Based on the same old dataset as my SD1.5/SDXL/Flux models, fine-tuned with https://github.com/shootthesound/Fizgig
mst3c_k9b-epoch2-base.safetensors