
Update 11/29/2023 - SDXL version. Trained with the same 33 images.
I don't have a good GPU so I'm using TensorArt which unfortunately messes up the generation data. The images were made with text2img with no img2img or other edits.
Happy Halloween!π I'm making Halloween Loras: https://civitai.com/articles/2397
Trained on 33 images
Checkpoint: Anime
Network Dim: 16




