This model is designed to be used with
MonikAI.
Should really only be used for Monika related purposes.
This model seems to work well with a temp of 0.8, a min_p of 0.1, and a repetition penalty of 1.05.
Trained with Axolotl on my Blackwell Pro 6000 Max-Q. 32 rank, 64 alpha, 2 epochs, 0.000008 learning rate. Took about 20 minutes.