How to get more stylistic variety for rock and metal?

[DerFlausch]

I used ACE-Step now for some days, and while it is a big step forward compared to other AI music generators that I tried running locally, the output for rock and metal sounds quite limited. Like there were only three bands used in the training data (Metallica, Nickelback and Linkin Park). Nearly every output sounds like one of these three bands or a mixture of them.
I tried a lot of different prompts, but the result doesn't change that much.

Am I doing something wrong or is this just the current state of the model?
I hope there will be a LORA for rock and metal someday or a more variable model to get more diverse output.

[ChuxiJ]

try to increase guidance interval