Revised Dataset. Much more flexible than the previous version imo. Has trouble producing "slit pupils" or "snake eyes" so you'd have to inpaint those. Kept the old version, since it’s more similar to the show’s style.
Recommended weight between 0.7-0.9; can go higher or lower depending on additional LoRAs, etc.
Generated Batch Size 6, and picked one seed. Then used Hi-Res Fix (Upscaler: 4x-UltraSharp) 2x —> Extras (Upscaler: 4x-UltraSharp) 2x
Description
trained on based95.
needed to improve backgrounds, overall compatibility with other LoRA art-styles.
FAQ
Comments (3)
fdacdb3dae
jinx_TT???
without this embedding the images do not even come close to your result.
Do RAVEN PLEASEEEE
this looks beautiufl, can you do a karin uzumaki lora too please <3