much. I tried it out and couldn’t get it to work, but it turned out that the reason ACE Step wasn’t working was unrelated to that and I only figured out what was wrong after I stopped experimenting with ADG. So I haven’t gone back to try it again.
I’m not really much of a music connoisseur, I just know what I like when I hear it. So mostly I just put together lyrics and then throw them at the wall to see what sounds good. :)
ADG for ace step is Adaptive Dual Guidance. try turning that on + make sure you have thinking enabled, I saw a big difference with both those on (more so thinking)
Ah. After poking around in the Gradio UI a bit, I found an “Enable ADG” but the tooltip says it’s “Angle Domain Guidance”, same thing?
I’m a programmer, but sometimes with AI I feel like a primitive tribesperson blindly attempting various rituals in an effort to appease the machine spirits. Eventually something works, and then I just keep on doing that.
Edit: I have angered the gods! My ritual failed! When I enabled ADG the spirits smote me with the following:
RuntimeError: The size of tensor a (11400) must match the size of tensor b (5700) at non-singleton dimension 1
Guess I won’t be trying that for now. :)
