KVR Audio

nosleep0 · Post by **nosleep0** » Mon Feb 03, 2025 11:58 am

Hello there! I'm looking for advice:

Let's say I want to train a ai-model to generate music tracks in some specific style, or in the style of some particular artist (type here the name of any author, in your opinion, that has its own unique features)...
And I have a library of reference tracks. And I want my model to be trained on this data, within the sound and stylistic peculiarities of the tracks I have collected. But, at the same time, so that the result of generation would not be a mess of different uncoordinated parts, but would obey the generally accepted norms of harmony and musical literacy.

Is there any script or ready-made solution or anything else that could fulfill this task? So far, all the ways of building a custom music-ai model I have found in the web (such as Musicgen collab) have turned out to be inoperable.

eassae · Post by **eassae** » Mon Feb 03, 2025 1:33 pm

jancivil · Post by **jancivil** » Mon Feb 03, 2025 6:55 pm

My advice is take a long walk off a short pier.

AIs are being trained daily to rip musicians off, we're onto you.
For example:

Custom AI training model based on your own audio data (?)