Sorry I took so long to reply, things have been a little hectic with personal/work life

So thats a super interesting feature which might be doable depending on whether or not I opensource this thing (which I might end up doing in the end)… For now I think it’s out of scope…

That being said, I ended up implementing an audience reaction feature today (think IG live emoji thing) along with an accompanying DJ tool (so that the DJ can “see” reactions to a song).

I’ll probably implement an audio quality thing next. I noticed that when I’m on mobile and in a bad coverage area (which is everywhere I live), music take a LOOONG time to actually begin playing… really dislike that…

Sent it to in a DM because of all the downvotes here

You can definitely give it a shot now, it’s done at this point! Keep in mind the llm can pick and choose between the various tools to call so news may not be called out each time

OK so, I’m almost done with the tool for the AI DJ. Basically we’re using searxng to search for audience locale news + using rss feeds to provide the llm with context about specific news categories (Technology for example)

Yeah that’s actually a tool I was already working on creating for the AI DJ. basically pulls current news info based on audience location

Saw your request and songs are queued for generation! The new station is called Cloud 9 Chill FM So it supports station jingles but the intent there was just for a “youre listening to so and so station” announcements (same as in regular radio).

I’m not sure about AI podcasts though… maybe like a news station would be useful

Thanks, I really appreciate it! Not all the stations have an AI DJ btw… I should probably add some visual queue in the station browser which shows whether or not one is set. Also, some of the AI DJ actions dont trigger if there arent enough people listening to a station (like polls for example)

Sent it to you in a DM because of all the downvotes here 🤕

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)

I’m curious, do you have ADG enabled at all? How many steps do you use generally?

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…

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

Yup, I have instrumental only stations (and they’re tagged as such so you can filter by those only)

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

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)

It’s one of the reasons I got solar!

My electric bill was higher than my loan payment so it just made sense for me.

I’m just gonna try vllm, seems like ik_llama.cpp doesnt have a quick docker method

IK sounds promising! Will check it out to see if it can run in a container

I’ll take a look at both tabby and vllm tomorrow

Hopefully there’s cpu offload in the works so I can test those crazy models without too much fiddling in the future (server also has 128gb of ram)

Unfortunately i didn’t set up nvlink, but ollama auto splits things for models which require it

I really just a “set and forget” model server lol (that’s why I keep mentioning the auto offload)

Ollama integrates nicely with OWUI