The previous version had training errors, and this version should make the 3d style much more prominent and with less watermarks. This lora was trained with 250 different images of high quality 3D images on NoobAI-XL (NAI-XL) V-Pred-0.75S-Version.

Enjoy!