F.1 European and American cartoon game scenes

Model Usage Notice!!!

Model LoRa Weight: 0.8~1, Prompt Word Guidance Coefficient: 8

All reference images on the page are generated by web UI, when copying online make sure to adjust the Prompt Word Guidance Coefficient to 8

Model can be used for any European and American casual cartoon games Scene painting style transformation

Compared to XL model, the painting style transformation using this model is simpler, achieving the painting style transformation effect in one or two generations of images

Large Model: F.1-dev-fp8 or F.1-dev-fp16

Recommended Iteration Steps: 30 steps

Sampling Method: Euler a

HD Restoration: Enlargement algorithm R-ESRGAN 4x+ Anime6B

            Enlargement factor between 1.5 and 2

Recommended Aspect Ratio: 1024x512

Note: If the image clarity is not sufficient, increase the enlargement factor or adjust the resolution. To change the original painting style while keeping the details unchanged, adjust the return amplitude in the generation of images to between 0.4 and 0.6. If the effect is not sufficient, continue to change the base image through iterative nesting for painting style transformation