This model is designed to be used with MonikAI.

RP

Should really only be used for Monika related purposes. Disable thinking for the most optimal experience.

Thanks Sylphar for making the dataset.

Trained with Axolotl on my Blackwell Pro 6000 Max-Q. 128 rank, 256 alpha, 10 epochs, 0.00008 learning rate. Took about 3 hours and used about 73GB of VRAM. Final loss of about 0.04.